|
|
|
20 June 2005, 01:31 pm
|
#1 |
|
Member
Join Date: Feb 2005
Posts: 7
|
low level clicks and noise
I have been using MPC for some time and have been very satisfied, but recently I have noticed click and distortion when copying some classical cd,s
This proble occures when when either foobar or winamp are used as players. I have done some tests and have determined that using a test cd static tones below -80db are replaced by clicks and noise. Is there a command line to introduce the correct dither or alternativly are ther better plugins for the players. This is a major problem as the artifacts are clearly audible on low noise classical music. |
|
|
|
20 June 2005, 04:53 pm
|
#2 |
|
Admin
Join Date: Jul 2004
Posts: 372
 |
Please provide us with a small sample (lossless file) that portrays this issue. You could either post a link to a file or send a file to us via our IRC network.
|
|
|
|
|
20 June 2005, 08:08 pm
|
#3 |
|
Member
Join Date: Feb 2005
Posts: 7
|
low level clicks
I have undertaken a few more tests and have gained more detail
On tone signals below -70 a number of spurious noises occur, these are made many times worse when dither is applied to the recording. You can imagine on a track which has a high replay gain this can be obvious. I have prepared a short 5 second dithered tone at -80dB if this wave file is encoded in mpc, then has replay gain applied you will hear many unusual noises. I have placed this wave file on our web site at http://www.wavecor.co.uk/minus80.wav I hope that this helps. Thanks for the support David McGhee |
|
|
|
|
20 June 2005, 08:45 pm
|
#4 |
|
Admin
Join Date: Jul 2004
Posts: 372
|
First of all, this is an artificial tone with background hiss, nothing which can be found in classical music. Such artificial samples are no indication of anything regarding quality of codecs. Secondly, -80db is very, very low volume, and no audio track in the world comes anywhere near such a volume, so your claim that replaygaining a track could make such artifacts noticable is false. In any normal case you would not notice artifacts that Musepack, Vorbis or LAME produce on your sample. If you extremely amplify it, you can't expect much. All lossy codecs would produce artifacts on such a sample, which are often not noticable unless you highly amplify it.
Here is is an archive containing flac files converted from Musepack @ q7, Vorbis @ q7 and LAME @ extreme amplified by 48db: minus80.zip |
|
|
|
|
20 June 2005, 09:31 pm
|
#5 |
|
Member
Join Date: Feb 2005
Posts: 7
|
low level noises
I find your reply most unhelpful, there are many instruments includig piano which will produce the condition which I have demonstrated.
My tests confirm that these artifacts do not occur when using ogg. or mp3 or wavpack or aac/mp4. I was trying to be helpful in high-lighting this problem with the mpc encoder, these very low level artifacts are only a problem if you intend mpc to be a quality coder. To suggest that they are unimportant does not help David McGhee |
|
|
|
|
20 June 2005, 09:39 pm
|
#6 |
|
Admin
Join Date: Jul 2004
Posts: 372
|
I have edited my post some seconds before you replied, please reread.
If you find my reply unhelpful, there's not much I can do. My "tests" confirm that other codecs produce artifacts as well, so I don't know what you're trying to say. You have provided no real world example that portrays a similar artifacting. And as I said before, only in very extreme, unnatural cases of amplification, such artifacts are noticable. The very design of lossy audio codecs is meant to change the audio signal without causing a change in perceptual quality. This is what Musepack does excellently. What you call a "problem" is not a real-world, natural condition problem. Until you provide any real world example of such artifacting, there's not much left to be said. |
|
|
|
|
21 June 2005, 01:31 pm
|
#7 |
|
Member
Join Date: Aug 2004
Posts: 4
|
The extreme low gain appear to trip the tonality estimation of mpc and mp3. Both artifact badly at standard and extreme. Higher settings sound fine.
I ran wavegain on the WAV sample and now mpc / mp3 encode correct even at standard settings. Shy is right about inducing artificial gain levels, but this is still interesting to me. |
|
|
|
|
22 June 2005, 05:56 am
|
#8 |
|
Member
Join Date: Aug 2004
Location: Strasbourg (France)
Posts: 13
|
Problems occurring at -80 dB shouldn’t be audible in any situation. Even with a dynamic compressor which would bring lowest part of the signal to a much higher one, I seriously doubt that audible problems might be revealed to the listeners. I have more than 1000 classical music CDs, and even ultra-quiet parts of wide-dynamic compositions (piano, full orchestra) aren’t at -80 dB. It’s typically -65...-50 dB. There is of course some information at -80 dB, but it’s usually noise, with very little information coming from instruments.
