I had a little fun with the Vorbis 1.0 libs, hacked a bit with it, and made some demonstration oggs of what it is capable of at the really low end.
Have a listen to:
http://sjeng.org/ftp/vorbis/scar.ogg (http://sjeng.org/ftp/vorbis/scar.ogg) (6.4kbps)
http://sjeng.org/ftp/vorbis/queen.ogg (http://sjeng.org/ftp/vorbis/queen.ogg) (5.3kbps)
http://sjeng.org/ftp/vorbis/ss.ogg (http://sjeng.org/ftp/vorbis/ss.ogg) (6.3kbps)
http://sjeng.org/ftp/vorbis/waitress.ogg (http://sjeng.org/ftp/vorbis/waitress.ogg) (6.7kbps)
Edit: also
http://sjeng.org/ftp/vorbis/daan.ogg (http://sjeng.org/ftp/vorbis/daan.ogg) (6.6kbps)
And keep in mind this is 1/20'th of an 128kbps file.
--
GCP
:jawdrop:
Holy f'n $#!+!!!! That's incredible!
Don't get me wrong, they sound like ass. They sound remarkably good for 8kb/s though.
How much encoder tuning have you done, and is more possible?
Originally posted by Ardax
They sound remarkably good for 8kb/s though.
How much encoder tuning have you done, and is more possible?
6kbps, not 8kbps
There is no real tuning in here; I spent about an hour re-familiarizing myself with the 1.0 infrastructure and made the new encoding mode on guesswork. Getting everything to work took another two hours. I need to talk to monty to get a few things cleared up before more can be done.
I would like to put a CD on a floppy. If you have 2.8M floppies, you can do so already.
--
GCP
and make Vorbis Floppy-Transparent to 99% users over 99% samples:
"OGG Vorbis is the only format to give You floppy-quality at ~6kbps"
you can imagine what can be possible for 64 kbps :eek:
Impressive Garf indeed! I think most of the pure tones in Scar sound particularly very impressive at 8 kHz telephony fidelity. Have you tested this out with any sort of speech signal yet? You didn't happen to by any chance use floor 0 did you for testing purposes?
Setup an ogg stream (q0 40~45kb/s 22khz )
rexit2.ath.cx:8000/q0.ogg
(seems to be a GREAT alternative to mp3 streaming @ lower bitrates)
Originally posted by HotshotGG
Impressive Garf indeed! I think most of the pure tones in Scar sound particularly very impressive at 8 kHz telephony fidelity. Have you tested this out with any sort of speech signal yet? You didn't happen to by any chance use floor 0 did you for testing purposes?
I've tested it with speech against Speex 5.1.
The size comes around 5-6kbps, which is lower than Speex quality 1 (goes from 1-10) and higher than Speex quality 0. Speex quality 0 artifacts quite badly and I wouldn't know if this mode or Speex is better. At quality 1, Speex wins hands down. The problem of this mode is that it doesn't window-switch, which (I think) is needed to prevent sibilants from echoing. Except for the echo's, the mode sounds quite reasonable on voice.
Based on this result, I think you could make a Vorbis 1.0 encoder that handles voice at least as well as the current Speex. Interesting result, considering Speex was designed for voice and Vorbis wasn't. You'd want to have someone with more knowledge about voice coding than me working on it though
I would have liked to try floor0 with this, as it _possibly_ could be a little more efficient, but the 1.0 code doesn't have any other mode that uses it! Without an example or prior codebooks to work from it is impossible for me to make one. I already had enough problems with the floor1 codebooks now.
--
GCP
Without an example or prior codebooks to work from it is impossible for me to make one. I already had enough problems with the floor1 codebooks now.
yes, in deed I had been looking over the floor 1 backend implementation myself and I still need to look over it a few more times to figure it out. I wouldn't argue with you about the codebooks as they look they are a major task to figure out. I had wanted to point that out with Floor 0, because floor 0 as you know uses a IIR LSP representation. This might be especially useful in speech coding. It seems as though floor 0 is deprecated for the time being though like you said. I had been looking through the documentation for it though anyway.
Originally posted by HotshotGG
I had wanted to point that out with Floor 0, because floor 0 as you know uses a IIR LSP representation. This might be especially useful in speech coding.
