The problem is not the AVI or Ogg; it's that the vast majority of AVI *players* cannot handle a VBR codec. These players ignore all the sync timestamping, assuming the audio is coming in CBR.
Have a look at ww.hydrogenaudio.org for discussion of players that work properly.
Re:Double Blind Listening Tests... Here!
on
Ogg Vorbis RC3 Released
·
· Score: 5, Informative
ff123 from the r3mix.net/hydrogenaudio.org forums is conducting automated ABX double blind tests comparing Vorbis, mp3 (several encoders), AAC, WMA and MPC. The best part of this is... everyone can participate.
>I've been waiting for the 1.0 release of Ogg >Vorbis for a few years now
Really? Development only began in 1998, and nothing was even announced to the world until 2000 (right here in slashdot, a few months before we'd have liked word to leak out). No one has even known about it 'for a few years'.:-)
>Yes, it's a nice CODEC, but the development >timeline has been less than ideal for commercial >adoption.
MPEG required ~10 years. Our code has been production grade since beta1, and every bitstream make since May 8th, 2000 will work forever. That's less than two years from beginning to frozen. The '1.0' label is just waiting on a paper list of features that has grown over time.
Hrumf. We should have just called 'rc1' 1.0 and no one would have known the difference.
> Ogg Tarkin is still in
> extremely early development,
very true.
> without even alpha code to show for the effort.
Running Tarkin code exists; we actually have three competing implementations, two in CVS, and the 'w3d' module at cvs.xiph.org is the current frontrunner (and the one we're actively developing).
My first thought when I saw this article was, "Oh boy... this should get ugly and yet remain light and fluffy" but all the posts I've seen (reading at +2) have been pretty good. I don't really have much of anything to add other than 'we have some really nice quality improvements in store for rc3', mainly new noise estimation metrics, lots of stereo fixes, and other random nicities (like 20kHz cutoff at 128...)
BTW, for more in depth discussion that has been ongoing, have a look at the forums at r3mix.net and the Ogg-specific forums at Hydrogen Audio. I keep up with both forums, and the folks there tend to make prerelease build binaries available for people to play with. For up-to-date detailed information without the overhead of the Vorbis-dev list, those are the places to go.
One more link for folks who want to know more: The beginning of the document describing Vorbis stereo discusses good terminology and qualification of subjective fidelity. It's nothing new to most posters I expect, but it might help keep the discussion consistent.
up to 96kbps lossless stereo to 4/6kHz (depending on block), point/6phase/lossless stereo above that
128 lossless stereo to 4/6kHz, 6phase/lossless to 11kHz, point/6phase/lossless stereo above
160 lossless stereo to 4/6kHz, 8phase/lossless to 11kHz, point/8phase/6phase/lossless above
192 lossless stereo to 4/6kHz, 8 phase/lossless above
256+ lossless coupling
The trigger between different stereo models is currently amplitude based (how far the spectral energy is from the noise floor). The trigger threshholds are also increasingly conservative as bitrate increases.
you can't yet (it's not a simple switch, it affects the entropy model to use). I *will* make user-specification of stereo model possible as soon as possible.
I ran across that story after a 36 hour sleepless (well, ok, a:20 nap) hacking session pushing out rc2. Gah. Nasty.
I was kinda surprised that it was Ogg that drove away my wife (here I thought it was my awful personality. What a relief),that I've been habitually unemployed, destitute, and that I don't even own a proper desk. Ah-heh.
Unfortunately, the green shag carpeting part is completely true. Damn you, the 1970's! Will you never die?!
Monty
(Oh well, at least it seemed to be positive on Ogg)
OK, since only about half of the mail we get is about the name 'Ogg Vorbis', it's clearly time to karma-whore a popular subject and open this can of worms one more time.
Our "The Name Sucks!"/"The name Rulez!" mail ratio is about 50/50. Some of you have threatened to kill us if we change the name, some of you have threatened to kill us if we don't. So you're gonna hear what I think about it. I'm not going to waste the opportunity my minor fame gives me for a healthy round of peer-mockery.
