Hydrogenaudio Forums

Lossy Audio Compression => Opus => Topic started by: jmvalin on 2017-12-22 01:05:52

Title: Opus 1.3-beta is here
Post by: jmvalin on 2017-12-22 01:05:52
This beta release of the upcoming Opus 1.3 includes:
Additionally, as a way to test the upcoming update to opus-tools, we’re providing Windows binaries built with 1.3-beta. These binaries are based on libopusenc, which means opusenc is finally able to make use of the Opus delayed-decision feature to make better speech/music transitions.

Source code: opus-1.3-beta.tar.gz (https://archive.mozilla.org/pub/opus/opus-1.3-beta.tar.gz)
Win32 binaries (experimental): opus-tools-test1.zip (https://archive.mozilla.org/pub/opus/win32/opus-tools-test1.zip)
Title: Re: Opus 1.3-beta is here
Post by: Fairy on 2017-12-22 10:14:55
Nice!

Any changes in fidelity in the 80-128 kbit range?
Title: Re: Opus 1.3-beta is here
Post by: includemeout on 2017-12-22 11:24:45
Thank you very much for such great a format!

From someone for whom Opus has become the exclusive low-bitrate choice for his (almost daily) walks, given that his ancient Moto X phone doesn't have an SD-card slot and storage space is therefore, always at a premium.
Title: Re: Opus 1.3-beta is here
Post by: Anakunda on 2017-12-23 06:13:44
Nice!

Any changes in fidelity in the 80-128 kbit range?

I'd like to know too
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-23 13:30:00
Nice!

Any changes in fidelity in the 80-128 kbit range?
the quality is already transparent for 2 channel stereo music at 48 kbps, but I wonder if they can make 5.1 or 7.1 audio sound transparent at 96 kbps, because in my tests, its transparent for 5.1 sound at 128kbps. I can't wait to see what's in store for the next year when it comes to efficiency at even lower bitrates than 48kbps!
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2017-12-23 19:02:17
Some results for 1.3 beta

Opus 1.2.1 - 1.3 beta vs HE-AAC  ( Test at 32 kbps) (https://docs.google.com/spreadsheets/d/1HUXHtAxNzMGdAstwHIwKWDhBllhRVGcpXXnUwFODTqg/edit#gid=0)
Files (https://drive.google.com/file/d/15OAI0bcyYSDXNFrtxuZrXjmR4sKCxjq7/view)

There are some improvements ( "Can't wait" and "Fatboy" samples) but also some regressions. Most notable regression is on sample 4º.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-24 03:01:19
the quality is already transparent for 2 channel stereo music at 48 kbps
Now is a good time for TOS 8,
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2017-12-24 04:50:52
the quality is already transparent for 2 channel stereo music at 48 kbps
Now is a good time for TOS 8,

Actually, TOS 8 is more about the opposite case -- people claiming they hear a difference between two signals. Anyone can produce the null hypothesis in an ABX test (through bad hearing, not caring, or just answering randomly) even when there's an obvious difference. OTOH, not everyone can produce significant results when two files are nearly identical. IOW, there's no point in saying "prove it" to someone claiming something is transparent to them.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2017-12-24 04:51:59
There are some improvements ( "Can't wait" and "Fatboy" samples) but also some regressions. Most notable regression is on sample 4º.

I personally cannot hear much difference for sample #4, so I'm curious what artefact is annoying you. Is it something at a specific time?
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2017-12-24 16:00:17
It's the same artifact for all 3 samples (Korean speech, French speech, and Sample #4) where 1.3 does worse than 1.2.1. It's extra sibilance which possibly maybe described as "ringing", "paper sound" etc...

It repeats in a form of constant large periods during all time. It's not at a specific time.

This artifact is  present in L channel on the sample of Korean speech. Maybe it's stereo savings which cause this issue

Edit: grammatical errors
Title: Re: Opus 1.3-beta is here
Post by: flowzy on 2017-12-24 16:58:14
Just did some random transcode test with a song to opus format at low bitrate around 20kbps, i think Opus 1.3-beta perform worse than 1.2.1 at these low bitrate. i can hear some noise/artifacts that do not present in version 1.2.1-28. Suspect the speech/music detector mess up at these low bit rate.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-25 21:36:33
Just did some random transcode test with a song to opus format at low bitrate around 20kbps, i think Opus 1.3-beta perform worse than 1.2.1 at these low bitrate. i can hear some noise/artifacts that do not present in version 1.2.1-28. Suspect the speech/music detector mess up at these low bit rate.

20kbps is way too low for music, maybe in a year or two it will reach the same quality as 48 kbps and up.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-25 21:37:41
Thank you very much for such great a format!

From someone for whom Opus has become the exclusive low-bitrate choice for his (almost daily) walks, given that his ancient Moto X phone doesn't have an SD-card slot and storage space is therefore, always at a premium.


what is the bitrate that you encode your music in?
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-25 22:42:22
Here is my ABX of opus 1.3 using 48kbps and a rock track. Very easy to hear the difference and the opus 48kbps sounds "less focused" in the highs and sounds mono every now and then. I had to use an online ABX because my Foobar2000 under Arch Linux and WINE does not want to output any sound.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-25 22:49:49
When I encoded the track at one time the opusenc produced weird result in the beginning. I had to convert it to FLAC from opus since the file I encoded was longer than thirty seconds and when I cut everything from 00:30 and onwards I thought that I might as well make it a FLAC.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-26 03:16:40
When I encoded the track at one time the opusenc produced weird result in the beginning. I had to convert it to FLAC from opus since the file I encoded was longer than thirty seconds and when I cut everything from 00:30 and onwards I thought that I might as well make it a FLAC.

make sure that you are using libopus 1.2.1 dirty. that artifact in the beginning is a relic from opus 1.2 beta and the final version of 1.2. i too noticed that artifact, which was not present in the previous opusenc . i emailed jmvalin about that issue and he fixed it and released libopus 1.2.1

if you use opus 1.2.1 dirty, it will no longer have that artifact
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-26 03:18:36
When I encoded the track at one time the opusenc produced weird result in the beginning. I had to convert it to FLAC from opus since the file I encoded was longer than thirty seconds and when I cut everything from 00:30 and onwards I thought that I might as well make it a FLAC.

do you have dbpoweramp? well you can read tags  and see exactly what version of libopus it was encoded in
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-26 03:28:10
Here is my ABX of opus 1.3 using 48kbps and a rock track. Very easy to hear the difference and the opus 48kbps sounds "less focused" in the highs and sounds mono every now and then. I had to use an online ABX because my Foobar2000 under Arch Linux and WINE does not want to output any sound.


download libopus 1.2.1 dirty and transcode that flac into opus, and you will not hear any artifacts at 48 kbps. i have thousands of songs in this format at that bitrate and i never had problems with that revision.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-26 03:43:01
here are some songs in 48 kbps  libopus 1.2.1
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-26 03:52:36
Here is my ABX of opus 1.3 using 48kbps and a rock track. Very easy to hear the difference and the opus 48kbps sounds "less focused" in the highs and sounds mono every now and then. I had to use an online ABX because my Foobar2000 under Arch Linux and WINE does not want to output any sound.


here is that same song you posted, but encoded in 1.2.1 dirty

it sounds pretty transparent to me
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-26 13:19:16
When I encoded the track at one time the opusenc produced weird result in the beginning. I had to convert it to FLAC from opus since the file I encoded was longer than thirty seconds and when I cut everything from 00:30 and onwards I thought that I might as well make it a FLAC.

make sure that you are using libopus 1.2.1 dirty. that artifact in the beginning is a relic from opus 1.2 beta and the final version of 1.2. i too noticed that artifact, which was not present in the previous opusenc . i emailed jmvalin about that issue and he fixed it and released libopus 1.2.1

if you use opus 1.2.1 dirty, it will no longer have that artifact
This artifact happened with b1.3 when I encoded the song but then didn't happen again after that. Really strange that it happened once but not again.
Title: Re: Opus 1.3-beta is here
Post by: madmax on 2017-12-26 13:24:18
ehh.... what is opus 1.2.1 dirty
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-26 13:55:22
Here is my ABX of opus 1.3 using 48kbps and a rock track. Very easy to hear the difference and the opus 48kbps sounds "less focused" in the highs and sounds mono every now and then. I had to use an online ABX because my Foobar2000 under Arch Linux and WINE does not want to output any sound.


here is that same song you posted, but encoded in 1.2.1 dirty

it sounds pretty transparent to me

Interesting that you found that song on another album because I thought it was quite rare. Anyway, the track is full of artifacts which makes me wonder what age you are and if you have suffered any ear damage. I promise, I am not at all being rude but I felt I had to ask since the quality is bad and you say it is transparent to you. Regards.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-26 14:11:49
the quality is already transparent for 2 channel stereo music at 48 kbps
Now is a good time for TOS 8,

Actually, TOS 8 is more about the opposite case -- people claiming they hear a difference between two signals. Anyone can produce the null hypothesis in an ABX test (through bad hearing, not caring, or just answering randomly) even when there's an obvious difference. OTOH, not everyone can produce significant results when two files are nearly identical. IOW, there's no point in saying "prove it" to someone claiming something is transparent to them.
The rule says to rule out subjective opinions and must therefore mean any opinion, transparent or not transparent. If that makes it difficult to know when someone just don't care which result in an ABX that says that the person could not distinguish between to files is then up the Hydrogenaudio to sort out. I can talk you into believing that I think 14kbps is transparent. I don't have to do one but by doing an ABX and not caring by just pressing randomly the result will back up my claim that 14kbps is transparent. But if I claim that 14kbps is not transparent I am enforced to do an ABX. This is weird.

Actually, shouldn't the claim transparent be a TOS #8 since we can not easily prove it? We can prove though that we can hear a difference by getting good results on an ABX and therefore the only thing we can prove is that we do hear a difference and not when it comes to not hearing a difference.

Please let me know if I am missing some crucial scientific evidence of the opposite or anything else. Regards.
Title: Re: Opus 1.3-beta is here
Post by: yourlord on 2017-12-26 15:48:57
Anyway, the track is full of artifacts which makes me wonder what age you are and if you have suffered any ear damage. I promise, I am not at all being rude but I felt I had to ask since the quality is bad and you say it is transparent to you. Regards.

I winced when I saw him claim transparency at 48kbps as well because to me just about any full bandwidth music is obviously degraded at that bitrate. But, I can play 48kbps mp3 files for my wife and kids and they just can't hear the artifacts. I don't know what it is but they swear the files sound just fine to them. To me they sound like someone drowning kittens.

I've come to believe that some people just don't have the capacity to detect/notice/care about audio faults and degradation. I'm kinda jealous because they can store 5 times the amount of music on their devices since it can be encoded at stupidly low rates. They can also get by perfectly fine on $5 garbage can quality headphones and think they are getting perfect fidelity.

I'm in my mid-40's and played bass in a metal band most of my life, so my hearing is far from what anyone would call good. That being said, I'm the only one in my family that ever notices and complains when audio plays which is loaded with artifacts (it's really bad on Sling TV for instance). For me, Opus tends to get near transparency around 96kbps. Around 112kbps it gets to be essentially impossible for me to ABX. I know there are people who can, but that is about the level my ears and brain tap out.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-26 17:16:25
Anyway, the track is full of artifacts which makes me wonder what age you are and if you have suffered any ear damage. I promise, I am not at all being rude but I felt I had to ask since the quality is bad and you say it is transparent to you. Regards.

I winced when I saw him claim transparency at 48kbps as well because to me just about any full bandwidth music is obviously degraded at that bitrate. But, I can play 48kbps mp3 files for my wife and kids and they just can't hear the artifacts. I don't know what it is but they swear the files sound just fine to them. To me they sound like someone drowning kittens.

I've come to believe that some people just don't have the capacity to detect/notice/care about audio faults and degradation. I'm kinda jealous because they can store 5 times the amount of music on their devices since it can be encoded at stupidly low rates. They can also get by perfectly fine on $5 garbage can quality headphones and think they are getting perfect fidelity.

I'm in my mid-40's and played bass in a metal band most of my life, so my hearing is far from what anyone would call good. That being said, I'm the only one in my family that ever notices and complains when audio plays which is loaded with artifacts (it's really bad on Sling TV for instance). For me, Opus tends to get near transparency around 96kbps. Around 112kbps it gets to be essentially impossible for me to ABX. I know there are people who can, but that is about the level my ears and brain tap out.


i have abx'ed Opus at different bitrates and i could not find a difference between 48 kbps and 64kbps, 80kbps, 96kbps, 112 kbps, and 128kbps for stereo music. i have done tests using quality Sennheiser headphones and mid range speakers.

 I do care about audio quality, and i have pretty good hearing.
Title: Re: Opus 1.3-beta is here
Post by: yourlord on 2017-12-26 18:48:31
I don't mean to imply you don't care. I'm just saying that you just may not be able to hear the difference for whatever reason. My wife can listen to files that sound like a garbled mess to me and not be able to distinguish it from the CD. It's not her ears because she can hear a gnat fart in the room next to us.  It just seems her brain masks out the artifacts or something. I'm not sure. I was just putting forth anecdotal evidence that some people legitimately can't hear or recognize the degradation. I'm not kidding when I say a 48kbps mp3 of full bandwidth music sounds like someone drowning kittens to me. It's almost painful to hear. But my wife and kids will play that though headphones with a usable frequency range of 3kHz and think it's CD quality.

Opus does do an amazing job at 48kbps though.
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2017-12-26 23:25:02
Well I believe that it matters the fact that people who can detect compression artifacts know what to expect when they get to hear compressed audio. Like your brain knows the pattern or better your brain knows where and on what the encoder will fail to do a transparent job. Most of the times its frequency balding, weird cymbals, and kinda grainy and crunchy "esses" on codecs with SBR. To say it better, it's like one of those illusions images that you don't quite get what you see or depending how you see it you see a different picture. Once you know what you are seeing the magic is broken. It's not a black box to you anymore.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-27 04:26:04
@king2101

Quote
i have abx'ed Opus at different bitrates and i could not find a difference between 48 kbps and 64kbps, 80kbps, 96kbps, 112 kbps, and 128kbps for stereo music. i have done tests using quality Sennheiser headphones and mid range speakers.

 I do care about audio quality, and i have pretty good hearing.

I am 38 years old and without spending too much time on ABX stuff, like doing what i call a 'quick ABX' test on some random songs, I found that once I hit 64kbps with Opus I can't easily notice it anymore, but i can notice the sound quality hit at 48kbps. like going from 48 to 64 to 80 and so on (basically increasing it 16kbps each bump). this was done on my computer speakers which are Klipsch Pro-Media speakers which I would say are better than your typical computer speakers for sure. I have had these since the early 2000's. I have not really tested headphones yet so who knows what would happen with those but I suspect it would be similar results, maybe a bit higher on the bit rate but i would GUESS probably not over 80kbps or so. but the headphones I got (Sony MDR-NC7) are nothing special but they are not total crap either.

but from what I have heard around these forums, Opus @ 80kbps is similar to Apple AAC @ 96kbps in overall sound quality.

still, I just use 128kbps (Apple AAC @ q64 TVBR) to play it safe as at that bit rate it's high enough to where no one will complain about sound quality and still low enough for efficient use of disc space.

p.s. i consider myself to have normal hearing as i think i can at least claim that my hearing is no worse than average. like i doubt i have any hearing loss outside of what's typical from the natural aging process as it appears there are certain frequencies that someone in my age bracket probably can't hear that say a teenager might be able to.

@yourlord ; Yeah, i am with you as i don't know how people can go down to 48kbps or so especially with MP3 as it's clearly flawed as i would have to assume they are capable of hearing a difference at that bit rate especially if they heard the lossless source file switched back and fourth between that and the 48kbps MP3. but even without that i am confident i would still be able to tell that it sounds like crap at that rate on MP3. but with Opus @ 48kbps, while i can tell the difference, i do agree that it sounds quite good considering the bit rate. hell even 32kbps on Opus is respectable considering the very low bit rate. Opus does great with speech at 14-15kbps which is mainly what i use Opus for as you can have a boatload of speech audio for barely any file size. i think one speech stuff i got was just shy of 4hours for about 25MB of storage space and the sound quality is not much worse than the much higher bit rate MP3 that the source file was. basically after some quick testing 14-15kbps is the lowest i like for for speech with Opus as while i can go lower the quality starts to really fizzle out even though it's still impressive considering the bit rate.

but basically when it comes to music... with Opus if it's at least 64kbps chances are i would not complain given my quick ABX testing with Opus/Apple AAC etc.
Title: Re: Opus 1.3-beta is here
Post by: Deathcrow on 2017-12-27 11:14:47
this was done on my computer speakers which are Klipsch Pro-Media speakers which I would say are better than your typical computer speakers for sure.

At 48kbps you don't even need decent speakers or headphones. The artifacts are easy to notice if you pay attention or take some time to find the difficult portions of a track. The nice thing with modern codecs like opus is just that they are rarely annoying. IMHO it depends a lot on the music too. I find it quite difficult in very loud music (metal and rock), where there's very little silence and lots of walls of sound, making it difficult to listen for ringing artifacts, etc.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-27 11:55:17
@Deathcrow

Quote
At 48kbps you don't even need decent speakers or headphones. The artifacts are easy to notice if you pay attention or take some time to find the difficult portions of a track.

Yeah, that's basically my point is that while it's fairly easy to spot the difference between FLAC and Opus @ 48kbps, once I hit 64kbps it's no longer easy for me. I just took some random tracks I got a while ago and that's how I came to that conclusion that in terms of a 16kbps boost in bit rate, 48kbps to 64kbps gives me the largest all around sound quality improvement for a small bit rate increase as it's pretty safe for me to say I won't see that kind of improvement for only 16kbps in any other thing like say 64kbps to 80kbps or say 32kbps to 48kbps. so basically if I don't want any more obvious sound quality hits I can use Opus as low as 64kbps, but I figure 80kbps is a bit safer. so I would say Opus @ 80kbps and Apple AAC @ 96kbps would be my suggested minimum if someone asked me what's the lowest bit rates you can use and not sacrifice the sound quality too much. although if your mainly concerned with file size and don't want any obvious sound quality drop offs then Opus @ 64kbps is what I would suggest.

because when I did my "quick ABX" test, even without finishing the tests(as I wanted to see the point where things were no longer easy for me to notice and spending not much time on ABXing to give me more of a real world situation where your not going to concentrate too hard), I just knew I was not going to fail the ABX @ 48kbps so I stopped. but once I hit 64kbps I lost that level of confidence in being able to easily detect the FLAC vs the Opus @ 64kbps. so I pretty much use 64kbps as a minimum suggestion for people wanting to use the Opus codec for music. but with Opus I would generally suggest using bit rates in the 80-128kbps range (basically one of four settings... 80/96/112/128) as I suspect this will make most people happy even though 64kbps is useable if you want maximum storage without any obvious sound quality drop offs.

but anyways, I don't even think you need all that much effort to ABX Opus @ 48kbps (like on random music tracks like Rock/Pop and the like) as it seems it does not take much at that bit rate for the overall sound to have a detectable difference from the FLAC file. but once I hit 64kbps it's no longer easy for me and I would not be surprised if I failed the ABX test on some typical Rock/Pop type of music. even if I can pass the ABX on Opus @ 64kbps it's going to require decent concentration for me which I figure at that point the music is pretty much transparent for me (but I would likely bump it to 80kbps to play it a bit safer if I wanted to use Opus at a lower bit rate).

Quote
The nice thing with modern codecs like opus is just that they are rarely annoying.

I pretty much agree as even at 32kbps the overall sound is pretty consistent I guess I could say.

p.s. Opus is less CPU intensive on my Sansa e250 (which I had for nearly 10 years now) than AAC-HE (and the like) as while both tax the CPU, Opus is useable (but navigating menu's are sluggish when playing Opus) unlike AAC-HE which pretty much freezes the device (too CPU intensive). but since typical AAC-LC works great on it and offers quality sound @ 128kbps there really ain't no reason for me to use Opus for music on that device outside of playing around with it occasionally.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-27 13:07:47
I used to encode all of my music into ogg vorbis at 64-80kbps, and I could not tell the difference between the original file.

