Re: [FFmpeg-user] differences between -vf and -filter_complex
Le primidi 21 floréal, an CCXXV, Paul B Mahol a écrit : > ffmpeg -f lavfi -i testsrc2=r=1:d=5,format=yuv422p10le -f framemd5 - This is still suboptimal. ffmpeg -lavfi testsrc2=r=1:d=5,format=yuv422p10le -f framemd5 - No need to wrap the frames in packets. Regards, -- Nicolas George signature.asc Description: Digital signature ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] differences between -vf and -filter_complex
On 5/10/17, Dave Ricewrote: > >> On May 10, 2017, at 2:13 PM, Paul B Mahol wrote: >> >> On 5/10/17, Dave Rice > wrote: >>> Hi ffmpeg-user, >>> >>> In some cases I've been using -vf and -filter_complex interchangeability >>> but >>> notice that some behavior is unexpectedly different. >>> >>> For instance if I use format=yuv422p10le within -vf and within >>> -filter_complex the output is different. Is there something else I should >>> add to the filterchain to get consistent results. >>> >>> Below is the output of: >>> ffmpeg -f lavfi -i testsrc2=r=1:d=5 -vf format=yuv422p10le -f >>> framemd5 - >>> and >>> ffmpeg -f lavfi -i testsrc2=r=1:d=5 -filter_complex format=yuv422p10le -f >>> framemd5 - >> >> Both commands are suboptimal, use: >> >> ffmpeg -f lavfi -i testsrc2=r=1:d=5,format=yuv422p10le -f framemd5 - > > I usually avoid lavfi formatted inputs when I can, because IIUC there's no > method to preserve frame interlacement data from the input through the > filterchain. See https://trac.ffmpeg.org/ticket/4057. How is that even related? format simply tells in what format testsrc2 will render its output. By your command it is doing slow scale work internally always. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] differences between -vf and -filter_complex
> On May 10, 2017, at 2:13 PM, Paul B Maholwrote: > > On 5/10/17, Dave Rice > wrote: >> Hi ffmpeg-user, >> >> In some cases I've been using -vf and -filter_complex interchangeability but >> notice that some behavior is unexpectedly different. >> >> For instance if I use format=yuv422p10le within -vf and within >> -filter_complex the output is different. Is there something else I should >> add to the filterchain to get consistent results. >> >> Below is the output of: >> ffmpeg -f lavfi -i testsrc2=r=1:d=5 -vf format=yuv422p10le -f >> framemd5 - >> and >> ffmpeg -f lavfi -i testsrc2=r=1:d=5 -filter_complex format=yuv422p10le -f >> framemd5 - > > Both commands are suboptimal, use: > > ffmpeg -f lavfi -i testsrc2=r=1:d=5,format=yuv422p10le -f framemd5 - I usually avoid lavfi formatted inputs when I can, because IIUC there's no method to preserve frame interlacement data from the input through the filterchain. See https://trac.ffmpeg.org/ticket/4057. Dave Rice ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] differences between -vf and -filter_complex
On 5/10/17, Dave Ricewrote: > Hi ffmpeg-user, > > In some cases I've been using -vf and -filter_complex interchangeability but > notice that some behavior is unexpectedly different. > > For instance if I use format=yuv422p10le within -vf and within > -filter_complex the output is different. Is there something else I should > add to the filterchain to get consistent results. > > Below is the output of: > ffmpeg -f lavfi -i testsrc2=r=1:d=5 -vf format=yuv422p10le -f > framemd5 - > and > ffmpeg -f lavfi -i testsrc2=r=1:d=5 -filter_complex format=yuv422p10le -f > framemd5 - Both commands are suboptimal, use: ffmpeg -f lavfi -i testsrc2=r=1:d=5,format=yuv422p10le -f framemd5 - ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] differences between -vf and -filter_complex
> On May 10, 2017, at 11:55 AM, Gyanwrote: > > On Wed, May 10, 2017 at 9:08 PM, Gyan wrote: > >> >> >> On Wed, May 10, 2017 at 8:17 PM, Dave Rice wrote: >> >>> >>> For instance if I use format=yuv422p10le within -vf and within >>> -filter_complex the output is different. Is there something else I should >>> add to the filterchain to get consistent results. >>> >>> Best Regards, >>> Dave Rice >>> >> >> These two commands produce the same result: > > ffmpeg -f lavfi -i testsrc2=r=1:d=5 -f lavfi -i testsrc2=r=1:d=5 > -filter_complex scale=flags=bicubic,format=yuv422p10le -f framemd5 - > > ffmpeg -f lavfi -i testsrc2=r=1:d=5 -f lavfi -i testsrc2=r=1:d=5 -vf > scale=flags=bicubic,format=yuv422p10le -f framemd5 - > > > 0, 0, 0,1, 307200, > b8ef3d6c4fc96d70fafbf20afd8f9cdf > 0, 1, 1,1, 307200, > dada36d572dfe9b75b799b207e54590b > 0, 2, 2,1, 307200, > f6b2bd8a2cafbe7261bd97cd41930ad3 > 0, 3, 3,1, 307200, > 5efbf76a4e8af59271de6a3e8cfb63f4 > 0, 4, 4,1, 307200, > f2289ce67d1f24230cee1219dc2557a5 Thank you so much. Adding `scale=flags=bicubic` into my workflow produced consistent results. I'm still a bit uncertain why these two routes opt for different defaults, but this method allows them to be forced to the same default. Thanks again. Dave Rice ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] differences between -vf and -filter_complex