<tongue-in-cheek>
<neeneer-neener>
I Like The Name. I Wrote the Software. The Name Stays.
</neener-neener>
But there's more to this story than 'nyah nyah'. The 'rename Ogg!' forces have provided me with some of my favorite mail ever. I recall fondly the guy who went on, in great detail, why 'Ogg Vorbis' sucks, and that I must adopt 'a cutting edge, truly kick-ass name like "FreeMP3"!!!!!'
As for 'Ogg Vorbis', I hadn't really meant the 'Vorbis' part to get tacked on. The name of the format is Ogg. Just Ogg. Vorbis happens to be the first codec. Had 'Vorbis' been perhaps one more syllable (like, say 'Sorensen'), we wouldn't have this problem. People would just call it 'Ogg' like God (that's me) intended. Of course, particularly obsessive people *do* occasionally say 'QuickTime Sorensen', but they don't get invited to parties much, and when invited, they are shunned. 'Course they're usually just arguing with the punch bowl so shunning is easy.
I don't want my users to be shunned at parties, so I'm gonna help you out here. Just call it 'Ogg'. Ogg is a good, simple, very satisfying word.
It makes a good noun, a better verb and can even be used effectively in a curse. It is a real word and contains no numbers. It has only two unique characters, making it simpler than mp3. It is only one syllable, making it shorter to say than mp3. If you still can't handle it, try reboot-reinstall.
</tongue-in-cheek>
I talked to him before the story went in the Post
(thus the little sentance about 1.0) and he said
it was neither blind nor structured (and I gave
him my concerns about that on Wednesday)...
In over a year of press coverage, we've gotten one relatively negative quality review... and Slashdot covers that one. OK, fine, I better say something about it.
I spoke to the reporter who did these 'listening tests'... the tests were not structured, were not controlled, were not blind or double blind... in short, the reviewers knew exactly which codec sample they were listening to at all times. The reviewers also came into the test with already formed opinions about which codec was best and-- surprise--- the test seemed to confirm what they already 'knew'.
In ABX blind testing, Vorbis usually surprises the tester, for example, from our happy maniac friends at r3mix forum (MPEGplus and Vorbis won this ABX test first and second place over mp3pro, AAC, WMA8, TwinVQ and ATRAC3):
"Also, I must admit, both Ogg Vorbis and MPEGplus suprised me. I wasn't expecting to see them both so near the top. I've gotta say, I really had to guess alot for both the of these tests. Pretty darn close[...]
For MPEGplus, I was guessing almost every time, but I 'knew for sure' on Vorbis."
That is, he knew for sure... incorrectly. A brilliant example of the power of suggestion.
In any case, after talking to the Post reporter he feels a little sheepish about the whole thing... he thought he balanced the article by mixing positive traits of the openness of the code with a critical quality review and has agreed to be more fair to the first 1.0 encoder release candidate.
is that they are trying to avoid anything which has been patented so far.Meaning some of the best algorithms.
Not really. All of the annoyances so far have been trivial. For a patent to be valid, it has to narrow to a specific innovation. Yes, that's been abused alot recently, but the research and commercial organizations we're talking about here (Fraunhofer, Dolby, AT&T, Philips, NTT, etc) don't actually have very abusive patents. it's just the way they're using them; it is nearly impossible to make an unencumbered MP3 derivative. But if you're not doing things the MP3 way (we're not), sidestepping the patent issue isn't that hard.
But there's also been other good lossless compression work done lately in the form of FLAC and Monkey's Audio. Squish will be updated after I've had some time to commisserate with the other projects and steal technology from them-- AHEM! I mean share what we know.
> MP3 encoded the fastest, in around 30 seconds.
> WMA took 50 seconds. Yet it took nearly 3
> minutes to encode the song into OGG format.
Umm.. then I don't think you were using beta 4. Since we're comparing against *beta 4* now (and the speed/quality improvements is brings...)
your review doesn't mean that much:-(
If you have problems with beta 4, please send us example samples.