Nevertheless, musepack has poor performances at mid/high profile (--standard, --extreme and sometimes --insane) with low volume tracks. This problem really becomes audible, and sometimes dramatic, after replaygaining. I have many tracks that are measured by RG at > +20 dB; some of them also reach the +30 dB floor: organ, orchestra and chamber music (especially contemporary one) are periodically implicated. Therefore, if you play these tracks with RG track mode enabled, there will be a terrible ringing. I suppose it could be highly improved, because other encoders are performing *much* better even with low setting: vorbis at –q4, faac at –q100 and also lame MP3 3.97a at 128 kbps offer all (much) better encoding than mpc --standard and sometimes –extreme. I’ve uploaded some encodings to illustrate this problem. Reference file is an organ piece, coming from complete Liszt organ music (5CD) played by Olivier Vernet. The RG value for this single track is +29.47 dB; the RG value for the short sample I’ve selected is +40 dB. But I’ve manually edited the RG value to match the reference one (+30 dB). http://guruboolez.free.fr/MPC/low_volume_encodings.zip PCM file could be downloaded in the same folder: http://guruboolez.free.fr/MPC/Liszt_Choral.wv The ringing, absolutely terrible at –q5 after RGing, is still annoying at –q6, but clearly less with –q7. With –q8 profile, ABXing wasn’t a real problem, but the annoyance is near zero (reference sound quality is more questionable). I’ve ABXed –q9, but failed at –q10. Code:
ABC/HR for Java, Version 0.5a, 21 juin 2005
Testname:
Tester:
1R = C:\tests\Liszt\Liszt - Choral MPC Q6.wav
2R = C:\tests\Liszt\Liszt - Choral MPC Q10.wav
3L = C:\tests\Liszt\Liszt - Choral MPC Q5.wav
4L = C:\tests\Liszt\Liszt - Choral MPC Q7.wav
5R = C:\tests\Liszt\Liszt - Choral MPC Q8.wav
6R = C:\tests\Liszt\Liszt - Choral MPC Q9.wav
---------------------------------------
General Comments:
---------------------------------------
1R File: C:\tests\Liszt\Liszt - Choral MPC Q6.wav
1R Rating: 1.5
1R Comment: awful ringing
---------------------------------------
2R File: C:\tests\Liszt\Liszt - Choral MPC Q10.wav
2R Rating: 4.8
2R Comment: Transparent this time. I've tried to focused on background details (organ mechanic), but obtained questionable pval on ABX (11% chance for guessing).
---------------------------------------
3L File: C:\tests\Liszt\Liszt - Choral MPC Q5.wav
3L Rating: 1.0
3L Comment: « below telephone ». 0/5 would be more realistic.
---------------------------------------
4L File: C:\tests\Liszt\Liszt - Choral MPC Q7.wav
4L Rating: 3.0
4L Comment: ringing, obvious but much less annoying than 1R and 3L
---------------------------------------
5R File: C:\tests\Liszt\Liszt - Choral MPC Q8.wav
5R Rating: 4.0
5R Comment: unconstant, fluctuant noise
---------------------------------------
6R File: C:\tests\Liszt\Liszt - Choral MPC Q9.wav
6R Rating: 4.5
6R Comment: ABXable, but variations of noise are very subtle
---------------------------------------
ABX Results:
Original vs C:\tests\Liszt\Liszt - Choral MPC Q10.wav
11 out of 16, pval = 0.105
Original vs C:\tests\Liszt\Liszt - Choral MPC Q7.wav
8 out of 8, pval = 0.0030
Original vs C:\tests\Liszt\Liszt - Choral MPC Q5.wav
8 out of 8, pval = 0.0030