But I switched to Opus @64kbps for encoding all of my music, and then lowered the bitrate to 48kbps after hearing no difference between the two bitrates.

At 32kbps though, there is definetly a difference, and that's the bitrate where I can hear the artifacts. Hopefully in the future, the engineers will be able to get the transparency level down to bitrates lower than 48kbps.

Just imagine having full bandwidth stereo music at only 12-32kbps with no artifacts, that would be insane!!
        
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-28 01:37:38
@king2101

I am 38 years old, yourlord is mid-40's... if you don't mind me asking how old are you?

I mainly ask because I wonder if why you seem to think Opus @ 48kbps is transparent is partially a age thing? ; but who knows, there could be other factors that others are probably more familiar with around here as maybe the age does not have much to do with it (assuming your not TOO old ;) ). but then again it seems like after a certain age I would guess through normal hearing loss as we age that eventually it's going to have a effect on what you can discern in sound quality tests etc. I am not exactly sure which age that would be but if i had to take a guess I would figure someone in their 60's (or older) would have noticeably worse hearing then say someone in their 30's or 40's and I am just assuming the only hearing loss they have is natural and not because of them being in noisy environments that damaged their hearing over the years.

I mean no offense or anything but I am a little surprised you don't notice 48kbps from the lossless source as, while I could be wrong, I would guess most people would be able to detect artifacts in Opus music without too much effort to at least 64kbps (assuming we stick to the 16kbps each increase standard, like... 32/48/64/80 and so on). but then again it's hard for me to say how many people even care about these things as I am aware that some people can deal with lower quality sound and be happy with it like yourlord basically mentioned in this topic. it does make me wonder if those people just don't care or do they really not notice it as it's hard for me to believe someone could literally not hear any difference, especially with MP3 @ 48kbps (and the like) that yourlord mentioned. but then again he did say 'wife and kids' and I suspect even though kids hearing is fine maybe their brains just ain't developed enough to notice or care about this stuff yet and I also would speculate that females on average care less about audio/video quality than us males do because I am willing to bet that some females I know would probably have a hard time telling SD from HD video (or even if they could, would probably not care all that much) even though I can clearly notice it (not a major difference but clearly detectable), especially when switching back and fourth between the two video files. but I will say that when it comes to 720p vs 1080p video on my 43" 1080p TV that I see no difference at all when testing the same movie which is why I typically stick to 720p to cut back on file size. NOTE: I am pretty sure my vision is not perfect (as in 20/20 vision) either(I don't wear glasses), but it's no where near bad.

just some thoughts ;)
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2017-12-28 01:52:14
Guys, nice discussion.  :)

But has someone tried this new 1.3 beta and compared it to 1.2.1 at some low bitrates?
Any findings?

My findings are that 1.3. has some improvements as well as some regressions as I posted before. It will be really helpful if someone will do (at least quick and dirty) tests on  a few samples or so and confirm these findings (or refute them).
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-28 08:52:15
@IgorC

This might be a bit of a noobie question... but how do you determine the number you rate the samples at? ; because I know 1 is horrible and 5 is transparent, but I am wondering how you (and others who posts these tests with the 1 through 5 scores) quantify this with a number.

because even if I downloaded the samples you posted on the first page of this topic and listened, and even assuming I could detect some differences, how does one determine the score? ; or would it be better off if I simply said, "I think Song A sounds better on Opus v1.3beta than v1.2.1" ? ; or lets say I can't notice any difference(or at least nothing I feel confident in claiming), I assume that I would be useless for confirming or denying your claim (hence, useless for helping the developer in the topic)?

I am just trying to get a feel for this stuff ;)

p.s. but then again I might be of some use since I mentioned the whole 48kbps to 64kbps thing where it's not difficult for me to detect flaws in the overall sound at 48kbps but once I hit 64kbps I no longer feel confident in claiming that. so since the test is at 32kbps, which is comfortably below 64kbps sound quality level, I might be possibly of some help here.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-28 12:21:28
@king2101

I am 38 years old, yourlord is mid-40's... if you don't mind me asking how old are you?

I mainly ask because I wonder if why you seem to think Opus @ 48kbps is transparent is partially a age thing? ; but who knows, there could be other factors that others are probably more familiar with around here as maybe the age does not have much to do with it (assuming your not TOO old ;) ). but then again it seems like after a certain age I would guess through normal hearing loss as we age that eventually it's going to have a effect on what you can discern in sound quality tests etc. I am not exactly sure which age that would be but if i had to take a guess I would figure someone in their 60's (or older) would have noticeably worse hearing then say someone in their 30's or 40's and I am just assuming the only hearing loss they have is natural and not because of them being in noisy environments that damaged their hearing over the years.

I mean no offense or anything but I am a little surprised you don't notice 48kbps from the lossless source as, while I could be wrong, I would guess most people would be able to detect artifacts in Opus music without too much effort to at least 64kbps (assuming we stick to the 16kbps each increase standard, like... 32/48/64/80 and so on). but then again it's hard for me to say how many people even care about these things as I am aware that some people can deal with lower quality sound and be happy with it like yourlord basically mentioned in this topic. it does make me wonder if those people just don't care or do they really not notice it as it's hard for me to believe someone could literally not hear any difference, especially with MP3 @ 48kbps (and the like) that yourlord mentioned. but then again he did say 'wife and kids' and I suspect even though kids hearing is fine maybe their brains just ain't developed enough to notice or care about this stuff yet and I also would speculate that females on average care less about audio/video quality than us males do because I am willing to bet that some females I know would probably have a hard time telling SD from HD video (or even if they could, would probably not care all that much) even though I can clearly notice it (not a major difference but clearly detectable), especially when switching back and fourth between the two video files. but I will say that when it comes to 720p vs 1080p video on my 43" 1080p TV that I see no difference at all when testing the same movie which is why I typically stick to 720p to cut back on file size. NOTE: I am pretty sure my vision is not perfect (as in 20/20 vision) either(I don't wear glasses), but it's no where near bad.

just some thoughts ;)
  

I just turned 30 years old. There is one strange thing about playing 48kbps files through a cell phone's internal speakers, it has these weird artifacts that are not present when playing them through headphones and regular speakers.

Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-28 13:00:42
How do these files sound to you? These are encoded at 48kbps

If they are not transparent,  I believe that they are at least comparable to 192-224 kbps mp3.

Even 64-80kbps vorbis had pretty great sound quality, and that is a way older codec.


Title: Re: Opus 1.3-beta is here
Post by: johnb on 2017-12-28 13:44:14
I believe you should not upload any song snippets longer than 30s to the forum for copyright reasons ...
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-28 14:09:51
How do these files sound to you? These are encoded at 48kbps [...] If they are not transparent,  I believe that they are at least comparable to 192-224 kbps mp3.
I am sorry king2101 but these are easily abxed and do in no way sound as good as mp3 around 192-224kbps. I feel bad for saying this but you are so wrong about this. Your tracks are full of artifacts but it is extremely impressive how well opus sounds at bitrate where almost no other codec can compete. But, at 48kbps it does not sound transparent.
Title: Re: Opus 1.3-beta is here
Post by: j7n on 2017-12-28 14:48:42
I find that "Water Runs Dry" and "Duncan's Toy Store" sound very objectionable, as expected with dynamic music with few instruments that can be isolated. Right at the start you should be able to hear what seems like a chorus of small metallic bells in both pieces and a dark room reverb, also the flute has acquired a digital breath from synthesized high frequencies.

"Escape From the City" and "NBA" are more difficult to analyze because they are loud and low level distortions are masked by other sounds. The start of "Ice Palace Car Chase" with the electronic drums also seems acceptable quality, but towards the end of the sample, we can hear the orchestra alone, and it suffers from the same problems as Duncan's Toy Store.

The amount of sharp treble in "Soul Makossa" gives a good first impression, yet there are artifacts at 1:10 and 3:09.

Maybe younger listeners are desensitized to artifacts because they hear transcoded music on the internet, digital radio and distortion used as an effect in popular music. Almost nobody leaves a critical comment when low bitrate transcodes get uploaded to the web.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2017-12-28 14:55:33
I am just about the same age as king2101s and heard an mp3 for the first time around 1996. I was so amazed at the sound quality and that it could almost fit onto a floppy disk. It didn't take long though before I started hearing things that were not sounding like it was part of the music. I do wish though that very low bitrates would be acceptable to me so that I could use only a tenth of the storage I use because I easily hear artifacts and to be on the safe side.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2017-12-28 21:37:42
It's the same artifact for all 3 samples (Korean speech, French speech, and Sample #4) where 1.3 does worse than 1.2.1. It's extra sibilance which possibly maybe described as "ringing", "paper sound" etc...

It repeats in a form of constant large periods during all time. It's not at a specific time.

This artifact is  present in L channel on the sample of Korean speech. Maybe it's stereo savings which cause this issue

Can you give this build a try: opus-tools-f3cff05.zip (https://jmvalin.ca/misc_stuff/opus-tools-f3cff05.zip)? It's somewhere between 1.2.1 and 1.3.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-29 07:19:31
@king2101

Quote
I just turned 30 years old. There is one strange thing about playing 48kbps files through a cell phone's internal speakers, it has these weird artifacts that are not present when playing them through headphones and regular speakers.

Okay, thanks for clearing this up as that explains things in why your saying 48kbps is transparent for you is because of the cellphone speaker handicapping your hearing abilities.

that makes sense because while I don't really have much experience with cell phone speakers I just kind of assumed they are below the typical decent set of headphones/PC speakers and your comment seems to further suggest this.

but with that said... when you quick ABX test (like I did without spending much effort) with a decent set of computer speakers or headphones at what rate (on Opus) do you need before it's no longer EASY for you to ABX? ; because like i was saying without spending much effort once I hit 64kbps (sticking with the 16kbps each increase standard (i.e. 32/48/64/80/96/112 and so on)) it's no longer easy for me to spot artifacts but at 48kbps I can still fairly easily spot them when playing random songs I got and this is based on my computer speakers which are Klipsch Pro-Media speakers (I had since early 2000's) which are better than your typical computer speakers.

Quote
I believe that they [Opus @ 48kbps] are at least comparable to 192-224 kbps mp3.

I am pretty sure many around here would disagree with this, including myself, based on this quote...

Quote
MP3-encoded files are generally considered artifact-free at bitrates at/above 192kbps.

so in other words... that pretty much means LAME v2 (i.e. 190kbps average) or higher is generally considered transparent with a lot of people and it's pretty safe for me to say Opus @ 48kbps is not at that standard of quality. so while Opus needs less bit rate to reach MP3 v2 (LAME v2) level of transparency, even with Opus, I would imagine that rate would be at least around 128kbps(?) to reach MP3 v2 level sound quality. but just on a personal level... I likely could use bit rates on both encoders at lower bit rates (i.e. with Opus lower quality than 128kbps and with MP3 lower quality than v2) and still be transparent for me, or at least largely.

but speaking of MP3, just doing those 'quick ABX' tests not all that long ago here is what I observed (which might vary a bit from person-to-person)...

-Opus/Apple AAC @ 64kbps average (NOTE: with these two encoders I was simply sticking to the default increase which was 16kbps between selectable options in Foobar2000 on encoder quality setting adjustments (i.e. 48/64/80/96 etc))

-MP3 @ v7(100kbps average) (NOTE: v8 (85kbps average) is the previous setting/point I could still fairly easily notice artifacts)

those two seem to be similar sound quality to me as that is the point where those encoders are no longer easy for me to ABX quickly. so this info tells me that Apple AAC/Opus are superior encoders which is why I generally avoid MP3 at this point in time.

NOTE: with MP3 (LAME) I simply started at a low rate (like how it's v9(lowest quality) to v0(highest quality)) and worked my way up til it was no longer easy for me to detect artifacts and the above is what conclusion I came up with. I am curious if you are similar to me or not? (I won't be surprised if you vary some though)

p.s. source for that quote above = http://wiki.hydrogenaud.io/index.php?title=Transparency
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-29 12:06:27
@king2101

Quote
I just turned 30 years old. There is one strange thing about playing 48kbps files through a cell phone's internal speakers, it has these weird artifacts that are not present when playing them through headphones and regular speakers.

Okay, thanks for clearing this up as that explains things in why your saying 48kbps is transparent for you is because of the cellphone speaker handicapping your hearing abilities.

that makes sense because while I don't really have much experience with cell phone speakers I just kind of assumed they are below the typical decent set of headphones/PC speakers and your comment seems to further suggest this.

but with that said... when you quick ABX test (like I did without spending much effort) with a decent set of computer speakers or headphones at what rate (on Opus) do you need before it's no longer EASY for you to ABX? ; because like i was saying without spending much effort once I hit 64kbps (sticking with the 16kbps each increase standard (i.e. 32/48/64/80/96/112 and so on)) it's no longer easy for me to spot artifacts but at 48kbps I can still fairly easily spot them when playing random songs I got and this is based on my computer speakers which are Klipsch Pro-Media speakers (I had since early 2000's) which are better than your typical computer speakers.

Quote
I believe that they [Opus @ 48kbps] are at least comparable to 192-224 kbps mp3.

I am pretty sure many around here would disagree with this, including myself, based on this quote...

Quote
MP3-encoded files are generally considered artifact-free at bitrates at/above 192kbps.

so in other words... that pretty much means LAME v2 (i.e. 190kbps average) or higher is generally considered transparent with a lot of people and it's pretty safe for me to say Opus @ 48kbps is not at that standard of quality. so while Opus needs less bit rate to reach MP3 v2 (LAME v2) level of transparency, even with Opus, I would imagine that rate would be at least around 128kbps(?) to reach MP3 v2 level sound quality. but just on a personal level... I likely could use bit rates on both encoders at lower bit rates (i.e. with Opus lower quality than 128kbps and with MP3 lower quality than v2) and still be transparent for me, or at least largely.

but speaking of MP3, just doing those 'quick ABX' tests not all that long ago here is what I observed (which might vary a bit from person-to-person)...

-Opus/Apple AAC @ 64kbps average (NOTE: with these two encoders I was simply sticking to the default increase which was 16kbps between selectable options in Foobar2000 on encoder quality setting adjustments (i.e. 48/64/80/96 etc))

-MP3 @ v7(100kbps average) (NOTE: v8 (85kbps average) is the previous setting/point I could still fairly easily notice artifacts)

those two seem to be similar sound quality to me as that is the point where those encoders are no longer easy for me to ABX quickly. so this info tells me that Apple AAC/Opus are superior encoders which is why I generally avoid MP3 at this point in time.

NOTE: with MP3 (LAME) I simply started at a low rate (like how it's v9(lowest quality) to v0(highest quality)) and worked my way up til it was no longer easy for me to detect artifacts and the above is what conclusion I came up with. I am curious if you are similar to me or not? (I won't be surprised if you vary some though)

p.s. source for that quote above = http://wiki.hydrogenaud.io/index.php?title=Transparency
@king2101

Quote
I just turned 30 years old. There is one strange thing about playing 48kbps files through a cell phone's internal speakers, it has these weird artifacts that are not present when playing them through headphones and regular speakers.

Okay, thanks for clearing this up as that explains things in why your saying 48kbps is transparent for you is because of the cellphone speaker handicapping your hearing abilities.

that makes sense because while I don't really have much experience with cell phone speakers I just kind of assumed they are below the typical decent set of headphones/PC speakers and your comment seems to further suggest this.

but with that said... when you quick ABX test (like I did without spending much effort) with a decent set of computer speakers or headphones at what rate (on Opus) do you need before it's no longer EASY for you to ABX? ; because like i was saying without spending much effort once I hit 64kbps (sticking with the 16kbps each increase standard (i.e. 32/48/64/80/96/112 and so on)) it's no longer easy for me to spot artifacts but at 48kbps I can still fairly easily spot them when playing random songs I got and this is based on my computer speakers which are Klipsch Pro-Media speakers (I had since early 2000's) which are better than your typical computer speakers.

Quote
I believe that they [Opus @ 48kbps] are at least comparable to 192-224 kbps mp3.

I am pretty sure many around here would disagree with this, including myself, based on this quote...

Quote
MP3-encoded files are generally considered artifact-free at bitrates at/above 192kbps.

so in other words... that pretty much means LAME v2 (i.e. 190kbps average) or higher is generally considered transparent with a lot of people and it's pretty safe for me to say Opus @ 48kbps is not at that standard of quality. so while Opus needs less bit rate to reach MP3 v2 (LAME v2) level of transparency, even with Opus, I would imagine that rate would be at least around 128kbps(?) to reach MP3 v2 level sound quality. but just on a personal level... I likely could use bit rates on both encoders at lower bit rates (i.e. with Opus lower quality than 128kbps and with MP3 lower quality than v2) and still be transparent for me, or at least largely.

but speaking of MP3, just doing those 'quick ABX' tests not all that long ago here is what I observed (which might vary a bit from person-to-person)...

-Opus/Apple AAC @ 64kbps average (NOTE: with these two encoders I was simply sticking to the default increase which was 16kbps between selectable options in Foobar2000 on encoder quality setting adjustments (i.e. 48/64/80/96 etc))

-MP3 @ v7(100kbps average) (NOTE: v8 (85kbps average) is the previous setting/point I could still fairly easily notice artifacts)

those two seem to be similar sound quality to me as that is the point where those encoders are no longer easy for me to ABX quickly. so this info tells me that Apple AAC/Opus are superior encoders which is why I generally avoid MP3 at this point in time.

NOTE: with MP3 (LAME) I simply started at a low rate (like how it's v9(lowest quality) to v0(highest quality)) and worked my way up til it was no longer easy for me to detect artifacts and the above is what conclusion I came up with. I am curious if you are similar to me or not? (I won't be surprised if you vary some though)

p.s. source for that quote above = http://wiki.hydrogenaud.io/index.php?title=Transparency

I said that the 48Kbps files exhibit artifacts only on the cell phone's internal speaker only, not on headphones or speakers.

Now, with opus at 32kbps, I can definitely hear artifacts, both on a smartphone's internal speakers, as well as headphones and normal speakers.

Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-29 12:27:54


p.s. source for that quote above = http://wiki.hydrogenaud.io/index.php?title=Transparency
[/quote]
@king2101[/member

Quote
I just turned 30 years old. There is one strange thing about playing 48kbps files through a cell phone's internal speakers, it has these weird artifacts that are not present when playing them through headphones and regular speakers.

Okay, thanks for clearing this up as that explains things in why your saying 48kbps is transparent for you is because of the cellphone speaker handicapping your hearing abilities.

that makes sense because while I don't really have much experience with cell phone speakers I just kind of assumed they are below the typical decent set of headphones/PC speakers and your comment seems to further suggest this.

but with that said... when you quick ABX test (like I did without spending much effort) with a decent set of computer speakers or headphones at what rate (on Opus) do you need before it's no longer EASY for you to ABX? ; because like i was saying without spending much effort once I hit 64kbps (sticking with the 16kbps each increase standard (i.e. 32/48/64/80/96/112 and so on)) it's no longer easy for me to spot artifacts but at 48kbps I can still fairly easily spot them when playing random songs I got and this is based on my computer speakers which are Klipsch Pro-Media speakers (I had since early 2000's) which are better than your typical computer speakers.

Quote
I believe that they [Opus @ 48kbps] are at least comparable to 192-224 kbps mp3.

I am pretty sure many around here would disagree with this, including myself, based on this quote...

Quote
MP3-encoded files are generally considered artifact-free at bitrates at/above 192kbps.

so in other words... that pretty much means LAME v2 (i.e. 190kbps average) or higher is generally considered transparent with a lot of people and it's pretty safe for me to say Opus @ 48kbps is not at that standard of quality. so while Opus needs less bit rate to reach MP3 v2 (LAME v2) level of transparency, even with Opus, I would imagine that rate would be at least around 128kbps(?) to reach MP3 v2 level sound quality. but just on a personal level... I likely could use bit rates on both encoders at lower bit rates (i.e. with Opus lower quality than 128kbps and with MP3 lower quality than v2) and still be transparent for me, or at least largely.

but speaking of MP3, just doing those 'quick ABX' tests not all that long ago here is what I observed (which might vary a bit from person-to-person)...

