Hi there In the past, i have used lame to encode high quality mp3-files (vbr 1, bitrate ~ 192kbit). I tend to switch to ogg with Quality 4 or 5, but i noticed, that many ogg-files tend to produce too much high frequencies response. In many cases, this is very noticeable. For my opinion, i cannot accept this worse frequence reponse. I have used latest version of ogg (OggEnc v1.0, precompiled binary). To demonstrate this, i have made a short sample: http://home.t-online.de/home/520022073876/highfreq.zip The length is 2MB. It includes the original wav-file (5sec length) and ogg-files from quality 0 up to 10. This effect is very noticable from q0 to q4, it is then reduced more and more from q5 up to q10. Quality 10 sound ok, the frequence response is flat. But the filesize is much too high for using it. I hope, this can be fixed in a further version of ogg. Sorry for crossposting in vorbis@xiph.org and vorbis-dev@xiph.org, but i don't know which list is the better place for this topic. Frank --- >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-dev-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, 09 Oct 2002 20:32:02 +0200 Frank Grotelueschen <fgro@gmx.de> wrote:> Hi there > > In the past, i have used lame to encode high quality mp3-files > (vbr 1, bitrate ~ 192kbit). I tend to switch to ogg with Quality 4 > or 5, but i noticed, that many ogg-files tend to produce too much > high frequencies response. In many cases, this is very noticeable. > For my opinion, i cannot accept this worse frequence reponse. > I have used latest version of ogg (OggEnc v1.0, precompiled binary). > > To demonstrate this, i have made a short sample: > http://home.t-online.de/home/520022073876/highfreq.zip > The length is 2MB. > > It includes the original wav-file (5sec length) and ogg-files from > quality 0 up to 10. This effect is very noticable from q0 to q4, > it is then reduced more and more from q5 up to q10. > Quality 10 sound ok, the frequence response is flat. But the > filesize is much too high for using it. > > I hope, this can be fixed in a further version of ogg. > > Sorry for crossposting in vorbis@xiph.org and vorbis-dev@xiph.org, > but i don't know which list is the better place for this topic. > > Frank >I can hear it too from -q0 up to -q5... -q6 isn't affected... Haven't ABX'd it yet but it's quite obvious The higher frequencies seems to be a bit higher as in the orginal Bye --- >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-dev-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.
Sebastian, <p>please don't reply to crossposts with crossposts. :P I think vorbis@ is the correct list for this thread. <p>Moritz --- >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.
Frank Grotelueschen wrote:> It includes the original wav-file (5sec length) and ogg-files from > quality 0 up to 10. This effect is very noticable from q0 to q4, > it is then reduced more and more from q5 up to q10. > Quality 10 sound ok, the frequence response is flat. But the > filesize is much too high for using it.I'd better not try to explain myself, or I'd likely come off sounding like a goof. :) Check out these: http://www.xiph.org/archives/vorbis/200202/0283.html http://www.xiph.org/archives/vorbis/200209/0026.html http://www.xiph.org/ogg/vorbis/doc/stereo.html AFAIK, this problem should not be observed above -q 6 in 1.0, where lossless stereo coupling takes effect. Here's another interesting thread I'll throw at you: http://www.xiph.org/archives/vorbis/200207/0178.html -- -:-:- David K. Gasaway -:-:- XNS : =David K Gasaway -:-:- Email: dave@gasaway.org -:-:- Web : dave.gasaway.org --- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>Ogg with at least quality 6 produces files with near the same filesizeas>lame (seetings above) or higher. So, there is no need to change to ogg.I found it quite amusing to read the 128 kbps results in the c't test. Both the professional panel (composers, musicians, sound engineers) and the reader panel had problems to find the original uncompressed sound file among the compressed samples. There were only a small difference in the rating of the uncompressed sample, Vorbis and MP3. <p>So you are basically right. For "archival quality" at the bitrates you are using, there is absolutely no reason to switch from MP3 to Vorbis. Stay with MP3, as you can assume that this format will be much more often supported than Vorbis. Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>Also, you're free to make/tune your own Vorbis encoder. :)How are you supposed to do that, as the file format is not correctly documented? Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>Because the file format *is* correctly documented. >http://www.xiph.org/ogg/vorbis/docs.htmlWell, then try to implement a decoder following the file format specification you refer to and see what you get. I've pointed out several points where the actual libvorbis implementation differs from the file format specification, but there doesn't seem to be much interest in correcting the mistakes, although Monty did correct the first two I reported. Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>Anyways, this all is not the point. You are _free_ to do your ownVorbis>implementation. You are _not free_ to implement MP3, for example, >because you *will* infringe patents. And, what do you expect from Xiph? >I don't think their priorities are unreasonably set. Everything takes >time; this includes documenting something that has been in developement >for about 10 years.The fact that I don't have to pay any licencing fees or tolls to swim across the Atlantic doesn't make it more feasible. And: The patents relevant to MP3 are probably only valid in the USA, and I could have chosen the easy way and use any of the several available MP3 libraries instead of at least trying to use (and support) the effort Xiph has placed into Vorbis. The problem is, that I cannot use libvorbis for a project I am working on, and I made the mistake to invest quite a lot of time on developing a new Vorbis decoder before I realized that it is not possible to get it to a production quality level. I did indeed expect from Xiph to take care about the correctnes of the specification when they are distributing one. I did also expect that there are reference files available to test new decoder implementations. There are not, and the available encoders don't make use of all parts of the specification, making self-generated vorbis files not very suitable for tests. Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:> Why can't you use libvorbis?Can not was perhaps a too strong statement, so I'll reduce it to will not. The reason is that I am developing a Java application, and using libvorbis would mean that I either loose platform independence, have to provide binaries myself for all platforms or expect from the user to compile a JNI-wrapper (a layer implemented in C for accessing libvorbis from Java). Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>isn't there a Java implementation of Vorbis? (or did that stop at >RC2?) Jorbis?jOrbis is not really a Java implementation, but a "translated" version of libvorbis. It uses rather nasty hacks to achieve C functionality in Java (e.g. emulation of pointers), it is not integrated in the Java IO API, it is rather slow, the API is not documented, it is most likely not thread safe and does not support seeking in Ogg files. <p>Tor <p><p><p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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, Oct 09, 2002 at 08:32:02PM +0200, Frank Grotelueschen wrote:> Hi there > > In the past, i have used lame to encode high quality mp3-files > (vbr 1, bitrate ~ 192kbit). I tend to switch to ogg with Quality 4 > or 5, but i noticed, that many ogg-files tend to produce too much > high frequencies response. In many cases, this is very noticeable. > For my opinion, i cannot accept this worse frequence reponse.I've been assuming for some time that this phenomenon (high frequencies seeming amplified) has been a combination of two things: 1) When point stereo rotates noise audio into the middle of the coupled image, the energy does not increase (the rotation is correct and elliptical), but because the sum total of diffuse noise is now focused in one place, it is more immediately noticable. 2) Quantization of very low level energy, unlike mp3, does not purposely minimize added quantization noise. It either allows the energy to creep up, or if things quantize to *much* lower than original energy, will purpousely artificially support it. The phase and presence of noise energy is much more important to the charachter of the noise than minimizing the naieve vector difference summation ala mp3. Both of these contribute to the impression of boosted highs. Rather 1) gives the impression, but 2) can actually result in a few dB of boost. This is why it seems like at -q 6 the problem mostly disappears, when lossy stereo is finally completely switched off. Note that these are actually very reasonable sounding theories I've only partially tested and justified to myself. They're not bugs so much as shortcomings in the encoding psy model we'll need to correct. Two steps here: First verify these are in fact the causes, then fix them. 1) is hard to deal with as it's somewhat inherent in low-level point stereo. 2) is relatively easy to fix conceptually, what remains to be seen is how annoying it will be to code. Most of the tests people have supplied to 'prove' what is going on have been making incorrect assumptions about the encoding model and thus are flawed in one way or another. I'd not mind gathering a few of the people who are *really* sensitive to this boost (I'll admit that I notice it, but it doesn't bug me) to run tests verifying what is going on. Frank, are you up for this? 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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>Excuse me? Speak your mind, but please refarin from putting words in >my mouth or inventing what I'm thinking.Where did I? Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>> The problem is, that I cannot use libvorbis for a project I amworking>> on, and I made the mistake to invest quite a lot of time on developing>> a new Vorbis decoder before I realized that it is not possible to >> get it to a production quality level. I did indeed expect from Xiph >> to take care about the correctnes of the specification when they >> are distributing one. > >At this point, it's possibly your one reasonable expectation. It's >easier to appreciate help coming in from someone who isn't injecting >"I hate you, I hate you, I hate you, you suck" between the lines.And it is much easier to help, if you get the impression that the help is appreciated. I stated in one of my first mails to this mail- list, that the file format specification contained so many errors, that it is not possible to use it as a basis for a new Vorbis implementation (and you have to admit that is was not), and got an answer from you in a tone like "hah, show me". I did, you corrected the errors I reported in my two first mails, but the following three (from 23rd September) didn't cause any action at all. I see now, that a few things I pointed out are indeed corrected (although the last changed date of the documents are before the 23rd), but there still are a couple of things I pointed out, which are not yet changed.>We're volunteers working our asses off to compete with the world's >largest software companies on their own terms.And with your attitude towards documentation and support, I honestly (and I regret that) don't believe that you have a chance to do so. <p>>I *have* appreciated all the correction mails you sent.>I handled each one when it came in. Did I miss one somewhere?Yes, several points in my mails from 23rd September.>I'd >rather fix it then spend time bowing, scraping and apologizing for >inadvertantly dropping a piece of mail.Actually I thought until today, that all three mails were dropped and ignored. I see now, that some of the points I mentioned have been changed in the documentation, but I relied on the "last changed" date on the top of the documents and assumed that nothing had happened. <p>>I don't give free blowjobs too. You should try it. It's more pleasing than when charging money for it. Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Torsdag, 10 oktober 2002, skrev du:>I just looked back in my archive, I assume you mean your mails dated >23rd of September. Reading these mails, I conclude that neither of >these errors change anything algorithmically.How do you get to that conclusion? One of the errors I reported made the floor calculation fail completely, the other ones would result in an incorrect floor.>Keep in mind that the >specification of the Vorbis algorithms is just that, a >specification of algorithms. It's written to be easy to understand, >not to include all the optimisations in libvorbis. So, often, >libvorbis won't follow the spec exactly.If the spec says subtract and libvorbis adds or uses different values for the calculation, it's not an optimization. If the spec includes steps, which are unneccesary, it won't make it easier to understand them.>My guess is that Monty labelled your bug reports as not-a-bug and >went back to work on Theora. IMHO a simple mail with a short >explanation would have prevented this entire discussion. > >Finally, keep in mind that Xiph.org isn't much more than a Free >Software project that happens to have a meatspace version of itself >in the form of a (non-profit!) foundation, so as to make it >possible to have a few (one I think actually, namely Monty) >developers working on it full-time, and to make it easier to have >corporate contacts. Still, resources are very very scarce, and I >think Xiph.org is doing the best it can right now.If your resources are so scarce, why don't you make it just a little bit easier for new people, who are interested in contributing to do so instead of leaving the entry threshold so high, that people just give up before getting anywhere? <p>Tor <p><p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
> If your resources are so scarce, why don't you make it just a little > bit easier for new people, who are interested in contributing to > do so instead of leaving the entry threshold so high, that people > just give up before getting anywhere?Congratulations, Tor. You have just volunteered to make an effort in lowering the barrier to entry on Xiph.org projects. I would like to discuss this with you further, and I'm interested to know what we can do to address the problem. Let's talk about this off-list, and put together a plan to help people out. Let's stop this list-bitching and get to work, shall we? Emmett Plant CEO, Xiph.org Foundation --- >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-dev-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.
