ffmpegsource - Project Hosting on Google Code

Today 14 hours ago issue 12 (Audio indexing fails first time round on too-long filenames) commented on by francois.visagie - Hmmm mod16.org, where TheFluff pointed me to for Aegisub builds after we met via http://devel.aegisub.org/ticket/899! r309 does indeed fix this issue, nice work done here, thanks guys! BTW check the nifty work-around I'd been using for this issue. Using the Avisynth script above: ffmpeg -y -i AVSCRIPT.AVS -an -vn NUL ffmpeg -y -i AVSCRIPT.AVS ... pass 1 output.m4v ffmpeg -y -i AVSCRIPT.AVS ... pass 2 output.m4v On the first ffmpeg invocation ffms2 indexes the input file. But ffmpeg exits immediately complaining of no video hence no time wasted :-). Hmmm mod16.org, where TheFluff pointed me to for Aegisub builds after we met via http://devel.aegisub.org/ticket/899! r309 does indeed fix this issue, nice work done here, thanks guys! BTW check the nifty work-around I'd been using for this issue. Using the Avisynth script above: ffmpeg -y -i AVSCRIPT.AVS -an -vn NUL ffmpeg -y -i AVSCRIPT.AVS ... pass 1 output.m4v ffmpeg -y -i AVSCRIPT.AVS ... pass 2 output.m4v On the first ffmpeg invocation ffms2 indexes the input file. But ffmpeg exits immediately complaining of no video hence no time wasted :-). Yesterday 34 hours ago issue 12 (Audio indexing fails first time round on too-long filenames) Status changed by kalle.bl...@gmail.com - This was most likely fixed in r309. Try this build and see if it fixes the issue: http://www.mod16.org/ffms2/ffms2-r309.7z Status: Fixed This was most likely fixed in r309. Try this build and see if it fixes the issue: http://www.mod16.org/ffms2/ffms2-r309.7z Status: Fixed Last 7 days May 12, 2010 r312 (Fix issues related to threading. Fix crashes caused by setti...) committed by kemuri9 - Fix issues related to threading. Fix crashes caused by setting threads to <= 0. Actually init the thread pools ffmpeg uses for threading video codecs. Fix issues related to threading. Fix crashes caused by setting threads to <= 0. Actually init the thread pools ffmpeg uses for threading video codecs. May 11, 2010 issue 12 (Audio indexing fails first time round on too-long filenames) reported by francois.visagie - What steps will reproduce the problem? 1. Download the sample video clip 2. Play the AVS script (mind the plugin path): Input="123456789012345678901234.avi" Cachefile=Input+".ffindex" LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\ffms2-r292\ffms2.dll") FFIndex(source=Input, cachefile=Cachefile, indexmask=-1, dumpmask=0, errorhandling=3, overwrite=false) Audio=FFAudioSource(source=Input, track=-1, cache=true, cachefile=Cachefile, adjustdelay=-1) Video=FFVideoSource(source=Input, track=-1, cache=true, cachefile=Cachefile, seekmode=1, rffmode=0, width=-1, height=-1, colorspace="") AudioDub(Video, Audio) What is the expected output? What do you see instead? Expected: video and audio to play Actual: No audio, and video displays "Audio track contains no audio frames... line 5" What version of the product are you using? On what operating system? ffmpegsource r292 with Avisynth 2.58 on Windows XP SP 3 Please provide any additional information below. 1. When the AVS script above is played above, the index file is correctly created 2. When the AVS script is played a second time with the index file still present, audio indexing and playback work 3. If the index file is deleted and the input filename shortened by one or more characters, audio indexing and playback work the first time round What steps will reproduce the problem? 