The usage is so different that success of LSP/LPC in speech coding has no bearing on Vorbis. I'd prefer floor0 because it may be more efficient for uncoupled modes (like mono) than floor1, even when not encoding speech.
--
GCP
The usage is so different that success of LSP/LPC in speech coding has no bearing on Vorbis. I'd prefer floor0 because it may be more efficient for uncoupled modes (like mono) than floor1, even when not encoding speech.
I see where you coming from. It might be useful from an experimental point. I had figured that LSP/LPC representation to have no bearing for speech coding in Vorbis. I had just been quite confused when it had been implemented and I was reading it was popular in speech coding. Now I understand everyone has there own flavors of it.
Uhm...could you give us the exact command line for these ultra-low settings? I'd like to do previews of entire albums for my band at cdstreet.com. We have 10 MB of space of samples of the CDs and with that bitrate (which sound damn good) I can fit a ton of stuff in there...
so, what q- setting is this???? -1? -2? -3?
OMG! LOOK HOW SMALL AND CUTE THEY ARE! 6KBPS!!! WOW! THEY ARE LIKE LITTLE BABIES!
Originally posted by Zaraza
Uhm...could you give us the exact command line for these ultra-low settings? I'd like to do previews of entire albums for my band at cdstreet.com. We have 10 MB of space of samples of the CDs and with that bitrate (which sound damn good) I can fit a ton of stuff in there...
so, what q- setting is this???? -1? -2? -3?
This is a custom version of the encoder libs. If I can get monty to help sort out a few remaining issues, I'll make it available.
--
GCP
Originally posted by Zaraza
We have 10 MB of space of samples of the CDs and with that bitrate (which sound damn good) I can fit a ton of stuff in there...
oggenc -q -1 --downmix --resample 8000
with the normal Vorbis 1.0 oggenc will already fit an entire CD in 10M
--
GCP
If I can get monty to help sort out a few remaining issues
What would you need sorted out might I ask?
Originally posted by HotshotGG
What would you need sorted out might I ask?
Why I get massively differing performance from one codebook training run to another.
Why some samples completely fall apart (as in: not remotely comparable to original) even though they're not particularly hard.
Why the codebook trainers sometimes don't produce all the residue training files I'd expect them to.
What the correct parameters for floor training are (I guessed them).
The codebook training part of Vorbis is one of the least intiutive (at least to me) and it's almost undocumented (the spec doesn't help either)
--
GCP
The codebook training part of Vorbis is one of the least intiutive (at least to me) and it's almost undocumented (the spec doesn't help either)
yes, when I was looking over the source code myself I didn't even make an effort to look over the codebooks as well I had the same basic idea in my mind. What codebooks are you using just the regular scalar huffman codebooks?
Originally posted by HotshotGG
What codebooks are you using just the regular scalar huffman codebooks?
IIRC, floor codebooks are basic huffman and residue is lattice VQ
--
GCP
IIRC, floor codebooks are basic huffman and residue is lattice VQ
Thank you for bringing that to my attention again. I was aware of the residue using lattice VQ codebooks, however I never really thought that the floor codebooks used huffman coding. I see the puzzle fits together now.
I was fooling around with this extreme low bitrate stuff last week. Mono, -q -1, 8KHz resampling, etc. I encoded George Winston's December CD, which is all solo piano. Even at 8kbps, it didn't sound half bad. I guess you could use this to stream music over a 28.8kbps modem connection...but who has those anymore?
<- 1.5Mbps broadband pipe
which is all solo piano
yes, strong tones sound particularly good it's unfortunate there is no window switching like Garf mentioned above to eliminate any echo's etc.
<- 1.5Mbps broadband pipe
My dad is heavenly into Broadband and from a growing response in that area I think we are moving in just the opposite direction. I can't even get my friend back on his 56.6 V. 90 dial-up. AT&T is ripping everyone off in this city because there is no competition, but that's another story that is more or less important for a discussion on this forum because physcoacoustics audio compression is more important.]