Fredag, 11 oktober 2002, skrev du:>Congratulations, Tor. You have just volunteered to make an effort in >lowering the barrier to entry on Xiph.org projects. I would like to >discuss this with you further, and I'm interested to know what wecan do>to address the problem. Let's talk about this off-list, and puttogether a>plan to help people out. > >Let's stop this list-bitching and get to work, shall we?Ok, then please make a suggestion on what i can do. I've already pointed out a few things I think is missing, and the most important (at least for my current needs) are: - a correct specification of the file format - reference files to test if a new decoder implements the specification completely I have not the knowledge to do the first step, neither do I know where to obtain it, and create a reference file based on an incorrect specification makes no sense. <p>Tor <p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Hi Monty, Frank and others. The high-frequency-boosting problem has been discussed for quite a while now. IMHO, it is the most prominent problem of this otherwise magnificent audio encoder. The problem is easily noticeable in the known applaud test sample at http://lame.sourceforge.net/download/samples/applaud.wav This is what led me to ask for advice on 2002-02-12 when I posted the message "[vorbis] APPLAUD.WAV problems". Some time has passed from then and RC3 has been replaced with 1.0. The artefact is now less pronounced but still noticeable. I am realizing that the problem is not so easy to solve. Monty wrote: ///////////////////////// I've been assuming for some time that this phenomenon (high frequencies seeming amplified) has been a combination of two things: .... Both of these contribute to the impression of boosted highs. Rather 1) gives the impression, but 2) can actually result in a few dB of boost. This is why it seems like at -q 6 the problem mostly disappears, when lossy stereo is finally completely switched off. \\\\\\\\\\\\\\\\\\\\\\\\\ Perhaps the short-term solution would be to implement the lossless-stereo-encoding-switch in the next official release. This would make content those of us, who like more the sample that -> sounds closer to the ORIGINAL but worse as STAND-ALONE than the one that -> sounds better as STAND-ALONE but is different from the ORIGINAL, which is what Monty prefers if I understood correctly. I appreciate your effort in makinf oggenc easy to use, but this switch would greatly increase its usefullness for some critical stereo samples. Thank you for your effort Monty and colleagues. Regards, Bostjan --- >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.
Fredag, 11 oktober 2002, skrev du:>To verify the spec, what we really need is an experienced programmer >without deep internal knowledge of vorbis (at least, one who can't >quote large swatchs of the libvorbis code from memory). This will not >only give a great practical critique of the code, but also proveide >the fully 'genetically-independent' implementation many standards >bodies require. The current Vorbis code was written for ease of >maintainence. A from-scratch implementation could focus on some other >goal as well, such as lower resource use.And this is exactly what I've been trying to do. I downloaded the specification, tried to use that as a basis for implementing a new decoder from scratch and found several points where the specification is incorrect (as compared to libvorbis), and also a few points where parts of the specification seems to be unneccesary or are rather confusing. Some of my suggestions have been implemented by you in the specification, but some of my questions are still left unanswered. I understand, that you have other things to do, but if noone has time to do work on the specification anyway, then I hope that you understand, that I have no interest in wasting my time writing emails, which are not read by anyone. Tor <p><p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Fredag, 11 oktober 2002, skrev du:>To verify the spec, what we really need is an experienced programmer >without deep internal knowledge of vorbis (at least, one who can't >quote large swatchs of the libvorbis code from memory). This will not >only give a great practical critique of the code, but also proveide >the fully 'genetically-independent' implementation many standards >bodies require. The current Vorbis code was written for ease of >maintainence. A from-scratch implementation could focus on some other >goal as well, such as lower resource use.And this is exactly what I've been trying to do. I downloaded the specification, tried to use that as a basis for implementing a new decoder from scratch and found several points where the specification is incorrect (as compared to libvorbis), and also a few points where parts of the specification seems to be unneccesary or are rather confusing. Some of my suggestions have been implemented by you in the specification, but some of my questions are still left unanswered. I understand, that you have other things to do, but if noone has time to do work on the specification anyway, then I hope that you understand, that I have no interest in wasting my time writing emails, which are not read by anyone. Tor <p><p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.
Tirsdag, 15 oktober 2002, skrev du:>It would be useful if you could provide an up-to-date summary of your >proposed corrections, in which you could outline which have notyet been>applied that you think should be, and a short paragraph of rationalefor>each of those. That would be a useful contribution to the project.Fair suggestion. I have collected all my requests and suggestions (some old, and some new), dated and numbered them and will be posting them in a separate mail. Tor <p><p><p>==================================================================EASY and FREE access to your email anywhere: http://Mailreader.com/ ================================================================== <p>--- >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-dev-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.