-Opus/Apple AAC @ 64kbps average (NOTE: with these two encoders I was simply sticking to the default increase which was 16kbps between selectable options in Foobar2000 on encoder quality setting adjustments (i.e. 48/64/80/96 etc))

-MP3 @ v7(100kbps average) (NOTE: v8 (85kbps average) is the previous setting/point I could still fairly easily notice artifacts)

those two seem to be similar sound quality to me as that is the point where those encoders are no longer easy for me to ABX quickly. so this info tells me that Apple AAC/Opus are superior encoders which is why I generally avoid MP3 at this point in time.

NOTE: with MP3 (LAME) I simply started at a low rate (like how it's v9(lowest quality) to v0(highest quality)) and worked my way up til it was no longer easy for me to detect artifacts and the above is what conclusion I came up with. I am curious if you are similar to me or not? (I won't be surprised if you vary some though)

p.s. source for that quote above = http://wiki.hydrogenaud.io/index.php?title=Transparency


For now on I am going to encode my music in 64kbps Opus, instead of 48kbps.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2017-12-29 16:11:12
well, you guys were right!!! 48kbps is not transparent, even though i could not tell the difference for 2 years. I used to encode all of my music into 64kbps, but i wanted to save space, and fit more music on my hard drive and cell phone.

The strange thing is when i play the 48kbps files on my android box using my headphones (they are pretty decent), and on my phone (using my headphones), and on my tv, i could not tell hear the artifacts, but i can only hear the artifacts on my headphones  on my computer, and on my cell phone's internal speaker.

Psychoacoustics are crazy!!! why is it that i cant detect the artifacts on certain equipment?

I am going to have to redownload hundreds and thousands of songs (possibly millions) again because of my mistake of bringing down the bitrate to 48 kbps instead of leaving it at 64 kbps!!!
Title: Re: Opus 1.3-beta is here
Post by: Makaki on 2017-12-30 00:14:21
ehh.... what is opus 1.2.1 dirty

This was never answered. What is "Dirty", and how does it differ from the windows binary provided at the official website.
Title: Re: Opus 1.3-beta is here
Post by: soundping on 2017-12-30 01:44:11
I noticed at the end of Radioheads song Feral the bass would make the encoder jump up to 115kbps. The rest of the song 58-74kbps.

I was encoding at 64kbps VBR.
Quote
Can you give this build a try: opus-tools-f3cff05.zip (https://jmvalin.ca/misc_stuff/opus-tools-f3cff05.zip)? It's somewhere between 1.2.1 and 1.3.
Title: Re: Opus 1.3-beta is here
Post by: sveakul on 2017-12-30 05:31:13
ehh.... what is opus 1.2.1 dirty

This was never answered. What is "Dirty", and how does it differ from the windows binary provided at the official website.


I believe this is referring to builds like at Chocobo1's site (https://github.com/Chocobo1/opus-tools_win32-build/releases) where the compiler has made additions/changes to the original code, as in this case for his 2017-07-12 build of opus-tools:

    Remove requirement of SSE2 instructions in x86 builds
    Add support for encoding from flac to opus directly (opusenc)

When using that build for encoding, the word "dirty" is included in the encoder description frame of the resultant file.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-30 07:23:23
@king2101

Quote
I said that the 48Kbps files exhibit artifacts only on the cell phone's internal speaker only, not on headphones or speakers.

Now, with opus at 32kbps, I can definitely hear artifacts, both on a smartphone's internal speakers, as well as headphones and normal speakers.

Weird, I would have thought being those cellphone speakers are generally of lower quality (at least I would imagine they are(?)) than a decent set of headphones/PC speakers it would seem if you heard artifacts on a cell phone speaker they would be present on headphones/PC speakers to or if they are not present on the headphones/PC speakers that they would not be there on the cell phone speaker.

but apparently your saying differently. it's funny how that sometimes works out.

Quote
For now on I am going to encode my music in 64kbps Opus, instead of 48kbps.

I was just giving my opinion on things as you can use whatever bit rate you want.

but if your definitely sticking with Opus, I would consider using 80kbps if I were you simply because it gives you a bit of a buffer beyond 64kbps in terms of sound quality and is barely more file size.

also, they say Opus @ 80kbps is similar to Apple AAC @ 96kbps and when it comes to Apple AAC I would rather not use bit rates lower than 96kbps, especially for music you care about. hell, you also could consider this... assuming your using Opus format, for music that's more important to you use 96-128kbps and music less important 64-80kbps and possibly a 3rd option of 64kbps-ish for music that's not that much concern for you.

also, for lossy music with modern hard drive prices... it's definitely not worth skimping THAT much on bit rate as you can get 2TB hard drives online for $50 or so (brand new) (or even about $90-100 for a cheaper 4TB hard drive) which is a pretty cheap price for a hard drive and plenty of space for music.

also, while I could understand wanting a really low bit rate for a cell phone, it's simply not worth it for computer usage unless your quite limited on storage space etc for whatever reason. plus, while I realize it's convenient to keep pretty much your whole music collection on whatever device your using (as I basically do that on my Sansa e250 device), I would imagine a large portion of that music won't see use all that much use. so assuming you can't fit your entire collection say on your cell phone etc(if that's what your trying to do?), you could simply use a higher/safer bit rate and while less music will fit on there it will probably be worth it if you can still store many albums(or songs) on the device.


p.s. If you got some time... I would be curious if you have a similar conclusion as myself with that 64kbps Opus/Apple AAC being similar to MP3 @ v7 (100kbps average), or not? ; like start off at a low bit rate for both and work your way up until it's no longer easy to ABX Opus(or AAC) vs MP3. I am assuming you got FLAC files here that you make the Opus/AAC/MP3 files from since you obviously need to test the lossless vs lossy files. NOTE: you can pretty much just test Opus against MP3 and skip Apple AAC if you want as based on my 'quick ABX' testing Opus and AAC are similar to each other on similar bit rates down low unlike MP3 where you can clearly tell you need more bit rate to clean up the sound when your on those low bit rates. the lowest starting point with MP3 is v9(65kbps average) then it goes to v8(85kbps), v7(100kbps), v6(115kbps),v5(130kbps) and then starts to ramp up. but anyways, with Opus you can start off with 32kbps then go to 48/64/80/96 etc until you hit that point where it's no longer easy for you to spot artifacts and then you have the point with Opus and MP3 that would be similar in overall sound to you and that's how i came up with the Opus @ 64kbps vs MP3 @ v7(100kbps) being similar. but it's possible you might feel differently, which is what I am a little curious about.

Quote
I am going to have to redownload hundreds and thousands of songs (possibly millions) again because of my mistake of bringing down the bitrate to 48 kbps instead of leaving it at 64 kbps!!!

Which is why the cardinal rule is to keep ALL of your important music in FLAC format in case you ever need to re-rip to a different lossy format etc and unless you got a boatload of music chances are it won't cost you too much $$$ to store it given hard drive prices are quite reasonable lately for a lot of storage space.

but personally I just like to use what I refer to as the 'sweet spot' and use Apple AAC @ q64 TVBR (128kbps average) and be done with it as AAC format (well I should say AAC-LC) is widely supported and since they have not updated sound quality in quite a few years now the format is mature and you likely won't need to re-rip and it's(128kbps) still quite efficient with storage space to, just not at the maximum. but if you ask me it's not worth lowering the 128kbps bit rate down to 96kbps or 64kbps given the sound is not as safe with a wider ranger of people and unless you got a ton of music your not going to save THAT much storage space going from 128kbps to 96kbps and while 128kbps down to 64kbps literally halves the storage space your going to take a decent hit in sound quality amongst a wider range of people at that point which makes the storage savings not worth it overall as while I like efficiency you can only take space saving so far before it's just not worth risking it on the sound quality and being storage space is cheap it makes those bit rates below 96-128kbps much less appealing.

------------------------------------------------

so to simplify things... just trust me, use 128kbps with Apple AAC or Opus and be done with it (unless storage space is at a premium) as those bit rates are high enough to where sound quality won't be a issue and yet are still small enough that they don't take up too much storage space. on Apple AAC I prefer TVBR over CVBR because apparently in sound quality tests around here there is no clear cut difference in sound quality between the two. so assuming that's true, then you might as well use TVBR due to they are generally smaller in size over CVBR.

if you want to roll the dice and live on the edge a bit... you might be able to go as low as 80kbps with Opus but I would suggest going no lower unless you don't care about the music all that much.

just some thoughts/suggestions ;)

p.s. I suspect you and others (hell, even myself), start to get a bit obsessed with tweaking bit rates etc. but I just say use 128kbps on Apple AAC (or Opus) and be done with it as those settings are efficient and I don't see too many people complaining about lack of sound quality either. one last thing... it should be pretty safe for me to say that most people, who care about storage space/efficiency, will want a bit rate with Opus or Apple AAC between 96-160kbps (with Opus you could probably include 80kbps). because those who don't care about efficiency and the like, and don't want to use a lossless audio format, can probably just use MP3 @ v0 or Apple AAC @ 256kbps as your pretty much guaranteed to have sound quality that will please pretty much everyone then.

NOTE: even on a poll I started a little while ago in these forums... currently out of the 39 total votes, 77% of those who voted preferred one of the following three settings as the 'sweet spot' for Apple AAC... 96/128/160 and out of those choices not far from half of the voters (i.e. 43.6%) choosing 128kbps, then 160kbps (23.1%) and the last of the three 96kbps (10.3%).
Title: Re: Opus 1.3-beta is here
Post by: imepv7 on 2017-12-30 09:17:53
Encoding at 20.0x max using Foobar 1.4 beta! If Im not mistaken it used to be around 100.0x. Much slower compared to FhG AAC @ 70.0x max.
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2017-12-30 13:44:41
I'm gonna fall into the category of most people here and say that I'm hitting pure transparency at (+-8kbps) 88kbps for Opus.
Impressive, especially considering where we were a decade/decade and a half ago.
Makes me wonder how things would have been if we had such polished encoders and codecs back then.
Like, how it would have been if MiniDiscs, and other portable music players had the ability to encode/decode audio of today's standars.
For example, how our world would have been if Opus would replace LP2 and LP4 modes of ATRAC3? At 132kbps and 66kbps we would have a completely different picture of their use.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2017-12-31 00:57:58
@Klimis

Quote
I'm gonna fall into the category of most people here and say that I'm hitting pure transparency at (+-8kbps) 88kbps for Opus.
Impressive, especially considering where we were a decade/decade and a half ago. Makes me wonder how things would have been if we had such polished encoders and codecs back then.

Yeah, that's the funny thing... like back in those days when storage space cost a arm and a leg (and even sound quality was somewhat so-so even at somewhat higher bit rates), which basically means we had limited storage space, we really needed these modern finely tuned encoders that have quality sound at low bit rates. but now the encoders are great and even pretty good at low bit rates storage space is much cheaper. so, while it's still nice to have these finely tuned encoders, they are not as important now to the common person. but then again, for the sake of efficiency it's always nice to lower the bit rate as much as possible, while retaining transparency (or close to that level), just to see the limits of sound encoder technology and Opus continues that trend.

but about the bit rate... while I never thoroughly tested Opus I would imagine I am close to that (i.e. 80kbps-96kbps) either way given my general quick testing that once I hit 64kbps it's no longer easy for me to ABX with confidence and then I really have to start paying attention etc based on Klipsch Pro-Media PC speakers (which are the best speakers I have). but speaking of the paying attention and focus etc... I assume it's normal(?) to be able to ABX early on at a certain bit rate that's not easy but you can pretty much do it, but as you keep on testing random songs etc it almost seems like fatigue sets in and then you start losing confidence in your ability to hear the difference.

but one of these days (who knows, maybe within the next week or two) I might try ABXing Opus on a Windows 10 laptop with Foobar2000 with my Sony MDR-NC7 headphones just to see if I can do any better than my main PC with the Klipsch Pro-Media speakers because all of my past ABX tests were on the Klipsch Pro-Media speakers.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2017-12-31 18:24:47
Can you give this build a try: opus-tools-f3cff05.zip (https://jmvalin.ca/misc_stuff/opus-tools-f3cff05.zip)? It's somewhere between 1.2.1 and 1.3.
Nothing has changed. The issue repeats exactly in same way.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2017-12-31 18:35:50
This might be a bit of a noobie question... but how do you determine the number you rate the samples at? ; because I know 1 is horrible and 5 is transparent, but I am wondering how you (and others who posts these tests with the 1 through 5 scores) quantify this with a number.
It's simple for me. I rate it as  a reviewed product.  Something like 5 stars (transparent), 4(perceptible but  not annoying), 3 (slightly annoying, 3.5 (something in the middle of 3 and 4) ...  and so on. Well, you get the point.

because even if I downloaded the samples you posted on the first page of this topic and listened, and even assuming I could detect some differences, how does one determine the score? ; or would it be better off if I simply said, "I think Song A sounds better on Opus v1.3beta than v1.2.1" ? ; or lets say I can't notice any difference(or at least nothing I feel confident in claiming), I assume that I would be useless for confirming or denying your claim (hence, useless for helping the developer in the topic)?
ABC/HR handles it pretty well.  You can read about it here http://listening-test.coresv.net/

Well, there isn't much to explain (or too much actually maybe). It's about trying the things.

Encoding at 20.0x max using Foobar 1.4 beta! If Im not mistaken it used to be around 100.0x. Much slower compared to FhG AAC @ 70.0x max.
I got  a slight slow down from ~300x to ~250x on my 3632QM (i7, Ivy Bridge). Not an issue for me.
What CPU do You have?
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2017-12-31 20:37:01
Can you give this build a try: opus-tools-f3cff05.zip (https://jmvalin.ca/misc_stuff/opus-tools-f3cff05.zip)? It's somewhere between 1.2.1 and 1.3.
Nothing has changed. The issue repeats exactly in same way.
I assume you mean the same issues as 1.3. In that case, can you give opus-tools-95c4871.zip (https://jmvalin.ca/misc_stuff/opus-tools-95c4871.zip) a try?
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2017-12-31 23:17:12
I assume you mean the same issues as 1.3.
Yes.

Now there are some audible interesting differences as for  opus-tools-95c4871.zip (I call it "31_12_2017")


Big picture is  1.2.1>95c4871>>1.3b1 on two speech samples (Korean and Sample #4)

1.Korean speech: 1.2.1 is  the best of all.  Good news, a nasty mosquito ringing is gone on 95c4871 but appears some amplitude variations.
2.Sample#4: 1.2.1 and 95c4871 are practically on par. Good.

 
 

Code: [Select]
ABC/HR Version 1.1 beta 2, 18 June 2004
Testname:

1L = E:\DETODO\Audio\Opus 1.3b1\04 Speech and Mix\04 Speech (He chased) Opus 1.3b1 32 kbps.wav
2R = E:\DETODO\Audio\Opus 1.3b1\04 Speech and Mix\04 Speech (He chased) Opus 1.2.1 32 kbps.wav
3R = E:\DETODO\Audio\Opus 1.3b1\04 Speech and Mix\04 Speech (He chased) Opus 1.3b1 32 kbps 31_12_2017.wav

---------------------------------------
General Comments:

---------------------------------------
1L File: E:\DETODO\Audio\Opus 1.3b1\04 Speech and Mix\04 Speech (He chased) Opus 1.3b1 32 kbps.wav
1L Rating: 2.8
1L Comment:
---------------------------------------
2R File: E:\DETODO\Audio\Opus 1.3b1\04 Speech and Mix\04 Speech (He chased) Opus 1.2.1 32 kbps.wav
2R Rating: 4.4
2R Comment:
---------------------------------------
3R File: E:\DETODO\Audio\Opus 1.3b1\04 Speech and Mix\04 Speech (He chased) Opus 1.3b1 32 kbps 31_12_2017.wav
3R Rating: 4.3
3R Comment:
---------------------------------------
ABX Results:

Code: [Select]
ABC/HR Version 1.1 beta 2, 18 June 2004
Testname:

1L = E:\DETODO\Audio\Opus 1.3b1\07 Korean speech\Greensleeves-Korean-male-speech Opus 1.2.1 32 kbps.wav
2R = E:\DETODO\Audio\Opus 1.3b1\07 Korean speech\Greensleeves-Korean-male-speech Opus 1.3b1 32 kbps 31_12_2017.wav
3L = E:\DETODO\Audio\Opus 1.3b1\07 Korean speech\Greensleeves-Korean-male-speech Opus 1.3b1 32 kbps.wav

---------------------------------------
General Comments:

---------------------------------------
1L File: E:\DETODO\Audio\Opus 1.3b1\07 Korean speech\Greensleeves-Korean-male-speech Opus 1.2.1 32 kbps.wav
1L Rating: 4.5
1L Comment:
---------------------------------------
2R File: E:\DETODO\Audio\Opus 1.3b1\07 Korean speech\Greensleeves-Korean-male-speech Opus 1.3b1 32 kbps 31_12_2017.wav
2R Rating: 3.5
2R Comment: Variable amplitude.
---------------------------------------
3L File: E:\DETODO\Audio\Opus 1.3b1\07 Korean speech\Greensleeves-Korean-male-speech Opus 1.3b1 32 kbps.wav
3L Rating: 3.0
3L Comment:
---------------------------------------
ABX Results:



Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-01 08:41:37
@IgorC

Quote
It's simple for me. I rate it as  a reviewed product.  Something like 5 stars (transparent), 4(perceptible but  not annoying), 3 (slightly annoying, 3.5 (something in the middle of 3 and 4) ...  and so on. Well, you get the point.

Hell, I thought you had some complicated math formula or something. lol

with that said, just to clear it up you just use .5 ratings? ; like 1 to 1.5 to 2 to 2.5 to 3 to 3.5 to 4 to 4.5 to 5 and then simply try your best to select things based on that general scale.

p.s. Happy New Year ;) ; It's about 4 hours into the new year here.

Quote
ABC/HR handles it pretty well.  You can read about it here http://listening-test.coresv.net/
Well, there isn't much to explain (or too much actually maybe). It's about trying the things.

Thanks for the info.

but I noticed on that site it says, "Headphones are a must-have". so when I do a test ill probably just use the older laptop I got recently for $20 (i.e. HP DV5-1002NR) instead of my main PC since it's easier to connect my headphones (Sony MDR-MC7) to it.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2018-01-01 15:49:39
what i found out about comparing bitrates with opus (48kbps vs 64kbps) is that violins and other classical instruments suffer at 48 kbps (they sound perfectly fine at 64kbps, however), while most other genre's of music sound somewhat similar at both bitrates.

But for now on all of music is going to be encoded in bitrates 64 kbps and up.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-02 08:01:32
@king2101

I have to correct some comments I made in here as I found the text file I made with the 64kbps vs MP3 v7(100kbps) being roughly equivalent. but the thing is that 64kbps setting was NOT Opus, it was Apple AAC(AAC-LC). sorry about that!

here is a straight quote from the text file I made late last year...

Quote
Some quick ABX testing (never finished actual test but i just based the following on my confidence level in hearing a obvious difference without spending a bunch of time on it) on Steppenwolf 'Desperation' song (i.e. FLAC vs MP3/AAC songs in ABX test)...

Apple AAC seems to hit a sweet spot for me when i want bare minimum bit rates on 64kbps as the next lowest setting of 56kbps i can obviously detect sound quality drop off with ABX testing in Foobar2000.

LAME MP3 equivalent to the above on that same song would be LAME v7(100kbps average)(not easy to detect) vs v8(85kbps average)(easy to detect).

bottom line... based on the info above LAME v7(100kbps average) is roughly equivalent to Apple AAC @ q27(64kbps average).

p.s. test was ran on my PC's Klipsch Pro-Media speakers in about Sep 2017.

but with that out of the way...

after some brief testing with Opus v1.2.1 with that same song (i.e. Steppenwolf - Desperation) I don't even feel confident in saying I noticed anything obvious between 48kbps(51kbps actual song bit rate) and 64kbps(68kbps actual song bit rate). but when dropped to 32kbps(34kbps actual song bit rate) I can definitely notice that.

