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: Who Here Has Their Albums At OGG?? (Read 11179 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Who Here Has Their Albums At OGG??

Anybody else keep there music collection at OGG.  I like OGG it sounds really dam good and it is like 40% to 65% smaller than FLACs.  What I do it is download albums in FLAC then I down convert them to Q10 500kbs with Mediamonkey.  I have like 8500 albums if they were all in FLAC they would take up 3tb maybe more space I not a fan of albums taking 450mb to 850mb.  Hardcore FLAC downloaders are really anal about FLACs all other audio codecs are garbage to them S M H.  I bet these FLAC downloaders have never listened to 1 to 3 500kbs OGG albums all the way through.  IMO I can not tell the deference between OGG and FLAC I have listened to both on decent speakers. 

Re: Who Here Has Their Albums At OGG??

Reply #1
I have converted most of my albums to OGG. Before, I stored them in FLAC, the amount of data was very large. Today, I have 225 GB in OGG. To my ears, I don't "see" differences. I convert the songs to q10.
wagner reatto

Re: Who Here Has Their Albums At OGG??

Reply #2
Bizarre thread. Keep the FLAC, it's lossless so you can convert to whichever lossy format you want to in future. 3TB is nothing as far as HDD space is concerned these days.

Re: Who Here Has Their Albums At OGG??

Reply #3
Wouldn't this count as spam?
TAPE LOADING ERROR

Re: Who Here Has Their Albums At OGG??

Reply #4
I think that 500 kbit/s is not enough for lossy files. I think that you need to use atleast 1000000 kbit/s. Just to be sure. /s
Main point of lossy files is transparency at smallest size. Vorbis (OGG is container, not format) achieves this at 192 kbit/s.
What you are doing makes zero sense.
gold plated toslink fan

Re: Who Here Has Their Albums At OGG??

Reply #5
When you purchase something in lossless form, you're getting full quality form; it's the full mix at that sample rate, to listen to or examine with as you please; you actually own it. Neither Vorbis nor Opus are designed for lossless encoding. Granted there are bad FLACs made from lossy encodes, but this idea of "good enough" is why MP3s are still sold for albums. It's up to you, the individual, whether or not to buy something less for the same price for the convenience of hard drive space. (And you wouldn't have to worry about hard drive space if you were able to get the CD, and rip to OGG.) But let's not kid ourselves that "good enough" is fully owning what you bought.

Re: Who Here Has Their Albums At OGG??

Reply #6
As with many of us, the original poster just wants some confirmation that they have done something acceptable.
In the past, I really wanted to keep all my music in a lossy format to save space, but loose the least.
My tests consisted of inverting the compressed result and mixing it into the original lossless source, then listening to the leftover sounds. Any sound left is what is either added to, or excluded from the original. For all formats, I used the highest bit rate offered. I concluded that that wma pro, was the cleanest. Then Vorbis. Then all the others had more noise, or cut off frequencies above 19kHz, so I didn't rank them.
In the end Vorbis won because the custom tagging worked where wma didn't. Now I just use it for all the songs I have that I have not been able to get a lossless copy for yet.
Processed audio in java and python.

Re: Who Here Has Their Albums At OGG??

Reply #7
My tests consisted of inverting the compressed result and mixing it into the original lossless source, then listening to the leftover sounds. Any sound left is what is either added to, or excluded from the original.
That test has very little to do with reality, and you should drop it completely. I mean, just turn down to 0.5; it will sound like you lost half the signal, but in reality you lost nothing (assuming a floating-point format like mp3).
A mosquito is inaudible next to full volume music, but audible by itself. Codecs have been smarter since ... ages.

As for Ogg Vorbis ... why choose Ogg Vorbis over Opus really, if you have compatibility? Oh, and there are regular polls of how many uses what format. Here is the most recent for lossies: https://hydrogenaud.io/index.php?topic=119927.0 .
WMA is largely abandonware as well.

Re: Who Here Has Their Albums At OGG??

Reply #8
Quote
As for Ogg Vorbis ... why choose Ogg Vorbis over Opus really, if you have compatibility? Oh, and there are regular polls of how many uses what format. Here is the most recent for lossies: https://hydrogenaud.io/index.php?topic=119927.0 .
WMA is largely abandonware as well.

When encoding to lossy I still prefer Ogg/Vorbis because it has better gapless playback support than Opus.


Re: Who Here Has Their Albums At OGG??

Reply #10
I have a couple of non-Android portable players around, so there is no compatibility for Opus.

Whenever possible, my music is stored lossless.

From there, it may be converted to AAC or Vorbis, to go to those portables.

Re: Who Here Has Their Albums At OGG??

Reply #11
Anybody else keep there music collection at OGG.  I like OGG it sounds really dam good and it is like 40% to 65% smaller than FLACs.  What I do it is download albums in FLAC then I down convert them to Q10 500kbs with Mediamonkey.  I have like 8500 albums if they were all in FLAC they would take up 3tb maybe more space I not a fan of albums taking 450mb to 850mb.  Hardcore FLAC downloaders are really anal about FLACs all other audio codecs are garbage to them S M H.  I bet these FLAC downloaders have never listened to 1 to 3 500kbs OGG albums all the way through.  IMO I can not tell the deference between OGG and FLAC I have listened to both on decent speakers. 

