Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
John L orionfyre at hotmail.com writes: Please test the following: $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.ac3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.mp3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 -acodec pcm_f32le out.wav I ran all three as requested, including '-loglevel debug'. All three resulting files resulted in poor quality audio as before. Now we are there;-) Hendrik says the option fixes audio for him, you report it does not fix the issue... Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
To: ffmpeg-user@ffmpeg.org From: ceho...@ag.or.at Date: Mon, 18 May 2015 08:48:32 + Subject: Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality John L orionfyre at hotmail.com writes: Please test the following: $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.ac3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.mp3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 -acodec pcm_f32le out.wav I ran all three as requested, including '-loglevel debug'. All three resulting files resulted in poor quality audio as before. Now we are there;-) Hendrik says the option fixes audio for him, you report it does not fix the issue... Carl Eugen I reviewed my work on this section and I was wrong; this does in fact work and solve the issue i was originally having. Perhaps I had my file browser pointed at the wrong working folder...? Thanks all, especially you Carl. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
John L wrote: Please test the following: $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.ac3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.mp3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 -acodec pcm_f32le out.wav I ran all three as requested, including '-loglevel debug'. All three resulting files resulted in poor quality audio as before. the filtergraph output does show something different however, but the resulting audio is still terrible and indistinguishable from before. That's strange, it seems to do the correct thing for me (only tested last one). I notice typical of loud movie dts the master sample is quite extreme to start with (and clipped a bit by studio I think). But ignoring that the downmix with -rematrix_maxval 1.0 is the same for me as using a signed wav. Looking with sox ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 -acodec pcm_f32le outf-1.wav sox outf-1.wav -n stats sox WARN wav: wave header missing FmtExt chunk Overall Left Right DC offset -0.000397 -0.000397 -0.000138 Min level -0.986728 -0.977729 -0.986728 Max level 0.988566 0.974450 0.988566 Pk lev dB -0.10 -0.20 -0.10 RMS lev dB-11.54-11.47-11.60 RMS Pk dB -4.73 -4.73 -4.97 RMS Tr dB -36.11-36.08-36.11 Crest factor - 3.66 3.76 Flat factor 0.00 0.00 0.00 Pk count 2 2 2 Bit-depth 32/32 32/32 32/32 Num samples1.44M Length s 29.995 Scale max 1.00 Window s 0.050 Without -rematrix_maxval 1.0 sox outf-0.wav -n stats sox WARN wav: wave header missing FmtExt chunk Overall Left Right DC offset -0.005071 -0.005071 -0.002057 Min level -1.00 -1.00 -1.00 Max level 1.00 1.00 1.00 Pk lev dB 0.00 0.00 0.00 RMS lev dB -5.62 -5.61 -5.63 RMS Pk dB -0.79 -0.79 -1.03 RMS Tr dB -28.46-28.42-28.46 Crest factor - 1.91 1.91 Flat factor45.34 46.05 44.52 Pk count222k 228k 216k Bit-depth 32/32 32/32 32/32 Num samples1.44M Length s 29.995 Scale max 1.00 Window s 0.050 sox WARN sox: `outf-0.wav' input clipped 443634 samples ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
John L wrote: Backstory: I have a system in place to automagically convert video files to smaller formats/versions on request to have a sort of mobile version for my father who travels extensively. The purpose is so that he can fit significantly more videos on his tablet than if they were the high quality rips. It all boils down to: ffmpeg -i [input-file] -ac 2 -c:v libx264 -c:a libmp3lame -b:v 1024k -preset fast [output-file] I was under the impression everything was hunky dory until I took a bunch of the shrunken movies on my phone on a roadtrip. A good many of the videos were as good as can be expected, and nothing was egregiously wrong. However on a few videos the audio was absolutely atrocious, blown out, clipping, and just noise from seemingly nowhere. One of the