I was even trying the Laptop with headphones (Sony MDR-NC7) and, at least based on that song with some fairly quick testing, I don't even think I can hear a difference with Opus v1.2.1 between 48kbps and 64kbps.

so basically... Opus is apparently superior to Apple AAC (AAC-LC) based on my quick testing here on that single song at these lower bit rates and it appears I am pretty much similar to King2101 in that I can't notice any obvious differences ('maybe' barely any differences(?)) between Opus @ 48kbps to 64kbps, at least on that single song I tested. but I feel confident I can notice it when doing some real quick ABX tests @ 32kbps (actual song bit rate is 34kbps). so assuming that much is true, at least based on the particular song I tested, my basic transparent point on it must be a bit rate setting somewhere between 32kbps to 48kbps. this clearly tops Apple AAC(AAC-LC) for me where that same song takes the 64kbps setting before I no longer feel confident in being able to ABX it.

I even re-ripped the song a moment ago to 56kbps (Apple AAC-LC) and still feel confident I can easily detect a difference as it's more muffled sounding which is why it's easy for me to detect as that is no longer there when using the 64kbps setting. the bit rate of that 56kbps setting(easy for me to detect) song is 73kbps in Apple AAC. when ripped to 64kbps setting (no longer easy for me to detect) the actual song bit rate is 83kbps. this was all done on my Klipsch Pro-Media PC speakers.

so based on that info above, on the song Desperation by Steppenwolf, Opus is pretty much transparent for me at setting no higher than about 48kbps(51kbps actual song bit rate), maybe even lower, where as with the Apple AAC(AAC-LC) it takes 64kbps (83kbps actual song bit rate) to get to basically the same point.

does anyone feel similar to myself with Opus v1.2.1 vs Apple AAC (AAC-LC)?
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2018-01-02 15:10:17
@king2101

I have to correct some comments I made in here as I found the text file I made with the 64kbps vs MP3 v7(100kbps) being roughly equivalent. but the thing is that 64kbps setting was NOT Opus, it was Apple AAC(AAC-LC). sorry about that!

here is a straight quote from the text file I made late last year...

Quote
Some quick ABX testing (never finished actual test but i just based the following on my confidence level in hearing a obvious difference without spending a bunch of time on it) on Steppenwolf 'Desperation' song (i.e. FLAC vs MP3/AAC songs in ABX test)...

Apple AAC seems to hit a sweet spot for me when i want bare minimum bit rates on 64kbps as the next lowest setting of 56kbps i can obviously detect sound quality drop off with ABX testing in Foobar2000.

LAME MP3 equivalent to the above on that same song would be LAME v7(100kbps average)(not easy to detect) vs v8(85kbps average)(easy to detect).

bottom line... based on the info above LAME v7(100kbps average) is roughly equivalent to Apple AAC @ q27(64kbps average).

p.s. test was ran on my PC's Klipsch Pro-Media speakers in about Sep 2017.

but with that out of the way...

after some brief testing with Opus v1.2.1 with that same song (i.e. Steppenwolf - Desperation) I don't even feel confident in saying I noticed anything obvious between 48kbps(51kbps actual song bit rate) and 64kbps(68kbps actual song bit rate). but when dropped to 32kbps(34kbps actual song bit rate) I can definitely notice that.

I was even trying the Laptop with headphones (Sony MDR-NC7) and, at least based on that song with some fairly quick testing, I don't even think I can hear a difference with Opus v1.2.1 between 48kbps and 64kbps.

so basically... Opus is apparently superior to Apple AAC (AAC-LC) based on my quick testing here on that single song at these lower bit rates and it appears I am pretty much similar to King2101 in that I can't notice any obvious differences ('maybe' barely any differences(?)) between Opus @ 48kbps to 64kbps, at least on that single song I tested. but I feel confident I can notice it when doing some real quick ABX tests @ 32kbps (actual song bit rate is 34kbps). so assuming that much is true, at least based on the particular song I tested, my basic transparent point on it must be a bit rate setting somewhere between 32kbps to 48kbps. this clearly tops Apple AAC(AAC-LC) for me where that same song takes the 64kbps setting before I no longer feel confident in being able to ABX it.

I even re-ripped the song a moment ago to 56kbps (Apple AAC-LC) and still feel confident I can easily detect a difference as it's more muffled sounding which is why it's easy for me to detect as that is no longer there when using the 64kbps setting. the bit rate of that 56kbps setting(easy for me to detect) song is 73kbps in Apple AAC. when ripped to 64kbps setting (no longer easy for me to detect) the actual song bit rate is 83kbps. this was all done on my Klipsch Pro-Media PC speakers.

so based on that info above, on the song Desperation by Steppenwolf, Opus is pretty much transparent for me at setting no higher than about 48kbps(51kbps actual song bit rate), maybe even lower, where as with the Apple AAC(AAC-LC) it takes 64kbps (83kbps actual song bit rate) to get to basically the same point.

does anyone feel similar to myself with Opus v1.2.1 vs Apple AAC (AAC-LC)?

I had no idea that AAC-LC had transparency at around 64-80kbps vbr! I thought it was only transparent at 96 kbps or higher!

You should do a test on a cell phone with a 48Kbps opus file and a 64kbps opus file playing through the internal speaker. like i mentioned before, the 48kbps file exhibits strange artifacts while being played back through the internal speaker, and is  not heard on headphones or other speakers, while, the 64kbps has no such artifacts


Title: Re: Opus 1.3-beta is here
Post by: Funkstar De Luxe on 2018-01-02 15:34:23
@king2101

I have to correct some comments I made in here as I found the text file I made with the 64kbps vs MP3 v7(100kbps) being roughly equivalent. but the thing is that 64kbps setting was NOT Opus, it was Apple AAC(AAC-LC). sorry about that!

here is a straight quote from the text file I made late last year...

Quote
Some quick ABX testing (never finished actual test but i just based the following on my confidence level in hearing a obvious difference without spending a bunch of time on it) on Steppenwolf 'Desperation' song (i.e. FLAC vs MP3/AAC songs in ABX test)...

Apple AAC seems to hit a sweet spot for me when i want bare minimum bit rates on 64kbps as the next lowest setting of 56kbps i can obviously detect sound quality drop off with ABX testing in Foobar2000.

LAME MP3 equivalent to the above on that same song would be LAME v7(100kbps average)(not easy to detect) vs v8(85kbps average)(easy to detect).

bottom line... based on the info above LAME v7(100kbps average) is roughly equivalent to Apple AAC @ q27(64kbps average).

p.s. test was ran on my PC's Klipsch Pro-Media speakers in about Sep 2017.

but with that out of the way...

after some brief testing with Opus v1.2.1 with that same song (i.e. Steppenwolf - Desperation) I don't even feel confident in saying I noticed anything obvious between 48kbps(51kbps actual song bit rate) and 64kbps(68kbps actual song bit rate). but when dropped to 32kbps(34kbps actual song bit rate) I can definitely notice that.

I was even trying the Laptop with headphones (Sony MDR-NC7) and, at least based on that song with some fairly quick testing, I don't even think I can hear a difference with Opus v1.2.1 between 48kbps and 64kbps.

so basically... Opus is apparently superior to Apple AAC (AAC-LC) based on my quick testing here on that single song at these lower bit rates and it appears I am pretty much similar to King2101 in that I can't notice any obvious differences ('maybe' barely any differences(?)) between Opus @ 48kbps to 64kbps, at least on that single song I tested. but I feel confident I can notice it when doing some real quick ABX tests @ 32kbps (actual song bit rate is 34kbps). so assuming that much is true, at least based on the particular song I tested, my basic transparent point on it must be a bit rate setting somewhere between 32kbps to 48kbps. this clearly tops Apple AAC(AAC-LC) for me where that same song takes the 64kbps setting before I no longer feel confident in being able to ABX it.

I even re-ripped the song a moment ago to 56kbps (Apple AAC-LC) and still feel confident I can easily detect a difference as it's more muffled sounding which is why it's easy for me to detect as that is no longer there when using the 64kbps setting. the bit rate of that 56kbps setting(easy for me to detect) song is 73kbps in Apple AAC. when ripped to 64kbps setting (no longer easy for me to detect) the actual song bit rate is 83kbps. this was all done on my Klipsch Pro-Media PC speakers.

so based on that info above, on the song Desperation by Steppenwolf, Opus is pretty much transparent for me at setting no higher than about 48kbps(51kbps actual song bit rate), maybe even lower, where as with the Apple AAC(AAC-LC) it takes 64kbps (83kbps actual song bit rate) to get to basically the same point.

does anyone feel similar to myself with Opus v1.2.1 vs Apple AAC (AAC-LC)?

I had no idea that AAC-LC had transparency at around 64-80kbps vbr! I thought it was only transparent at 96 kbps or higher!

You should do a test on a cell phone with a 48Kbps opus file and a 64kbps opus file playing through the internal speaker. like i mentioned before, the 48kbps file exhibits strange artifacts while being played back through the internal speaker, and is  not heard on headphones or other speakers, while, the 64kbps has no such artifacts




My Spidey sense tells me that there's something wrong with Android's implementation of Opus, because it does sounds drastically different through the speaker.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2018-01-02 18:51:35
@king2101

I have to correct some comments I made in here as I found the text file I made with the 64kbps vs MP3 v7(100kbps) being roughly equivalent. but the thing is that 64kbps setting was NOT Opus, it was Apple AAC(AAC-LC). sorry about that!

here is a straight quote from the text file I made late last year...

Quote
Some quick ABX testing (never finished actual test but i just based the following on my confidence level in hearing a obvious difference without spending a bunch of time on it) on Steppenwolf 'Desperation' song (i.e. FLAC vs MP3/AAC songs in ABX test)...

Apple AAC seems to hit a sweet spot for me when i want bare minimum bit rates on 64kbps as the next lowest setting of 56kbps i can obviously detect sound quality drop off with ABX testing in Foobar2000.

LAME MP3 equivalent to the above on that same song would be LAME v7(100kbps average)(not easy to detect) vs v8(85kbps average)(easy to detect).

bottom line... based on the info above LAME v7(100kbps average) is roughly equivalent to Apple AAC @ q27(64kbps average).

p.s. test was ran on my PC's Klipsch Pro-Media speakers in about Sep 2017.

but with that out of the way...

after some brief testing with Opus v1.2.1 with that same song (i.e. Steppenwolf - Desperation) I don't even feel confident in saying I noticed anything obvious between 48kbps(51kbps actual song bit rate) and 64kbps(68kbps actual song bit rate). but when dropped to 32kbps(34kbps actual song bit rate) I can definitely notice that.

I was even trying the Laptop with headphones (Sony MDR-NC7) and, at least based on that song with some fairly quick testing, I don't even think I can hear a difference with Opus v1.2.1 between 48kbps and 64kbps.

so basically... Opus is apparently superior to Apple AAC (AAC-LC) based on my quick testing here on that single song at these lower bit rates and it appears I am pretty much similar to King2101 in that I can't notice any obvious differences ('maybe' barely any differences(?)) between Opus @ 48kbps to 64kbps, at least on that single song I tested. but I feel confident I can notice it when doing some real quick ABX tests @ 32kbps (actual song bit rate is 34kbps). so assuming that much is true, at least based on the particular song I tested, my basic transparent point on it must be a bit rate setting somewhere between 32kbps to 48kbps. this clearly tops Apple AAC(AAC-LC) for me where that same song takes the 64kbps setting before I no longer feel confident in being able to ABX it.

I even re-ripped the song a moment ago to 56kbps (Apple AAC-LC) and still feel confident I can easily detect a difference as it's more muffled sounding which is why it's easy for me to detect as that is no longer there when using the 64kbps setting. the bit rate of that 56kbps setting(easy for me to detect) song is 73kbps in Apple AAC. when ripped to 64kbps setting (no longer easy for me to detect) the actual song bit rate is 83kbps. this was all done on my Klipsch Pro-Media PC speakers.

so based on that info above, on the song Desperation by Steppenwolf, Opus is pretty much transparent for me at setting no higher than about 48kbps(51kbps actual song bit rate), maybe even lower, where as with the Apple AAC(AAC-LC) it takes 64kbps (83kbps actual song bit rate) to get to basically the same point.

does anyone feel similar to myself with Opus v1.2.1 vs Apple AAC (AAC-LC)?

how does Opus at 48-64kbps sound on different speakers/headphones other than your PC speakers and  Sony MDR-NC7?
Title: Re: Opus 1.3-beta is here
Post by: lvqcl on 2018-01-02 18:58:09
There's no need to quote an entire post just to write a single line
Title: Re: Opus 1.3-beta is here
Post by: madmax on 2018-01-02 20:24:58
so... after all it's recommended to stick to v1.2.1 final. is that right?
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-03 00:04:37
@king2101

Quote
I had no idea that AAC-LC had transparency at around 64-80kbps vbr! I thought it was only transparent at 96 kbps or higher!

Well in general it seems Apple AAC @ 96kbps is the lowest people would suggest if you don't want to gamble and someone could easily disagree with me on that Apple AAC @ 64kbps claim, and I would expect them to, because I was simply doing quick ABX tests til I reached the point it was no longer easy for me to spot artifacts and keep in mind that was only one song I tested but I feel that should be a good ball park guideline for me when it comes to THE minimum bit rates I would consider using for Apple AAC.

p.s. but like my signature says, Apple AAC @ 128kbps seems to be the 'sweet spot' as, while I personally could go lower than that and probably not notice it, I just like to play it safe and given user tests around here at the 128kbps setting on Apple AAC your going to have solid sound quality with a wider range of people. sorta that balance of minimal file size but still strong sound quality. but your minimum suggestion of 96kbps (i.e. q45 TVBR) is still what I would use as a general fairly safe minimal guideline for most people.

so sorta like this with Apple AAC (AAC-LC)...

-256kbps = if your paranoid about sound quality and still want to use lossy audio and don't care about storage space or efficiency.
-128kbps = this is my general use suggestion if one has some concern with efficiency, but is not paranoid about sound quality. it's what I refer to as the 'sweet spot' of AAC.
-96kbps = my minimum suggested setting if one don't want to gamble with sound quality too much but storage space is of higher concern than sound quality.
-64kbps = THE minimum I would even consider using as anything lower than this sacrifices sound quality too much. this would be only for those who's primary concern is maximum songs on a device and but still don't want any fairly major hits to the sound quality.

Quote
You should do a test on a cell phone with a 48Kbps opus file and a 64kbps opus file playing through the internal speaker.

Ill just take your word for it as I don't have access to one I can test as, to be honest, I don't even own a cell phone. I know that's rare but I don't really use phones much in general.

Quote
how does Opus at 48-64kbps sound on different speakers/headphones other than your PC speakers and  Sony MDR-NC7?

Well the thing is I don't have much to test it on in general. basically just my main PC (with the Klipsch Pro-Media speakers) and the Laptop (HP DV5-1002NR) and I can use those Sony headphones on both. but I just used the headphones on the Laptop since it's easier. but I figure trying the headphones on the main PC would be mostly pointless(?) so I did not bother as it's somewhat of a pain to use the headphones on the main PC because of lack of cord length on the headphones.

@lvqcl

Quote
There's no need to quote an entire post just to write a single line

Exactly. I just do the @ symbol with the users name immediately next to it and then do "[.quote] text here [./quote]" (remove the "." and the ").

makes things a lot less clutter that way.
Title: Re: Opus 1.3-beta is here
Post by: eahm on 2018-01-03 02:18:16
ThaCrip, that suggestion list is useless if you don't say at what bitrate the codec is considered transparent IMO, in fact, that's all I want to know then I'll test myself. AAC = Ogg = Opus = transparent at around 160kbps.

Also suggesting 256 then the next one half the size it's a bit too much don't you think? What about 160? 192? Anyway, this is just my opinion, you have yours of course, you sound like a really nice guy actually and agree with AAC really good at ~96, in fact that what I use for mobile.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-03 08:41:20
@eahm

Quote
ThaCrip, that suggestion list is useless if you don't say at what bitrate the codec is considered transparent IMO, in fact, that's all I want to know then I'll test myself. AAC = Ogg = Opus = transparent at around 160kbps.

If your concerned with Transparency... http://wiki.hydrogenaud.io/index.php?title=Transparency ; or just to quote the important info...

Quote
Transparency at the lowest possible bitrate also seems to be used as a measure of the quality or degree of sophistication and tuning of a lossy compression algorithm:

    MP3-encoded files are generally considered artifact-free at bitrates at/above 192kbps.
    Vorbis ogg files are supposedly artifact-free at bitrates at/above 160kbps.
    AAC- and Opus-encoded files, depending on the particular encoder implementation, are claimed to be artifact-free at lower bitrates than both Vorbis ogg and MP3.

Which is basically inline what your 160kbps or so suggestion as beyond this bit rate things definitely lose efficiency.

but speaking of efficiency as I sorta commented on before, it seems most people will likely find one of the following three settings to be their sweet spot for Apple AAC (and probably Opus to)...

-96kbps (q27 TVBR) (more concerned about file size than sound quality)
-128kbps (q64 TVBR)  (my suggested default setting for most people)
-160kbps (q82 TVBR) (more concerned with sound quality than file size)

NOTE: I just stick to TVBR over CVBR simply because, given what I have read around here, there is no clear difference between the two from a sound quality point of view and TVBR generally uses less bit rate which makes it a bit more efficient than CVBR. hence, the better choice in my opinion.

NOTE #2: even on the poll I started (i.e. https://hydrogenaud.io/index.php/topic,114862.0.html ) nearly 8 out of 10 voters (or 77% (from 39 votes)) basically agree with the three basic bit rates above when it comes to the 'sweet spot' of Apple AAC.

Quote
Also suggesting 256 then the next one half the size it's a bit too much don't you think? What about 160? 192? Anyway, this is just my opinion, you have yours of course, you sound like a really nice guy actually and agree with AAC really good at ~96, in fact that what I use for mobile.

Yeah, I agree 256kbps is overkill but some people are paranoid and as I assume you already know Apple uses that setting. so that's why I just suggested this for those who are paranoid about sound and don't really care about encoder efficiency and the like.

so to correct(add to) my previous post above, it would now look something like this with Apple AAC (and Opus probably ain't much different(maybe say Opus @ 80kbps instead of AAC's 96kbps etc, but you get the gist :) ))...

-256kbps = if your paranoid about sound quality and still want to use lossy audio and don't care about storage space or efficiency.
-192kbps = almost no one would need to use a bit rate higher than this it seems.
-160kbps = seems to be the high side of what would be considered efficient use of bit rate and it's a very safe setting from a sound quality point of view as it will please a high percentage of people.
-128kbps = this is my general use/default suggestion if one has some concern with efficiency, but is not somewhat paranoid about sound quality. it's what I refer to as the 'sweet spot' of AAC as it keeps file size minimal without sacrificing sound much.
-96kbps = my minimum suggested setting if one don't want to gamble with sound quality too much but storage space is of higher concern than sound quality.
-64kbps = THE minimum I would even consider using as anything lower than this sacrifices sound quality too much. this would be only for those who's primary concern is maximum songs on a device and but still don't want any fairly major hits to the sound quality.

that should make things a bit better for you (and others) now I suspect ;)
Title: Re: Opus 1.3-beta is here
Post by: eahm on 2018-01-03 09:31:03
You didn't get my point at all.
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2018-01-03 12:19:45
I would not consider anything below 96kbps on AAC-LC near to transparent, the lowpass and the artifacting starts to hit alot. Maybe you mean AAC-HE, at higher bitrates (always below 96kbps) it's much more acceptable, maybe even transparent than encoding at the same bitrate on AAC-LC.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-03 12:41:25
@eahm

Quote
You didn't get my point at all.

I just simply responded to what was said and corrected my post based on what you said and then put in those six settings as sorta a general guideline for someone who don't want to think about this stuff too much. I try to keep things fairly simple.

@Klimis

Quote
I would not consider anything below 96kbps on AAC-LC near to transparent, the lowpass and the artifacting starts to hit alot.

