Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Opus 1.3-rc (Read 53694 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Opus 1.3-rc

I just released Opus 1.3-rc. This is a release candidate for the upcoming Opus 1.3. Changes include:
  • Making it possible to use SILK down to bitrates around 5 kb/s
  • Using wideband encoding down to 9 kb/s
  • Improving security (including a new --enable-hardening option)
  • Minor quality improvement on tones
  • Improving Ambisonics support (still experimental)
  • Minor bug fixes
Again, we’re providing experimental Windows binaries built with 1.3-rc and the upcoming opus-tools and libopusenc.

Source code: opus-1.3-rc.tar.gz
Win32 binaries: opus-tools-test-1.3-rc.zip

Re: Opus 1.3-rc

Reply #1
Header gain +5.00 dB?

Re: Opus 1.3-rc

Reply #2
Good.
Anyway I don't recommend using less than 64 kbps VBR I for encoding music even if the tonality analysis has improved. I prefer 64 kbps bitrate (Stereo).

Re: Opus 1.3-rc

Reply #3
I tested it from windows and there were no apparent bugs.

Encoding on 9kbps-13kbps range was a bit slower than the other rates.
I don't believe this is a serious issue, but anyway.

Other than the bitrate, the default settings were used.
Code: [Select]
opus-tools-test-1.3-rc\opusenc --bitrate %v %i %o

Re: Opus 1.3-rc

Reply #4
It might not be the right place to report this bug but creating files with the parameter --framesize 60, creates files that are not playable in Android 6.0+ where there is native Opus support. It simply outputs silence. The same files work on everything else I tried so it's only on Android. Google claims that their encoder is based on libopus 1.2.1 so whatever is going on it may have to do with that.

Thread with further discussion on the issue: https://hydrogenaud.io/index.php/topic,116065.msg957834/topicseen.html#new

Re: Opus 1.3-rc

Reply #5
I just released Opus 1.3-rc. This is a release candidate for the upcoming Opus 1.3. Changes include:
  • Making it possible to use SILK down to bitrates around 5 kb/s
  • Using wideband encoding down to 9 kb/s
  • Improving security (including a new --enable-hardening option)
  • Minor quality improvement on tones
  • Improving Ambisonics support (still experimental)
  • Minor bug fixes
Again, we’re providing experimental Windows binaries built with 1.3-rc and the upcoming opus-tools and libopusenc.

Source code: opus-1.3-rc.tar.gz
Win32 binaries: opus-tools-test-1.3-rc.zip

I wonder how 1.3 compare now against EVS at low bitrate.

Re: Opus 1.3-rc

Reply #6
1.3 is an important update.  I've no time to post all results but here is a preview  ( Korean speech sample). General findings are in line with this sample.
https://listening-test.coresv.net/tracks/24-Greensleeves-Korean-male-speech.441.wav

1.3RC1 is pretty in line  with 1.3 post-beta 1 (opus-tools-9cfdcba), if not even better!


To Opus developers, great work and thank you for 1.3.

Re: Opus 1.3-rc

Reply #7
Would it still hold true that setting --framesize to anything bigger than 20 is either pointless or harmful? In case the delay doesn't matter (it's for offline playback from some kind of disk), and if small differences in decoder CPU or memory usage also do not matter.
a fan of AutoEq + Meier Crossfeed

Re: Opus 1.3-rc

Reply #8
Encoding on 9kbps-13kbps range was a bit slower than the other rates.
I don't believe this is a serious issue, but anyway.

As far as I can tell, this is due to three different modes being used: SILK narrowband up to 9 kb/s, SILK wideband (the slowest) between 9 and 13 kb/s, then CELT from there on. SILK typically decodes faster than CELT, but it's slower to encode, especially if you want to do a good job (i.e. high complexity setting).

Re: Opus 1.3-rc

Reply #9
(Somewhat off-topic.)

Is the opus-tools GitHub repo being looked at? None of the 1.3 stuff seems to be there. Also, I submitted a (trivial) patch to make the opusdec resampler quality configurable, but the place seems very inactive.

Re: Opus 1.3-rc

Reply #10
Foobar 1.4 beta 16 is now showing Header gain 0.00 dB.


Re: Opus 1.3-rc

Reply #12
what do new parameter do? --no-phase-inv


Re: Opus 1.3-rc

Reply #14
This is phenomenal! You guys are really challenging the limits of audio codec efficiency! I mean, I'd listen to a 9kbps Opus audio without complaining, this is so exciting!

Really, really thankful for this. Not just from me, but from everyone who uses lossy audio encoding technology. This is history. So happy to be living in this generation!

Re: Opus 1.3-rc

Reply #15
I hope there'll be improvements at 32 kbps and less artifacts at low bitrates.