Original vs C:\tests\Liszt\Liszt - Choral MPC Q8.wav
7 out of 8, pval = 0.035
Original vs C:\tests\Liszt\Liszt - Choral MPC Q6.wav
8 out of 8, pval = 0.0030
Original vs C:\tests\Liszt\Liszt - Choral MPC Q9.wav
8 out of 8, pval = 0.0030
---- Detailed ABX results ----
Original vs C:\tests\Liszt\Liszt - Choral MPC Q10.wav
Playback Range: 01.041 to 21.638
9:22:16 PM p 1/1 pval = 0.5
9:22:24 PM p 2/2 pval = 0.25
9:22:30 PM f 2/3 pval = 0.5
Playback Range: 06.556 to 08.678
9:22:46 PM p 3/4 pval = 0.312
9:22:52 PM p 4/5 pval = 0.187
9:23:01 PM p 5/6 pval = 0.109
9:23:05 PM p 6/7 pval = 0.062
9:23:09 PM p 7/8 pval = 0.035
9:23:13 PM f 7/9 pval = 0.089
9:23:18 PM p 8/10 pval = 0.054
9:23:22 PM f 8/11 pval = 0.113
9:23:26 PM p 9/12 pval = 0.072
9:23:30 PM p 10/13 pval = 0.046
9:23:34 PM p 11/14 pval = 0.028
9:23:39 PM f 11/15 pval = 0.059
9:23:47 PM f 11/16 pval = 0.105
Original vs C:\tests\Liszt\Liszt - Choral MPC Q7.wav
Playback Range: 01.041 to 21.638
9:18:36 PM p 1/1 pval = 0.5
9:18:39 PM p 2/2 pval = 0.25
9:18:41 PM p 3/3 pval = 0.125
9:18:43 PM p 4/4 pval = 0.062
9:18:45 PM p 5/5 pval = 0.031
9:18:48 PM p 6/6 pval = 0.015
9:18:50 PM p 7/7 pval = 0.0070
9:18:53 PM p 8/8 pval = 0.0030
Original vs C:\tests\Liszt\Liszt - Choral MPC Q5.wav
Playback Range: 01.041 to 21.638
9:17:54 PM p 1/1 pval = 0.5
9:17:55 PM p 2/2 pval = 0.25
9:17:56 PM p 3/3 pval = 0.125
9:17:58 PM p 4/4 pval = 0.062
9:17:59 PM p 5/5 pval = 0.031
9:18:00 PM p 6/6 pval = 0.015
9:18:07 PM p 7/7 pval = 0.0070
9:18:09 PM p 8/8 pval = 0.0030
Original vs C:\tests\Liszt\Liszt - Choral MPC Q8.wav
Playback Range: 01.041 to 21.638
9:19:28 PM f 0/1 pval = 1.0
9:19:31 PM p 1/2 pval = 0.75
9:19:33 PM p 2/3 pval = 0.5
9:19:37 PM p 3/4 pval = 0.312
9:19:39 PM p 4/5 pval = 0.187
9:19:42 PM p 5/6 pval = 0.109
9:19:46 PM p 6/7 pval = 0.062
9:19:50 PM p 7/8 pval = 0.035
Original vs C:\tests\Liszt\Liszt - Choral MPC Q6.wav
Playback Range: 01.041 to 21.638
9:17:26 PM p 1/1 pval = 0.5
9:17:27 PM p 2/2 pval = 0.25
9:17:29 PM p 3/3 pval = 0.125
9:17:30 PM p 4/4 pval = 0.062
9:17:31 PM p 5/5 pval = 0.031
9:17:33 PM p 6/6 pval = 0.015
9:17:35 PM p 7/7 pval = 0.0070
9:17:37 PM p 8/8 pval = 0.0030
Original vs C:\tests\Liszt\Liszt - Choral MPC Q9.wav
Playback Range: 01.041 to 21.638
9:20:21 PM p 1/1 pval = 0.5
9:20:31 PM p 2/2 pval = 0.25
9:20:38 PM p 3/3 pval = 0.125
9:20:49 PM p 4/4 pval = 0.062
9:21:00 PM p 5/5 pval = 0.031
9:21:21 PM p 6/6 pval = 0.015
9:21:40 PM p 7/7 pval = 0.0070
9:21:49 PM p 8/8 pval = 0.0030
Of course, this situation is not very common, and the big artefact only appears with RG enabled, and with trackgain only. But performance is definitively poor (compared to other tools), and the bad surprises comes from an exceptional low performance of musepack at -q5, used to be a champion with this encoding profile (it was ranked first after a blind listening test involving 18 classical music samples I’ve performed last summer), and which introduces here terrible artefacts that even competitors at 100 kbps could avoid. Working in this issue (IMO the biggest quality flaw of the format at –q5/-q6) might be a direction for further tunings of the current encoder. |
|
|
|
|
22 June 2005, 07:36 pm
|
#9 |
|
Member
Join Date: Feb 2005
Posts: 7
|
Low Level Cilcks
I'm glad that someone else has noticed this problem.