Every Ogg + XMMS crash bug we know about is fixed... in XMMS CVS. Because there have been no major releases of XMMS recently, most people are still downloading and using buggy 'official' XMMS packages. Thus, folks are unwittingly reporting bugs that have been fixed for weeks or months. A number of the crash bugs were in the Ogg plugin, but a number have also been in XMMS itself, which Ogg simply had the bad luck to tickle. Upgrading the plugin alone can't save you.
So, first grab and build XMMS from XMMS CVS. It actually builds cleanly with minimum fuss. At that point, if you get a crash playing Ogg, the XMMS developers and we would very much like to hear about it.
Napster pledged support early on but doesn;t seem to have gotten around to it.
OpenNap, however, does support Ogg. Just use a client (AudioGnome or a recent Gnapster) that supports.ogg, and OpenNap servers will happily take them. And, yes, I've used it so it does work, not just heresay;-)
In beta 2, oggenc's default encoding mode was 160kbps. In beta 4, the default is 128kbps. There *is* an audible difference, even with the improvements to 128kbps since earlier betas.
Actually, that sentence was out of a longer letter that was an RMS reply to Jack. There was some cut-and-thrust debate going on, and RMS's quote was a sensible reply to an earlier assertion made by Jack. The whole "Linux operating system" thing didn't just come out of the blue, the interview just ended up with it edited in a somewhat unfortunate sounding way for RMS;-)
Ogg Vorbis is free and unencumbered, and although I have no real idea how the lawsuit is going, assuming the Ogg side wins, we will have a high quality free [speech|beer]
audio compression format. No mp3, no Fraunhofer lawsuits.
The lawsuit!?!? What lawsuit?!?! We haven't been sued by anyone, let alone Fraunhofer or Thomson. The only thing that's happened so far (aside from minor industry FUD) is that a Thomson VP shot his mouth off on something about which he wasn't well informed, and a spokesman later backed off on the statement.
No one has sued anybody! Geez, the rumors Slashdot can start!
Slashdot Mirror
For chatting with developers real-time (but no guarantee when we'll be there), catch us on #vorbis at irc.openprojects.net.
Monty
The problem is not the AVI or Ogg; it's that the vast majority of AVI *players* cannot handle a VBR codec. These players ignore all the sync timestamping, assuming the audio is coming in CBR.
Have a look at ww.hydrogenaudio.org for discussion of players that work properly.
Monty
Monty
If you want to take the listening test yourself, read the instructions and jump in. For now, there's also a page of interim results, but to quote ff123, "Major conclusion: I need more listeners!"
Monty
>I've been waiting for the 1.0 release of Ogg >Vorbis for a few years now
:-)
Really? Development only began in 1998, and nothing was even announced to the world until 2000 (right here in slashdot, a few months before we'd have liked word to leak out). No one has even known about it 'for a few years'.
>Yes, it's a nice CODEC, but the development >timeline has been less than ideal for commercial >adoption.
MPEG required ~10 years. Our code has been production grade since beta1, and every bitstream make since May 8th, 2000 will work forever. That's less than two years from beginning to frozen. The '1.0' label is just waiting on a paper list of features that has grown over time.
Hrumf. We should have just called 'rc1' 1.0 and no one would have known the difference.
> Ogg Tarkin is still in
> extremely early development,
very true.
> without even alpha code to show for the effort.
Running Tarkin code exists; we actually have three competing implementations, two in CVS, and the 'w3d' module at cvs.xiph.org is the current frontrunner (and the one we're actively developing).
But this is not release grade code.
Monty
BTW, for more in depth discussion that has been ongoing, have a look at the forums at r3mix.net and the Ogg-specific forums at Hydrogen Audio. I keep up with both forums, and the folks there tend to make prerelease build binaries available for people to play with. For up-to-date detailed information without the overhead of the Vorbis-dev list, those are the places to go.
One more link for folks who want to know more: The beginning of the document describing Vorbis stereo discusses good terminology and qualification of subjective fidelity. It's nothing new to most posters I expect, but it might help keep the discussion consistent.
