IPB

Welcome Guest ( Log In | Register )

CELT 0.9.1 is out!, Help wanted
jmvalin
post Nov 16 2010, 02:56
Post #1


Xiph.org Speex developer


Group: Developer
Posts: 475
Joined: 21-August 02
Member No.: 3134



Hi,

I'd like to announce CELT version 0.9.1. There have been many quality enhancements since 0.8.x and even more so since the last version announced on HA. You can get it from the CELT website. Also, CELT is now a component of the Opus codec, which is in the process of being standardized by the IETF as a lossy audio codec for interactive applications.

Also, for those who would like to help, we are looking for volunteers to help tune the codec. The bit-stream is (finally) about to be frozen, so any quality improvements we can get before then is useful. No highly specialized skills required, just good critical listening abilities. As a first round, I'd be interested in comments and rankings of the following four audio files:

fileA.wav
fileB.wav
fileC.wav
fileD.wav

The bit-rate is fairly low (64 kb/s), so artefacts are easy to hear. This is the original (uncompressed) file. I'm interested in a quality ranking of all these four files (especially the ones that sound similar). I'll reveal the contents of these files after people have responded.
Go to the top of the page
+Quote Post
 
Start new topic
Replies
IgorC
post Dec 10 2010, 04:30
Post #2





Group: Members
Posts: 1533
Joined: 3-January 05
From: ARG/RUS
Member No.: 18803




Sample X did better on speech but still has low freq echo on wood-sound drums.
Oh and I encoded Vorbis with Aotuv 5.7 -q0. Nothing changes.

Now looking into Fatboy and the rest of samples....

Fatboy
CODE
ABC/HR Version 1.1 beta 2, 18 June 2004
Testname:

1R = D:\Samples\fatboy\Apple 64 HEAAC CVBR fatboy_30sec.wav
2R = D:\Samples\fatboy\fatboy_celt64c.wav
3L = D:\Samples\fatboy\nero 64 fatboy_30sec.wav
4R = D:\Samples\fatboy\aotuv 5.7 fatboy_30sec.wav

---------------------------------------
General Comments:

---------------------------------------
1R File: D:\Samples\fatboy\Apple 64 HEAAC CVBR fatboy_30sec.wav
1R Rating: 2.0
1R Comment:
---------------------------------------
2R File: D:\Samples\fatboy\fatboy_celt64c.wav
2R Rating: 4.0
2R Comment:
---------------------------------------
3L File: D:\Samples\fatboy\nero 64 fatboy_30sec.wav
3L Rating: 2.2
3L Comment:
---------------------------------------
4R File: D:\Samples\fatboy\aotuv 5.7 fatboy_30sec.wav
4R Rating: 4.5
4R Comment:
---------------------------------------
ABX Results:


CELT did transparent on Since Always.
CODE
ABC/HR Version 1.1 beta 2, 18 June 2004
Testname:

1R = D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\SinceAlways_celt64.wav
2L = D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\nero 64 kbps SinceAlways.wav
3R = D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\apple 64 HEAAC CVBR SinceAlways.wav
4R = D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\aotuv 5.7 SinceAlways.wav

---------------------------------------
General Comments:

---------------------------------------
2L File: D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\nero 64 kbps SinceAlways.wav
2L Rating: 1.7
2L Comment:
---------------------------------------
3R File: D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\apple 64 HEAAC CVBR SinceAlways.wav
3R Rating: 2.8
3R Comment:
---------------------------------------
4R File: D:\Samples\Herbien hancokc and Since alwyas\Nueva carpeta\aotuv 5.7 SinceAlways.wav
4R Rating: 4.7
4R Comment:
---------------------------------------
ABX Results:



CELT did very well on Hancock too.
CODE
ABC/HR Version 1.1 beta 2, 18 June 2004
Testname:

1R = D:\Samples\Herbien hancokc and Since alwyas\nero q025 64 kbps Herbie_Hancock.wav
2R = D:\Samples\Herbien hancokc and Since alwyas\aotuv 5.7 Herbie_Hancock.wav
3R = D:\Samples\Herbien hancokc and Since alwyas\Herbie_Hancock_celt64.wav
4R = D:\Samples\Herbien hancokc and Since alwyas\apple 64 heaac cvbr Herbie_Hancock.wav