worst was Intersteller which was completely unwatchable after the first two minutes with all the blown out crescendos, pops, cracks, static, and voices of the deep adulterating the audio stream. All video files affected by this were 5.1DTS sources, but not all 5.1DTS were affected. When talking with my father he said it was a frequent enough occurrence that he suspected it was just because I had shrunk the file so small and was an artifact of that. He did confirm that most videos that were affected weren't as bad as the Interstellar conversion. ~/testing$ ffmpeg -version ffmpeg version 2.5.6-0ubuntu0.15.04.1 Copyright (c) 2000-2015 the FFmpeg developers built with gcc 4.9.2 (Ubuntu 4.9.2-10ubuntu13) configuration: --prefix=/usr --extra-version=0ubuntu0.15.04.1 --build-suffix=-ffmpeg --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --shlibdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --enable-gpl --enable-shared --disable-stripping --enable-avresample --enable-avisynth --enable-ladspa --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libmodplug --enable-libmp3lame --enable-libopenjpeg --enable-libopus --enable-libpulse --enable-libschroedinger --enable-libshine --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvorbis --enable-libwavpack --enable-libwebp --enable-libxvid --enable-opengl --enable-x11grab --enable-libdc1394 --enable-libiec61883 --enable-libzvbi --enable-libzmq --enable-frei0r --enable-libvpx --enable-libx264 --enable-libsoxr --enable-gn ut ls --enable-openal --enable-libopencv --enable-librtmp --enable-libx265 libavutil 54. 15.100 / 54. 15.100 libavcodec 56. 13.100 / 56. 13.100 libavformat56. 15.102 / 56. 15.102 libavdevice56. 3.100 / 56. 3.100 libavfilter 5. 2.103 / 5. 2.103 libavresample 2. 1. 0 / 2. 1. 0 libswscale 3. 1.101 / 3. 1.101 libswresample 1. 1.100 / 1. 1.100 libpostproc53. 3.100 / 53. 3.100 To troubleshoot I copied out a particularly bad snippet of audio ffmpeg -i Int*.mkv -vn -c copy -ss 1:30 -t 0:30 inter.dts This audio clip is confirmed to be a good 5.1dts stream ffmpeg -i inter.dts -ac 2 -c libmp3lame inter-test.mp3 This audio sample has the exact same audio defects as in the shrunken video Converting it to a stereo wave format, and then converting into an mp3: ffmpeg -i inter.dts -ac 2 -c pcm_s32le inter.wav ffmpeg -i inter.wav -c libmp3lame inter.mp3 both inter.wav and inter.mp3 are confirmed to be GOOD stereo copies of the audio with no defects. https://www.dropbox.com/s/tru46zo07gcr8ve/testing.tar.gz?dl=0 This is a link to the files in question to my testing above. inter.dts : 30 second rip of audio from video inter-test : dts-mp3 conversion inter.mp3 : dts-wav-mp3 conversion I apologize if I'm missing something glaring, but I've been unable to find any other instances of this issue with my google-fu. Until I have a solution I've already edited my services to perform this intermediary wave step work-around on all conversions. Thank you for your time. If your target output file is .mp4 (no output type mentioned in your commands) then consider fdkaac, which many times produces smaller files than .mp3 + is stellar quality + plays anywhere an .mp4 file plays. Downmixing from 5.1 to stereo is automagick. To get .mp3 sized files use the he2 profile as with this snippet... -c:a libfdk_aac -profile:a aac_he_v2 -afterburner 1 -signaling explicit_sbr -vbr 5 -ac 2 -ar 44100 HE2 takes more CPU cycles than HE + produces far smaller files. Another trick is using vbr encoding which also reduces file size. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sun, 17 May 2015 02:16:56 -0500, John L orionf...@hotmail.com wrote: - the resulting wav file is significantly distorted, but qualitatively doesn't 'feel' as harsh Just FYI John, _some_ of those channels in the 5.1 are already flattening at peak levels so the sound overall, will never be great. Of course, this is not an FFMPEG thing or even related to your transcode issue. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Please test the following: $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.ac3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 out.mp3 $ ffmpeg -i inter.dts -rematrix_maxval 1.0 -ac 2 -acodec pcm_f32le out.wav I ran all three as requested, including '-loglevel debug'. All three resulting files resulted in poor quality audio as before. the filtergraph output does show something different however, but the resulting audio is still terrible and indistinguishable from before. [AVFilterGraph @ 0x2387f20] query_formats: 4 queried, 6 merged, 3 already done, 0 delayed 0.414214 0.00 0.292893 0.00 0.292893 0.00 0.00 0.414214 0.292893 0.00 0.00 0.292893 I do appreciate your help in resolving this issue. Just to see if it wasn't my system causing the issue I loaded up a windows XP vm and used the 32-bit windows binary from the ffmpeg homepage, resulting in the same outputs. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sun, 17 May 2015 07:41:31 + (UTC), Carl Eugen Hoyos ceho...@ag.or.at wrote: Not a problem about misunderstanding. Hope I make sense sometimes too -) I did now and the question now is: Is the issue reproducible with: $ ffmpeg -i inter.dts -ac 2 -acodec pcm_f32le outf.wav This (above) produces the the same overload/over-modulation/ distortion as in the original raised issue. You were expecting it? OP seems to think its an fltp issue. If not, what about the following? $ lame outf.wav Sorry. This last makes no sense to me. I don't have lame installed. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
John L orionfyre at hotmail.com writes: But I'm still curious why it would behave in such a way Me too! So please test the following and report back: ffmpeg -i inter.dts -acodec pcm_f32le -ac 2 out.wav Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Bazza lamia at jeack.com.au writes: The simple command 'ffmpeg -i inter.dts -ac 2 -ab 320k out.mp3', for an MP3 file, will produce this result. Plainly distorted. -) You told that and John told us. I did neither deny it nor asked for another example for the same comamnd line. When I originally asked for testing I had not yet completely understood what FFmpeg does internally (sorry!). I did now and the question now is: Is the issue reproducible with: $ ffmpeg -i inter.dts -ac 2 -acodec pcm_f32le outf.wav If not, what about the following? $ lame outf.wav If there is also a bug in the ac3 encoder, it will need additional testing. Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
To: ffmpeg-user@ffmpeg.org From: ceho...@ag.or.at Date: Sat, 16 May 2015 11:15:29 + Subject: Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality Moritz Barsnick barsnick at gmx.net writes: $ ffmpeg -loglevel debug -i inter.dts -ac 2 -c:a libmp3lame -f null - $ ffmpeg -loglevel debug -i inter.dts -ac 2 -c:a pcm_s16le -f null - You can insert other arbitrary codecs at will. The former shows a matrix: 1.00 0.00 0.707107 0.00 0.707107 0.00 0.00 1.00 0.707107 0.00 0.00 0.707107 [auto-inserted resampler 0 at 0xb713840] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:fltp r:48000Hz while the latter shows: 0.414214 0.00 0.292893 0.00 0.292893 0.00 0.00 0.414214 0.292893 0.00 0.00 0.292893 [auto-inserted resampler 0 at 0xb3b55c0] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:s16 r:48000Hz I think this may be the described behavior. If this really is the issue, it should be reproducible with at least one of the command lines I proposed (namely for -acodec pcm_f32le) and it is possible to work-around the issue by forcing s16p as the mp3 encoding format. The mp3 encoder accepts fltp, s16p and s32p. Thanks Everyone. You've fixed the problem for me. When I cycle through the available format options for libmp3lame, the only one that makes bad audio is 'fltp', both 's16p' and 's32p' produce a good file But I would really appreciate if somebody can confirm that the issue is reproducible with pcm_f32le (and neither with s16le nor s32le). Carl Eugen I ran the following command line and received the following: ffmpeg -loglevel debug -i inter.dts -c:a pcm_f32le -ac 2 -y inter-f32le.wav [AVFilterGraph @ 0xca9d20] query_formats: 4 queried, 6 merged, 3 already done, 0 delayed 1.00 0.00 0.707107 0.00 0.707107 0.00 0.00 1.00 0.707107 0.00 0.00 0.707107 [auto-inserted resampler 0 @ 0xc965c0] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:flt r:48000Hz - the resulting wav file is significantly distorted, but qualitatively doesn't 'feel' as harsh And as you expected, both s16le nor s32le result in acceptable files for both: ffmpeg -loglevel debug -i inter.dts -c:a pcm_s16le -ac 2 -y inter-s16le.wav ffmpeg -loglevel debug -i inter.dts -c:a pcm_s32le -ac 2 -y inter-s32le.wav I quickly re-encoded movies that were known to suffer from this issue using '-c:a libmp3lame -sample_fmt s16p' and they all result in acceptable audio levels. So OFFICIALLY my problem is resolved. But I'm still curious why it would behave in such a way that fltp-fltp would be allowed to blow out levels so badly (i'm also scared to go down that rabbit hole) Thank you to all who have contributed. I appreciate your help tremendously. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Instead, take the 5.1 and _DOWNMIX_ all tracks to a single stereo for the phone/tablet by declaring -acodec -ac 2. No intermediate steps should be required. Consider also - Do you need pcm_s32le ? pcm_s16le is usual. I fail to see how that is any different than what I am doing now. I was under the impression that the flags -acodec and -c:a were the same. Regardless using -acodec reults in identical clipping and noise generation dts-mp3. For reference here is the command I used: ffmpeg -i inter.dts -acodec libmp3lame -ac 2 inter-new.mp3 The resulting mp3 file still has horrendous crackling and noise. What I thought I stated quite clearly in my OP is the following: 5.1DTS-2.0MP3 results in horrible noise and clipping in the resulting mp3 file 5.1DTS2.0PCM-2.0MP3 does not generate the same atrocious noise. please reference the files I've included in the dropbox link in my OP. inter.dts was the ripped 5.1 audio inter-test.mp3 was encoded to 2.0mp3 format directly from 5.1dts and when played back on my laptop(s) (Windows, Linux, Mac; in Windows MP, ffplay, mplayer, vlc, xine, and more), Phones (s5,s3,iphone,htc one), tablet, ipod and my Sansa MP3 player all has horrific noise. inter.mp3 was generated by converting the 5.1DTS to 2.0PCM and then the 2.0PCM to 2.0MP3, sounds just fine when played back on all of my devices. I am fully aware that there should be NO NEED to use an intermediary wave format to downsample to stereo audio from 5.1 for a conversion to mp3. But that's exactly why I'm writing this problem into the group because it is NOT working as expected. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
John L orionfyre at hotmail.com writes: ffmpeg version 2.5.6-0ubuntu0.15.04.1 Please test current FFmpeg git head, see http://ffmpeg.org/download.html (Please do not test 2.6) ffmpeg -i inter.dts -ac 2 -c libmp3lame inter-test.mp3 This audio sample has the exact same audio defects as in the shrunken video Converting it to a stereo wave format, and then converting into an mp3: ffmpeg -i inter.dts -ac 2 -c pcm_s32le inter.wav ffmpeg -i inter.wav -c libmp3lame inter.mp3 both inter.wav and inter.mp3 are confirmed to be GOOD stereo copies of the audio with no defects. Please test the following: $ ffmpeg -i inter.dts -ac 2 out16.wav $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 Iiuc, out32.wav sounds ok. Do the other three files sound ok for you or not? Until I have a solution I've already edited my services to perform this intermediary wave step work-around on all conversions. You can do the intermediate step within FFmpeg, just play with -af aformat=s32. Thank you, Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sat, 16 May 2015 07:52:12 + (UTC), Carl Eugen Hoyos ceho...@ag.or.at wrote: Please test the following: $ ffmpeg -i inter.dts -ac 2 out16.wav $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 Carl, I did test some of this stuff. It seemed that attempts to downmix via -ac 2 would work OK for WAV (pcm_s16le) but anytime converting to other formats (AC3, or MP3s) generated significant signal overload. However, the channel mixing DID occur. A workaround was to include -af volume=-10dB. However, although I'm not the poster with the problem, I've just done your suggestions and, for me ... - MP3 suffers overload - WAVs are OK - AC3 suffers overload - AAC are OK - MP2 are OK But I'll let him speak ... ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Bazza lamia at jeack.com.au writes: Please test the following: $ ffmpeg -i inter.dts -ac 2 out16.wav $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 Carl, I did test some of this stuff. - MP3 suffers overload - WAVs are OK - AC3 suffers overload - AAC are OK - MP2 are OK Sorry, I am apparently extremely dim-witted: Did you test the four lines above? Which of them sound ok, which of them do not sound ok? Thank you, Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