I would not say 64kbps is much worse than 96kbps though but anything less I would start to consider bad as it's got more of a obvious muffled sound to the overall sound which is cleared up once I hit 64kbps from 56kbps.

but who knows, maybe with certain equipment even the 64kbps AAC-LC setting is obvious like how I can obviously notice the 56kbps setting(?). If you don't mind me asking, at what bit rate is it no longer easy for you to ABX a random song pretty quickly on some decent PC speakers etc?

but like I was saying, I would not suggest using below 96kbps for most people even though I feel you can use as low as 64kbps if your really hard up for storage space.

p.s. I am referring to AAC-LC.

Quote
Maybe you mean AAC-HE, at higher bitrates (always below 96kbps) it's much more acceptable, maybe even transparent than encoding at the same bitrate on AAC-LC.

When I talk Apple AAC I always mean AAC-LC until specified otherwise as AAC-HE, while clearly helps at lower bit rates, is too much of a performance hit especially on older hardware like Sansa e200 series etc as it goes from playing AAC-LC without issue to basically freezing the device (too much CPU) with AAC-HE.

plus, even those older devices aside... ain't once you go away from AAC-LC (like to AAC-HE and the like) then hardware support starts to get a bit more flaky?
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2018-01-03 18:43:49
I would not consider anything below 96kbps AAC-LC transparent. There's just too much frequency balding happening, it's a joke to even speak about it. It needs proper SBR after that thresshold. Well it's not like Opus doesn't do something similar at these bitrates, there's always so much you can do with MDCT based codecs, maybe if we move away from MDCT (ghost audio maybe?) you might be able to get greater band replication without the need of SBR but until then there's AAC-HE and Opus.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-04 02:04:59
@Klimis

Quote
I would not consider anything below 96kbps AAC-LC transparent.

I would pretty much agree with you here as it seems like that 96kbps is the minimum that most people would want to use without rolling-the-dice a bit given general tests etc I have read around here when looking at a wider range of people and wider range of music etc. so I am basically forced to agree with you on this much given that general info.

but with that said, ill say this for AAC-LC... in terms of overall sound I think 64kbps is much closer to 96kbps level of sound quality than it is 32kbps as I see 64kbps as what I would call the bottom-of-the-barrel of usable settings for music as any lower and your in the higher-tier of the garbage range and then rapidly declines from there.

I might even go as far as to claim, at least based on the single song I tested (with my Klipsch Pro-Media PC speakers), that there is more of a drop off in the overall sound quality with AAC-LC(Apple AAC) from 64kbps down to 56kbps than there is from 96kbps to 64kbps.

but who knows, maybe that's not a completely fair comparison because with the lower bit rates these bigger drop offs in overall sound quality are probably expected with little decreases in bit rate (i.e. 8kbps difference in setting) where as after you hit a certain point with bit rate the sound quality no longer has more obvious increases or drop offs when going 32kbps or so at a time with say 96kbps to 128kbps etc. but I guess I am not really saying anything new here either as it seems like with AAC-LC you start off at a low bit rate, say something less than 64kbps, and as you increase it gives rather large increases to sound quality as you bump up the bit rate a little bit but once you hit the 96kbps setting (maybe 128kbps-ish tops) the increases in overall sound quality start to decline a lot, and for those who can still hear artifacts etc, need to boost the bit rate a good amount to clean up some fairly minor stuff and seems to roughly top out around 160kbps setting before efficiency goes out the window going beyond 160kbps.

so in short... I basically agree with your general comment there but I just thought I would give my opinion on AAC-LC at the lower bit rates below 96kbps.

Ill stop babbling now ;)
Title: Re: Opus 1.3-beta is here
Post by: sPeziFisH on 2018-01-06 12:16:43
always funny if people want to encode music and don't want to sacrifice sound much  :D
The progress with the encoders is awesome and I really like the discussions. But I think, as storage space is cheap and small in handling size nowadays, it's mostly about streaming and temporarily use for music-OTG, so in masking environments.
I claim (!) that most of the longtime HA followers wouldn't encode with less than 160kbps-192kps if it's about 1-time-encoding for their lossy-archive..

I really hope that OPUS might bring a change and evolve into a new de-facto-standard as it is able to handle a wide spectrum of usecases while ensuring quality.
Title: Re: Opus 1.3-beta is here
Post by: ashtikuno on 2018-01-07 04:21:16
Good Day Sirs, I've just learned all about Opus and I'm planning to do a massive migration previously from AAC to Opus of about 2000 tracks from my library. Seeing that Opus 1.3 is coming, should I be waiting for 1.3 or the recent stable version a way to go?

Thanks in advance.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-07 08:24:01
@sPeziFisH

Quote
I claim (!) that most of the longtime HA followers wouldn't encode with less than 160kbps-192kps if it's about 1-time-encoding for their lossy-archive..

Personally I would not go more than those settings for a '1-time' encode especially given Apple AAC has not had any sound quality adjustments in years now (say roughly at the start of this decade, give or take) and it seems once you go beyond 160kbps (or so) efficiency really starts to go out the window with Apple AAC and beyond 192kbps seems like a bit overkill.

Plus, I feel that's even more of the case if your using that music on-the-go where listening conditions likely won't be all that optimal which makes it even that much less likely you will notice the small differences that may be there if your using say 128kbps instead of something like 160-192kbps and when just enjoying your music your probably much less likely to notice tiny little artifacts in the sound assuming you can detect these in the first place at the 128kbps setting.

p.s. I am not the 'longtime HA followers' group as I only had my account recently but I lurked around here from time to time as some years passed.

@ashtikuno

Quote
I've just learned all about Opus and I'm planning to do a massive migration previously from AAC to Opus of about 2000 tracks from my library. Seeing that Opus 1.3 is coming, should I be waiting for 1.3 or the recent stable version a way to go?

Honestly, if you already got AAC at a high enough bit rate it's probably not worth dumping that given it's wide support as while Opus seems to generally do better than AAC at lower bit rates, it seems once you reach a certain point Opus pretty much loses it's advantages over AAC. so I think for the most part if your going to use Opus it's more of the 128kbps or lower bit rates (generally lower is where you see it's benefits) is where it's best as if your going to use higher bit rates (say something like 128kbps+) your probably best off with AAC overall because of better support.

so if you want to use a lossy audio format 1-time and forget about it... Apple AAC @ 128kbs, maybe 160kbps if you would rather play it a bit safer.

but I guess we all have our preferences, just some thoughts ;)
Title: Re: Opus 1.3-beta is here
Post by: halb27 on 2018-01-07 08:55:41
Well, it's always personal taste where to put the sweet spot. But Opus @128 kbps has a higher probability of encoding even hard to encode tracks fine than AAC @128 kbps.For regular music usually everything is fine @128 kbps even with a good mp3 encoder.
Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-07 13:10:53
@halb27

Quote
Well, it's always personal taste where to put the sweet spot.

But given the info we have around here ain't there sorta some general guideline where to put that 'sweet spot'?

like while I know that technically it's personal preference, one would have to think there is at least a general area (i.e. say 96-160kbps for Apple AAC for example) where you will hit a pretty good standard of sound quality but while still having a efficient use of bit rate?

because say for example... a person claimed the 'sweet spot' for Apple AAC was like 256kbps. wouldn't it be easy to say they are wrong because of lack of efficiency at that bit rate for likely very little gains(or no gains(?)) in sound quality over bit rates a fair amount lower? ; so if that's true, it seems there has to be at least a ball park figure on the ranges of the sweet spot would be with Apple AAC(based on the info we got around here) before you either start to lose too much sound quality or use too much bit rate to clean up a really minor artifact and that's kinda why I see, based on the info I can find around these forums, that with Apple AAC claiming 'sweet spot' outside of the 96-160kbps range seems like someone could question it even though obviously someone could use whatever bit rate they want at the end of the day. but given the poll in here a little while ago nearly 8 out of 10 users (only 39 votes though) agreed the 'sweet spot' of Apple AAC was basically between 96-160kbps.

then you could apply that to Opus etc. but I would imagine Opus 'sweet spot' might vary a bit from Apple AAC as I would guess it would generally be lower bit rate ranges like say something like 80kbps-128kbps (or maybe a bit over 128kbps (or less than say 160kbps)) for the Opus equivalent to Apple AAC's 96-160kbps thing.

Quote
But Opus @128 kbps has a higher probability of encoding even hard to encode tracks fine than AAC @128 kbps.

Okay, so that was true then. because I thought I read a little on that, but I was not sure.

thanks for clearing that up.

so assuming Opus @ 128kbps can top Apple AAC @ 128kbps... I just wonder what bit rate would be required for Apple AAC to get on par with Opus in those situations? ; or is that simply too difficult to answer?

Quote
For regular music usually everything is fine @128 kbps even with a good mp3 encoder.

But ain't MP3 @ 128kbps generally considered more risky than Opus/AAC at that same bit rate across a wide range of music? ; but once you get to say something like LAME v2 (190kbps) (or maybe even v3 (175kbps)) then at that point it don't really matter whether you use Opus/AAC/MP3 since they are all basically top notch at that point, correct?
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-07 13:52:23
Guys, you are way off topic here. Please open a new topic and discuss comparison of different codecs there.
This topic is for development of  Opus 1.3 beta. Until now there isn't much help from you for Opus beta (actually no help at all).  So any suggestions or obsrervations for 1.3 beta?  

Title: Re: Opus 1.3-beta is here
Post by: Seren on 2018-01-07 19:31:17
So any suggestions or obsrervations for 1.3 beta?  

I haven't done any proper testing yet but have you discovered any other disparities between the original 1.3 beta and 95c4871?
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-07 20:06:00
Yes, see my previous posts in this topic.
1.3b and 95c4871 are no-go at 32 kbps.  1.2.1  hasn't perfect speech/music switching but  right now is the best (as for my ears)

1.2.1/1.3/"95c4871" at 32 kbps. (https://docs.google.com/spreadsheets/d/1vuf9TDaA6nOa2_mO0dmrAf-EacYInv9WiXw35ppRO4M/edit?usp=sharing)

Title: Re: Opus 1.3-beta is here
Post by: ThaCrip on 2018-01-09 08:34:38
@IgorC

Quote
Until now there isn't much help from you for Opus beta (actually no help at all).  So any suggestions or obsrervations for 1.3 beta?

Sorry about getting way off topic. Ill stop commenting on others posts on this topic with Apple AAC bit rates etc. but anyways...

as for the v1.3 beta... I don't think ill be of much help from what I can tell testing Opus v1.2.1 with 48-64kbps bit rate ranges because given the song I quickly tested (i.e. 'Desperation' by Steppenwolf) with Opus v1.2.1 @ 48kbps and 64kbps settings I could not say I noticed any obvious differences as I only notice obvious differences with Apple AAC (AAC-LC) when going to the setting just below the 64kbps (i.e. q27 TVBR) one which is q18 TVBR (i.e. 56kbps) and Opus at those ranges is still not obvious to detect artifacts for me (at least based on that single song I tested fairly quickly) as it does not have that muffled sound that's easy for me to detect on the AAC-LC @ 56kbps. so Opus at low bit rates must be clearly better than AAC-LC (Apple) at those low bit rates.

but being you tested Opus @ 32kbps, which I can tell sounds noticeably worse than even the 48kbps+ range, would I be correct in saying that any improvements the developer is attempting to make at those low rates would be difficult to detect between v1.2.1 and v1.3 beta? ; because if so, it makes me wonder how one improves things if someone don't have any pretty reliable information to provide the developer. or is there simply some people they know and trust to make these observations at those low bit rates etc?

but I did get that "Java ABC/HR" program from RareWares loaded onto the Windows 10 Laptop but recently I installed Mint Linux v18.3 Cinnamon x64 because maybe you have heard that recently after the Windows 10 update from 16299.125 to 16299.192 it stopped peoples PC's running some AMD CPU's from booting into Windows 10. but apparently Microsoft is aware of this issue and is working on a fix. but I might just leave Linux Mint on that 10 year old laptop I got as it seems to work well enough for what I use it for (as a bonus the touchpad fully works on Mint Linux where as even when Windows 10 was working the right side of the touchpad used to scroll web pages etc simply did not work on Windows 10 on my Laptop) as it's just my backup computer. but since that ABC/HR program works on Java I probably could get that working on Mint Linux I suspect assuming it plays the audio files fine etc (I am pretty much a Linux noobie).
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-10 01:11:55
but being you tested Opus @ 32kbps, which I can tell sounds noticeably worse than even the 48kbps+ range, would I be correct in saying that any improvements the developer is attempting to make at those low rates would be difficult to detect between v1.2.1 and v1.3 beta? ;
Not sure if I follow you but 1.2.1 was much better than 1.1.x at bitrates <= 48 kbps (and just was slightly better at higher bitrates).  Also it’s important to mention that 1.3’s scope is low bitrates (approx. ~24-48 kbps stereo). There probably won’t be much (if anything!)  for >48kbps range.

because if so, it makes me wonder how one improves things if someone don't have any pretty reliable information to provide the developer. or is there simply some people they know and trust to make these observations at those low bit rates etc?
That’s the idea. People submit their results, observations, reports etc. and devs decide whether it makes sense or not.
A good number of results from several listeners is better than any synthetic metric of audio quality (though some of metrics can be quite good) .

Anyway I have a feeling and I can be wrong but Opus 1.3 can be the last serious release. Audio codecs are pretty done today. AAC, MP3 and Vorbis are already overdeveloped and their developments are abandoned at this point. Add to it that we won't see newer generations formats (xHE-AAC, 3DA) in internet community anymore.  Development of video formats  is much more interesting ($) at this point. 
It's pretty close to good bye.
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2018-01-10 01:46:14
Anyway I have a feeling and I can be wrong but Opus 1.3 can be the last serious release. Audio codecs are pretty done today. AAC, MP3 and Vorbis are already overdeveloped and their developments are abandoned at this point. Add to it that we won't see newer generations formats (xHE-AAC, 3DA) in internet community anymore.  Development of video formats  is much more interesting ($) at this point. 
It's pretty close to good bye.


I hope that's not true. That would be a future I'd hate to live. Also considering the advancement in developing markets (say India) with people relying on lower bandwidths, anything that is of better quality is constantly giving the need to big companies (like Google) to push the boundaries further and further. It's not a coincidence they started supporting Opus natively at a period they started to give more attention to such markets. And India and Google is only one of the examples, you can find more companies drifting their focus on such markets with similar needs.

In short, as long there is a need for better quality with less bandwidth, there will always be advancements on the field. We may not live the days of yesteryear that the needs for better efficiency were dead obvious (like smaller capacity storage and portable media) but we still have the same need for a different reasons that came with a whole new market (relying alot on streaming media, bandwidth saving for servers etc).
Title: Re: Opus 1.3-beta is here
Post by: Phanton_13 on 2018-01-10 12:22:29
Actually Opus has a big limitation that can make its low bitrate quality improvement to stagnate beyon Opus 1.3, is that it is a low delay codec, but is very posible that a non low delay codec based in Opus can inprove the low bitrate quality furder.
Title: Re: Opus 1.3-beta is here
Post by: Nikaki on 2018-01-10 21:24:24
Actually Opus has a big limitation that can make its low bitrate quality improvement to stagnate beyon Opus 1.3, is that it is a low delay codec, but is very posible that a non low delay codec based in Opus can inprove the low bitrate quality furder.
Isn't the delay configurable at encoding time? opusenc uses a very high delay setting by default. Only realtime encoders (for voice communication) use a low delay setting. Right?
Title: Re: Opus 1.3-beta is here
Post by: Phanton_13 on 2018-01-11 00:18:11
Isn't the delay configurable at encoding time? opusenc uses a very high delay setting by default. Only realtime encoders (for voice communication) use a low delay setting. Right?

Yes for the first queation and no for the second question.

Opus can have a delay betwen 5ms to 66.5ms (there is a special case for 126.5ms delay), Opus by default have a delay of 26.5ms that is the optimal one as using bigger delays don't offer an advantage or it is minimal. Explaining the delay is basically lookahead+framesize, the lookahead is 2.5ms for a pure CELT mode or 6.5ms for the normal mode due to Silk, the framesize can be 2.5, 5, 10, 20, 40 and 60 ms but nativelly for the CELT layer the frames sizes are the ones from 2.5 to 20ms and as CELT Mode is the one used for music the bigger frame sizes dont offer practical advantage for music, and if I'm not wrong silk support nativelly the frames sizes from 10 to 60ms. The default delay is 6.5ms+20ms, basically bigest lookahead delay as it lets Opus use Silk. hybrid or CELT mode and 20 ms the biggest framesize for the celt layer.

Compare this to the delay for 48kHz in other codecs:

AAC-LC       +150ms
HE-AAC       +300ms
HE-AACv2   +400ms
AAC-LD        +20ms <- this codec sound like crap compared to Opus.
AAC-ELD      +20ms <- not much better than AAC-ELD

PD. If anyone know ther real world delays for the other codecs can they tell me those as the internet fonts are confusing.
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2018-01-11 15:32:49
So would an opus variant with higher frame sizes and longer look ahead make sense for applications where both don't matter or it seems like we have exploited already it's limits as a format?
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-11 17:34:23
So would an opus variant with higher frame sizes and longer look ahead make sense for applications where both don't matter or it seems like we have exploited already it's limits as a format?

Opus is -- right now -- able to exploit longer look-ahead for the purpose of speech/music detection and tonality analysis. Should it be useful for anything else, it can always be added with no change to the format. Larger frame sizes or window would require a change to the bit-stream and make a (slight) improvement to quality on highly tonal content (while probably making transients slightly worse), but the improvement is definitely not worth creating a new format. As for the limits of the format in general, it's hard to say. I doubt there's any particular large improvement to be made, but I still expect a series of small improvements for particular cases. Who knows what those will add up to.
Title: Re: Opus 1.3-beta is here
Post by: imepv7 on 2018-01-16 16:15:12
Encoding at 20.0x max using Foobar 1.4 beta! If Im not mistaken it used to be around 100.0x. Much slower compared to FhG AAC @ 70.0x max.

I got  a slight slow down from ~300x to ~250x on my 3632QM (i7, Ivy Bridge). Not an issue for me.
What CPU do You have?


I do have a very cheap AMD A8 Quad core CPU! Makes a lot of difference! Thanks for the input, though.
Title: Re: Opus 1.3-beta is here
Post by: Dynamic on 2018-01-19 09:53:18
I was just wondering, and perhaps @jmvalin would know, with the music/speech detection getting near-perfect as you mentioned in another thread a month or two ago, is it good enough that there would scope to:

A. Specify a low bitrate for speech-only and a higher bitrate for music (or mixed music and speech) to optimize better for podcasts or radio shows, say. (For example 16 or 24 kbps during speech-only is very good quality, while 40-64 kbps CELT would be good for theme tunes and incidental music (and perhaps even applause and certain other non-speech sounds, especially if in stereo), and perhaps 96 kbps for music would be better if music fidelity was highly important - e.g. a music documentary playing clips. Perhaps a preset -podcast-low would be (16 speech, 40 music), -podcast-medium would be (24 speech, 64 music), and -podcast-high would be (24 speech, 96 music)

B. Output a signal or a file indicating the switching points between speech and music. This is probably a niche use case, but I listen to a number of podcasts at around 2.0x speed in PodcastAddict, sometimes 2.4x if the people speak really slowly. In some cases they play a piece of music which really needs to be at 1.0x speed. I could imagine a gradual slow down as the music is about to start then a speed-up after the speech resumes might work.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-19 23:11:40
Yes, see my previous posts in this topic.
1.3b and 95c4871 are no-go at 32 kbps.  1.2.1  hasn't perfect speech/music switching but  right now is the best (as for my ears)
Note that the 95c4871 build is just some point early in the 1.3 development, before some of the speech-related changes. I'll get back to you later on this. In the mean time, I've come across some TF-related issues and made some builds for testing:
This is intended to improve on samples that mix tones and transients, e.g. the #41 intro. I'd be interested in feedback about how it improves and/or makes things worse.
Title: Re: Opus 1.3-beta is here
Post by: fabiorug on 2018-01-21 08:33:46
32 kbps opus-tools-newtf.zip:
High frequencies especially synth with (fart like sound, I don't know what sample it is) I noticed a good improvement compared to Opus 1.3 Beta:  the sound is less muddy, more clear.
But with this improvement I think  vocal lost a bit of presence  and detail (warmth).    (I tested with a Rock track, not acapella)..
Sure, an optimization could further improve the quality.
Good job.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-21 19:18:25
32 kbps opus-tools-newtf.zip:
High frequencies especially synth with (fart like sound, I don't know what sample it is) I noticed a good improvement compared to Opus 1.3 Beta:  the sound is less muddy, more clear.
But with this improvement I think  vocal lost a bit of presence  and detail (warmth).    (I tested with a Rock track, not acapella)..
Sure, an optimization could further improve the quality.
Can you provide the samples you're talking about here so I can listen to them? Both the one that's better and the one that's worse (ideally the original and the coded files to make sure everything's right).
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-21 21:28:45
Note that the 95c4871 build is just some point early in the 1.3 development, before some of the speech-related changes. I'll get back to you later on this.
Yes, I know. Previously I understood You correctly.