---------------------------------------
General Comments:

---------------------------------------
1R File: D:\Samples\Herbien hancokc and Since alwyas\nero q025 64 kbps Herbie_Hancock.wav
1R Rating: 3.8
1R Comment: left channel noise
---------------------------------------
2R File: D:\Samples\Herbien hancokc and Since alwyas\aotuv 5.7 Herbie_Hancock.wav
2R Rating: 3.2
2R Comment: distortion on trumpet
---------------------------------------
3R File: D:\Samples\Herbien hancokc and Since alwyas\Herbie_Hancock_celt64.wav
3R Rating: 4.3
3R Comment:
---------------------------------------
4R File: D:\Samples\Herbien hancokc and Since alwyas\apple 64 heaac cvbr Herbie_Hancock.wav
4R Rating: 2.7
4R Comment: intermitent noise in left channel
---------------------------------------
ABX Results:


This post has been edited by IgorC: Dec 10 2010, 05:05
Go to the top of the page
+Quote Post
jmvalin
post Dec 10 2010, 15:38
Post #3


Xiph.org Speex developer


Group: Developer
Posts: 475
Joined: 21-August 02
Member No.: 3134



Wow, thanks very much for all these results. I'm glad to see that I managed to fix the quality on fatboy without making the other cases worse. Also nice to see that the rich stereo samples didn't cause problems either, though I think they weren't too bad for intensity stereo because there's a lot of pan, but the image itself isn't that wide.
Go to the top of the page
+Quote Post
SebastianG
post Dec 10 2010, 16:10
Post #4





Group: Developer
Posts: 1317
Joined: 20-March 04
From: Göttingen (DE)
Member No.: 12875



Hi Jean-Marc,

I just skimmed through some parts of the source code and noticed in vq.c the "scrambling" (exprotation1 etc). It looks like this is roughly equivalent to an all-pass filter. Since you apply this on the spectral coefficients and due to the time/frequency duality this is equivalent to a time-dependent frequency shift within a frame. I know the original motivation for this processing (reducing metallic artefacts) and it seems to be doing what it's supposed to but it sure is an odd thing to do. On the downside you smear strong tonal components over a larger spectrum which kind of defeats the purpose of an MDCT in terms of energy compaction w.r.t. tonal components (MDCT as opposed to, say, a PQMF with fewer subbands). Maybe this is why the guitar sample doesn't work that well... The encoded coefficients correspond to some kind of chirps (due to the time-dependent frequency shift) and not a windowed cosine. Have you checked the impulse response of a single one surrounded by zeros in X followed by inverse exprotation + inverse MDCT? Might be interesting to see what it looks like...

...just wanted to share this perspective...

Cheers and congrats for the impressive 64kbps performance!
SG

This post has been edited by SebastianG: Dec 10 2010, 16:19
Go to the top of the page
+Quote Post
jmvalin
post Dec 10 2010, 19:34
Post #5


Xiph.org Speex developer


Group: Developer
Posts: 475
Joined: 21-August 02
Member No.: 3134



QUOTE (SebastianG @ Dec 11 2010, 00:10) *
I just skimmed through some parts of the source code and noticed in vq.c the "scrambling" (exprotation1 etc). It looks like this is roughly equivalent to an all-pass filter. Since you apply this on the spectral coefficients and due to the time/frequency duality this is equivalent to a time-dependent frequency shift within a frame. I know the original motivation for this processing (reducing metallic artefacts) and it seems to be doing what it's supposed to but it sure is an odd thing to do. On the downside you smear strong tonal components over a larger spectrum which kind of defeats the purpose of an MDCT in terms of energy compaction w.r.t. tonal components (MDCT as opposed to, say, a PQMF with fewer subbands). Maybe this is why the guitar sample doesn't work that well... The encoded coefficients correspond to some kind of chirps (due to the time-dependent frequency shift) and not a windowed cosine. Have you checked the impulse response of a single one surrounded by zeros in X followed by inverse exprotation + inverse MDCT? Might be interesting to see what it looks like...