I've shaved off another 40-50%, so they're around 3.8kbps now.
http://sjeng.org/ftp/vorbis/scarres2.ogg (http://sjeng.org/ftp/vorbis/scarres2.ogg) (4.0kbps)
http://sjeng.org/ftp/vorbis/ssres2.ogg (http://sjeng.org/ftp/vorbis/ssres2.ogg) (4.3kbps)
http://sjeng.org/ftp/vorbis/qnres2.ogg (http://sjeng.org/ftp/vorbis/qnres2.ogg) (3.2kps)
Not much worse than the previous ones due to better tuning & some hints from Klemm.
--
GCP
It's hard to believe ! Absolutely amazing (but really awful, too)
Can you go below ?
Amazing. Though there are a few codecs designed for voice transmission in the 1000-6000bps range, I don't recall any of them ever being able to handle a musical signal this well when playing around with them (voxware, 1-bit ADPCM, etc). An application for such a low bitrate could include streaming audio for low-bandwidth digital cell phones (9-14.4kbps). Is it possible to get an 8Khz stereo encode at bitrates under 10kbps?
If this demonstration were run by Micro$oft, they would probably claim "CD Quality At 3kbps!"
Originally posted by guruboolez
It's hard to believe ! Absolutely amazing (but really awful, too)
Can you go below ?
Going lower is no problem, keeping the quality acceptable is
I still need monty to help me out with a few things, I expect another reduction after that.
My goal is 3.2kbps average. (CD on a floppy)
--
GCP
Originally posted by Cygnus X1
Amazing. Though there are a few codecs designed for voice transmission in the 1000-6000bps range, I don't recall any of them ever being able to handle a musical signal this well when playing around with them (voxware, 1-bit ADPCM, etc). An application for such a low bitrate could include streaming audio for low-bandwidth digital cell phones (9-14.4kbps). Is it possible to get an 8Khz stereo encode at bitrates under 10kbps?
Stereo? That should be possible. It's 4kbps now, and using stereo should not double the bitrate.
--
GCP
An even more ambitous goal, one which I doubt is possible, would be able to get a signal down to 300 baud (.3kbps) as to theoretically be able to stream music over a 1970's era modem. Would this have any practical use? Of course not. However, it would be cool to know that it is theoretically possible. I used such a modem as kid in the 1980's and always wondered if music could be sent over it, even back then, before I knew what a CD or audio compression was
well.. about the 4Kbps .ogg's.. I preffer the 6kbps ones by far.. (there is more musical content!)
Also, I suppose that in these bitrates, the Header of vorbis does really influenciate (see http://www.hydrogenaudio.org/forums/showth...54&pagenumber=3 (http://www.hydrogenaudio.org/forums/showthread.php?s=&threadid=2854&pagenumber=3) ).
So, that 4kbps .ogg is quite in the limit of information.
An even more ambitous goal, one which I doubt is possible, would be able to get a signal down to 300 baud (.3kbps) as to theoretically be able to stream music over a 1970's era modem. Would this have any practical use? Of course not. However, it would be cool to know that it is theoretically possible.
Almost certainly not music, but there are experimental voice codecs which are aiming for ultra-low bitrates. Of course, they sound nothing like the original - but even getting something comprehensible out of bitrates this low is amazing.
For example this (http://citeseer.nj.nec.com/153082.html) discusses a 500 bps voice codec.
If one had REALLY crappy speakers (like the builtin laptop speakers from older laptops, or those unpowered little ones) this wouldn't sound like complete crap. As they are, they sound very nice at such a small bitrate.
It would be interesting to see what MP3Pro, LAME, and WMA could produce at similiar bitrates.
Originally posted by greenirft
It would be interesting to see what MP3Pro, LAME, and WMA could produce at similiar bitrates.
A 12kbps WMA sounds like crap already... So I'd hate to find out what 4-6kbps would do.
Ok, I just did a test with both WMA and OGG at 12kbps and the ogg sounds flatter but with less artifacts while the wma seems more dynamic but has the typical metal sound, but extremely magnified. Choose your poison.
Stereo versions:
http://sjeng.org/ftp/vorbis/scarres3.ogg (http://sjeng.org/ftp/vorbis/scarres3.ogg) (4.1kbps)
http://sjeng.org/ftp/vorbis/ssres3.ogg (http://sjeng.org/ftp/vorbis/ssres3.ogg) (4.5kbps)
http://sjeng.org/ftp/vorbis/qnres3.ogg (http://sjeng.org/ftp/vorbis/qnres3.ogg) (3.8kbps)
--
GCP
Ok, there's a good thing to tell about those last samples, and a bad thing aswell,
The good thing is that the stereo version sounds (quality wise) almost like the mono one, with not much more bits (although the stereo image is different from the original, but at least gives a feel of space).