John L wrote: Instead, take the 5.1 and _DOWNMIX_ all tracks to a single stereo for the phone/tablet by declaring -acodec -ac 2. No intermediate steps should be required. Consider also - Do you need pcm_s32le ? pcm_s16le is usual. I fail to see how that is any different than what I am doing now. I was under the impression that the flags -acodec and -c:a were the same. Regardless using -acodec reults in identical clipping and noise generation dts-mp3. For reference here is the command I used: ffmpeg -i inter.dts -acodec libmp3lame -ac 2 inter-new.mp3 The resulting mp3 file still has horrendous crackling and noise. What I thought I stated quite clearly in my OP is the following: 5.1DTS-2.0MP3 results in horrible noise and clipping in the resulting mp3 file 5.1DTS2.0PCM-2.0MP3 does not generate the same atrocious noise. please reference the files I've included in the dropbox link in my OP. inter.dts was the ripped 5.1 audio inter-test.mp3 was encoded to 2.0mp3 format directly from 5.1dts and when played back on my laptop(s) (Windows, Linux, Mac; in Windows MP, ffplay, mplayer, vlc, xine, and more), Phones (s5,s3,iphone,htc one), tablet, ipod and my Sansa MP3 player all has horrific noise. inter.mp3 was generated by converting the 5.1DTS to 2.0PCM and then the 2.0PCM to 2.0MP3, sounds just fine when played back on all of my devices. I am fully aware that there should be NO NEED to use an intermediary wave format to downsample to stereo audio from 5.1 for a conversion to mp3. But that's exactly why I'm writing this problem into the group because it is NOT working as expected. IIRC this has come up before. The issue seems to be that sometimes -ac 2 normalises and sometimes it doesn't (depending on what codec is used). You can see whether or not the matrix is normalised with -loglevel debug. FWIW dts may contain meta data specifying a matrix typically (and by default if there is none) this will be partially normalised. To use this you put -request_channels 2 before -i infile.dts. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sat, 16 May 2015 09:17:59 + (UTC), Carl Eugen Hoyos ceho...@ag.or.at wrote: Bazza lamia at jeack.com.au writes: Please test the following: $ ffmpeg -i inter.dts -ac 2 out16.wav $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 Carl, I did test some of this stuff. - MP3 suffers overload - WAVs are OK - AC3 suffers overload - AAC are OK - MP2 are OK Sorry, I am apparently extremely dim-witted: Did you test the four lines above? Which of them sound ok, which of them do not sound ok? Thank you, Carl Eugen Tested them all Carl. It's not quite a case of sounding OK it's that they are patently just wrong. I'm using a Windows /Zeranoe build. Observing results through Adobe's Audition. Results are independant of bit rate (as you specify),I tried those (and some other rates). All fail. Also, it appears to be independant of the codecs 16 Vs 32 etc. The audio signals in John's sample have 6 mono. Most of those are up to the clipping level when viewed separately. When he combines them (let's say via declaring -ac 2) the channels (all 6) do indeed mix but the 'numbers' are summed greater than the streams capability - hence severe clipping and overload. This appears to happen in the case of AC3 and MP3 (which is John's complaint). In doing a bit of reading, it seems to be the case that L,R, FL and FR are usually attenuated by 3dB per signal. Maybe that 'routine' is being bypassed (or not even called) in AC3 or MP3 situations. This is a pure guess but the numbers do add up when we drop a volume by 10 dB. Luckily his levels nearly clip. Now, I must add, Adobe's Audition does not like viewing AC3 and AAC stuff directly so I do a re-convert from the AC3 and AAC back into WAV (just to view) but, no doubts about it, does not look good. Output in the non-behaving file format is just too high a level. FFMPEG generates no complaints so that's good. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sat, 16 May 2015 09:17:59 + (UTC), Carl Eugen Hoyos ceho...