Hi Sebastian,

Actually, the spreading in vq.c is not equivalent to an all-pass filter. It's a non-linear operation (when viewed from a time-domain signal) because it actually creates frequency content that wouldn't be there otherwise. The idea is to avoid "birdie artefacts", aka musical noise. One thing to note is that the amount of spreading depends on two things: 1) the bit-rate (less spreading as the bitrate goes up), and 2) a global per-frame parameter. It's possible with the current bit-stream to use no spreading at all. In fact, there's a (still a bit simple) function that decides how much spreading to apply based on how tonal the audio is. That being said, I think the importance of the spreading has probably gone down recently since we're no longer allowing codebooks of just one pulse. In any case, you can hear the difference and decide for yourself what the effect is:

with spreading
without spreading

What do you think?
Go to the top of the page
+Quote Post

Posts in this topic
- jmvalin   CELT 0.9.1 is out!   Nov 16 2010, 02:56
- - punkrockdude   I think, after several listenings, that A sounds b...   Nov 19 2010, 19:18
- - Primius   In file A i noticed warbling distortion on guitars...   Nov 20 2010, 00:21
- - IgorC   logs h*tp://www.mediafire.com/?7ed142tcipntet1   Nov 21 2010, 02:25
- - jmvalin   Thanks very much guys. I guess I should start by t...   Nov 21 2010, 03:45
- - IgorC   logs h*tp://www.mediafire.com/?t8ko84f34bz8b5f Th...   Nov 22 2010, 20:32
|- - jmvalin   QUOTE (IgorC @ Nov 23 2010, 05:32) The ...   Nov 23 2010, 02:16
|- - IgorC   QUOTE (jmvalin @ Nov 22 2010, 23:16) Than...   Nov 23 2010, 04:20
|- - jmvalin   QUOTE (IgorC @ Nov 23 2010, 12:20) QUOTE ...   Nov 23 2010, 05:19
- - IgorC   All right, I will test fileL later. There was an...   Nov 23 2010, 03:38
- - IgorC   Sample I is still preferable but there is no stati...   Nov 27 2010, 04:51
- - IgorC   Description of artifacts. 1st sample. Samples H a...   Nov 28 2010, 05:00
|- - jmvalin   Thanks again for taking the time to listen to thos...   Dec 8 2010, 15:29
- - IgorC   What encoders did you used for Vorbis and HE-AAC? ...   Dec 8 2010, 20:37
|- - jmvalin   QUOTE (IgorC @ Dec 9 2010, 04:37) What en...   Dec 9 2010, 02:25
|- - jmvalin   QUOTE (jmvalin @ Dec 9 2010, 10:25) Regar...   Dec 9 2010, 02:36
- - IgorC   Chris has made tremendous work on compilation of ...   Dec 9 2010, 03:51
|- - jmvalin   QUOTE (IgorC @ Dec 9 2010, 11:51) Chris h...   Dec 9 2010, 06:01
- - IgorC   Logs h*tp://www.mediafire.com/?c54du2c2b95mjsn Ma...   Dec 9 2010, 05:07
|- - jmvalin   Thanks again for the results. I'll investigate...   Dec 9 2010, 14:47
|- - jmvalin   QUOTE (jmvalin @ Dec 9 2010, 22:47) Thank...   Dec 9 2010, 21:31
- - IgorC   Sample X did better on speech but still has low fr...   Dec 10 2010, 04:30
|- - jmvalin   Wow, thanks very much for all these results. I...   Dec 10 2010, 15:38
|- - SebastianG   Hi Jean-Marc, I just skimmed through some parts o...   Dec 10 2010, 16:10
|- - jmvalin   QUOTE (SebastianG @ Dec 11 2010, 00:10) I...   Dec 10 2010, 19:34