The bad one is... I CANNOT seek! (plugin fault?). When I jump after second 25, the player jumps to the next song.
(btw... this was with ssres2/3 )
Edit: Ok, now I can seek.
Try again. I tried rehuff on the files, which still seems to have a few problems. The ones that are up now have fixed headers. If there's still problems, I'll have to do without rehuff.
--
GCP
I have the same seeking problems with Tobias DirectShow Filters, when I go to later than 1:38 (scarres2) it plays again from the beginning. scarres3 doesn´t seek at all. Winamp (2.80a, in_vorbis 1.2) seeks perfect scarres2, but in sacrres3 it also doesn´t work at all.
Are there sources/binaries of these "20020717 'Floggy"-oggenc somewhere downloadable? I searched the vorbis/vorbis-dev mailing list but I couldn´t find anything.
Originally posted by S_O
I have the same seeking problems with Tobias DirectShow Filters, when I go to later than 1:38 (scarres2) it plays again from the beginning. scarres3 doesn´t seek at all. Winamp (2.80a, in_vorbis 1.2) seeks perfect scarres2, but in sacrres3 it also doesn´t work at all.
Are there sources/binaries of these "20020717 'Floggy"-oggenc somewhere downloadable? I searched the vorbis/vorbis-dev mailing list but I couldn´t find anything.
Did you download the current files?
The source/binaries for the library aren't available yet - it's unfinished. (On certain songs, it magically degrades into the sound of passing spaceships)
--
GCP
Yes.
Both encoded with Xiph.Org libVorbis I 20020717 'Floggy'
scarres2.ogg: 111.835 Bytes
scarres3.ogg: 116.354 Bytes
Also the library is unstable and unfinished yet, it would be great to test it. Tarkin is also avaible yet (and it works great at high bitrates, compresses images better than Jpeg2000!!)
Originally posted by Garf
I've shaved off another 40-50%, so they're around 3.8kbps now.
http://sjeng.org/ftp/vorbis/scarres2.ogg (http://sjeng.org/ftp/vorbis/scarres2.ogg) (4.0kbps)
http://sjeng.org/ftp/vorbis/ssres2.ogg (http://sjeng.org/ftp/vorbis/ssres2.ogg) (4.3kbps)
http://sjeng.org/ftp/vorbis/qnres2.ogg (http://sjeng.org/ftp/vorbis/qnres2.ogg) (3.2kps)
Not much worse than the previous ones due to better tuning & some hints from Klemm.
--
GCP
Bit packing is still suboptimal. gzip (a zero knowledge compressor) still reduces
size by 6..7 %.
Originally posted by S_O
Yes.
Both encoded with Xiph.Org libVorbis I 20020717 'Floggy'
scarres2.ogg: 111.835 Bytes
scarres3.ogg: 116.354 Bytes
Also the library is unstable and unfinished yet, it would be great to test it. Tarkin is also avaible yet (and it works great at high bitrates, compresses images better than Jpeg2000!!)
I uploaded a new version of scarres3.ogg (http://sjeng.org/ftp/vorbis/scarres4.ogg (http://sjeng.org/ftp/vorbis/scarres4.ogg)) which isn't rehuffed. But scarres2.ogg and scarres3 should work! I've tested it here and xmms handles them fine.
The original scarres3 that I uploaded was corrupted and not seekable, but the one that is up now should be. Maybe these are player bugs? XMMS for example thinks they're sampled at 60khz. I guess most players aren't tested with this kind of files
--
GCP
Originally posted by Frank Klemm
Bit packing is still suboptimal. gzip (a zero knowledge compressor) still reduces
size by 6..7 %.
I know! They are using untrained codebooks. Whenever I try to use my own ones, the quality drops noticeably. This is where I need monty's help.
--
GCP
scarres4 is seekable in winamp, but with tobias ds-filters only to the half. Because this problem is also in scarres2 which also seeks perfect in winamp it seems to be a problem in the ds-filters. I´ll report this problems to Tobias.