@ag.or.at wrote: Sorry, I am apparently extremely dim-witted: Did you test the four lines above? Which of them sound ok, which of them do not sound ok? Seem to have not explicitly answered the Q. Sorry. $ ffmpeg -i inter.dts -ac 2 out16.wav $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 1 = Good 2 = Good 3 = Bad 4 = Good and 1 you didn't list, AC3 = Bad ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Moritz Barsnick barsnick at gmx.net writes: $ ffmpeg -loglevel debug -i inter.dts -ac 2 -c:a libmp3lame -f null - $ ffmpeg -loglevel debug -i inter.dts -ac 2 -c:a pcm_s16le -f null - You can insert other arbitrary codecs at will. The former shows a matrix: 1.00 0.00 0.707107 0.00 0.707107 0.00 0.00 1.00 0.707107 0.00 0.00 0.707107 [auto-inserted resampler 0 at 0xb713840] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:fltp r:48000Hz while the latter shows: 0.414214 0.00 0.292893 0.00 0.292893 0.00 0.00 0.414214 0.292893 0.00 0.00 0.292893 [auto-inserted resampler 0 at 0xb3b55c0] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:s16 r:48000Hz I think this may be the described behavior. If this really is the issue, it should be reproducible with at least one of the command lines I proposed (namely for -acodec pcm_f32le) and it is possible to work-around the issue by forcing s16p as the mp3 encoding format. The mp3 encoder accepts fltp, s16p and s32p. But I would really appreciate if somebody can confirm that the issue is reproducible with pcm_f32le (and neither with s16le nor s32le). Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sat, May 16, 2015 at 13:07:01 +0200, Moritz Barsnick wrote: [auto-inserted resampler 0 @ 0xb713840] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:fltp r:48000Hz [auto-inserted resampler 0 @ 0xb3b55c0] ch:6 chl:5.1(side) fmt:fltp r:48000Hz - ch:2 chl:stereo fmt:s16 r:48000Hz (Unfortunately, my math tells me that the former matrix is bound to cause overflows with value-restricted numerical formats. There's something I don't seem to understand there. And is LFE really ignored when downmixing?) Ah, I see the difference now. In one case, it's outputting to floating point, so no clipping is expected. In the other case (depending on encoder obviously), it's assuming integer output. Moritz ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Bazza lamia at jeack.com.au writes: $ ffmpeg -i inter.dts -ac 2 out16.wav $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 1 = Good 2 = Good 3 = Bad 4 = Good and 1 you didn't list, AC3 = Bad Since your answer makes no sense (is ac3 doubly bad?), maybe you could map 1, 2, 3, 4 to out.mp2, out.ac3 and the two wav files? Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Sat, 16 May 2015 10:25:09 + (UTC), Carl Eugen Hoyos ceho...@ag.or.at wrote: Since your answer makes no sense (is ac3 doubly bad?), maybe you could map 1, 2, 3, 4 to out.mp2, out.ac3 and the two wav files? They say a picture is worth 1000 words etc so I'll do it this way. This is John's problem. Bad sound when 'downing' 5.1 to stereo. In his case, an attempt to use libmp3lame. Ignoring bitrates, the problem seems to boil down to whether the '-ac 2' option is being fully honoured (in channel numbers AND levels). The simple command 'ffmpeg -i inter.dts -ac 2 -ab 320k out.mp3', for an MP3 file, will produce this result. Plainly distorted. -) The links are captured PNG image files from Adobe's Audition. http://www.datafilehost.com/d/ef36d01a You asked for 4 tests to be carried out. Here they are. In order. 1). $ ffmpeg -i inter.dts -ac 2 out16.wav Conversions to WAV types are OK and will downmix. This one did. Sounds are fine http://www.datafilehost.com/d/85290f50 2). $ ffmpeg -i inter.dts -ac 2 -acodec pcm_s32le out32.wav Likewise for this other (32) codec. Performed as expected. http://www.datafilehost.com/d/fb55d8fe 3). $ ffmpeg -i inter.dts -ac 2 -ab 640k out.ac3 The process FAILS in this case(s). For AC3 and AAC. For these types of outputs, Audition does not recognize the file types so, to get around the problem of display, the AC3 and AAC types were converted back into a WAV type. The channels are there. The output level is just overloaded. The resultant output is virtually identical to John's initial problem wherein (I guess) the input channels are 'unweighted' http://www.datafilehost.com/d/932be2b6 4). $ ffmpeg -i inter.dts -ac 2 -ab 320k out.mp2 I initially thought that the MP2 was similarly irregular but now, I think I got myself confused with all the test files. This MP2 conversion appears to be OK. It too, is not importable into Audition so like (3) above, it had to be rendered to WAV. However, it seems to behave properly as expected. http://www.datafilehost.com/d/859a9402 This is as far as I can go. I'd like to believe that '-ac 2' was universal -) It is, in the sense that all channels are mixed. The volume option is a workaround. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
