After doing an informal (128k) listening test, I have concluded that I prefer Beta4 over RC2. The 16kHz low-pass on the RC2 encoder makes it sound like FM radio. Both encoders SEEM to have a couple of dB bump at 10kHz. JT --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
On Wed, Aug 15, 2001 at 07:39:15PM -0400, thobois@mindspring.com wrote:> The 16kHz low-pass on the RC2 encoder makes it sound like FM radio.You will be free to turn it off soon; lowpass will be user configurable. Personally, I find the least bit of high end instability (swishiness) to be far more objectionable to a lowpass. You'd likely never hear the swishiness on speakers, but on good cans, it's more obvious. Thus the decision to make the audio <16kHz rock solid as opposed to trying hard to encode the whole range. A number of rc2 users find 64kbps quite good. Even though it's my own software, I find it nearly unlistenable (but still better than WMA ;-) Anyway, that's my rationale, and my own specific subjective coloration that led to the decision for a 16kHz lowpass at 128. It's not necessary, I just like it better.> Both > encoders SEEM to have a couple of dB bump at 10kHz.Not quite. You'll see a slight bump in rc2 at about 11kHz due to unnormalized point stereo. b4 did not perform coupling. Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
16 kHz low-pass... I prefer Beta 4 and RC1 than RC2, too... I didn't test low and middle frequency but when high is similar to mp3 that I think it is going in wrong side... Fix it :> ----- Original Message ----- From: <thobois@mindspring.com> To: <vorbis@xiph.org> Sent: Thursday, August 16, 2001 1:39 AM Subject: Re: [vorbis] RC2 worse than RC1 and Beta4> After doing an informal (128k) listening test, I have concluded that I > prefer Beta4 over RC2. > The 16kHz low-pass on the RC2 encoder makes it sound like FM radio. Both > encoders SEEM to have a couple of dB bump at 10kHz. > > JT > > > --- >8 ---- > List archives: http://www.xiph.org/archives/ > Ogg project homepage: http://www.xiph.org/ogg/ > To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' > containing only the word 'unsubscribe' in the body. No subject is needed. > Unsubscribe messages sent to the list will be ignored/filtered.--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>You will be free to turn it off soon; lowpass will be user configurable.Not sure how difficult or practical this would be, but a couple of different slope options on the low-pass would be welcome. M.r. Rage --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>> Both>> encoders SEEM to have a couple of dB bump at 10kHz. >Not quite. You'll see a slight bump in rc2 at about 11kHz due to >unnormalized point stereo. b4 did not perform coupling. I did a little test to try to figure out what I'm hearing. Took a wav file I made with 8kHz, 10kHz and 12kHz sine wave test tones (all the same amplitude) and converted it to ogg 128k b4. Then I flipped the phase of the ogg and added it to the original wav, figuring this would show the artifacts induced in the compression. Looking at the resulting waveform, 10k was about twice the average peak amplitude of the 8k, with 12k about halfway in between. --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>That methodology is flawed. You will see an increase in noise in a>ramp at about 11kHz, following the ATH, because the ear is less >sensitive there. This does *not* imply a net increase in >spectral energy. Fine. My example was never intended to be anything more than an illustration of what I'm hearing. I've been a professional audio engineer since 1992, so I think I have a pretty good idea about what I'm hearing. Based on what I'm hearing in the upper sibilance range, I think there's another bump at around 7k. (Still using b4 128k) --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
I haven't got special test programs or hardware tools, but in my listening tests the RC2 seems to be worse than the beta4 (I'm missing the sharp/clear high sounds from the RC2 using the highest bitrates) But in the latest CVS (aug. 17.) there are some corrections in the psy.c and floor1.c and it seems with these modifications the sound quality is better (tested on high bitrates, but now the highest bitrate is 280kbit... you should calculate again the encoder tables...). I hope the developer(s) will find and correct every quality bugs... and we will get a perfect encoder in the far future... Attila --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>I'm not arguing you're not hearing it actually... just that the>proffered explanation is not correct. The fact that you identify 7kHz >and 10kHz matches the code in fact; those two places are exactly >where the code is relaxing the resolution given to noise (the bands >above 7-8kHz, and again at the bands above ~11kHz). >rc2 *does* currently increase the