Both programs (winmap and the ds-filters) say that the files are 6kHz sampled, is that the right samplerate?
Is this "Floggy"-Vorbis yust playing around or real work on a very-low-bitrate vorbis mode which will be implemented in the official encoder?
Originally posted by S_O
scarres4 is seekable in winamp, but with tobias ds-filters only to the half. Because this problem is also in scarres2 which also seeks perfect in winamp it seems to be a problem in the ds-filters. I´ll report this problems to Tobias.
Both programs (winmap and the ds-filters) say that the files are 6kHz sampled, is that the right samplerate?
For the 4kbps files yes, the 6kbps files are 8kHz sampled.
Is this "Floggy"-Vorbis yust playing around or real work on a very-low-bitrate vorbis mode which will be implemented in the official encoder?
It's just playing around, but when it's more or less working there's no reason it can't go into the offical codec. It doesn't interfere with the standard stuff.
--
GCP
No seeking problems here... Quite amazing quality, even though it sounds like crap...
Garf,
This is excellent stuff! Please please keep up the good work. All other codec's must be trounced upon with great force! You are the great force!
What I'd really love is something like 32khz Stereo tuned for 32kbps. I mean /proper/ tuned. I think that currently there is a window for improvement. Any ideas? (Frank?)
Ruairi
Yes, this really is amazing for 3kbps! I remember thinking a couple of years ago that it would be impossible to get a signal with musical content lower than 16kbps (though I think some terrrestrial phone systems use 8kbps?). I think a quality 20kbps, 22.05Khz stream might be doable with Ogg (if it isn't already, haven't tested that bitrate) for internet streaming, or dare I say 32Khz, 32-40kbps stereo audio? Excellent work!
This is impressive!
The idea of carying around a whole album (read 10 songs) on a floppy is simply amazing. What about building an Ogg Vorbis hardware player with a floppy and a CFI/II drive. Floppy would be an ideal cheap and easy to use media...
dev0
I played around with the ultralowbitrate clips and I was quite amazed. In some of them there was quite a lot of signalclipping (I did not activate the hard-limiter in the decoder). So I decided to run vorbisgain on the files.
This was one of the results:
D:vorbisbackupultralowbitrate>vorbisgain *.ogg
Processing directory '.':
Analyzing files...
Gain | Peak | Scale | New Peak | Track
----------+--------+-------+----------+------
-11.45 dB | 51988 | 0.27 | 13912 | scar.ogg
-1.29 dB | 25355 | 0.86 | 21856 | ss.ogg
+0.72 dB | 27163 | 1.09 | 29511 | queen.ogg
-7.17 dB | 43151 | 0.44 | 18901 | waitress.ogg
-7.12 dB | 52190 | 0.44 | 22993 | daan.ogg
Couldn't initialize gain analysis (nonstandard samplerate?) for 'scarres4.ogg'
Does this mean that 6kHz sampling rate does not comply to the standard or is Vorbisgain wrong here ?
I use Vorbisgain 0.32 dated 2002-07-24 from RareWares
Originally posted by dev0
What about building an Ogg Vorbis hardware player with a floppy and a CFI/II drive.
Eheh. Come on. The results are indeed amazing, but the quality is so bad anyway that I would throw my player through the window after listening to it for 10 minutes.
Originally posted by Hanky
Couldn't initialize gain analysis (nonstandard samplerate?) for 'scarres4.ogg'
Does this mean that 6kHz sampling rate does not comply to the standard or is Vorbisgain wrong here ?
You should read nonstandard as 'not commonly used enough that we put the code to handle it into vorbisgain'.
--
GCP
Originally posted by Hanky
I played around with the ultralowbitrate clips and I was quite amazed. In some of them there was quite a lot of signalclipping (I did not activate the hard-limiter in the decoder). So I decided to run vorbisgain on the files.
This was one of the results:
D:vorbisbackupultralowbitrate>vorbisgain *.ogg
Processing directory '.':
Analyzing files...