|- - SebastianG   QUOTE (jmvalin @ Dec 10 2010, 19:34) QUOT...   Dec 10 2010, 20:40
|- - jmvalin   QUOTE (SebastianG @ Dec 11 2010, 05:40) I...   Dec 10 2010, 20:48
- - IgorC   Jean-Marc, How does CELT scale with higher bitrat...   Dec 10 2010, 19:01
|- - jmvalin   QUOTE (IgorC @ Dec 11 2010, 03:01) How do...   Dec 10 2010, 19:25
|- - jmvalin   QUOTE (IgorC @ Dec 10 2010, 13:01) How do...   Feb 15 2011, 02:04
- - jmvalin   OK, so I've worked a bit on improving some of ...   Dec 10 2010, 23:36
- - IgorC   Jean-Marc, What do you plan for test of CELT? If...   Dec 13 2010, 22:01
|- - jmvalin   QUOTE (IgorC @ Dec 13 2010, 16:01) What d...   Dec 13 2010, 22:25
- - IgorC   Jean-Marc, I see that version 0.10.0 has been rel...   Dec 24 2010, 00:34
|- - NullC   QUOTE (IgorC @ Dec 23 2010, 16:34) Jean-M...   Dec 24 2010, 02:02
- - IgorC   The differences are small. The results with HD650...   Dec 24 2010, 19:35
|- - C.R.Helmrich   QUOTE (IgorC @ Dec 24 2010, 20:35) Fatboy...   Dec 25 2010, 21:23
- - IgorC   Hi, Chris. Actually the question was about the po...   Dec 26 2010, 04:32
- - IgorC   I think it can be worth to redo the last test or d...   Dec 26 2010, 09:39
- - IgorC   I've tried several time with HD650 and HD447. ...   Dec 29 2010, 23:26
- - IgorC   Aparently CELT scales good with higher bitrates. ...   Dec 30 2010, 20:55
|- - jmvalin   QUOTE (IgorC @ Dec 30 2010, 14:55) Aparen...   Dec 31 2010, 05:34
|- - IgorC   QUOTE (jmvalin @ Dec 31 2010, 01:34) If y...   Dec 31 2010, 06:00
- - IgorC   96 kbps: LC-AAC info: iTunes 9.0.0.70, 96 kbps, VB...   Dec 31 2010, 08:43
- - IgorC   Maybe CELT should compete not only with LD-AAC but...   Dec 31 2010, 18:44
- - rt87   I wonder if someone can do 48kbps tests comparing ...   Jan 1 2011, 01:59
- - IgorC   You can have an idea basing on results of 64 kbps ...   Jan 1 2011, 14:32
- - NullC   Anyone following along with the CELT technical dev...   Jan 26 2011, 02:09
- - jmvalin   OK, so we've been a bit quiet lately but worki...   Jan 28 2011, 00:36
- - IgorC   The artifacts of both files have very similar natu...   Jan 28 2011, 14:48
|- - jmvalin   QUOTE (IgorC @ Jan 28 2011, 08:48) The ar...   Jan 29 2011, 06:24
- - IgorC   0.11.0 has been released. Good news. I wonder tha...   Feb 5 2011, 02:50
- - sauvage78   Personnaly I have some interest in it, but I need ...   Feb 5 2011, 04:54
|- - jmvalin   A few things I'd like to say here about CELT t...   Feb 5 2011, 05:20
|- - IgorC   QUOTE (jmvalin @ Feb 5 2011, 01:20) If bi...   Feb 5 2011, 18:39
|- - NullC   QUOTE (IgorC @ Feb 5 2011, 09:39) PS. It...   Feb 7 2011, 05:42
|- - jmvalin   QUOTE (NullC @ Feb 6 2011, 23:42) I would...   Feb 7 2011, 05:53
|- - Ljubo44   QUOTE (NullC @ Feb 7 2011, 06:42) In any ...   May 23 2011, 20:40
- - sauvage78   One of the reason why I wait for a CLI encoder/dec...   Feb 5 2011, 06:20
|- - IgorC   QUOTE (sauvage78 @ Feb 5 2011, 02:20) Sor...   Feb 5 2011, 06:34
|- - jmvalin   QUOTE (sauvage78 @ Feb 5 2011, 00:20) Sor...   Feb 5 2011, 12:53
- - [JAZ]   @sauvage: I don't understand you. First, you...   Feb 5 2011, 11:40
- - sauvage78   QUOTE you put doubts on how much it has improved s...   Feb 5 2011, 13:28