Since starting this thred I have gained more background info. A typical test CD will have a range of fixed test tone levels from 0dB down to -120dB, this low level(well below the 16 bit minimun code level) is possible because of dither noise applied at the recording stage. My recent tests show that the current mpc coder starts to produce noise artifacts at -70dB, inpractice this means that a low level classical recording of say piano or organ will produce audible errors. this is in clear comparisson to other coders which will reproduce the entire test CD rangewithout trouble. Though I have not come to a firm conclusion I suspect that the mpc coder does better without record dither. I have an mpc library of over 4k. files and have built a good confidence level in mpc. The problem only came to light when I play some recent cd copys which were not as good as they should be. |
|
|
|
|
22 June 2005, 07:59 pm
|
#10 |
|
Musepack Nanny
Join Date: Jul 2004
Posts: 168
|
I have forwarded the URL to Frank Klemm through CiTay. I don't know what he'll say. I just know the current developers are incapable of handling this.
|
|
|
|
|
23 June 2005, 09:26 am
|
#11 |
|
Member
Join Date: Aug 2004
Location: Strasbourg (France)
Posts: 13
|
I've uploaded additional samples, and make a report on HA.org:
http://www.hydrogenaudio.org/forums/...howtopic=35030 |
|
|
|
|
23 June 2005, 10:28 am
|
#12 |
|
Member
Join Date: Nov 2004
Posts: 3
|
Guruboolez,
Could you try --standard --ltq 250 --xtreme --ltq 260 --insane --ltq 270 ... and so on? This uses the very sensitive "Filburt" scale in quiet. Musepack will label them as "Below Telephone"-profile encodings, but this is obviously a bug, because --ltq 2xx is higher quality then the default --ltq 5xx. Cheers, Tim |
|
|
|
|
23 June 2005, 10:38 am
|
#13 |
|
Member
Join Date: Aug 2004
Location: Strasbourg (France)
Posts: 13
|
It's not a bug, but a feature introduced by Frank Klemm to discourage the use of personal command line (they're used to lower the quality).
|
|
|
|
|
23 June 2005, 01:52 pm
|
#14 |
|
Member
Join Date: Feb 2005
Posts: 7
|
LowlLevel Clicks
I have just tried --xtreme --ltq 260 and it seems to work ok.
This leaves me very cunfused as I took the avilable preset lines as the best option, it now appears that my library will need recoding to get back to the expecte quality level. Are there other command line options that will allow me to choose the best coding level fo a particular music type. I also have many mono files is there a downmix to mono switch? Many thanks for your help. |
|
|
|
|
23 June 2005, 02:19 pm
|
#15 |
|
Member
Join Date: Nov 2004
Posts: 3
|
In my experience, mono files don't need extra swithes for extra bitrate-savings, i.e. adding --mx xx (ex. --ms 10) will not save some kbps on _mono_ files, but you could possibly save a few kbps by downmixing a 2-channel mono wav, to a 1-channel mono wav, before encoding to musepack. Musepack itself has no switch for downmixing to mono.
As for using --ltq 2xx, expect about 5% larger filesizes in general, IIRC. Cheers, Tim |
|
|
|
|
24 June 2005, 12:17 am
|
#16 |
|
Member
Join Date: Nov 2004
Posts: 3
|
I have the same results here with the Liszt sample. That is to say, without being the slightest familiar with the Listz_choral sample, --quality 6 --ltq 260 seemed at first hearing quite transparent to me, where --qualty 5, --quality 5 --ltq 250 and --quality 6, obviously were not transparent.