This makes little sense. Is it magic mushroom season in your part of the World. ???

Re: Who Here Has Their Albums At OGG??

Reply #12
It makes some sense. The OP will take 3:1 compression (ogg) over lossless. Likely as a single
collection as the advantage . Its similar to 320k mp3 encodes.

Re: Who Here Has Their Albums At OGG??

Reply #13
I had to read the OP's posting twice, but then it made sense to me as well.
You could argue at least just as well for lossless being 'shrooms. Especially if not backed up by listening tests.

Anyway, I have other priorities: ripping is to be done once and never again, a lossless archive can be retro-verified and sometimes repaired with CUETools, I know I don't have to do listening tests to find settings that work for every purpose, blah blah blah. Any lossy transcode of a lossless rip is disposable, which is also one reason that I use .opus - how many originals are delivered as .opus eh?

Re: Who Here Has Their Albums At OGG??

Reply #14
@DJboutit2
If you want to spend 500k on your encodes, I think you should consider WavPack in 500k range. It's better option (unless you need Vorbis for compatibility reasons).

Re: Who Here Has Their Albums At OGG??

Reply #15
@Porcus @Jaybee @ajp9
Sorry guys. I don't understand where my logic is going wrong. If the waveform produced by one codec, can more perfectly cancel out the original waveform than another, it will always be that much more imperceptibly different to the original than the other. Any change in volume is a result of the codec. Sure I'm dismissing results that I probably can't pick out with my own ears. I'm just trying to answer the question of: What codec keeps the most frequency and waveform while still being lossy?
I may be suffering from paranoia stemming from my teenage years when I couldn't tell the difference between low bit rate mp3 and a CD. Now over the years I seem to have developed an aversion to compression artifacts and am worried that, though I can't hear them now, I might start to hear them in the future. Just like what has happened to me every five years or so.
Processed audio in java and python.

Re: Who Here Has Their Albums At OGG??

Reply #16
Sorry guys. I don't understand where my logic is going wrong. If the waveform produced by one codec, can more perfectly cancel out the original waveform than another

"more perfectly" is not a matter of listening to the difference. A difference signal at volume v is "heard" together with the original at volume V, and if V is loud you won't hear v. Codecs and encoding algorithms are much smarter than just minimizing the difference. ("smarter" meaning: employ algorithms that give much better results)

Three examples:

1:  Imagine you got seven seconds of the following:
It is dead quiet and all you can hear is that single mosquito around your head - until all for sudden an elephant screams LOUD for three seconds.

Codec #1 understands that the mosquito takes X bits to encode losslessly, and allocates x bits for a lossy encoding of the mosquito; the elephant takes Y losslessly and gets y for a lossy version. The codec makes the best out of the mosquito in x bits and the best of the elephant in y, and merges them.
Codec #2 "knows" - i.e. is programmed to work such that - while the elephant screams, it will drown the mosquito sound completely. It is there, but you wouldn't know it. So it just cuts out the mosquito for those three seconds, and uses the bits for a loud and clear elephant.
Let us for the sake of the argument assume - though it does not work that way in practice - that it has enough bits then to give you a lossless elephant.
Codec #2 sounds better, but if you diff it from the original, there will be three seconds of magnificent mosquito in the diff, and you will be tricked into thinking it has discarded something essential. It hasn't. It knows that it is not essential, and the difference signal cannot tell that it isn't!


2: Movie scene. Some reporter speaks on the radio with noisy background.
Codec #1: speech takes X bits losslessly, gets x bits lossy. Noise takes Y bits, gets y. Make the best out of it.
Codec #2: noise is noise, if you got the frequency response correct. Do not actually encode it, just give the instructions: "produce noise with this frequency characteristic and put it on". That is cheap. More bits for speech. For the sake of the argument, assume that the speech is lossless.
If you listen to the diff to codec 1 you will get speech artefacts on top of noise artefacts.
Diff to codec 2: you will get loud noise. Possibly much louder than diff #1.


3: This codec #3 is just trying to be an asshole to you.
Codec #1:  does its best and does damn well because this time you have set it to high bitrate.
Codec #3: does like codec #1, but lowers the overall volume by 6 dB and shifts the entire stream 0.001 seconds because the original signal starts and ends with silence.
If you listen to the encoded signals with volume correction (ReplayGain) there is no way for you to tell the difference - they are equally good. But the diff signals will be vastly different.

Re: Who Here Has Their Albums At OGG??

Reply #17
What I Like in Ogg is the bass and the space in the sound. The reverb also. The high frequencies are not enough to satisfy my ears.

Re: Who Here Has Their Albums At OGG??

Reply #18
Thanks for taking the time to explain.
Now I know clearly where you're coming from. For example: You're saying that I shouldn't discredit codecs like opus just because they cut all frequency above 20kHz.
Here's where I'm coming from:
I like to do signal or sound processing on my music. One example: Slowing down music benefits form containing frequencies as high as possible.
Vorbis is one of the two lossy codecs I have tried that can keep the full frequency spectrum of a 48kHz recording. I figure my method would tell me if those high frequencies are just being filled with noise or actually reproducing them faithfully—Hence my post trying to agree with the original poster about how good vorbis is, and trying to explain how I came to that conclusion, rather than just stating it as an opinion.
Processed audio in java and python.