Gain | Peak | Scale | New Peak | Track
----------+--------+-------+----------+------
-11.45 dB | 51988 | 0.27 | 13912 | scar.ogg
-1.29 dB | 25355 | 0.86 | 21856 | ss.ogg
+0.72 dB | 27163 | 1.09 | 29511 | queen.ogg
-7.17 dB | 43151 | 0.44 | 18901 | waitress.ogg
-7.12 dB | 52190 | 0.44 | 22993 | daan.ogg
Couldn't initialize gain analysis (nonstandard samplerate?) for 'scarres4.ogg'
Does this mean that 6kHz sampling rate does not comply to the standard or is Vorbisgain wrong here ?
I use Vorbisgain 0.32 dated 2002-07-24 from RareWares
Replaygain (so also Vorbisgain) has some precomputed weighting filters.
They exist for the following frequencies:
8, 11.025, 12
16, 22.05, 24
32, 44.1, 48 kHz
// for each filter:
// [0] 48 kHz, [1] 44.1 kHz, [2] 32 kHz, [3] 24 kHz, [4] 22050 Hz, [5] 16 kHz, [6] 12 kHz, [7] is 11025 Hz, [8] 8 kHz
const Float_t AYule [9] [11] = {
{ 1., -3.84664617118067, 7.81501653005538,-11.34170355132042, 13.05504219327545,-12.28759895145294, 9.48293806319790, -5.87257861775999, 2.75465861874613, -0.86984376593551, 0.13919314567432 },
{ 1., -3.47845948550071, 6.36317777566148, -8.54751527471874, 9.47693607801280, -8.81498681370155, 6.85401540936998, -4.39470996079559, 2.19611684890774, -0.75104302451432, 0.13149317958808 },
{ 1., -2.37898834973084, 2.84868151156327, -2.64577170229825, 2.23697657451713, -1.67148153367602, 1.00595954808547, -0.45953458054983, 0.16378164858596, -0.05032077717131, 0.02347897407020 },
{ 1., -1.61273165137247, 1.07977492259970, -0.25656257754070, -0.16276719120440, -0.22638893773906, 0.39120800788284, -0.22138138954925, 0.04500235387352, 0.02005851806501, 0.00302439095741 },
{ 1., -1.49858979367799, 0.87350271418188, 0.12205022308084, -0.80774944671438, 0.47854794562326, -0.12453458140019, -0.04067510197014, 0.08333755284107, -0.04237348025746, 0.02977207319925 },
{ 1., -0.62820619233671, 0.29661783706366, -0.37256372942400, 0.00213767857124, -0.42029820170918, 0.22199650564824, 0.00613424350682, 0.06747620744683, 0.05784820375801, 0.03222754072173 },
{ 1., -1.04800335126349, 0.29156311971249, -0.26806001042947, 0.00819999645858, 0.45054734505008, -0.33032403314006, 0.06739368333110, -0.04784254229033, 0.01639907836189, 0.01807364323573 },
{ 1., -0.51035327095184, -0.31863563325245, -0.20256413484477, 0.14728154134330, 0.38952639978999, -0.23313271880868, -0.05246019024463, -0.02505961724053, 0.02442357316099, 0.01818801111503 },
{ 1., -0.25049871956020, -0.43193942311114, -0.03424681017675, -0.04678328784242, 0.26408300200955, 0.15113130533216, -0.17556493366449, -0.18823009262115, 0.05477720428674, 0.04704409688120 }
};
const Float_t BYule [9] [11] = {
{ 0.03857599435200, -0.02160367184185, -0.00123395316851, -0.00009291677959, -0.01655260341619, 0.02161526843274, -0.02074045215285, 0.00594298065125, 0.00306428023191, 0.00012025322027, 0.00288463683916 },
{ 0.05418656406430, -0.02911007808948, -0.00848709379851, -0.00851165645469, -0.00834990904936, 0.02245293253339, -0.02596338512915, 0.01624864962975, -0.00240879051584, 0.00674613682247, -0.00187763777362 },
{ 0.15457299681924, -0.09331049056315, -0.06247880153653, 0.02163541888798, -0.05588393329856, 0.04781476674921, 0.00222312597743, 0.03174092540049, -0.01390589421898, 0.00651420667831, -0.00881362733839 },
{ 0.30296907319327, -0.22613988682123, -0.08587323730772, 0.03282930172664, -0.00915702933434, -0.02364141202522, -0.00584456039913, 0.06276101321749, -0.00000828086748, 0.00205861885564, -0.02950134983287 },