Re: Opus 1.3-rc

Reply #16
Opus is the future.  ;D


Re: Opus 1.3-rc

Reply #18
Opus is the future.  ;D
For those who aren't locked to some software/hardware which doesn't support Opus, it also could very well be the present.
(and also for a lot of present VoIP implementations and whatnot.)
it is possible for him and the future but not now, this is clearly audible on the samples this https://drive.google.com/file/d/1qsRU3LhzKXZzngVWXaOW7yEKfATTOnVg/view?usp=sharing
Hmm so an other currently available encoder works better than Opus on this sample? Or what do you mean by this?
a fan of AutoEq + Meier Crossfeed

Re: Opus 1.3-rc

Reply #19
what do you mean by this?
I am not a professional in these matters, but this sample displays contemporary art, and since the opus has become a global standard, it must be transparent not only in classical music



 

Re: Opus 1.3-rc

Reply #22
I'd listen to a 9kbps Opus audio without complaining

as far as speech goes... 14kbps is about the lowest I would suggest as while you can go lower and have it still passable (i.e. still understand what's said etc) I feel sound quality starts to really drop off going below the 14kbps setting, at least based on my Klipsch Pro-Media speakers for the PC. because while I can notice the 14kbps being a hit to the overall sound compared to the 128kbps original MP3 speech file I got, it's not that much worse for a tiny fraction the original files bit rate.

basically I sort of feel 14kbps, for speech, is that sweet spot where one wants smallest possible file size but without sacrificing sound quality too much. I know the 'sweet spot' might vary from person-to-person though as some might favor sound quality over bit rate a bit and some might be more of the other end where they favor super small file size at the expense of sound quality but I suspect many would find my suggested 14kbps as solid enough as while you can definitely notice the difference in sound drop off, it's not that much considering the really low bit rate.

NOTE: I have not messed with Opus v1.3RC much yet but I suspect things will remain the same on what I said above.

I hope there'll be improvements at 32 kbps and less artifacts at low bitrates.

If I recall correctly, from what I have read around here not long ago, it seems like there will be mainly adjustments in the 48kbps and lower ranges as I would not expect much, if anything, on about 64kbps and probably next to nothing, if not nothing, in the 96kbps+ ranges since 96kbps is already pretty darn good given the listening tests around here.

but even assuming what I said above is correct... I would imagine we are talking very small differences in sound quality to where when casually listening I would not expect to hear any obvious differences between v1.2.1 and v1.3RC at bit rates of 32-48kbps (give or take). so basically what I am saying is, don't expect much as I doubt there will be any obvious differences from v1.2.1 to v1.3 as we are probably talking faint differences I would imagine for the most part.

while adjustments at the really low bit rates are nice, personally, I would like to see how far they can lower the bit rate while still 'approaching transparency' (or close enough to that standard) when it comes to music.  so in this regard, while 96kbps is pretty much that 'approaching transparency' level already with Opus v1.2.1 it would be nice if that could be lowered a bit to say improve 64kbps (and the like) as I imagine much under 64kbps is just going to be difficult to get near that standard and even 64kbps is probably asking a lot already given it's a 1/3rd less bit rate than the already strong 96kbps, which is already a pretty low bit rate.

but with that said... it seems Opus is already a pretty strong performer at those low bit rates (say under 96kbps) as say something like 10-15 years ago the sound quality Opus has at 64kbps was overall better than bit rates quite a bit higher than that. so it's obvious they (general audio encoders) made fairly major progress since then. but it's nice in that, at least with AAC/MP3, any files encoded in this decade (assuming the person used the newest encoders at the time) you don't have to worry about sound quality since it's pretty much the same. but in the previous decade as the years passed they made solid progress as the years passed. but this might not entirely apply to Opus given that seems to have made decent progress in this current decade where as AAC/MP3 peaked something like 8-ish years ago(?).
For music I suggest (using Foobar2000)... MP3 (LAME) @ V5 (130kbps). NOTE: using on AGPTEK-U3 as of Mar 18th 2021. I use 'fatsort' (on Linux) so MP3's are listed in proper order on AGPTEK-U3.

Re: Opus 1.3-rc

Reply #23
what do you mean by this?
I am not a professional in these matters, but this sample displays contemporary art, and since the opus has become a global standard, it must be transparent not only in classical music
I tried to find any art in that file and did not succeed.
The sample is also longer than 30 seconds which are allowed here for copyrighted samples.
Anyway Opus seems to do fine with this (with --bitrate 140, at least), do you have different test results? If you claim to be able to hear the difference, you need to prove it, that's in the rules of HA.
a fan of AutoEq + Meier Crossfeed

Re: Opus 1.3-rc

Reply #24
I tried to find any art in that file and did not succeed.
Probably you have a musical education?
The sample is also longer than 30 seconds which are allowed here for copyrighted samples.
I reduced to 10 seconds in accordance with the requirements and copyright
Anyway Opus seems to do fine with this (with --bitrate 140, at least), do you have different test results?
--vbr --bitrate 180

foo_abx 2.0.4 report
foobar2000 v1.4 beta 17
2018-06-08 13:41:39

File A: precision010.flac
SHA1: a1b28d8473317a6ee814598e9542d3aa63e0472f
File B: precision010.opus
SHA1: 8ae8807b027516fb9748b710f7451c4cce005535

Output:
DS : Первичный звуковой драйвер
Crossfading: NO

13:41:39 : Test started.
13:42:49 : 01/01
13:43:30 : 02/02
13:43:50 : 03/03
13:44:24 : 04/04
13:45:05 : 05/05
13:45:21 : 06/06
13:45:39 : 07/07
13:45:57 : 08/08
13:46:22 : 09/09
13:46:32 : 10/10
13:46:55 : 11/11
13:47:15 : 12/12
13:47:34 : 13/13
13:47:59 : 14/14
13:48:18 : 15/15
13:48:53 : 16/16
13:48:53 : Test finished.

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

 -- signature --
bdb235586eed63116588b7b4dd615c7ddfb6a855


my audio hardware: DAC ES9018K2M + TPA6133A2