In the mean time, I've come across some TF-related issues and made some builds for testing: ...
Great. I'll try them later these days and will report.
Question. AFAIK your plans about 1.3 cover mainly low bitrates as You have already mentioned it here (https://hydrogenaud.io/index.php/topic,114360.msg943386.html#msg943386) 
So I will test mainly 32 kbps, stereo and maybe 48 kbps, all right?  Opus is already very good at 48 kbps and higher.

Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-22 00:44:19
Question. AFAIK your plans about 1.3 cover mainly low bitrates as You have already mentioned it here (https://hydrogenaud.io/index.php/topic,114360.msg943386.html#msg943386) 
So I will test mainly 32 kbps, stereo and maybe 48 kbps, all right?  Opus is already very good at 48 kbps and higher.
Most of the 1.3 changes are indeed for low bitrates, but the newtf stuff applies to all bitrates -- at least in theory. I'm not sure at what bitrate it would have the most impact though.
Title: Re: Opus 1.3-beta is here
Post by: Nott on 2018-01-22 04:57:32
I want to convert my ~2k files with AC3 5.1 and AC3 2.0, should I wait for opus 1.3? What settings do you guys recommend to opus 5.1 and opus 2.0 to achieve transparency ? File size isn't really a issue here, I just want to have a library fully open source
Title: Re: Opus 1.3-beta is here
Post by: maikmerten on 2018-01-22 07:18:00
I want to convert my ~2k files with AC3 5.1 and AC3 2.0, should I wait for opus 1.3? What settings do you guys recommend to opus 5.1 and opus 2.0 to achieve transparency ? File size isn't really a issue here, I just want to have a library fully open source

If file size isn't the problem: Keep the your files as-is, in AC3. You've got nothing to gain quality wise, only to lose. AC3 is so old it *may* already be "free" in terms of patents having expired or being close to expiring and there are open-source implementations.
Title: Re: Opus 1.3-beta is here
Post by: Nott on 2018-01-22 07:50:42
I want to convert my ~2k files with AC3 5.1 and AC3 2.0, should I wait for opus 1.3? What settings do you guys recommend to opus 5.1 and opus 2.0 to achieve transparency ? File size isn't really a issue here, I just want to have a library fully open source

If file size isn't the problem: Keep the your files as-is, in AC3. You've got nothing to gain quality wise, only to lose. AC3 is so old it *may* already be "free" in terms of patents having expired or being close to expiring and there are open-source implementations.

I still care about streaming my stuff, and AC3 isn't supported on most platforms I stream : browsers, phone.  Only my streaming boxes like chromecast/roku can accept AC3 but I think plex still transcode it lol...  and opus is "future proof"

And I thought there is a transparent quality where the quality will be identical?
Title: Re: Opus 1.3-beta is here
Post by: fabiorug on 2018-01-22 12:18:53
I
Title: Re: Opus 1.3-beta is here
Post by: maikmerten on 2018-01-22 20:38:52
I still care about streaming my stuff, and AC3 isn't supported on most platforms I stream : browsers, phone.  Only my streaming boxes like chromecast/roku can accept AC3 but I think plex still transcode it lol...  and opus is "future proof"

And I thought there is a transparent quality where the quality will be identical?

Making your content more compatible, yes, that's a good reason for a format change.

The problem with converting from one lossy format to another is that although each encoding step "itself" should be "transparent" given a proper bitrate, the errors build up - so there's no guarantee that after two "transparent" encoding rounds the end result would still be "transparent" compared to the original file. This is why it may be preferrable, provided you have the storage space, to keep the best quality version you have stored somewhere even if you convert to another format - just in case you may decide to use something different yet again in the future.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2018-01-22 22:41:50
I want to convert my ~2k files with AC3 5.1 and AC3 2.0, should I wait for opus 1.3? What settings do you guys recommend to opus 5.1 and opus 2.0 to achieve transparency ? File size isn't really a issue here, I just want to have a library fully open source

t To me 128 Kbps opus is Transparent for 5.1 tracks and 64 Kbps is transparent for 2.0 tracks
Title: Re: Opus 1.3-beta is here
Post by: maikmerten on 2018-01-23 16:35:02
To me 128 Kbps opus is Transparent for 5.1 tracks and 64 Kbps is transparent for 2.0 tracks

Opus is certainly doing a great job at these bitrates, but not universally transparent at such low bitrates IMO. It depends a lot on the content. I attached an example that should certainly be ABX-able at 64 kbps and show very obvious differences at even lower bitrates, just in case you want to find out what sort of artifacts to listen for.
Title: Re: Opus 1.3-beta is here
Post by: Trace on 2018-01-24 01:39:35
To me 128 Kbps opus is Transparent for 5.1 tracks and 64 Kbps is transparent for 2.0 tracks

Opus is certainly doing a great job at these bitrates, but not universally transparent at such low bitrates IMO. It depends a lot on the content. I attached an example that should certainly be ABX-able at 64 kbps and show very obvious differences at even lower bitrates, just in case you want to find out what sort of artifacts to listen for.

At what bitrate would you consider Opus universally transparent?

I went for 160 kbps on my whole library for that reason, didn't want to have to encode twice.
Title: Re: Opus 1.3-beta is here
Post by: maikmerten on 2018-01-24 07:25:45
To me 128 Kbps opus is Transparent for 5.1 tracks and 64 Kbps is transparent for 2.0 tracks

Opus is certainly doing a great job at these bitrates, but not universally transparent at such low bitrates IMO. It depends a lot on the content. I attached an example that should certainly be ABX-able at 64 kbps and show very obvious differences at even lower bitrates, just in case you want to find out what sort of artifacts to listen for.

At what bitrate would you consider Opus universally transparent?

I went for 160 kbps on my whole library for that reason, didn't want to have to encode twice.

Difficult question. It's "easy" to prove that something is not transparent with a given sample using ABX, but it may be impossible to say "at this bitrate things will always be transparent, for any sample, for every ear".

Having said that, at 160 kbps I'd feel pretty confident that quality should be "universally really good", with no nastyness to be expected. If you find a real-life example that exhibits audible artifacting at 160 kbps, I'd guess jmvalin may be interested to investigate.
Title: Re: Opus 1.3-beta is here
Post by: KozmoNaut on 2018-01-24 09:01:55
At what bitrate would you consider Opus universally transparent?

I went for 160 kbps on my whole library for that reason, didn't want to have to encode twice.

It's an individual assessment. For me it is probably fully transparent at 96kbps and certainly at 128kbps.

Meaning I've not been able to ABX any differences to lossless, even on tough problem samples, such as harpsichord and castanets. If there are differences, they're extremely subtle to my ears.

Then again, I have been accused of being a cloth-eared git by MQA and hi-res proponents, so take it with a grain of salt ;-)
Title: Re: Opus 1.3-beta is here
Post by: Klimis on 2018-01-25 01:01:19
In what way does OPUS encode multichannel audio?
I mean do all channels share somekind of similar data using somekind of new psychoacoustic model that is tuned to more-than-2-channels-joint-mode, or they are simply encoded in pairs of two (or something else)?
I'm trying to ABX multichannel audio and I have noticed that it does a way better job than any other codec by a bigger margin than stereo content results.
Title: Re: Opus 1.3-beta is here
Post by: kode54 on 2018-01-25 04:02:58
The format is limited to pairs of channels anyway, but as far as I know, the multi-channel setups involve strategically selected pairs for left/right, and whichever stream handles the LFE is tuned to only carry low frequency information, which allows it to be encoded with a lower amount of bandwidth.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-28 06:21:18
I've come across some TF-related issues and made some builds for testing:
  • opus-tools-oldtf.zip (https://jmvalin.ca/misc_stuff/opus-tools-oldtf.zip)
  • opus-tools-newtf.zip (https://jmvalin.ca/misc_stuff/opus-tools-newtf.zip)
This is intended to improve on samples that mix tones and transients, e.g. the #41 intro. I'd be interested in feedback about how it improves and/or makes things worse.

Results for 96 kbps (https://docs.google.com/spreadsheets/d/1K7hTuBldF1fcNmIRaM2gDVYB5oba2HMkKXfj5pNOZDI/edit#gid=0)

Files (https://drive.google.com/file/d/1qvSfNIcYzbUgWRA9-q9Sdts0tUYY5dHA/view?usp=sharing)

Only 5 samples because the differences are rather small and not in favor of a new TF.

It would be good if someone  would confirm/refute these findings or try it at different bitrate.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-28 07:24:19
Only 5 samples because the differences are rather small and not in favor of a new TF.
It would be good if someone  would confirm/refute these findings or try it at different bitrate.
Thanks. Any particular "pattern" you noticed in the differences? e.g. was it transients or tones getting slightly worse? in what way? Also, can you give it a try at lower bitrates (e.g. 48 kb/s)?
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-28 16:43:25
Main pattern is that the differences are microscopic, heh. So take my observations with a grain of salt.

For the first three samples new TF has transients issue. Not sure about what I heard on 'Angels fall' as it's a blend of transients and tonal parts. And the last sanmple had no difference at all.

As for 48 kbps no matter as hard as I try I can't catch any audible difference between new/old TF.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-01-28 23:58:24
e.g. was it transients or tones getting slightly worse? in what way?
1. 41_30 sample
maracas-like noise/distortion (hess?) is present on both new and old TF. But it's slightly more pronounced on a new TF.
Not sure if I interpeted it correctly as a transients issues early.

2. Linchpin
I can only say that new TF has more pronounced "temporal errors" on both guitars and drums. Not sure if it's just transients. 

New TF has more harsh artifacts like Vorbis with its colored-noise artifacts (on both 41_30 and Linchpin samples). I hope You can draw a picture.

P.S. All that at 96 kbps.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-30 08:11:08
Main pattern is that the differences are microscopic, heh. So take my observations with a grain of salt.

For the first three samples new TF has transients issue. Not sure about what I heard on 'Angels fall' as it's a blend of transients and tonal parts. And the last sanmple had no difference at all.
OK, could be a tuning issue. Can you give this updated one a try:
opus-tools-newtf2.zip (https://jmvalin.ca/misc_stuff/opus-tools-newtf2.zip)
Title: Re: Opus 1.3-beta is here
Post by: soundping on 2018-01-30 10:07:37
OK, could be a tuning issue. Can you give this updated one a try:
opus-tools-newtf2.zip (https://jmvalin.ca/misc_stuff/opus-tools-newtf2.zip)
I'm seeing the 'uper' channels getting overemphasized. (48 kb/s)
Title: Re: Opus 1.3-beta is here
Post by: meowchin on 2018-01-30 17:54:31
I suppose we won't be getting constant quality option anytime soon, therefore I have a question to ask: how does one calculate what bitrate to use in multichannel? I was able to determine that 160k is transparent for stereo for me, but how do I transfer this to multichannel sound?
In wiki it says:
Quote
In the case of the reference release, libopus, the target bitrate is calibrated against the internal constant quality targets so that over a typical music collection, something very close to the target bitrate will be achieved
Is there a way to look into the code to know what these internal constant quality targets are?
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-01-30 18:09:52
I suppose we won't be getting constant quality option anytime soon, therefore I have a question to ask: how does one calculate what bitrate to use in multichannel? I was able to determine that 160k is transparent for stereo for me, but how do I transfer this to multichannel sound?
You'll have to see for yourself what you consider transparent. Transparency varies a lot from one person to another. On top of that, I assume that the 160 kb/s you're quoting is based on headphones. Since surround is generally loudspeakers, it changes everything. So really just try it and see for yourself.

In wiki it says:
Quote
In the case of the reference release, libopus, the target bitrate is calibrated against the internal constant quality targets so that over a typical music collection, something very close to the target bitrate will be achieved
Is there a way to look into the code to know what these internal constant quality targets are?
The calibration here is just making sure that when asking for X kb/s, the encoder will indeed produce X kb/s over a large collection. There's no quality aspect involved in there.
Title: Re: Opus 1.3-beta is here
Post by: radorn on 2018-02-01 18:56:01
The calibration here is just making sure that when asking for X kb/s, the encoder will indeed produce X kb/s over a large collection. There's no quality aspect involved in there.

I guess this must have been discussed before in many an occasion, but I still must ask: What's the rationale for that approach?
Obtaining an average bitrate across a large varied collection doesn't seem like such an useful goal or metric to me, and neither has it ever worked for me. Most of the music I've ever encoded turned out above that average and only infrequently does it go slighty under it. It seems I don't have such a large and varied collection. I'm curious just what such a collection would look like...

I remember someone saying that the B argument was supposed to work the same way as the classic Q argument, but with a different scale. But now jmvalin seems to contradict that notion, which sends me back to square one.

I remember reading that the Opus team didn't want to use the term "quality" to avoid comparisons to other popular codecs' "quality" scales, so, maybe, it could be called something else: fidelity, transparency, crystalinity, happy-ear-factor, ...

I don't care about the name or the scale. I just care that I can put a number that gives me a reasonably consistent level of perceptual transparency for anything I throw at it and uses whatever bitrate it needs for that.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-01 19:17:24
The calibration here is just making sure that when asking for X kb/s, the encoder will indeed produce X kb/s over a large collection. There's no quality aspect involved in there.
I guess this must have been discussed before in many an occasion, but I still must ask: What's the rationale for that approach?
Obtaining an average bitrate across a large varied collection doesn't seem like such an useful goal or metric to me, and neither has it ever worked for me. Most of the music I've ever encoded turned out above that average and only infrequently does it go slighty under it. It seems I don't have such a large and varied collection. I'm curious just what such a collection would look like...
The important thing isn't the exact collection itself, it's having some calibration. Otherwise, I could simply "improve" Opus by just increasing the bitrate for the same command-line arguments. With the calibration, I ensure that any increase in bitrate for a particular file is compensated by a decrease in other files. The result is that different versions should give about the same overall bitrate on your collection, even if it differs from mine (and even if your average bitrate differs from mine).

I remember reading that the Opus team didn't want to use the term "quality" to avoid comparisons to other popular codecs' "quality" scales, so, maybe, it could be called something else: fidelity, transparency, crystalinity, happy-ear-factor, ...
Whatever you read was incorrect or misleading.

I don't care about the name or the scale. I just care that I can put a number that gives me a reasonably consistent level of perceptual transparency for anything I throw at it and uses whatever bitrate it needs for that.
You already have a scale, it's called "bitrate". Instead of having a 0-10 scale that maps to different bitrates (because Vorbis also calibrates its quality scale on an average bitrate), the scale is 0-510 you don't even have to look up what value you need to get a certain average bitrate.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-02 01:39:49
Can you give this updated one a try:
opus-tools-newtf2.zip (https://jmvalin.ca/misc_stuff/opus-tools-newtf2.zip)
Yes, I will try as soon as I'll get some spare time.

Sincerely, there isn't any real feedback from members who use Opus (there are approx. 50 of them here on forum according to some polls). 

Guys, somebody?
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-02 14:40:44
Can you give this updated one a try:
opus-tools-newtf2.zip (https://jmvalin.ca/misc_stuff/opus-tools-newtf2.zip)
Yes, I will try as soon as I'll get some spare time.

Sincerely, there isn't any real feedback from members who use Opus (there are approx. 50 of them here on forum according to some polls). 

Guys, somebody?
I rarely sit on Windows so if you want to enable more of us to give you opinions then start with supplying at least some Linux compiles, thanks.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-02 16:48:40
The OP has a link to source code of 1.3.beta. It's still not conclusive whether there are any improvements comparing to stable release.
You can try it and report here. 

Thank You in advance.

P.S. That build had at least one bug. You should use an external resampler to avoid it. But it's still usable to inform audible differences comparing to stable release.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-03 10:02:48
Easy one!
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 10:44:46

File A: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3.flac
SHA1: 16861a51329079cd36767bb7d2d0ed68c6ced1e7
Gain adjustment: -8.25 dB
File B: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3_48kbps_newtf2.opus
SHA1: a5517aefce61051808905233099d0758c70cb88a
Gain adjustment: -8.27 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

10:44:46 : Test started.
10:46:20 : 01/01
10:46:33 : 02/02
10:46:57 : 03/03
10:47:33 : 04/04
10:47:47 : 05/05
10:48:10 : 06/06
10:48:48 : 07/07
10:49:37 : 08/08
10:50:01 : 09/09
10:50:22 : 10/10
10:50:40 : 11/11
10:50:59 : 12/12
10:51:33 : 13/13
10:52:21 : 13/14
10:52:44 : 14/15
10:53:02 : 15/16
10:53:02 : Test finished.

 ----------
Total: 15/16
Probability that you were guessing: 0.0%

 -- signature --
3de5ef2eb819edccab659d7f77b55c9644f79771



Even easier than 48kbps when I now knew better what to listen for (open hi-hats).
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 11:04:22

File A: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3.flac
SHA1: 16861a51329079cd36767bb7d2d0ed68c6ced1e7
Gain adjustment: -8.25 dB
File B: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3_64kbps_newtf2.opus
SHA1: c949103b3c214fe1e9177c3b520d058d777ef275
Gain adjustment: -8.26 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

11:04:22 : Test started.
11:05:59 : 01/01
11:06:09 : 02/02
11:06:28 : 03/03
11:06:42 : 04/04
11:06:59 : 05/05
11:07:18 : 06/06
11:07:34 : 07/07
11:07:46 : 08/08
11:07:57 : 09/09
11:08:12 : 10/10
11:08:30 : 11/11
11:08:46 : 12/12
11:09:02 : 13/13
11:09:10 : 14/14
11:09:21 : 15/15
11:09:33 : 16/16
11:09:33 : Test finished.

 ----------
Total: 16/16
Probability that you were guessing: 0.0%

 -- signature --
2d7f3b019eaa4b9922114341eeb30700177928b2



This amount of bitrate started to make it harder to distinguish the two tracks but I quite quickly found a passage, again with hi-hats, that had a strange transitioning effect. During a whole measure it sounded like a lowpass being automated and gradually going down to a lower frequency during a few seconds. It was very subtle though.
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 11:16:35

File A: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3.flac
SHA1: 16861a51329079cd36767bb7d2d0ed68c6ced1e7
Gain adjustment: -8.25 dB
File B: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3_80kbps_newtf2.opus
SHA1: dbe4987e3aba164f4165e089e4b368ebb0264d22
Gain adjustment: -8.26 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

11:16:35 : Test started.
11:17:43 : 01/01
11:17:51 : 02/02
11:18:15 : 03/03
11:18:27 : 04/04
11:18:40 : 05/05
11:18:53 : 06/06
11:19:03 : 07/07
11:19:13 : 08/08
11:19:24 : 09/09
11:19:45 : 10/10
11:20:00 : 11/11
11:20:23 : 12/12
11:20:37 : 13/13
11:20:47 : 14/14
11:21:02 : 15/15
11:21:24 : 16/16
11:21:24 : Test finished.