1. Download the sample video clip 2. Play the AVS script (mind the plugin path): Input="123456789012345678901234.avi" Cachefile=Input+".ffindex" LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\ffms2-r292\ffms2.dll") FFIndex(source=Input, cachefile=Cachefile, indexmask=-1, dumpmask=0, errorhandling=3, overwrite=false) Audio=FFAudioSource(source=Input, track=-1, cache=true, cachefile=Cachefile, adjustdelay=-1) Video=FFVideoSource(source=Input, track=-1, cache=true, cachefile=Cachefile, seekmode=1, rffmode=0, width=-1, height=-1, colorspace="") AudioDub(Video, Audio) What is the expected output? What do you see instead? Expected: video and audio to play Actual: No audio, and video displays "Audio track contains no audio frames... line 5" What version of the product are you using? On what operating system? ffmpegsource r292 with Avisynth 2.58 on Windows XP SP 3 Please provide any additional information below. 1. When the AVS script above is played above, the index file is correctly created 2. When the AVS script is played a second time with the index file still present, audio indexing and playback work 3. If the index file is deleted and the input filename shortened by one or more characters, audio indexing and playback work the first time round Last 30 days May 03, 2010 issue 9 (Index file has an incorrect name) Status changed by kalle.bl...@gmail.com - Status: Fixed Status: Fixed May 03, 2010 issue 11 ('UINT64_C' was not declared in this scope) Status changed by kalle.bl...@gmail.com - Was fixed in r311. Status: Fixed Was fixed in r311. Status: Fixed May 03, 2010 issue 11 ('UINT64_C' was not declared in this scope) commented on by rotmer - Uh, just got here from google, seems like its not the best place to report this issue.. Anyway :). Uh, just got here from google, seems like its not the best place to report this issue.. Anyway :). May 03, 2010 issue 11 ('UINT64_C' was not declared in this scope) commented on by rotmer - The same issue i'm getting here when compiling chromium with ffmpeg from svn: In file included from /usr/include/libavutil/avutil.h:81:0, from /usr/include/libavcodec/avcodec.h:30, from out/Release/obj.target/geni/ffmpeg_stubs.cc:16: /usr/include/libavutil/common.h: In function 'int32_t av_clipl_int32(int64_t)': /usr/include/libavutil/common.h:154:47: error: 'UINT64_C' was not declared in this scope make: * [out/Release/obj.target/geni/ffmpeg_stubs.o] Error 1 The same issue i'm getting here when compiling chromium with ffmpeg from svn: In file included from /usr/include/libavutil/avutil.h:81:0, from /usr/include/libavcodec/avcodec.h:30, from out/Release/obj.target/geni/ffmpeg_stubs.cc:16: /usr/include/libavutil/common.h: In function 'int32_t av_clipl_int32(int64_t)': /usr/include/libavutil/common.h:154:47: error: 'UINT64_C' was not declared in this scope make: * [out/Release/obj.target/geni/ffmpeg_stubs.o] Error 1 May 02, 2010 r311 (add __STDC_CONSTANT_MACROS define to fix a compilation error...) committed by kemuri9 - add __STDC_CONSTANT_MACROS define to fix a compilation error. error caused by ffmpeg now using UINT64_C in libavutil/common.h, a header used by everything ffmpeg. autotools update by diogomfranco@gmail.com add __STDC_CONSTANT_MACROS define to fix a compilation error. error caused by ffmpeg now using UINT64_C in libavutil/common.h, a header used by everything ffmpeg. autotools update by diogomfranco@gmail.com May 02, 2010 issue 11 ('UINT64_C' was not declared in this scope) reported by yxml...@gmail.com - Configuration information there is detailed information in the info.txt In file included from D:/WinLinux/Minsys/local/include/libavutil/ avutil.h:81, from D:/WinLinux/Minsys/local/include/libavcodec/ avcodec.h:30, from D:/WinLinux/Minsys/local/include/libavformat/ avformat.h:56 , from src/core/audiosource.h:25, from src/core/audiosource.cpp:21: D:/WinLinux/Minsys/local/include/libavutil/common.h: In function 'int32_t av_clipl_int32(int64_t)': D:/WinLinux/Minsys/local/include/libavutil/common.h:154: error: 'UINT64_C' was not declared in this scope make: * [src/core/audiosource.lo] Error 1 Configuration information there is detailed information in the info.txt In file included from D:/WinLinux/Minsys/local/include/libavutil/ avutil.h:81, from D:/WinLinux/Minsys/local/include/libavcodec/ avcodec.h:30, from D:/WinLinux/Minsys/local/include/libavformat/ avformat.h:56 , from src/core/audiosource.h:25, from src/core/audiosource.cpp:21: D:/WinLinux/Minsys/local/include/libavutil/common.h: In function 'int32_t av_clipl_int32(int64_t)': D:/WinLinux/Minsys/local/include/libavutil/common.h:154: error: 'UINT64_C' was not declared in this scope make: * [src/core/audiosource.lo] Error 1 May 02, 2010 issue 11 ('UINT64_C' was not declared in this scope) reported by yxml...