Maybe that's because you're converting lossy audio to another lossy audio format? ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Input 5.1 DTS, output 2.0 MP3: Atrocious Quality
On Fri, 15 May 2015 23:04:06 -0500, John L orionf...@hotmail.com wrote: Backstory: I have a system in place to automagically convert video files to smaller formats/versions on request to have a sort of mobile version for my father who travels extensively. The purpose is so that he can fit significantly more videos on his tablet than if they were the high quality rips. It all boils down to: ffmpeg -i [input-file] -ac 2 -c:v libx264 -c:a libmp3lame -b:v 1024k -preset fast [output-file] I was under the impression everything was hunky dory until I took a bunch of the shrunken movies on my phone on a roadtrip. A good many of the videos were as good as can be expected, and nothing was egregiously wrong. However on a few videos the audio was absolutely atrocious, blown out, clipping, and just noise from seemingly nowhere. One of the worst was Intersteller which was completely unwatchable after the first two minutes with all the blown out crescendos, pops, cracks, static, and voices of the deep adulterating the audio stream. All video files affected by this were 5.1DTS sources, but not all 5.1DTS were affected. When talking with my father he said it was a frequent enough occurrence that he suspected it was just because I had shrunk the file so small and was an artifact of that. He did confirm that most videos that were affected weren't as bad as the Interstellar conversion. ~/testing$ ffmpeg -version ffmpeg version 2.5.6-0ubuntu0.15.04.1 Copyright (c) 2000-2015 the FFmpeg developers built with gcc 4.9.2 (Ubuntu 4.9.2-10ubuntu13) configuration: --prefix=/usr --extra-version=0ubuntu0.15.04.1 --build-suffix=-ffmpeg --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --shlibdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --enable-gpl --enable-shared --disable-stripping --enable-avresample --enable-avisynth --enable-ladspa --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libmodplug --enable-libmp3lame --enable-libopenjpeg --enable-libopus --enable-libpulse --enable-libschroedinger --enable-libshine --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvorbis --enable-libwavpack --enable-libwebp --enable-lib xvid --enable-opengl --enable-x11grab --enable-libdc1394 --enable-libiec61883 --enable-libzvbi --enable-libzmq --enable-frei0r --enable-libvpx --enable-libx264 --enable-libsoxr --enable-gnut ls --enable-openal --enable-libopencv --enable-librtmp --enable-libx265 libavutil 54. 15.100 / 54. 15.100 libavcodec 56. 13.100 / 56. 13.100 libavformat56. 15.102 / 56. 15.102 libavdevice56. 3.100 / 56. 3.100 libavfilter 5. 2.103 / 5. 2.103 libavresample 2. 1. 0 / 2. 1. 0 libswscale 3. 1.101 / 3. 1.101 libswresample 1. 1.100 / 1. 1.100 libpostproc53. 3.100 / 53. 3.100 To troubleshoot I copied out a particularly bad snippet of audio ffmpeg -i Int*.mkv -vn -c copy -ss 1:30 -t 0:30 inter.dts Yes. That snippet has about 6 tracks. Some of them are clipping (all on their own). This audio clip is confirmed to be a good 5.1dts stream Good? Well OK -) ffmpeg -i inter.dts -ac 2 -c libmp3lame inter-test.mp3 This audio sample has the exact same audio defects as in the shrunken video Correct. All tracks (some of which had reached maximum encodable levels) are now being added/summed into 1 single (now) overloaded stream. Instead, take the 5.1 and _DOWNMIX_ all tracks to a single stereo for the phone/tablet by declaring -acodec -ac 2. No intermediate steps should be required. Consider also - Do you need pcm_s32le ? pcm_s16le is usual. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