 ----------
Total: 16/16
Probability that you were guessing: 0.0%

 -- signature --
c9972433884924455484fbd6b6ddb12ff1616e32
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-03 11:33:01
This one was very difficult. I could not in a short time identify obvious artifacts in the high frequencies so I started looking for "wobbling" sounds and I thought I found it at 02:13 in the bass frequencies. All I can describe it is that the lossless version had an extremely subtle "wooOO wooOO" effect and the opus 96kbps had one with an additional one creating "wooOO wooOO wooOO". I know, this is getting us into probably placebo effect but I did score only 10% guessing chance after all. I will try this bitrate again but this time look if I can find other artifacts that are a lot easier to spot like the earlier hi-hats. It is extremely difficult though using this song. The high frequencies are not so sharp and extended as other music so it is more diffficult to spot high frequency or transient artifacts.
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 12:09:55

File A: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3.flac
SHA1: 16861a51329079cd36767bb7d2d0ed68c6ced1e7
Gain adjustment: -8.25 dB
File B: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3_96kbps_newtf2.opus
SHA1: acd4ad2dfc44f925f4daf8ede9b49a993fb468ef
Gain adjustment: -8.26 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

12:09:55 : Test started.
12:14:18 : 00/01
12:14:47 : 01/02
12:15:11 : 02/03
12:15:32 : 03/04
12:15:55 : 03/05
12:16:17 : 04/06
12:16:33 : 05/07
12:16:51 : 06/08
12:17:04 : 07/09
12:17:19 : 07/10
12:17:47 : 07/11
12:18:03 : 08/12
12:18:38 : 09/13
12:18:57 : 10/14
12:19:22 : 10/15
12:19:53 : 11/16
12:19:53 : Test finished.

 ----------
Total: 11/16
Probability that you were guessing: 10.5%

 -- signature --
17592eacb0570e931df8c61324fce11074b2ac3d



This time I concentrated on the high frequencies again and it was easier even though I can feel that my ears are getting tired and I am overloading real soon. What I find easy to hear is that around 2:13 and about ten seconds onwards the cymbals, when they are hit together with the kick, sound like a compressor on the upper frequency attenuate not exactly on the hits but very short there after and quickly recovers to zero gain reduction. In the lossless file the cymbals hits have more force and aggression to them. 
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 12:33:27

File A: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3.flac
SHA1: 16861a51329079cd36767bb7d2d0ed68c6ced1e7
Gain adjustment: -8.25 dB
File B: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3_96kbps_newtf2.opus
SHA1: acd4ad2dfc44f925f4daf8ede9b49a993fb468ef
Gain adjustment: -8.26 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

12:33:27 : Test started.
12:35:22 : 01/01
12:35:57 : 02/02
12:36:08 : 03/03
12:36:28 : 04/04
12:36:40 : 05/05
12:37:00 : 06/06
12:37:22 : 07/07
12:37:36 : 08/08
12:37:57 : 09/09
12:38:17 : 10/10
12:38:32 : 11/11
12:39:17 : 12/12
12:40:04 : 12/13
12:40:38 : 13/14
12:41:10 : 14/15
12:41:41 : 15/16
12:41:41 : Test finished.

 ----------
Total: 15/16
Probability that you were guessing: 0.0%

 -- signature --
7fe818d80840ab1f332b990eed412b9de91c7826



I concentrated on the high frequencies again and followed the same pattern as in the last 96 kbps abx test. I didn't nail the test as I was only one successful identification away from it but this time I also heard the compressor effect on the upper frequencies. I have to say that this one demanded extreme concentration.
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 12:53:12

File A: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3.flac
SHA1: 16861a51329079cd36767bb7d2d0ed68c6ced1e7
Gain adjustment: -8.25 dB
File B: 02 Rich Kids On LSD - Beaultiful Feeling Pt. 3_112kbps_newtf2.opus
SHA1: 3ce4b5f4339a43085482271cab681bfe6950e59f
Gain adjustment: -8.25 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

12:53:12 : Test started.
12:54:36 : 01/01
12:54:59 : 02/02
12:55:40 : 03/03
12:56:05 : 03/04
12:56:41 : 04/05
12:57:26 : 05/06
12:57:49 : 05/07
12:58:32 : 06/08
12:58:50 : 07/09
12:59:15 : 07/10
12:59:38 : 08/11
13:00:16 : 08/12
13:00:45 : 09/13
13:01:30 : 10/14
13:02:08 : 11/15
13:02:38 : 11/16
13:02:38 : Test finished.

 ----------
Total: 11/16
Probability that you were guessing: 10.5%

 -- signature --
82ae98ccf348a140a7c032db793c12ac10805d29
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-03 12:13:55
I do not think I will do a 128 kbps abx test of the Rich Kids On LSD track since the 112 kbps version was so difficult. I am thinking of testing a voice recording with full frequency response but quite a lot of ambience and some recording noise. I encoded one, a mono file, a week ago and had to use extremely high bitrate to make some artifacts go away. To avoid any TOS I will not say what bitrate at this point in time.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-03 13:06:58
I am using headphones this time because I think my neightbours are tired enough of me playing loud music and if they would have to listen to a recording of me talking for an extended period of time they would probably freak out. Sounds like the high frequencies (the noise) are being tuned the first couple of seconds to later calm down. The noise contains more high frequency than the lossless file. Sounds like things are "tearing up".

Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 13:48:04

File A: voice_noise_ambience.flac
SHA1: 7d21e164e4e46fe7cd081a49e51b2a536e08d615
Gain adjustment: +5.65 dB
File B: voice_noise_ambience_24kbps_newtf2.opus
SHA1: 327f8afc57d203b2c4132a311565d8db8f71dc1d
Gain adjustment: +5.88 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

13:48:04 : Test started.
13:48:21 : 01/01
13:48:31 : 02/02
13:48:38 : 03/03
13:48:43 : 04/04
13:48:50 : 05/05
13:48:57 : 06/06
13:49:04 : 07/07
13:49:10 : 08/08
13:49:18 : 09/09
13:49:23 : 10/10
13:49:29 : 11/11
13:49:36 : 12/12
13:49:41 : 13/13
13:49:46 : 14/14
13:49:50 : 15/15
13:49:56 : 16/16
13:49:56 : Test finished.

 ----------
Total: 16/16
Probability that you were guessing: 0.0%

 -- signature --
0b7e9328c4cdd393acff1e090db2ef9055d24a76



Still sounds like the high frequencies (the noise) are being tuned the first couple of seconds to later calm down. The noise contains more high frequency than the lossless file. Still sounds like things are "tearing up".
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 13:51:45

File A: voice_noise_ambience.flac
SHA1: 7d21e164e4e46fe7cd081a49e51b2a536e08d615
Gain adjustment: +5.65 dB
File B: voice_noise_ambience_48kbps_newtf2.opus
SHA1: cf8ca6bacbda09b4789cff4e510c79356b4cea75
Gain adjustment: +5.77 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

13:51:45 : Test started.
13:52:26 : 01/01
13:52:31 : 02/02
13:52:45 : 03/03
13:52:56 : 04/04
13:53:10 : 05/05
13:53:31 : 06/06
13:53:36 : 07/07
13:53:43 : 08/08
13:53:50 : 09/09
13:53:59 : 10/10
13:54:58 : 11/11
13:55:06 : 12/12
13:55:13 : 13/13
13:55:22 : 14/14
13:55:29 : 15/15
13:55:41 : 16/16
13:55:41 : Test finished.

 ----------
Total: 16/16
Probability that you were guessing: 0.0%

 -- signature --
8504d77d408aa14cbaac0be0397596cec764cfdb



This time I failed! I still (tried to )listen to the increased high frequency content the first couple of seconds. After this ABX tried to listen to other things but I can't find any.
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-03 14:10:33

File A: voice_noise_ambience.flac
SHA1: 7d21e164e4e46fe7cd081a49e51b2a536e08d615
Gain adjustment: +5.65 dB
File B: voice_noise_ambience_64kbps_newtf2.opus
SHA1: f05b0d74fa4b5c88f8377a85330de74b5f0c3f7c
Gain adjustment: +5.59 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

14:10:33 : Test started.
14:11:31 : 00/01
14:11:40 : 01/02
14:11:55 : 02/03
14:12:14 : 02/04
14:12:25 : 03/05
14:12:33 : 04/06
14:13:27 : 05/07
14:13:35 : 05/08
14:13:47 : 06/09
14:14:03 : 07/10
14:14:16 : 07/11
14:14:39 : 07/12
14:15:06 : 07/13
14:15:19 : 08/14
14:15:36 : 09/15
14:15:51 : 09/16
14:15:51 : Test finished.

 ----------
Total: 9/16
Probability that you were guessing: 40.2%

 -- signature --
06a177cadba1f3dbadda8009d10867e73a7362f4
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-03 15:25:07
 You have done a lot of interesting tests. Thank you.

The thing is that developers are interested in comparison between oldtf vs newtf2.
It’s already know that newtf2 or any other lossy files won’t be transparent at those rates but still can have high quality.

ABC/HR should be used in order to compare “old tf” vs “newtf2”. Not ABX comparator.
Here is some info on ABC/HR https://web.archive.org/web/20110611013208/http://ff123.net/64test/practice.html

And other thing. 5/5 (probability of guessing, p=3%) is already enough for ABX log in testing lossy codecs. It’s even more true for ABC/HR. Nobody pretends to think that you fake a results or try multiple times in order to get “valid” results.

You can probably download some ABC/HR app for your Linux distro. 
Here is one http://listening-tests.hydrogenaud.io/igorc/aac-96-a/ABC-HR_bin.zip  
Some documentation http://listening-tests.hydrogenaud.io/igorc/aac-96-a/readme.txt
(Well, Java can give some headache)

Testing Opus newtf2 vs oldtf is hard because differences between them are rather small at least in this moment.

 :)
Title: Re: Opus 1.3-beta is here
Post by: butrus on 2018-02-04 01:08:07
The important thing isn't the exact collection itself, it's having some calibration. Otherwise, I could simply "improve" Opus by just increasing the bitrate for the same command-line arguments. With the calibration, I ensure that any increase in bitrate for a particular file is compensated by a decrease in other files. The result is that different versions should give about the same overall bitrate on your collection, even if it differs from mine (and even if your average bitrate differs from mine).

I would appreciate if I got (more or less) constant quality for different encoder versions no matter what the bitrate is :-(
Title: Re: Opus 1.3-beta is here
Post by: Dynamic on 2018-02-05 13:29:58
Focussing on the default setting, unconstrained VBR --vbr, and ignoring the little-used constrained CBR --cvbr or strict CBR --hard-cbr

The developers stuck to the default philosophy of unconstrained VBR in letting the bitrate fluctuate greatly and fall where it may for a specific track or section of audio, depending on the difficulty of encoding with the same target quality.

However, they differ in philosophy from many popular encoders like Ogg Vorbis, LAME, etc. as the Opus developers calibrate the quality scale to achieve a certain bitrate specified by the user, averaged over a wide and diverse collection of audio. Most encoders using unconstrained VBR have some arbitrary quality metric, though I get the impression that iTunes TVBR AAC quality setting is calibrated to the equivalent bitrate per audio channel so would be closer to the Opus approach.  Any independently developed Opus encoders could take the approach of their choice.

Then when a new version comes out with no regressions (and this thread indicates how careful they are to avoid regressions), the bitrate at the same setting will be the same but the quality will be at least the same or higher. This might be the same in iTunes TVBR AAC if improved versions were released. In other codecs such as LAME or Vorbis with unconstrained VBR, the same setting in an improved version is likely to have the same quality but at the same or slightly lower bitrate overall.

The exception to this can be problem samples, where for that sample alone (but not the diverse collection of audio), the fix may recognise that a higher bitrate is needed for the affected frames, whereas before the fix it didn't allocate enough bits to the right features to encode adequately and caused an artifact.

For me, this is just as valid and bitrate is easier to understand than some arbitrary quality scale, and I appreciate the significant improvements made in the sub-64kbps area and the artifact fixes in the 64kbps+ area, but I can understand the other point of view too.

If you typically use 64 or 96 kbps or higher, I'd say the quality is only marginally changed except that more problem samples have been fixed. If you typically use bitrates of 48kbps or lower, I'd say bitrate is more important than quality and you'd accept some differences from the original that can be noticed but aren't annoying, but you could choose to lower bitrate significantly compared to v1.0 or v1.1 and still achieve music that's good to listen to with unannoying differences from the original.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-05 17:22:33
Opus has very wide VBR. It's comparable to Vorbis -q.

Bitrate distribution:
http://listening-test.coresv.net/s/bitrate_distribution_en.png

Quality distribution:
http://listening-test.coresv.net/s/scores_by_tracks_en.png

Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-05 22:02:34
I rarely sit on Windows so if you want to enable more of us to give you opinions then start with supplying at least some Linux compiles, thanks.
I also develop on Linux. In fact, I even build these Windows binaries from Linux. If you'd like to test, the easiest is probably to grab commits from Git. The "old tf" commit is edaafe3 and the "newtf2" commit is 195ea42.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-05 22:26:53
However, they differ in philosophy from many popular encoders like Ogg Vorbis, LAME, etc. as the Opus developers calibrate the quality scale to achieve a certain bitrate specified by the user, averaged over a wide and diverse collection of audio. Most encoders using unconstrained VBR have some arbitrary quality metric, though I get the impression that iTunes TVBR AAC quality setting is calibrated to the equivalent bitrate per audio channel so would be closer to the Opus approach.  Any independently developed Opus encoders could take the approach of their choice.
Actually, the Vorbis -q values are calibrated to give specific bitrates on a large collection of music. You can see them on the Vorbis Wikipedia page (https://en.wikipedia.org/wiki/Vorbis).

Then when a new version comes out with no regressions (and this thread indicates how careful they are to avoid regressions), the bitrate at the same setting will be the same but the quality will be at least the same or higher. This might be the same in iTunes TVBR AAC if improved versions were released. In other codecs such as LAME or Vorbis with unconstrained VBR, the same setting in an improved version is likely to have the same quality but at the same or slightly lower bitrate overall.
So keep in mind here that all of the improvements between 1.0 and 1.2 probably (my opinion, not measured) amount to ~10% reduction in bitrate for the same quality over a period of ~5 years. To keep some kind of equal-quality option properly calibrated across version, we would need very accurate listening tests at multiple bitrates. The effort involved in conducting those tests would probably be greater than the effort involved in making the quality improvements in the first place. It would also delay any new release by a year or so. That's why nobody does that.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-06 13:54:32
So keep in mind here that all of the improvements between 1.0 and 1.2 probably (my opinion, not measured) amount to ~10% reduction in bitrate for the same quality over a period of ~5 years.
I'd prefer Opus 1.2.1@96 kbps over  Opus 1.0@112 kbps.  That's ~15%
1.2.1@64 kbps over 1.0@72-75 kbps.  That's ~15%
1.2.1@32 kbps over 1.0@40-42.  That's 25-30%+!

Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-06 14:23:49
I just want to let you know that I was able to ABX at 160kbps opus 1.3b (not the newtf's). In this file I was detected that a ride cymbal that was played with the tip and mostly in the left channel and was delay to the right to create Haas effect had the transients on the right channel weakened compare to the unaltered left channel. Imagine that you would make some of the hits pianossimo (pp) instead of them all being forte.
Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-06 14:54:33

File A: mixa04.flac
SHA1: 3409e6a76cf6fc66bc64a7fbb8fa9cb8b182bb9c
Gain adjustment: +1.51 dB
File B: mixa04_160kbps_cvbr.opus
SHA1: f2d5313ece1132f54ccb97771291678ba97f5863
Gain adjustment: +1.51 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

14:54:33 : Test started.
14:55:21 : 00/01
14:55:35 : 01/02
14:56:02 : 02/03
14:56:18 : 03/04
14:56:32 : 04/05
14:56:58 : 05/06
14:57:18 : 06/07
14:57:35 : 07/08
14:58:12 : 08/09
14:58:37 : 09/10
14:58:37 : Test finished.

 ----------
Total: 9/10
Probability that you were guessing: 1.1%

 -- signature --
05b99a0810f8db11ca5ae89782960e0354826703

I attached a sample of the problematic ride cymbal. Encode to opus and listen between 00:01 and 00:05. Some hits always sound let their initial attacks have been removed. I tried mp3 at V0 and ogg at high bitrates and they didn't manage it either. I could not successfully ABX the sample when I used opus and 192kbps.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-07 02:39:45
Until now I could ABX a difference between newTF2 vs oldTF only on one sample.

Right now I'm at the point to start to listen white noise and hear some human voices .... from hell 
Yes, differences are that small.

Files (https://drive.google.com/file/d/1sNFuw2moD5jT2-f0C3IYxfhWrxgd6kXX/view?usp=sharing)

Code: [Select]
ABC/HR Version 1.1 beta 2, 18 June 2004
Testname:

1R = E:\DETODO\Audio\When The Levee Breaks\When The Levee Breaks Opus 96k old TF.wav
2L = E:\DETODO\Audio\When The Levee Breaks\When The Levee Breaks Opus 96k NEW TF2.wav

---------------------------------------
General Comments:

---------------------------------------
1R File: E:\DETODO\Audio\When The Levee Breaks\When The Levee Breaks Opus 96k old TF.wav
1R Rating: 4.0
1R Comment:
---------------------------------------
2L File: E:\DETODO\Audio\When The Levee Breaks\When The Levee Breaks Opus 96k NEW TF2.wav
2L Rating: 4.1
2L Comment:
---------------------------------------
ABX Results:


---------------------------------------
MY NOTES. ABX training log of "old TF.wav" vs "NEW TF2.wav"
---------------------------------------
1 of   1, p = 0.500
  2 of   2, p = 0.250
  3 of   3, p = 0.125
  3 of   4, p = 0.313
  4 of   5, p = 0.188
  5 of   6, p = 0.109
  6 of   7, p = 0.063
  7 of   8, p = 0.035
TRAINING MODE -- not written to file

P.S. As for this sample I think newTF2 is just a little bit more constant on backround, less splashy (less wavy).
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-08 12:02:18
Igorc, the couple of seconds in the beginning I mentioned when it sounds like it is tuning the high frequencies. What is that? It goes from very audible in the beginning during to noise calming down after a few seconds.
Title: Re: Opus 1.3-beta is here
Post by: Leo 69 on 2018-02-08 12:18:58
@ punkrockdude

Bro, could you please test this sample against Apple AAC @160 kbps as well please.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-08 12:39:35
@ Leo 69
No problem at all but do you mean the sample with the ride having a Haas effect or the voice sample containing noise? I don't have Apple AAC Encoder and rarely use Windows except for when I record or mix and now that is rare since Reaper has unofficial versions for Linux.

Wait, since you said 160kbps I suspect you mean the ride sample since I could almost ABX it using opus. If you encode it and attach it in a message then I will ABX it for you.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-08 17:12:27
Igorc, the couple of seconds in the beginning I mentioned when it sounds like it is tuning the high frequencies. What is that? It goes from very audible in the beginning during to noise calming down after a few seconds.
Provide a link to your original message and/or sample which You are talking about because honestly I can't follow every word in 6-page disscussion.

Code: [Select]
foo_abx 2.0.4 report
foobar2000 v1.4 beta 2
2018-02-06 14:54:33

File A: mixa04.flac
SHA1: 3409e6a76cf6fc66bc64a7fbb8fa9cb8b182bb9c
Gain adjustment: +1.51 dB
File B: mixa04_160kbps_cvbr.opus
SHA1: f2d5313ece1132f54ccb97771291678ba97f5863
Gain adjustment: +1.51 dB

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

14:54:33 : Test started.
14:55:21 : 00/01
14:55:35 : 01/02
14:56:02 : 02/03
14:56:18 : 03/04
14:56:32 : 04/05
14:56:58 : 05/06
14:57:18 : 06/07
14:57:35 : 07/08
14:58:12 : 08/09
14:58:37 : 09/10
14:58:37 : Test finished.

 ----------
Total: 9/10
Probability that you were guessing: 1.1%

 -- signature --
05b99a0810f8db11ca5ae89782960e0354826703

mixa04_160kbps_cvbr.opus

Any particular reason for CVBR instead of VBR?

Opus CVBR is equivalent to AAC CBR (ABR with very little variation/window)
Title: Re: Opus 1.3-beta is here
Post by: 2012 on 2018-02-08 17:45:00
Quote
Wait, since you said 160kbps I suspect you mean the ride sample since I could almost ABX it using opus. If you encode it and attach it in a message then I will ABX it for you.

It's worth noting that the output of libopus 1.2.1 and 1.3 beta(with fixes (https://git.xiph.org/?p=opus.git;a=summary)) is identical when encoding at --bitrate 160.