@gmail.com - Configuration information there is detailed information in the info.txt In file included from D:/WinLinux/Minsys/local/include/libavutil/ avutil.h:81, from D:/WinLinux/Minsys/local/include/libavcodec/ avcodec.h:30, from D:/WinLinux/Minsys/local/include/libavformat/ avformat.h:56 , from src/core/audiosource.h:25, from src/core/audiosource.cpp:21: D:/WinLinux/Minsys/local/include/libavutil/common.h: In function 'int32_t av_clipl_int32(int64_t)': D:/WinLinux/Minsys/local/include/libavutil/common.h:154: error: 'UINT64_C' was not declared in this scope make: * [src/core/audiosource.lo] Error 1 Configuration information there is detailed information in the info.txt In file included from D:/WinLinux/Minsys/local/include/libavutil/ avutil.h:81, from D:/WinLinux/Minsys/local/include/libavcodec/ avcodec.h:30, from D:/WinLinux/Minsys/local/include/libavformat/ avformat.h:56 , from src/core/audiosource.h:25, from src/core/audiosource.cpp:21: D:/WinLinux/Minsys/local/include/libavutil/common.h: In function 'int32_t av_clipl_int32(int64_t)': D:/WinLinux/Minsys/local/include/libavutil/common.h:154: error: 'UINT64_C' was not declared in this scope make: * [src/core/audiosource.lo] Error 1 Apr 21, 2010 r310 (Correct the timebase if it is invalid with respect to the fr...) committed by kemuri9 - Correct the timebase if it is invalid with respect to the framerate. Correct the timebase if it is invalid with respect to the framerate. Earlier this year Apr 07, 2010 issue 10 (FFAudioSource doesn't like lossless formats) commented on by wiia...@gmail.com - Also per Lord's suggestion, I tested with ffmpeg (r22716 last build from arrozcru's autobuilds). Both .flac and .wv, even inside Matroska, were bit-exact decoded. TTA had some kind of error and got the first second or so cut, same inside Matroska. Also uploaded the result files from decoding using ffaudiosource, because I thought it might be of interest to fix the issue: http://fdp.fansubban.org/rand/ffaudio/extra-ffms.wv http://fdp.fansubban.org/rand/ffaudio/extra-ffms.tta http://fdp.fansubban.org/rand/ffaudio/extra-ffms-mka.flac Also per Lord's suggestion, I tested with ffmpeg (r22716 last build from arrozcru's autobuilds). Both .flac and .wv, even inside Matroska, were bit-exact decoded. TTA had some kind of error and got the first second or so cut, same inside Matroska. Also uploaded the result files from decoding using ffaudiosource, because I thought it might be of interest to fix the issue: http://fdp.fansubban.org/rand/ffaudio/extra-ffms.wv http://fdp.fansubban.org/rand/ffaudio/extra-ffms.tta http://fdp.fansubban.org/rand/ffaudio/extra-ffms-mka.flac Apr 07, 2010 issue 10 (FFAudioSource doesn't like lossless formats) commented on by wiia...@gmail.com - As per Lord's suggestions, the source is: Format : PCM Codec ID : A_PCM/INT/LIT Channel(s) : 2 channels Sampling rate : 48.0 KHz Resolution : 16 bits As per Lord's suggestions, the source is: Format : PCM Codec ID : A_PCM/INT/LIT Channel(s) : 2 channels Sampling rate : 48.0 KHz Resolution : 16 bits Apr 06, 2010 issue 10 (FFAudioSource doesn't like lossless formats) commented on by wiia...@gmail.com - Tried now without the trims. Results are almost the same except: .flac in .mka = Length mismatch : 1:31.920000 (source) vs 1:31.562667 (ffaudio), 4412160 vs 4395008 samples .wv = Differences found: 8661386 sample(s), starting at 1.0000000 second(s), peak: 0.5469360 at 11.8185417 second(s), 1ch .tta = Differences found: 8657868 sample(s), starting at 1.0448958 second(s), peak: 0.6888428 at 34.3765625 second(s), 1ch Tried now without the trims. Results are almost the same except: .flac in .mka = Length mismatch : 1:31.920000 (source) vs 1:31.562667 (ffaudio), 4412160 vs 4395008 samples .wv = Differences found: 8661386 sample(s), starting at 1.0000000 second(s), peak: 0.5469360 at 11.8185417 second(s), 1ch .tta = Differences found: 8657868 sample(s), starting at 1.0448958 second(s), peak: 0.6888428 at 34.3765625 second(s), 1ch Apr 06, 2010 issue 10 (FFAudioSource doesn't like lossless formats) commented on by wiia...