Re: Who Here Has Their Albums At OGG??

Reply #19
Now I know clearly where you're coming from. For example: You're saying that I shouldn't discredit codecs like opus just because they cut all frequency above 20kHz.
No, I didn't really speak about that. By "mosquito" I did not mean anything ultrasonic, I meant a sound you clearly hear in complete silence, but which you have no chance at hearing when it is masked by something loud. An good encoder will spend bits on a mosquito in silence but not a mosquito behind a loud scream. Hence the difference signal shall depend upon what you actually try to encode.
Also there are obvious issues with differences. For example, volume changes. If you just encode to half the volume the difference sounds big, but you have not really done much wrong. (No, you have not cut off 8 bits, most lossy codecs don't work that way at all.)

The bottom line is: Listening to the difference file gives you a completely wrong impression of what the encoded signal sounds like. You can think of it as a very first crudest method that was obsolete long before MP3.




Vorbis is one of the two lossy codecs I have tried that can keep the full frequency spectrum of a 48kHz recording.

AAC can sample at 192 kHz (though I think most implementations stop at 96?)
WMA Pro can sample at 96.
WavPack lossy can go .... much higher.
I don't know about OptimFrog hybrid.

And if by "48kHz" you mean sampledI at 48 kHz, several can. AC3, Musepack, MP2, MP3, ...


I figure my method would tell me if those high frequencies are just being filled with noise or actually reproducing them faithfully
Filling the highest frequencies with noise is actually a centuries-old method - makers of musical instruments have done so! In the presence of lower frequencies, "noise is noise"; replacing it with similar-sounding noise is quite forgiving.

But I think you are getting it very wrong. You should probably not hear anything if you try the following:
* Take your difference file (orig minus encoded)
* Apply a very aggressive high-pass filter at 20 kHz. Gives you a new treble signal.
* Create three new files: orig plus treble, orig minus treble, encode plus treble.
. The "orig + treble" should sound much brighter than the original. ABX them to hear if it does.
. The "encode plus treble" should have the same treble content as the original. The "orig minus treble" should have the same treble content as the encode. But can you ABX out any differences here?
. How loud is the "treble" file? What volume must you play it back at, to distinguish it from silence?

Re: Who Here Has Their Albums At OGG??

Reply #20
What codec keeps the most frequency and waveform while still being lossy?
There is no one-size-fits-all best lossy, especially when different encoders preserve parts of the spectrum very differently with different management of transients no matter how high you try to push the bit rate.

In terms of spectrum preservation, I can say that AC3 was very good at covering the entire spectrum for all channels at 320-448k, but it wasn't very good at temporal resolution/transients; AAC is far better at temporal resolution, but AAC encoders typically fall short at stereo fidelity (there can be a lot of content forced mono to keep spectrum); Vorbis and Opus can be just as good if not better at transients when smaller block sizes are used, but in practice aren't as good as AAC when it comes to synth or near-infrasonic frequencies, or seamless/gapless transitions between tracks.

I'm not surprised that Vorbis was suggested. It is still one of the higher ranking codecs in overall quality; my mobile collection is in OGG. But unfortunately aoTuV will still remove a lot of the higher frequencies at the highest setting. Even if the removed frequencies aren't audible due to masking, it's still lost information.

To the other point, good lossy encoders are never focused on matching the waveform. Psychoacoustic codecs generally don't achieve more than 8 bits of lossless accuracy to the original. You will always get better results with WavPack -b10. (But, then again WavPack is slow enough that I personally never bother using it to encode albums.)

Re: Who Here Has Their Albums At OGG??

Reply #21
What I Like in Ogg is the bass and the space in the sound. The reverb also. The high frequencies are not enough to satisfy my ears.

Nice TOS#8 violation from a moderated user.

Re: Who Here Has Their Albums At OGG??

Reply #22
What I Like in Ogg is the bass and the space in the sound. The reverb also. The high frequencies are not enough to satisfy my ears.

Nice TOS#8 violation from a moderated user.
I didn't write that codec x is better than codec y.
Science can sometimes take advantage of intuitions which, however, must be verified or disproved.
  Music gives us emotions that differ depending on the moment of our life. Some songs remind us of sad or happy moments in our life. Emotions cannot be quantified, except maybe by measuring the dopamine that we get from listening to music. Let me express my feelings once in 20 years since I've been following this website from the start, and for me it is the most reliable source for audio codec knowledge.
By the way, happy birthday HA and long live HA.

 

Re: Who Here Has Their Albums At OGG??

Reply #23
You will always get better results with WavPack -b10. (But, then again WavPack is slow enough that I personally never bother using it to encode albums.)

No need to go that high. You can get excellent results using WavPack lossy -b4 to -b6. You can also avoid slow encoding by using extra high mode (-hh). It's also high quality and it's very fast.