|- - jmvalin   If you want to see how far CELT has come, there...   Feb 5 2011, 15:18
- - list   any future improve towards 32kbps range?   Feb 6 2011, 19:25
|- - jmvalin   QUOTE (list @ Feb 6 2011, 13:25) any futu...   Feb 7 2011, 04:00
- - list   I found some samples in which @64 kbps,48 khz , CE...   Feb 7 2011, 12:52
|- - jmvalin   QUOTE (list @ Feb 7 2011, 06:52) I found ...   Feb 7 2011, 13:03
|- - NullC   QUOTE (jmvalin @ Feb 7 2011, 04:03) QUOTE...   Feb 7 2011, 16:30
- - IgorC   I have a job to do until 10th of March. After that...   Feb 9 2011, 10:26
- - IgorC   Very short test on difficult eig sample Big surp...   Feb 13 2011, 21:09
- - IgorC   Single result for one random sample. Test of scala...   Feb 13 2011, 23:37
- - IgorC   The first sample that wasn't transparent at 19...   Feb 14 2011, 16:32
- - IgorC   CELT performs not enough well and worse than Vorbi...   Feb 14 2011, 22:21
- - IgorC   ~67 kbps test: CELT 0.11.2 (complexity 10, bitrate...   Mar 15 2011, 16:02
|- - NullC   QUOTE (IgorC @ Mar 15 2011, 07:02) ~67 kb...   Mar 16 2011, 13:27
- - LaserSokrates   Hey jmvalin, thank you and the team behind CELT fo...   Mar 16 2011, 14:27
- - Anakunda   Hello there, tried to compile celtenc with library...   Jun 2 2011, 14:35
|- - 2304p   QUOTE (Anakunda @ Jun 2 2011, 15:35) Hell...   Jun 3 2011, 07:16
- - Anakunda   Thanks much, that's work. I'd be intereste...   Jun 3 2011, 07:26
- - klonuo   You can get 0.11.2 also from here: http://listenin...   Jun 3 2011, 08:01
- - lvqcl   version 0.11.4 available at http://git.xiph.org/?...   Jun 3 2011, 10:43
|- - klonuo   QUOTE (lvqcl @ Jun 3 2011, 11:43) version...   Jun 3 2011, 11:35
|- - 2304p   QUOTE (lvqcl @ Jun 3 2011, 11:43) version...   Jun 4 2011, 15:09
|- - klonuo   QUOTE (2304p @ Jun 4 2011, 16:09) I have ...   Jun 4 2011, 19:37
|- - 2304p   QUOTE (klonuo @ Jun 4 2011, 20:37) QUOTE ...   Jun 4 2011, 22:42
- - Anakunda   celt.c violates the C standard several times in th...   Jun 3 2011, 11:45
|- - lvqcl   QUOTE (Anakunda @ Jun 3 2011, 14:45) celt...   Jun 3 2011, 11:57
|- - Anakunda   QUOTE (lvqcl @ Jun 3 2011, 12:57) QUOTE (...   Jun 3 2011, 12:09
- - lvqcl   Either I cannot compile CELT 0.11.4 correctly, or ...   Jun 4 2011, 16:28
- - klonuo   run automation bash script provided and then as re...   Jun 4 2011, 23:03
- - 2304p   can you compile celt-0.11.4.tar.gz and upload as c...   Jun 5 2011, 12:46
- - Anakunda   I can compile but for me it doesnot encode anythin...   Jun 5 2011, 13:38
- - lvqcl   Celt 2011-04-21: (re-uploaded: compiled with CUS...   Jun 5 2011, 16:38
|- - 2304p   QUOTE (lvqcl @ Jun 5 2011, 17:38) Celt 20...   Jun 5 2011, 17:59
- - darkbyte   Hi! First, thanks for this truly amazing code...   Nov 13 2011, 10:00
|- - NullC   QUOTE (darkbyte @ Nov 13 2011, 02:00) I...   Nov 14 2011, 09:23
- - darkbyte   QUOTE (NullC @ Nov 14 2011, 10:23) For fi...   Nov 14 2011, 17:57
- - Gainless   Can someone re-compile the celt version 0.11.4 to ...   Dec 12 2011, 22:37


Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 25th July 2014 - 20:32