@gmail.com - Oh, haven't tried without the trims, so that might make some difference, I dunno? Oh, haven't tried without the trims, so that might make some difference, I dunno? Apr 06, 2010 issue 10 (FFAudioSource doesn't like lossless formats) reported by wiiaboo - What steps will reproduce the problem? 1. Convert .wav to .flac, .wv, .tta, even muxing them to .mka 2. FFAudioSource() on the resulting files 3. On those that work, use Bit Compare plugin in Foobar2000 What is the expected output? What do you see instead? Expected output is bit-exact decoding. In some, it doesn't work at all (.flac and .tta inside .mka). On the ones that do work, it isn't bit-exact, except in one single case. What version of the product are you using? On what operating system? Avisynth 2.5.8; FFMS2 r309 (mod16.org); MkvToolNix 3.3.0; Foobar2000 1.0.2 beta 1 Windows 7 64-bit Please provide any additional information below. .flac = Error "Invalid initial pts and dts" .tta in .mka = Error "Could not open audio codec" .flac in .mka = Differences found: 1965 sample(s), starting at 0.0000000 second(s), peak: 0.0014343 at 0.0006042 second(s), 1ch .tta = Differences found: 932551 sample(s), starting at 1.0448958 second(s), peak: 0.5542603 at 4.6865208 second(s), 1ch .wv = Differences found: 936851 sample(s), starting at 1.0000000 second(s), peak: 0.6018066 at 8.0636042 second(s), 1ch .wv in .mka = No differences in decoded data found. Also tested input with uncompressed data: .wav, .w64, .wav in .mka = No differences in decoded data found. .avs used in testing: http://pastebin.com/vmJZ0vEK Video: http://fdp.fansubban.org/rand/ffaudio/extra.mkv (737KB) Trimmed Video: http://fdp.fansubban.org/rand/ffaudio/extra-trimmed.mkv (65,4KB) Audio: http://fdp.fansubban.org/rand/ffaudio/extra.wv (8,96MB) What steps will reproduce the problem? 1. Convert .wav to .flac, .wv, .tta, even muxing them to .mka 2. FFAudioSource() on the resulting files 3. On those that work, use Bit Compare plugin in Foobar2000 What is the expected output? What do you see instead? Expected output is bit-exact decoding. In some, it doesn't work at all (.flac and .tta inside .mka). On the ones that do work, it isn't bit-exact, except in one single case. What version of the product are you using? On what operating system? Avisynth 2.5.8; FFMS2 r309 (mod16.org); MkvToolNix 3.3.0; Foobar2000 1.0.2 beta 1 Windows 7 64-bit Please provide any additional information below. .flac = Error "Invalid initial pts and dts" .tta in .mka = Error "Could not open audio codec" .flac in .mka = Differences found: 1965 sample(s), starting at 0.0000000 second(s), peak: 0.0014343 at 0.0006042 second(s), 1ch .tta = Differences found: 932551 sample(s), starting at 1.0448958 second(s), peak: 0.5542603 at 4.6865208 second(s), 1ch .wv = Differences found: 936851 sample(s), starting at 1.0000000 second(s), peak: 0.6018066 at 8.0636042 second(s), 1ch .wv in .mka = No differences in decoded data found. Also tested input with uncompressed data: .wav, .w64, .wav in .mka = No differences in decoded data found. .avs used in testing: http://pastebin.com/vmJZ0vEK Video: http://fdp.fansubban.org/rand/ffaudio/extra.mkv (737KB) Trimmed Video: http://fdp.fansubban.org/rand/ffaudio/extra-trimmed.mkv (65,4KB) Audio: http://fdp.fansubban.org/rand/ffaudio/extra.wv (8,96MB) Apr 03, 2010 issue 9 (Index file has an incorrect name) commented on by lordd.csc - "Does it read the index correctly when you open the same .avs again?" Yes, .avs opens normally, regardless of the index name. "Try this build and see if it fixes the issue: http://www.mod16.org/ffms2/ffms2- r302.7z" Now the index name is correct, thank you. Sorry that took so long to reply. "Does it read the index correctly when you open the same .avs again?" Yes, .avs opens normally, regardless of the index name. "Try this build and see if it fixes the issue: http://www.mod16.org/ffms2/ffms2- r302.7z" Now the index name is correct, thank you. Sorry that took so long to reply. Mar 26, 2010 r309 (Fix an invalid memory access that could cause crashes when o...) committed by kalle.bl...