real spectral noise energy at 11kHz >plus. This is due to point stereo not having the code to actually do >the correct trig when rotating phase. It fudges it for speed and adds >up to a factor of sqrt(2) to each spectral line it rotates. Many, >many picky people are hearing that and I'm going to do the rotation >correctly in rc3. >b4 may be giving a similar impression, but the perceived increase in >energy is false. What many people describe as an artificial crispness >in the b4 high end is actually due to the opposite reason than users >claim. It is not 'crisping' noise in the high spectrum, it is >actually smearing it. Rather than hearing exact harmonics above 10k >(for example, in voice where such harmonics are important), one hears >the harmonics somewhat fuzzed. Because the harmonics are not clean, >they're more noticable (not exactly mathing lower harmonics or >fundamentals), and thus a user could think that's a boost. >Careful spectral analysis should show this explanantion to be correct. Thanks for the detailed response, Monty. What you describe is exactly what I'm hearing. I was going to mention that the artifacts sounded similar to the "Aural Exciter" effect. I do need to mention that the fuzzy harmonics are far less objectionable than the over-ring heard with other codecs. Vorbis is a great sounding codec and I appreciate you addressing my complaints. Viva Vorbis! --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Gian-Carlo Pascutto wrote:>If you are using high bitrates you are just seeing placebo >effects. There are changes in the point stereo calculations >in CVS, but point stereo is not used in the higher bitrate >modes.1. Interesting... encoding the same song in 350kbit/s bitrate mode I get: with the RC2: 273 kbit/s OGG file with the latest CVS: 261 kbit/s OGG file The sound quality is also different, better in the 2. case. 2. a kind of channel coupling (phase stereo?) is used on the high bitrates too (or not?), and I hear that it also filters out a part of the high sounds from the music and I lose from the stereo effect too a little bit. (I encoded with the non-coupled mode too). Maybe it's not totally lossless yet (if I remember well, you said that high bitrates use a lossless channel coupling mode, but I hear that the coupled and non-coupled songs are different (or the 'lossless' doesn't mean this?)). 3. there is noise masking in RC2 on the high bitrates, the beta4 doesn't use it on high bitrates... I know more filters/maskings reduce the bitrate, gives more dynamic and frequency range, but maybe we lose some sharp/nice sounds on this way. BUT I will repeat my all tests, and if these were mistakes, I will beg your pardon... (the differences/quality bugs are very small, I need a longer time to make sure of these bugs, but maybe these are compiler/optimizer differences only (I use Win32/MSVC))>If you do listening tests, always do a blind test. Your >mind _will_ play tricks on you otherwise. In fact, RC2 >sounding worse than beta4 for you may very well >be just because you read on this list that is what other >people thought. There are others who percieve it just >the opposite.I did dare to share my own opinion, because I saw that other people also think this. And we said that in the listening test the beta4 sounds nicer, even if you said (and perhaps we hear it too) that it produces a false output... And I don't want to argue with you/developers, you know better what you do... Aleksandar Dovnikovic wrote:>the only big problem in this mode is pre-echo.When will be corrected this? I forget to say that I (and I think so a lot of other people) need a near-lossless audio encoder (in 1:5 compression), and I've thought that this will be the Ogg Vorbis.... regards Attila --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
There seems to be another multiple format test @192kbps on r3mix. If you got better luck than I, you could download the files from: http://66.96.216.160/cgi-bin/YaBB.pl?board=c&action=display&num=998264509 R3mix Forum (r3mix.net/forum or cd-rw.org/forum or r3mixforum.cjb.net)-Technical issues, Tests & Experiments-'Archive quality' test with other codecs (192kbps) On Sun, 19 Aug 2001, Attila Padar wrote: Attila> I know more filters/maskings reduce the bitrate, gives more Attila> dynamic and frequency range, but maybe we lose some sharp/nice Attila> sounds on this way. Yes, I agree with what Attila says: RC2 doesn't sound as sharp as LAME. However, even beta4 didn't maintain the original sound at high notes, neither. Taking whether LAME or Vorbis is still remained as the matter of tase, I think, though. I can't tell which is better in all the aspects. (After beta 4, I just simply compare Ogg and LAME and take which I like). On Sun 19 Aug 2001 Gian-Carlo Pascutto <gcp@sjeng.org> wrote: Gian> :) I had to fiddle with it in my tuned mode because I