Code:
ABC/HR for Java, Version 0.5a, 24 juni 2005
Testname: mpc --ltq 2xx test
Tester:
1R = C:\Test\003 Liszt_Choral.q5ltq250.wav
2R = C:\Test\002 Liszt_Choral.q5.wav
3L = C:\Test\004 Liszt_Choral.q6.wav
4L = C:\Test\005 Liszt_Choral.q6ltq260.wav
---------------------------------------
General Comments:
---------------------------------------
1R File: C:\Test\003 Liszt_Choral.q5ltq250.wav
1R Rating: 2.1
1R Comment:
---------------------------------------
2R File: C:\Test\002 Liszt_Choral.q5.wav
2R Rating: 1.0
2R Comment:
---------------------------------------
3L File: C:\Test\004 Liszt_Choral.q6.wav
3L Rating: 3.0
3L Comment:
---------------------------------------
ABX Results:
Original vs C:\Test\005 Liszt_Choral.q6ltq260.wav
2 out of 8, pval = 0.964
Original vs C:\Test\003 Liszt_Choral.q5ltq250.wav
8 out of 8, pval = 0.0030
Original vs C:\Test\004 Liszt_Choral.q6.wav
8 out of 8, pval = 0.0030
---- Detailed ABX results ----
Original vs C:\Test\005 Liszt_Choral.q6ltq260.wav
Playback Range: 00.000 to 21.638
1:56:04 AM f 0/1 pval = 1.0
1:56:14 AM f 0/2 pval = 1.0
1:56:32 AM f 0/3 pval = 1.0
1:56:38 AM p 1/4 pval = 0.937
1:56:53 AM p 2/5 pval = 0.812
1:57:00 AM f 2/6 pval = 0.89
1:57:17 AM f 2/7 pval = 0.937
1:57:24 AM f 2/8 pval = 0.964
Original vs C:\Test\003 Liszt_Choral.q5ltq250.wav
Playback Range: 00.000 to 21.638
1:51:13 AM p 1/1 pval = 0.5
1:51:19 AM p 2/2 pval = 0.25
1:51:22 AM p 3/3 pval = 0.125
1:51:26 AM p 4/4 pval = 0.062
1:51:29 AM p 5/5 pval = 0.031
1:51:34 AM p 6/6 pval = 0.015
1:51:38 AM p 7/7 pval = 0.0070
1:51:47 AM p 8/8 pval = 0.0030
Original vs C:\Test\004 Liszt_Choral.q6.wav
Playback Range: 00.000 to 21.638
1:52:57 AM p 1/1 pval = 0.5
1:53:04 AM p 2/2 pval = 0.25
1:53:23 AM p 3/3 pval = 0.125
1:53:32 AM p 4/4 pval = 0.062
1:53:40 AM p 5/5 pval = 0.031
1:53:46 AM p 6/6 pval = 0.015
1:53:51 AM p 7/7 pval = 0.0070
1:53:54 AM p 8/8 pval = 0.0030
 ops: As for the suggestion as posted on HA.org to use --ath_gain -14, or something in that style, for normal classical music with the occasional pianissimo passage, I don't think that is a wise swith to use, because it would lower the ath_scale with 14 db in all subbands equally, even at 20 khz, which would not lead to audible gain, but to bitrate bloat in frequency-regions where it makes no audible difference, imho. Cheers, Tim (Tim Mervielde on HA.org) |
|
|
|
|
27 June 2005, 09:49 am
|
#17 |
|
Member
Join Date: Aug 2004
Location: Strasbourg (France)
Posts: 13
|
I've quickly tried both switches, and --ath_gain -14 offers better improvments than --ltq 250 with the samples I've posted.
|
|
|
|
|
27 June 2005, 07:04 pm
|
#18 |
|
Musepack developer
Join Date: Nov 2004
Location: seattle, washington usa
Posts: 111
 |
mmm i recall ath_gain -14 being used back in the day for soft/silent/quiet/whatever sources
is that right? |
|
|
|
|
28 June 2005, 07:31 am
|
#19 |
|
Member
Join Date: Aug 2004
Location: Strasbourg (France)
Posts: 13
|
I've also tried --ath_gain with --quality 4 profile. Bitrate jumps from ~140 to ~160 kbps and reach a comparable value than --quality 4.5 (I obtained these numbers by encoding 160 classical music samples; RG from -10 dB to +30 dB).
--quality 4 --ath_gain -8 [I've also tried -11] considerably reduces the level of distortion with various samples compared to --quality 4.5. The progress is audible even on tracks at normal volume (89 dB). It may be interesting to compare --quality 5 --ath_gain -14 to --quality 5.x with tracks at normal volume to check possible progress. The regression I've noticed between 1.01j and 1.15 is maybe a consequence of different ATH threshold or something directly linked to ATH. It' just a possibility, tests are needed to confirm or infirm that. |
|
|
|
|
28 June 2005, 07:09 pm
|
#20 | |
|
Musepack Nanny
Join Date: Jul 2004
Posts: 168
|
Here is what Frank Klemm said about the issue and several other related points. Translated by CiTay:
Quote:
|
|
|
|
|
 |
| Thread Tools | Search this Thread |
| Display Modes | |
|
|
Hybrid Mode