Happy hacking,
Monty
xiph.org
call it what you like. I reserve the right the say 'neener' at you if you complain it's too long.
Monty
Coupling is there.... you just can't override the hardwired stereo model selection right now.
Monty
His beta 4 files will continue to work forever... Nothing about this release makes old files incompatable.
:-)
...Or shall I just stop improving things at this point?
Monty
Ooo, point. I have a little code cleanup to do for 68k...
Monty
in rc 2, the hardwired stereo models are:
up to 96kbps lossless stereo to 4/6kHz (depending on block), point/6phase/lossless stereo above that
128 lossless stereo to 4/6kHz, 6phase/lossless to 11kHz, point/6phase/lossless stereo above
160 lossless stereo to 4/6kHz, 8phase/lossless to 11kHz, point/8phase/6phase/lossless above
192 lossless stereo to 4/6kHz, 8 phase/lossless above
256+ lossless coupling
The trigger between different stereo models is currently amplitude based (how far the spectral energy is from the noise floor). The trigger threshholds are also increasingly conservative as bitrate increases.
Monty
you can't yet (it's not a simple switch, it affects the entropy model to use). I *will* make user-specification of stereo model possible as soon as possible.
Monty
I ran across that story after a 36 hour sleepless (well, ok, a :20 nap) hacking session pushing out rc2. Gah. Nasty.
I was kinda surprised that it was Ogg that drove away my wife (here I thought it was my awful personality. What a relief),that I've been habitually unemployed, destitute, and that I don't even own a proper desk. Ah-heh.
Unfortunately, the green shag carpeting part is completely true. Damn you, the 1970's! Will you never die?!
Monty
(Oh well, at least it seemed to be positive on Ogg)
Our "The Name Sucks!"/"The name Rulez!" mail ratio is about 50/50. Some of you have threatened to kill us if we change the name, some of you have threatened to kill us if we don't. So you're gonna hear what I think about it. I'm not going to waste the opportunity my minor fame gives me for a healthy round of peer-mockery.
<tongue-in-cheek>
<neeneer-neener>
I Like The Name. I Wrote the Software. The Name Stays. </neener-neener>
But there's more to this story than 'nyah nyah'. The 'rename Ogg!' forces have provided me with some of my favorite mail ever. I recall fondly the guy who went on, in great detail, why 'Ogg Vorbis' sucks, and that I must adopt 'a cutting edge, truly kick-ass name like "FreeMP3"!!!!!'
As for 'Ogg Vorbis', I hadn't really meant the 'Vorbis' part to get tacked on. The name of the format is Ogg. Just Ogg. Vorbis happens to be the first codec. Had 'Vorbis' been perhaps one more syllable (like, say 'Sorensen'), we wouldn't have this problem. People would just call it 'Ogg' like God (that's me) intended. Of course, particularly obsessive people *do* occasionally say 'QuickTime Sorensen', but they don't get invited to parties much, and when invited, they are shunned. 'Course they're usually just arguing with the punch bowl so shunning is easy.
I don't want my users to be shunned at parties, so I'm gonna help you out here. Just call it 'Ogg'. Ogg is a good, simple, very satisfying word.
It makes a good noun, a better verb and can even be used effectively in a curse. It is a real word and contains no numbers. It has only two unique characters, making it simpler than mp3. It is only one syllable, making it shorter to say than mp3. If you still can't handle it, try reboot-reinstall.
</tongue-in-cheek>
Monty
xiph.org
???
I talked to him before the story went in the Post
(thus the little sentance about 1.0) and he said
it was neither blind nor structured (and I gave
him my concerns about that on Wednesday)...
OK, I better mail him again and see what's up.
Monty
I spoke to the reporter who did these 'listening tests'... the tests were not structured, were not controlled, were not blind or double blind... in short, the reviewers knew exactly which codec sample they were listening to at all times. The reviewers also came into the test with already formed opinions about which codec was best and-- surprise--- the test seemed to confirm what they already 'knew'.