Apple AAC encoding at a slightly higher bitrate (https://archive.org/download/unsorted_files/mixa04_sample_qaac_V70.m4a).
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-09 12:37:34
It's worth noting that the output of libopus 1.2.1 and 1.3 beta(with fixes (https://git.xiph.org/?p=opus.git;a=summary)) is identical when encoding at --bitrate 160.
I ABXed lossless vs opus and not one version of opus to another version.
Title: Re: Opus 1.3-beta is here
Post by: 2012 on 2018-02-09 15:25:12
It's worth noting that the output of libopus 1.2.1 and 1.3 beta(with fixes (https://git.xiph.org/?p=opus.git;a=summary)) is identical when encoding at --bitrate 160.
I ABXed lossless vs opus and not one version of opus to another version.

I know. I just wanted to make it clear that there is no beta regressions here.

I didn't notice your opus encode is CVBR though (why?). Luckily, the output is also identical between 1.2.1 and 1.3-beta with --cvbr --bitrate 160.

Here (https://archive.org/download/unsorted_files/mixa04_sample_qaac_v160.m4a) is the sample encoded with qaac -v 160 (actual bitrate is 176kbps). Although, I agree with Igor that this is not a very useful comparison.

Side note: You can use qaac in GNU/Linux via wine. That's what I do.
Title: Re: Opus 1.3-beta is here
Post by: punkrockdude on 2018-02-10 15:31:14
Why I used CVBR in this case was just to try out the different options and see if it made any audible reason. I didn't detect any. I will ABX the AAC file asap.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-10 21:51:57
OK, could be a tuning issue. Can you give this updated one a try:
opus-tools-newtf2.zip (https://jmvalin.ca/misc_stuff/opus-tools-newtf2.zip)
I have sent You PM with my results.
Title: Re: Opus 1.3-beta is here
Post by: king2101 on 2018-02-12 13:10:22
@jmvalin the main problems with the opus codec at bitrates of 48kbps is with violins and certain other strings (real & electronic) . Would it be possible for your team to fix that issue? Every other song sounds transparent at that bitrate except those ones
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-12 16:54:23
@jmvalin the main problems with the opus codec at bitrates of 48kbps is with violins and certain other strings (real & electronic) . Would it be possible for your team to fix that issue? Every other song sounds transparent at that bitrate except those ones
Yes, I'm aware of the problem. So far I haven't found any way to improve that, but I'm still looking. I already have a few samples, but if you have a short clip that makes the problem particularly obvious, it's always useful.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-15 20:33:55
Some results for 1.3 beta
Opus 1.2.1 - 1.3 beta vs HE-AAC  ( Test at 32 kbps) (https://docs.google.com/spreadsheets/d/1HUXHtAxNzMGdAstwHIwKWDhBllhRVGcpXXnUwFODTqg/edit#gid=0)
Files (https://drive.google.com/file/d/15OAI0bcyYSDXNFrtxuZrXjmR4sKCxjq7/view)
There are some improvements ( "Can't wait" and "Fatboy" samples) but also some regressions. Most notable regression is on sample 4º.

OK, I'm back tracking this regression with two builds:
These will only affect the speech samples, so no need to test the music ones. In fact, the "regression" is due to an improvement in the speech/music detector. Previously the speech samples would get mostly coded with CELT and in 1.3-beta they're being coded with hybrid. It just turns out that hybrid could do better, which I'm trying to fix here. I couldn't really hear much artefacts on sample 4, but the ones on sample 7 (Korean speech) were pretty obvious. Last detail, make sure you decode with the opusdec I'm providing (either is fine) because they have the "recent" decoder update (RFC 8251) enabled, which may not be the case with all players. Let me know if either or both these builds address what you were hearing.
(For non-Win32 users, just checkout the Git hash in the file name and build that)
Title: Re: Opus 1.3-beta is here
Post by: juliobbv on 2018-02-17 01:26:03
Hi everybody,

I'm Julio. I've been lurking HA for a looong time (back from the "Project Mayhem" era in the early 2000s, with the orange board). I finally created an account today because I like Opus and recently have found some free time to volunteer with listening tests. I think I can help with the low-bitrate tuning effort @jmvalin has been working on.

As somebody who has never done any ABC/HR or ABX tests (ironically!): can you help me with recommended resources on how to set up everything, in terms of installing software and preparing the encoded samples for the listening tests? I use Windows 10. Also, let me know if there are any samples you want me to focus on first, and which Opus encoders should I use to encode and compare the samples. I'll give it a shot over the long weekend.
Title: Re: Opus 1.3-beta is here
Post by: Leo 69 on 2018-02-17 19:52:07
Hi everybody,

I'm Julio. I've been lurking HA for a looong time (back from the "Project Mayhem" era in the early 2000s, with the orange board). I finally created an account today because I like Opus and recently have found some free time to volunteer with listening tests. I think I can help with the low-bitrate tuning effort @jmvalin has been working on.

As somebody who has never done any ABC/HR or ABX tests (ironically!): can you help me with recommended resources on how to set up everything, in terms of installing software and preparing the encoded samples for the listening tests? I use Windows 10. Also, let me know if there are any samples you want me to focus on first, and which Opus encoders should I use to encode and compare the samples. I'll give it a shot over the long weekend.

Hi Julio! And, welcome to the club. What's your audio gear?

@punkrockdude

Hi man, did you have a chance to test that AAC sample?
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-17 20:21:42
Leo '69'

What part of TOS5 (https://hydrogenaud.io/index.php/topic,3974.msg149478.html#msg149478) You don't understand
Quote
Submission of a thread that is not related to any of the topics covered by the Hydrogenaudio forums and subforums, or posts which are not related to the thread that they are submitted to, is against the rules. If you wish to discuss Off-Topic issues, please use our Off-Topic forum. Repeated violation of this rule after one notification may lead to administrative action including banishment.
Title: Re: Opus 1.3-beta is here
Post by: Leo 69 on 2018-02-17 21:41:04
Leo '69'

What part of TOS5 (https://hydrogenaud.io/index.php/topic,3974.msg149478.html#msg149478) You don't understand
Quote
Submission of a thread that is not related to any of the topics covered by the Hydrogenaudio forums and subforums, or posts which are not related to the thread that they are submitted to, is against the rules. If you wish to discuss Off-Topic issues, please use our Off-Topic forum. Repeated violation of this rule after one notification may lead to administrative action including banishment.

So not even being a moderator here, You decided to publicly charge me with some imaginary violations? While I DO respect your reputation on these forums and couple of others, you should stop going overboard.

You'd better answer the guy, who sincerely wants to help improve the already great codec and devote a lot of time to this. Instead of being a douche. Sorry for being so straightforward, but that's me.
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-17 22:07:47
Leo '69'

What part of TOS5 (https://hydrogenaud.io/index.php/topic,3974.msg149478.html#msg149478) You don't understand
Quote
Submission of a thread that is not related to any of the topics covered by the Hydrogenaudio forums and subforums, or posts which are not related to the thread that they are submitted to, is against the rules. If you wish to discuss Off-Topic issues, please use our Off-Topic forum. Repeated violation of this rule after one notification may lead to administrative action including banishment.
So not even being a moderator here, You decided to publicly charge me with some imaginary violations? While I DO respect your reputation on these forums and couple of others, you should stop going overboard.

You'd better answer the guy, who sincerely wants to help improve the already great codec and devote a lot of time to this - instead of being a douche. Sorry for being straightforward, but that's me.

@IgorC, while not obvious at first sight (I was also puzzled originally), I believe Leo 69's post was (sufficiently) on topic. I believe he was asking about audio gear because juliobbv looked like he wanted to do some listening tests. As for his reply to punkrockdude it seems to be related to an earlier comment in the thread about AAC vs Opus comparison.

@Leo 69, please no need for personal insults here. Your post looked a bit off topic at first sight and I believe IgorC was acting in good faith, even if a little harsh.

Now anyone has any comments on the latest Opus builds?  ;)
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-17 22:26:02
jmvalin,

No, no, no
 He wasn't on topic here.  Leo just asked to test AAC absolutely nothing to do with "Opus vs AAC" . I followed topic.

He was off-topic 100%.

So not even being a moderator here, You decided to publicly charge me with some imaginary violations? While I DO respect your reputation on these forums and couple of others, you should stop going overboard.
And yes, I will indicate you to stop off-topic.
I'm a member of this forum and I have right to indicate to You to stop off-topic. You're violating rules of this forum.


You'd better answer the guy, who sincerely wants to help improve the already great codec and devote a lot of time to this. Instead of being a douche. Sorry for being so straightforward, but that's me.
Irony.  :)   Do you understand that I was around and involved all this time in every freaking Opus version, testing it, organinizing tests etc.....
I don't remeber You being involved .... again, Who are You?  And what makes you think You can tell me "You'd better answer" ..... F.O.

Do You understand that my collaboration with Jean-Marc has moved to private mail because absolutely noone here have submit one single meaninful result. This is thread is full nonsense and off-topic. 

The only guys who here "sincerely want" (as you said) to help improve the codec is me and JM himself. I'm not trying to be a superman but  there isn't much help really. We don't see much  help from You or any other person.

If You're not helping at all.... at least stop off-topic.
Title: Re: Opus 1.3-beta is here
Post by: Leo 69 on 2018-02-17 22:55:35
@jmvalin
Thank you very much for your comments. They are 100% on point. Someone just can't read between the lines..my goodness.

Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-18 00:58:43
As somebody who has never done any ABC/HR or ABX tests (ironically!): can you help me with recommended resources on how to set up everything, in terms of installing software and preparing the encoded samples for the listening tests? I use Windows 10. Also, let me know if there are any samples you want me to focus on first, and which Opus encoders should I use to encode and compare the samples. I'll give it a shot over the long weekend.
Here are some hopefully useful links https://hydrogenaud.io/index.php/topic,115156.msg952288.html#msg952288

Also You can take a look here https://hydrogenaud.io/index.php/topic,16295.0.html
Title: Re: Opus 1.3-beta is here
Post by: juliobbv on 2018-02-18 01:08:46
Hi Julio! And, welcome to the club. What's your audio gear?

Hi Leo. Thanks for the welcome! As for my audio gear, I have these Audio-Technica ATH-M30x headphones: https://www.amazon.com/gp/product/B00HVLUQW8.

I read ff123's "Practice With ABC/HR" guide (https://web.archive.org/web/20110611013208/http://ff123.net/64test/practice.html) to get ABC-HR set up on my machine, and went over the suggested "newkid" session. After that, I used one of @IgorC's samples to create my own "smoke-test" session to verify that I did set up everything right.

Code: [Select]
1L = smoketest\02 fatboy_30sec Opus 1.2.1 32 kbps.wav
2R = smoketest\02 fatboy_30sec Opus 1.3b1 32 kbps.wav

---------------------------------------
1L File: smoketest\02 fatboy_30sec Opus 1.2.1 32 kbps.wav
1L Rating: 3.0
1L Comment: Very odd voice change at 1s. 'Ringing' artifacts appear to be sligtly worse than sample 2.
---------------------------------------
2R File: smoketest\02 fatboy_30sec Opus 1.3b1 32 kbps.wav
2R Rating: 3.8
2R Comment: First part (with the solo robotic voice) almost perfect. Second part contains some 'ringing' artifacts (esp. at 16.5s).
---------------------------------------

ABX Results:


It looks like I should be good by now -- I'm ready to try out more samples. What interesting samples/encoder combinations would you like for me to try out? I'm open on ideas.

Edit:

Here are some hopefully useful links https://hydrogenaud.io/index.php/topic,115156.msg952288.html#msg952288

Also You can take a look here https://hydrogenaud.io/index.php/topic,16295.0.html

Thanks for the links -- I'll go take a look at them.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-18 02:54:38
You can try any of your own samples  or try from these sources

http://listening-tests.hydrogenaud.io/igorc/aac-96-a/all_samples.zip
http://www.rarewares.org/test_samples/
http://www.mp3-tech.org/tests/aac_48/samples/
http://listening-test.coresv.net/results.htm


At this point it's worth to try these builds vs stable 1.2.1 at 32-48 kbps.
https://hydrogenaud.io/index.php/topic,115156.msg952996.html#msg952996
Title: Re: Opus 1.3-beta is here
Post by: Moni on 2018-02-18 16:06:29
I have also been following Opus development for a long time but have not participated in any sort of listening test I could submit to JMV for improvement of the codec.
I am the director of concert recordings at a conservatory and we are currently in the initial planning stages for what our next digital platform will be to make these recordings available. I would very much like for Opus to be one of the flavors available. I realize I have an opportunity with the large body of recordings I work with, and my calibrated office with good equipment, to contribute. Classical and opera can be difficult to encode during quiet or solo portions, as they lack the large masking effect provided by most modern music recordings, and the importance of preserving the timbre.
Anyhow, thanks to the initial question from @juliobbv and helpful response from @IgorC , I will start some testing in the next few weeks.
Title: Re: Opus 1.3-beta is here
Post by: IgorC on 2018-02-23 00:00:16
I read ff123's "Practice With ABC/HR" guide (https://web.archive.org/web/20110611013208/http://ff123.net/64test/practice.html) to get ABC-HR set up on my machine, and went over the suggested "newkid" session. After that, I used one of @IgorC's samples to create my own "smoke-test" session to verify that I did set up everything right.

Code: [Select]
1L = smoketest\02 fatboy_30sec Opus 1.2.1 32 kbps.wav
2R = smoketest\02 fatboy_30sec Opus 1.3b1 32 kbps.wav

...

Just for the record, your test is  in-line with my results. Good.
At this point it would be great to try speech or mixed speed/music samples at 16-32-40 kbps. https://hydrogenaud.io/index.php/topic,115156.msg952996.html#msg952996

I don't want to spoil a surprise (this is a phrase when actually a surprise will be spoiled   :)  ) but  'speech/music classification using an RNN' (https://people.xiph.org/~jm/demo/rnnoise/) starts to work after ironing out issues.
At least for speech, Opus starts to wipe the floor with a modern codecs like xHE-AAC/USAC, EVS. (leave alone outdated HE-AAC v1/v2 at this point) . Well, all modern codecs are already on par on music at 48 kbps and higher.
Something big is coming.

Title: Re: Opus 1.3-beta is here
Post by: butrus on 2018-02-24 17:31:03
Do You understand that my collaboration with Jean-Marc has moved to private mail because absolutely noone here have submit one single meaninful result.

That's really pity, I was really interrested in reading about the progress of Opus and the tests - unfortunatelly I cannot contribute just now as I didn't find any suitable programm for Linux to do the ABX and I don't know if it would make sense with my current equipment (chep laptop speakers)...
Title: Re: Opus 1.3-beta is here
Post by: maikmerten on 2018-02-24 21:16:10
Do You understand that my collaboration with Jean-Marc has moved to private mail because absolutely noone here have submit one single meaninful result.

That's really pity, I was really interrested in reading about the progress of Opus and the tests - unfortunatelly I cannot contribute just now as I didn't find any suitable programm for Linux to do the ABX and I don't know if it would make sense with my current equipment (chep laptop speakers)...

For Linux, use squishyball, which is the Xiph.org ABX tool.

For Ubuntu and Debian, a "sudo apt install squishyball" should do the job.

Of course, laptop speakers are most likely not useful for the job... but even cheap headphones can yield interesting results.
Title: Re: Opus 1.3-beta is here
Post by: Moni on 2018-02-24 23:49:36
After the inital test in abc-hr, I was dismayed to see how much work goes into creating one's own tests. I found abc-hr for Java here (http://www.rarewares.org/others.php) due to a link in this document: www.rarewares.org/rja/ListeningTest.pdf
I just wanted to check that results from this program are considered valid as well.

I am going to start assembling some of my own samples from audiobooks and lectures on music (alternating speech and music) and see what I get, using the builds referenced by IgorC.
Title: Re: Opus 1.3-beta is here
Post by: fabiorug on 2018-02-25 09:19:20
I  think that opus new tf 2 sounds better at 24 kbps than newtf.
64 kbps is similar ( not so big improvement,  maybe it's better in some samples.)
NEW TF 2 it's not that bad.
Title: Re: Opus 1.3-beta is here
Post by: LoRd_MuldeR on 2018-02-25 13:57:18
Suggested patch to enable progress output (percentage) in "opusdec" tool (new libopusfile-based version):
https://gist.github.com/anonymous/2ea78e68a251f7fdb6e46822575af2eb

This also limits the update frequency a bit, because current version writes updates very frequently, which can be slow on Windows console...
Title: Re: Opus 1.3-beta is here
Post by: 2160p on 2018-02-25 17:35:19
This beta release of the upcoming Opus 1.3 includes:
  • Enabling by default the spec fixes in RFC 8251
  • Improvements to the VAD and speech/music classification using an RNN
  • Improvements to stereo speech coding at low bitrate
  • Added support for ambisonics projection using mapping 3 (disabled by default)
  • Fixes to the CELT PLC
Additionally, as a way to test the upcoming update to opus-tools, we’re providing Windows binaries built with 1.3-beta. These binaries are based on libopusenc, which means opusenc is finally able to make use of the Opus delayed-decision feature to make better speech/music transitions.

Source code: opus-1.3-beta.tar.gz (https://archive.mozilla.org/pub/opus/opus-1.3-beta.tar.gz)
Win32 binaries (experimental): opus-tools-test1.zip (https://archive.mozilla.org/pub/opus/win32/opus-tools-test1.zip)

Moin,
What's happen if add two-pass option?
Title: Re: Opus 1.3-beta is here
Post by: LoRd_MuldeR on 2018-02-26 12:12:22
Suggested patch to enable progress output (percentage) in "opusdec" tool (new libopusfile-based version):
https://gist.github.com/anonymous/2ea78e68a251f7fdb6e46822575af2eb

This also limits the update frequency a bit, because current version writes updates very frequently, which can be slow on Windows console...

Small update with some clean-up:
https://gist.github.com/anonymous/103196c96a7f795c37f2cabffe7dbf03
Title: Re: Opus 1.3-beta is here
Post by: jmvalin on 2018-02-26 18:29:50
What's happen if add two-pass option?
If you're referring to the delayed-decision option, it's not two-pass, it's really just a 2-second look-ahead in the analysis that makes it possible to make better speech/music decisions. It also helps a little with tone onsets. The advantage of this being a delay and not a two-pass option is that it can still be used for real-time streaming if a short delay is acceptable (it can also be configured to be less than 2 seconds). Of course, for VoIP and videoconferencing there's no delayed decision at all since any delay is bad.
Title: Re: Opus 1.3-beta is here
Post by: Seymour on 2018-04-02 19:08:20
Found a bug with v1.3beta. High frequencies suffer for a noticeable period, compared to v1.2.
Please download: http://wdfiles.ru/5lXj (big blue single button in the middle of the page).
Use the phrase "haopusbeta" to open.
Apologies for full composition, couldn't find a way to show the bug other way.
Title: Re: Opus 1.3-beta is here
Post by: soundping on 2018-05-25 20:17:09
Is there an update to OPUS beta?  :)
Title: Re: Opus 1.3-beta is here
Post by: meowchin on 2018-05-31 22:16:08
You already have a scale, it's called "bitrate". Instead of having a 0-10 scale that maps to different bitrates (because Vorbis also calibrates its quality scale on an average bitrate), the scale is 0-510 you don't even have to look up what value you need to get a certain average bitrate.
I know I'm VERY late with the answer, but I completely forgot I asked this question.

I don't have a problem with this approach when it comes to stereo, the problem is achieving the same quality in multichannel files. Currently I have pretty good heaphones and a mediocre 5.1 setup, and I would like to encode a futureproof 5.1 file (in case I update my setup to something better) - with the same quality level that I was able to determine is transparent on headphones. There is really no other way to do that than to guess what bitrate will be appropriate, or take ABX tests again - which I don't want to do, because, as I said, my setup isn't the greatest.

With qaac it's easy - I know that TVBR 109 is transparent for me, so I can just encode any file at this setting, no matter how many channels it has, and I will always end up with bitrate that's appropriate for the encoded content.
SimplePortal 1.0.0 RC1 © 2008-2018