@gmail.com - Fix an invalid memory access that could cause crashes when opening audio (and potentially video too). Fix an invalid memory access that could cause crashes when opening audio (and potentially video too). Mar 18, 2010 r308 (fix a bug in FFVideoSource colorspace selection.) committed by kemuri9 - fix a bug in FFVideoSource colorspace selection. fix a bug in FFVideoSource colorspace selection. Mar 14, 2010 r307 (initial import of the avisynth C interface code. due to how ...) committed by kemuri9 - initial import of the avisynth C interface code. due to how finicky/broken autoconf and related tools are me doing anything with them ends up with something broken. As such the code is there and the end user can try autoreconf'ing it until this is fixed by someone else. TL;DR I really really hate autoconf... initial import of the avisynth C interface code. due to how finicky/broken autoconf and related tools are me doing anything with them ends up with something broken. As such the code is there and the end user can try autoreconf'ing it until this is fixed by someone else. TL;DR I really really hate autoconf... Mar 10, 2010 r306 (branch for avisynth C plugin interface to get a x86_64 plugi...) committed by kemuri9 - branch for avisynth C plugin interface to get a x86_64 plugin branch for avisynth C plugin interface to get a x86_64 plugin Mar 05, 2010 r305 (Move includes used by UTF-8 hack to UTF-8 ifdef) committed by diogomfr...@gmail.com - Move includes used by UTF-8 hack to UTF-8 ifdef Move includes used by UTF-8 hack to UTF-8 ifdef Mar 05, 2010 r304 (Define _FILE_OFFSET_BITS=64 Silences warnings about fseeko ...) committed by diogomfr...@gmail.com - Define _FILE_OFFSET_BITS=64 Silences warnings about fseeko being implicitly defined. Define _FILE_OFFSET_BITS=64 Silences warnings about fseeko being implicitly defined. Mar 05, 2010 r303 (Fix nix build) committed by diogomfr...@gmail.com - Fix nix build Fix nix build Mar 04, 2010 issue 9 (Index file has an incorrect name) commented on by kalle.bl...@gmail.com - Try this build and see if it fixes the issue: http://www.mod16.org/ffms2/ffms2-r302.7z Try this build and see if it fixes the issue: http://www.mod16.org/ffms2/ffms2-r302.7z Mar 04, 2010 r302 (fix the docs) committed by kalle.bl...@gmail.com - fix the docs fix the docs Mar 04, 2010 r301 (Rework the filename string handling a bit. Probably fixes is...) committed by kalle.bl...@gmail.com - Rework the filename string handling a bit. Probably fixes issue #9 , and FFMS_USE_UTF8_PATHS does no longer require patching ffmpeg. Rework the filename string handling a bit. Probably fixes issue #9 , and FFMS_USE_UTF8_PATHS does no longer require patching ffmpeg. Mar 04, 2010 issue 9 (Index file has an incorrect name) commented on by kalle.bl...@gmail.com - Does it read the index correctly when you open the same .avs again? I haven't actually tried it but I suspect the cause is that if you use a std::fstream() with a char instead of a wchar_t it will always assume the data is ISO8859-1 or CP1252 instead of the local codepage. The entire ffms2 API uses char's everywhere, but I guess that on windows we could do a MultiByteToWideChar() using the local codepage when writing/reading the index, if FFMS_USE_UTF8_PATHS isn't defined. It'd be a hack, but it would probably fix this bug. I'm not sure what it would break, though. Does it read the index correctly when you open the same .avs again? I haven't actually tried it but I suspect the cause is that if you use a std::fstream() with a char instead of a wchar_t it will always assume the data is ISO8859-1 or CP1252 instead of the local codepage. The entire ffms2 API uses char*'s everywhere, but I guess that on windows we could do a MultiByteToWideChar() using the local codepage when writing/reading the index, if FFMS_USE_UTF8_PATHS isn't defined. It'd be a hack, but it would probably fix this bug. I'm not sure what it would break, though. Feb 28, 2010 r300 (Add .so version information Breaks dynamic linking ABI with...) committed by diogomfr...