In ABX blind testing, Vorbis usually surprises the tester, for example, from our happy maniac friends at r3mix forum (MPEGplus and Vorbis won this ABX test first and second place over mp3pro, AAC, WMA8, TwinVQ and ATRAC3):
That is, he knew for sure... incorrectly. A brilliant example of the power of suggestion.In any case, after talking to the Post reporter he feels a little sheepish about the whole thing... he thought he balanced the article by mixing positive traits of the openness of the code with a critical quality review and has agreed to be more fair to the first 1.0 encoder release candidate.
Monty
We talked via email, this is beta 4, just to make sure every one else sees it (ie, my previous response was wrong).
The low speed is a little surprising, but not as big a deal as the reported artifacts. I'm looking at it.
Monty
xiph.org
Not really. All of the annoyances so far have been trivial. For a patent to be valid, it has to narrow to a specific innovation. Yes, that's been abused alot recently, but the research and commercial organizations we're talking about here (Fraunhofer, Dolby, AT&T, Philips, NTT, etc) don't actually have very abusive patents. it's just the way they're using them; it is nearly impossible to make an unencumbered MP3 derivative. But if you're not doing things the MP3 way (we're not), sidestepping the patent issue isn't that hard.
Monty
xiph.org
Squish will reappear.
But there's also been other good lossless compression work done lately in the form of FLAC and Monkey's Audio. Squish will be updated after I've had some time to commisserate with the other projects and steal technology from them-- AHEM! I mean share what we know.
(sorry been reading Microsoft rants)
Monty
xiph.org
Evil Anti-Capitalist Anti-Innovation Anti-American Open Source Pinko Peckerhead Freak
> MP3 encoded the fastest, in around 30 seconds.
:-(
> WMA took 50 seconds. Yet it took nearly 3
> minutes to encode the song into OGG format.
Umm.. then I don't think you were using beta 4. Since we're comparing against *beta 4* now (and the speed/quality improvements is brings...)
your review doesn't mean that much
If you have problems with beta 4, please send us example samples.
Monty
xiph.org
Every Ogg + XMMS crash bug we know about is fixed... in XMMS CVS. Because there have been no major releases of XMMS recently, most people are still downloading and using buggy 'official' XMMS packages. Thus, folks are unwittingly reporting bugs that have been fixed for weeks or months. A number of the crash bugs were in the Ogg plugin, but a number have also been in XMMS itself, which Ogg simply had the bad luck to tickle. Upgrading the plugin alone can't save you.
So, first grab and build XMMS from XMMS CVS. It actually builds cleanly with minimum fuss. At that point, if you get a crash playing Ogg, the XMMS developers and we would very much like to hear about it.
Monty
xiph.org
Napster pledged support early on but doesn;t seem to have gotten around to it.
.ogg, and OpenNap servers will happily take them. And, yes, I've used it so it does work, not just heresay ;-)
OpenNap, however, does support Ogg. Just use a client (AudioGnome or a recent Gnapster) that supports
Monty
xiph.org
In beta 2, oggenc's default encoding mode was 160kbps. In beta 4, the default is 128kbps. There *is* an audible difference, even with the improvements to 128kbps since earlier betas.
Monty
xiph.org
Actually, that sentence was out of a longer letter that was an RMS reply to Jack. There was some cut-and-thrust debate going on, and RMS's quote was a sensible reply to an earlier assertion made by Jack. The whole "Linux operating system" thing didn't just come out of the blue, the interview just ended up with it edited in a somewhat unfortunate sounding way for RMS ;-)
Monty
xiph.org
The lawsuit!?!? What lawsuit?!?! We haven't been sued by anyone, let alone Fraunhofer or Thomson. The only thing that's happened so far (aside from minor industry FUD) is that a Thomson VP shot his mouth off on something about which he wasn't well informed, and a spokesman later backed off on the statement.
No one has sued anybody! Geez, the rumors Slashdot can start!
Monty
http://www.xiph.org/