didn't like Gian> the results. Well, what you meant was only masking part, then. Sorry for the dummy question. Attila> And I don't want to argue with you/developers, Attila> you know better what you do... Gian> I twiddle with numbers in the vauge illusion I know Gian> what I'm doing ;) Developers know theories but we know what we hear. Some theories are convincing but the others are not. We ignorant mortals only can be subjective. Well, I am sorry but I do not seem to reach the files. :-( Someone who has got better ears? ___________________________________________________________________ To get your own FREE ZDNet Onebox - FREE voicemail, email, and fax, all in one place - sign up today at http://www.zdnetonebox.com --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
---- On Mon 20 Aug 2001, Gian-Carlo Pascutto <gcp@sjeng.org> wrote:> I have put the files on a faster and much more reliable server > since. (and updated the links in the link above)Thanx. ...let's see...> My interest is getting an idea of how many people can > tell the codecs on a hard sample at high bitrates, and > some data about how the codecs perform there.How about making some more detailed options for Ogg? Although I once suggested the programmers do this, I’ll repeat it. The reason is below. 1.Promotion purpose. This would naturally introduce hardcore codec fans to Ogg. So, give them something to play with. ;-) 2.Development purpose. This is related with the reason above. Once hardcore fans are accustomed to switches, they will try then out and find the flaws with some options, which would allow the developers to identify the cause of it and save them from wild goose chases. It is needless to say that the above is based on the success of LAME project. For example, if I need to state one of them at all, the ‘nspsytune’ switch, which was introduced to the format about a year ago, is now the part of official r3mix option. As long as I am concerned, I don’t think it is completely useless to tell programmers what we ordinary computer users want in this mailing list. The reasons are simple: Open source project programmers and small companies cannot afford comprehensive researches like gigantic enterprises can do. Media Jukebox has a unique message board system at their home page, which seems to be quite successful. Therefore, if we write what we want here, not only programmers but also private companies who are thinking of using Ogg on their software will know what they need to improve to attract users. Increasing users means increasing fund and increasing the possibility of the success of the project. I wrote about problems with Peter’s plug-ins here while ago since it must be one of the key programs for the popularisation of Ogg. Now, if the plug-in (1.14b) works on the other peoples’ computers as well as on my computer, it seems to be already prepared to commercial usage. Really a good job, Peter. Honestly, I am impressed. Some people also wrote about the possibility of All-in-One software for Windows, which, I think, is important as well. For not so many people rip their CDs into wav, encode it with Oggdrop and edit comments. At least, a program allows the automation of the procedure above is needed. For the future improvement of Ogg, if the program also has a comprehensive net radio function, it should be more desirable. So, I don’t think speaking up what we want is a purely selfish act. Some programmers on Linux do not know much about what Windows users think. They know much about computers, which makes it difficult for them to understand many ignorant computer users like me. We are ignorant of their thoughts and they are ignorant of our thoughts. (They couldn’t understand why Windows were successful despite the technical flaws. It could be said that Microsoft merchandised this gap between the recognition of technical experts and that of ordinary users. Reducing the gap would make Microsoft uncomfortable, then.) This is why the user mailing lists exist. Different from spyware style on-sided patent enterprises communication, we can communicate through open source projects. In the world of open source, programmers are not slave workers under the pressure of marketing and users are not consumer cattle for gigantic enterprises anymore. In this way, we can relate ourselves with the future against marketing-centred economy, in which we are excluded to participate in except as slaves. (Some people seem to be talking about Warez, here. However, let’s forget about Warez guys, since it just makes and keeps them pity thieves, of whom monopolists are not afraid.) ___________________________________________________________________ To get your own FREE ZDNet Onebox - FREE voicemail, email, and fax, all in one place - sign up today at http://www.zdnetonebox.com --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.