{ 0.33642304856132, -0.25572241425570, -0.11828570177555, 0.11921148675203, -0.07834489609479, -0.00469977914380, -0.00589500224440, 0.05724228140351, 0.00832043980773, -0.01635381384540, -0.01760176568150 },
{ 0.44915256608450, -0.14351757464547, -0.22784394429749, -0.01419140100551, 0.04078262797139, -0.12398163381748, 0.04097565135648, 0.10478503600251, -0.01863887810927, -0.03193428438915, 0.00541907748707 },
{ 0.56619470757641, -0.75464456939302, 0.16242137742230, 0.16744243493672, -0.18901604199609, 0.30931782841830, -0.27562961986224, 0.00647310677246, 0.08647503780351, -0.03788984554840, -0.00588215443421 },
{ 0.58100494960553, -0.53174909058578, -0.14289799034253, 0.17520704835522, 0.02377945217615, 0.15558449135573, -0.25344790059353, 0.01628462406333, 0.06920467763959, -0.03721611395801, -0.00749618797172 },
{ 0.53648789255105, -0.42163034350696, -0.00275953611929, 0.04267842219415, -0.10214864179676, 0.14590772289388, -0.02459864859345, -0.11202315195388, -0.04060034127000, 0.04788665548180, -0.02217936801134 }
};
const Float_t AButter [9] [3] = {
{ 1., -1.97223372919527, 0.97261396931306 },
{ 1., -1.96977855582618, 0.97022847566350 },
{ 1., -1.95835380975398, 0.95920349965459 },
{ 1., -1.95002759149878, 0.95124613669835 },
{ 1., -1.94561023566527, 0.94705070426118 },
{ 1., -1.92783286977036, 0.93034775234268 },
{ 1., -1.91858953033784, 0.92177618768381 },
{ 1., -1.91542108074780, 0.91885558323625 },
{ 1., -1.88903307939452, 0.89487434461664 }
};
const Float_t BButter [9] [3] = {
{ 0.98621192462708, -1.97242384925416, 0.98621192462708 },
{ 0.98500175787242, -1.97000351574484, 0.98500175787242 },
{ 0.97938932735214, -1.95877865470428, 0.97938932735214 },
{ 0.97531843204928, -1.95063686409857, 0.97531843204928 },
{ 0.97316523498161, -1.94633046996323, 0.97316523498161 },
{ 0.96454515552826, -1.92909031105652, 0.96454515552826 },
{ 0.96009142950541, -1.92018285901082, 0.96009142950541 },
{ 0.95856916599601, -1.91713833199203, 0.95856916599601 },
{ 0.94597685600279, -1.89195371200558, 0.94597685600279 }
};
Originally posted by rjamorim
Eheh. Come on. The results are indeed amazing, but the quality is so bad anyway that I would throw my player through the window after listening to it for 10 minutes.
Sounds shit but better than most 64kbps mp3's I've heard which lots of people (not on this forum) find to be acceptable quality. At this speed/quality OGG is going to kill-out all audio codecs in no time.
Setup an ogg stream (q0 40~45kb/s 22khz )
rexit2.ath.cx:8000/q0.ogg
(seems to be a GREAT alternative to mp3 streaming @ lower bitrates)
I like this stream. Which setting uses? -q0 --resample 22050?
// for each filter:
//[0] 48 kHz, [1] 44.1 kHz, [2] 32 kHz, [3] 24 kHz, [4] 22050 Hz, [5] 16 kHz, [6] 12 kHz, [7] is 11025 Hz, [8] 8 kHz
how about extending the filters' table with coefficients that relevant for 96k, 88.2k & 64k ?
this will be useful in my AACGain..
I would like to put a CD on a floppy. If you have 2.8M floppies, you can do so already.
I think it would be cool if you could take one or more tracks of audio compress them with Vorbis and fit them onto a floppy optionally with a simple player app. Ideally depending on the amount of music you want to put onto disk it'll select the highest bitrate that can be fitted onto the media.
A disk that holds a CD is novel but potentially useful idea, could be used for giving audio previews. Potentially this low bandwidth Vorbis could be multiplexed on modems for multiple simultanious streams of audio.