@gmail.com - Add .so version information Breaks dynamic linking ABI with previous revisions. Add .so version information Breaks dynamic linking ABI with previous revisions. Feb 28, 2010 r299 (Add alloca.h check Some systems need to include it for the ...) committed by diogomfr...@gmail.com - Add alloca.h check Some systems need to include it for the matroska parser. Add alloca.h check Some systems need to include it for the matroska parser. Feb 27, 2010 r298 (Fix some warnings on matroskaparser.c Remove IsWritingApp f...) committed by diogomfr...@gmail.com - Fix some warnings on matroskaparser.c Remove IsWritingApp function, it's unused outside of an avi-mux gui hack. Add initializers to some variables. Fix some warnings on matroskaparser.c Remove IsWritingApp function, it's unused outside of an avi-mux gui hack. Add initializers to some variables. Feb 27, 2010 r297 (Add -std=c99 to CFLAGS All .c files are valid C99 and the m...) committed by diogomfr...@gmail.com - Add -std=c99 to CFLAGS All .c files are valid C99 and the matroska parser uses C99-specific features. Add -std=c99 to CFLAGS All .c files are valid C99 and the matroska parser uses C99-specific features. Feb 27, 2010 r296 (Delete src/core/libffms2.la) committed by diogomfr...@gmail.com - Delete src/core/libffms2.la Delete src/core/libffms2.la Feb 27, 2010 r295 (Add .gitignore to svn:ignore ) committed by diogomfr...@gmail.com - Add .gitignore to svn:ignore Add .gitignore to svn:ignore Feb 27, 2010 r294 (Remove recursive makefiles from svn:ignore Add ignore for n...) committed by diogomfr...@gmail.com - Remove recursive makefiles from svn:ignore Add ignore for newly generated files. Delete accidentally committed backup file. Remove recursive makefiles from svn:ignore Add ignore for newly generated files. Delete accidentally committed backup file. Feb 26, 2010 r293 (Replace recursive make with single Makefile) committed by diogomfr...@gmail.com - Replace recursive make with single Makefile Replace recursive make with single Makefile Feb 25, 2010 ffms2-r292.7z (FFmpegsource r292, Pthreads auto-initializationn test, inclu...) file uploaded by asmith5...@gmail.com - Labels: Type-Archive OpSys-Windows Featured Labels: Type-Archive OpSys-Windows Featured Feb 20, 2010 issue 9 (Index file has an incorrect name) reported by lordd.csc - What steps will reproduce the problem? 1. Take a file, the name of which contains non-Latin characters from one of the Windows codepages. For example, "Проверка.avi" (cp1251); 2. Create a AviSynth script in the same codepage with FFVideoSource ("Проверка.avi"); 3. Open script in any video editor. What is the expected output? What do you see instead? Everything will work fine, but the index file has an incorrect name: Ïðîâåðêà.avi.ffindex. What version of the product are you using? On what operating system? FFMS 2.13 on Windows 7 x64. Please provide any additional information below. What steps will reproduce the problem? 1. Take a file, the name of which contains non-Latin characters from one of the Windows codepages. For example, "Проверка.avi" (cp1251); 2. Create a AviSynth script in the same codepage with FFVideoSource ("Проверка.avi"); 3. Open script in any video editor. What is the expected output? What do you see instead? Everything will work fine, but the index file has an incorrect name: Ïðîâåðêà.avi.ffindex. What version of the product are you using? On what operating system? FFMS 2.13 on Windows 7 x64. Please provide any additional information below. Feb 17, 2010 issue 6 (Corrupted frames while randomly seeking with FFVideoSource) commented on by wiiaboo - The only difference besides GDSMux working is that both using FFVideoSource("bd.m2ts") and FFVideoSource("gdsmux.mkv") give a certain number of frames and using DGAVC and the other remuxed mkv, i get the +2 frames. In this m2ts's case, the last 20 or so frames are black so I can't really see any problem with getting less two frames. Checked the timecodes when using FFVideoSource and indeed, there are a couple frames extra at the end of the gdsmux remux but the rest have pretty much the same timestamps. Reported FPS are the same for m2ts and remuxed mkvs (24000/1001) and DGAVC gives 240000/10010. No problem there. Is there any way to know for sure what the correct frame count is? The only difference besides GDSMux working is that both using FFVideoSource("bd.m2ts") and FFVideoSource("gdsmux.mkv") give a certain number of frames and using DGAVC and the other remuxed mkv, i get the +2 frames. In this m2ts's case, the last 20 or so frames are black so I can't really see any problem with getting less two frames. Checked the timecodes when using FFVideoSource and indeed, there are a couple frames extra at the end of the gdsmux remux but the rest have pretty much the same timestamps. Reported FPS are the same for m2ts and remuxed mkvs (24000/1001) and DGAVC gives 240000/10010. No problem there. Is there any way to know for sure what the correct frame count is? Feb 17, 2010 issue 6 (Corrupted frames while randomly seeking with FFVideoSource) commented on by wiiaboo - Still happens with BD .m2ts, if FFVideoSource is used directly with them. And no, still happens with remuxes using eac3to, eac3to .264 muxed to mkv with mmg and tsmuxer .264 remuxed to mkv with mmg. GDSMux is still the only one that doesn't present corrupted frames. Still happens with BD .m2ts, if FFVideoSource is used directly with them. And no, still happens with remuxes using eac3to, eac3to .264 muxed to mkv with mmg and tsmuxer .264 remuxed to mkv with mmg. GDSMux is still the only one that doesn't present corrupted frames. Feb 17, 2010 issue 6 (Corrupted frames while randomly seeking with FFVideoSource) commented on by wiiaboo - Still happens with BD .m2ts, if FFVideoSource is used directly with them. And yes, still happens with remuxes using eac3to, eac3to .264 muxed to mkv with mmg and tsmuxer .264 remuxed to mkv with mmg. GDSMux is still the only one that doesn't present corrupted frames. Still happens with BD .m2ts, if FFVideoSource is used directly with them. And yes, still happens with remuxes using eac3to, eac3to .264 muxed to mkv with mmg and tsmuxer .264 remuxed to mkv with mmg. GDSMux is still the only one that doesn't present corrupted frames. Feb 17, 2010 issue 6 (Corrupted frames while randomly seeking with FFVideoSource) Status changed by kalle.bl...@gmail.com - This has probably been fixed in 2.13, can you confirm that it is gone now? Status: Accepted This has probably been fixed in 2.13, can you confirm that it is gone now? Status: Accepted Feb 17, 2010 ProjectsThatUseFFMS2 (List of projects that use FFMS2) Wiki page edited by kalle.bl...@gmail.com - Revision r292 Edited wiki page through web user interface. Revision r292 Edited wiki page through web user interface. Feb 16, 2010 r291 (Silence warnings when using clang++ -Wswitch-enum clang++ i...) committed by diogomfr...@gmail.com - Silence warnings when using clang++ -Wswitch-enum clang++ is still unable to compile ffms2 though. Silence warnings when using clang++ -Wswitch-enum clang++ is still unable to compile ffms2 though. Feb 12, 2010 ProjectsThatUseFFMS2 (List of projects that use FFMS2) Wiki page commented on by flebber.crue - StaxRip http://forum.doom9.org/showthread.php?t=102652 (Discussion) http://www.stax76.bplaced.net/ (Project Home) StaxRip http://forum.doom9.org/showthread.php?t=102652 (Discussion) http://www.stax76.bplaced.net/ (Project Home) Feb 12, 2010 ProjectsThatUseFFMS2 (List of projects that use FFMS2) Wiki page edited by kalle.bl...