I was looking into using MELP for allowing low bandwidth phone conversations over the net (without distributing your ordinary traffic much) as it only requires 300 bytes/sec for quality speech but I fear there are many patent issues with it. Ogg Vorbis and speex could be better here.
Anyway keep up the good work Garf, low bandwidth Vorbis can hold many potentially useful possibilities.
Doug
Sounds shit but better than most 64kbps mp3's I've heard which lots of people (not on this forum) find to be acceptable quality.
Hummm... well I agree it may sound better than MP3 at 64kbps - but only if you use Blade. (Or dist10)
IMO, FhG at 64 sound waaaay better than vorbis at 6 (and that shouldn't be any surprise)
// for each filter:
//[0] 48 kHz, [1] 44.1 kHz, [2] 32 kHz, [3] 24 kHz, [4] 22050 Hz, [5] 16 kHz, [6] 12 kHz, [7] is 11025 Hz, [8] 8 kHz
how about extending the filters' table with coefficients that relevant for 96k, 88.2k & 64k ?
this will be useful in my AACGain..
// for each filter:
//[0] 48 kHz, [1] 44.1 kHz, [2] 32 kHz, [3] 24 kHz, [4] 22050 Hz, [5] 16 kHz, [6] 12 kHz, [7] is 11025 Hz, [8] 8 kHz
how about extending the filters' table with coefficients that relevant for 96k, 88.2k & 64k ?
this will be useful in my AACGain..
- Filter coeffs are calculated by 2bdecided
- 64k, 88.2k, 96k, 128k, 176.4k, 192k are possible
- to my mind also the filters are suboptimal. They underestimate the loudness
between 100 and 200 Hz, overestimate <25 Hz
- at home I have a list of 30 used sample frequencies which should be supported
- best would be to derive loundness from the psycho model in the encoder
- Filter coeffs are calculated by 2bdecided
- 64k, 88.2k, 96k, 128k, 176.4k, 192k are possible
- to my mind also the filters are suboptimal. They underestimate the loudness
between 100 and 200 Hz, overestimate <25 Hz
- at home I have a list of 30 used sample frequencies which should be supported
Can't you design and recalculate 'corrected' filter coeffs for the missing frequencies?
I mean...this should be exactly your 'cup of tea' Frank
Indeed, coefficients for different sampling rates would be very welcome.
I played a bit with an alpha version of AACgain, by DSPguru, and the program is coming out nicely.
So, could you provide us these coeffs, Frank?
I had a little fun with the Vorbis 1.0 libs, hacked a bit with it, and made some demonstration oggs of what it is capable of at the really low end.
Have a listen to:
http://sjeng.org/ftp/vorbis/scar.ogg (http://sjeng.org/ftp/vorbis/scar.ogg) (6.4kbps)
http://sjeng.org/ftp/vorbis/queen.ogg (http://sjeng.org/ftp/vorbis/queen.ogg) (5.3kbps)
http://sjeng.org/ftp/vorbis/ss.ogg (http://sjeng.org/ftp/vorbis/ss.ogg) (6.3kbps)
http://sjeng.org/ftp/vorbis/waitress.ogg (http://sjeng.org/ftp/vorbis/waitress.ogg) (6.7kbps)
Edit: also
http://sjeng.org/ftp/vorbis/daan.ogg (http://sjeng.org/ftp/vorbis/daan.ogg) (6.6kbps)
And keep in mind this is 1/20'th of an 128kbps file.
--
GCP
Gosh....yes Ogg is really the best if you want a lossy codec WITH ability to keep a good sound!
theres quite a bit of distortion but with a multiband dynamic compressor enabled (im using audiostocker 0.08 for winamp) the samples sound astonishing to say the least.
Amazing...
Hmm, intresting "serial number" (winamp 2.81, "File info"). Normal ogg has eg. 2345 or 4257, 13582, these ones: 1426917371, 190683057.
Your oggs have serials with 4 numbers? my serials have normaly 8 numbers. Yust the one created with the directshow-filters have always the serial 0.
All my oggs have serials with four digits. They were created with oggdropXPd v1.31.
Whatever happened to this "fit a CD on a floppy" idea? I think it was interesting, to say the least.
Very low bitrate encoding also has potential for web radio & such. "The first Internet music radio where you can listen in on a 14.4K modem & still browse the web!"