@gmail.com - Revision r290 Edited wiki page through web user interface. Revision r290 Edited wiki page through web user interface. Feb 12, 2010 ProjectsThatUseFFMS2 (List of projects that use FFMS2) Wiki page added by kalle.bl...@gmail.com - Revision r289 Created wiki page through web user interface. Revision r289 Created wiki page through web user interface. Feb 12, 2010 ffms2-2.13_src.7z (FFmpegSource 2.13 source code) file uploaded by kalle.bl...@gmail.com - Labels: Featured Type-Source OpSys-All Labels: Featured Type-Source OpSys-All Feb 12, 2010 ffms2-2.13.7z (FFmpegSource 2.13) file uploaded by kalle.bl...@gmail.com - Labels: Featured Type-Archive OpSys-Windows Labels: Featured Type-Archive OpSys-Windows

issue 10 (FFAudioSource doesn't like lossless formats) reported by wiiaboo - What steps will reproduce the problem? 1. Convert .wav to .flac, .wv, .tta, even muxing them to .mka 2. FFAudioSource() on the resulting files 3. On those that work, use Bit Compare plugin in Foobar2000 What is the expected output? What do you see instead? Expected output is bit-exact decoding. In some, it doesn't work at all (.flac and .tta inside .mka). On the ones that do work, it isn't bit-exact, except in one single case. What version of the product are you using? On what operating system? Avisynth 2.5.8; FFMS2 r309 (mod16.org); MkvToolNix 3.3.0; Foobar2000 1.0.2 beta 1 Windows 7 64-bit Please provide any additional information below. .flac = Error "Invalid initial pts and dts" .tta in .mka = Error "Could not open audio codec" .flac in .mka = Differences found: 1965 sample(s), starting at 0.0000000 second(s), peak: 0.0014343 at 0.0006042 second(s), 1ch .tta = Differences found: 932551 sample(s), starting at 1.0448958 second(s), peak: 0.5542603 at 4.6865208 second(s), 1ch .wv = Differences found: 936851 sample(s), starting at 1.0000000 second(s), peak: 0.6018066 at 8.0636042 second(s), 1ch .wv in .mka = No differences in decoded data found. Also tested input with uncompressed data: .wav, .w64, .wav in .mka = No differences in decoded data found. .avs used in testing: http://pastebin.com/vmJZ0vEK Video: http://fdp.fansubban.org/rand/ffaudio/extra.mkv (737KB) Trimmed Video: http://fdp.fansubban.org/rand/ffaudio/extra-trimmed.mkv (65,4KB) Audio: http://fdp.fansubban.org/rand/ffaudio/extra.wv (8,96MB)

What steps will reproduce the problem? 1. Convert .wav to .flac, .wv, .tta, even muxing them to .mka 2. FFAudioSource() on the resulting files 3. On those that work, use Bit Compare plugin in Foobar2000 What is the expected output? What do you see instead? Expected output is bit-exact decoding. In some, it doesn't work at all (.flac and .tta inside .mka). On the ones that do work, it isn't bit-exact, except in one single case. What version of the product are you using? On what operating system? Avisynth 2.5.8; FFMS2 r309 (mod16.org); MkvToolNix 3.3.0; Foobar2000 1.0.2 beta 1 Windows 7 64-bit Please provide any additional information below. .flac = Error "Invalid initial pts and dts" .tta in .mka = Error "Could not open audio codec" .flac in .mka = Differences found: 1965 sample(s), starting at 0.0000000 second(s), peak: 0.0014343 at 0.0006042 second(s), 1ch .tta = Differences found: 932551 sample(s), starting at 1.0448958 second(s), peak: 0.5542603 at 4.6865208 second(s), 1ch .wv = Differences found: 936851 sample(s), starting at 1.0000000 second(s), peak: 0.6018066 at 8.0636042 second(s), 1ch .wv in .mka = No differences in decoded data found. Also tested input with uncompressed data: .wav, .w64, .wav in .mka = No differences in decoded data found. .avs used in testing: http://pastebin.com/vmJZ0vEK Video: http://fdp.fansubban.org/rand/ffaudio/extra.mkv (737KB) Trimmed Video: http://fdp.fansubban.org/rand/ffaudio/extra-trimmed.mkv (65,4KB) Audio: http://fdp.fansubban.org/rand/ffaudio/extra.wv (8,96MB)

Today

Yesterday

Last 7 days

Last 30 days

Earlier this year