On Wed, May 10, 2017 at 9:08 PM, Gyanwrote: > > > On Wed, May 10, 2017 at 8:17 PM, Dave Rice wrote: > >> >> For instance if I use format=yuv422p10le within -vf and within >> -filter_complex the output is different. Is there something else I should >> add to the filterchain to get consistent results. >> >> Best Regards, >> Dave Rice >> > > These two commands produce the same result: ffmpeg -f lavfi -i testsrc2=r=1:d=5 -f lavfi -i testsrc2=r=1:d=5 -filter_complex scale=flags=bicubic,format=yuv422p10le -f framemd5 - ffmpeg -f lavfi -i testsrc2=r=1:d=5 -f lavfi -i testsrc2=r=1:d=5 -vf scale=flags=bicubic,format=yuv422p10le -f framemd5 - 0, 0, 0,1, 307200, b8ef3d6c4fc96d70fafbf20afd8f9cdf 0, 1, 1,1, 307200, dada36d572dfe9b75b799b207e54590b 0, 2, 2,1, 307200, f6b2bd8a2cafbe7261bd97cd41930ad3 0, 3, 3,1, 307200, 5efbf76a4e8af59271de6a3e8cfb63f4 0, 4, 4,1, 307200, f2289ce67d1f24230cee1219dc2557a5 ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] differences between -vf and -filter_complex
On Wed, May 10, 2017 at 8:17 PM, Dave Ricewrote: > > For instance if I use format=yuv422p10le within -vf and within > -filter_complex the output is different. Is there something else I should > add to the filterchain to get consistent results. > > Best Regards, > Dave Rice > Does not affect conversions to RGB formats of any bit depth. Does not affect conversions if chroma sampling scheme is the same i.e. -f lavfi -i testsrc2=r=1:d=5,format=yuv422p or yuv422p16le. Also does not affect if only the luma plane is converted. So this has to do with converting the chroma sampling frequency in swscale. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-user] differences between -vf and -filter_complex
Hi ffmpeg-user, In some cases I've been using -vf and -filter_complex interchangeability but notice that some behavior is unexpectedly different. For instance if I use format=yuv422p10le within -vf and within -filter_complex the output is different. Is there something else I should add to the filterchain to get consistent results. Below is the output of: ffmpeg -f lavfi -i testsrc2=r=1:d=5 -vf format=yuv422p10le -f framemd5 - and ffmpeg -f lavfi -i testsrc2=r=1:d=5 -filter_complex format=yuv422p10le -f framemd5 - The output framemd5 do not match and there is not minor differences in the pixel values. Is this expected? Any recommendations to get the same result when using format=yuv422p10le in both -vf and -filter_complex? Which output is correct? ffmpeg -f lavfi -i testsrc2=r=1:d=5 -vf format=yuv422p10le -f framemd5 - ffmpeg version N-47211-g164e277326 Copyright (c) 2000-2017 the FFmpeg developers built with Apple LLVM version 8.1.0 (clang-802.0.38) configuration: --prefix=/usr/local/Cellar/ffmpeg/HEAD-164e277 --enable-shared --enable-pthreads --enable-gpl --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-ffplay --enable-libfreetype --enable-libmp3lame --enable-libopus --enable-libtesseract --enable-libvpx --enable-libx264 --enable-libxvid --enable-opencl --disable-lzma --enable-vda libavutil 55. 62.100 / 55. 62.100 libavcodec 57. 95.100 / 57. 95.100 libavformat57. 72.101 / 57. 72.101 libavdevice57. 7.100 / 57. 7.100 libavfilter 6. 88.100 / 6. 88.100 libavresample 3. 6. 0 / 3. 6. 0 libswscale 4. 7.101 / 4. 7.101 libswresample 2. 8.100 / 2. 8.100 libpostproc54. 6.100 / 54. 6.100 [opencl @ 0x10fe94d68] Same kernel code has been registered Last message repeated 1 times Input #0, lavfi, from 'testsrc2=r=1:d=5': Duration: N/A, start: 0.00, bitrate: N/A Stream #0:0: Video: rawvideo (I420 / 0x30323449), yuv420p, 320x240 [SAR 1:1 DAR 4:3], 1 fps, 1 tbr, 1 tbn, 1 tbc Stream mapping: Stream #0:0 -> #0:0 (rawvideo (native) -> rawvideo (native)) Press [q] to stop, [?] for help #format: frame checksums #version: 2 #hash: MD5 #software: Lavf57.72.101 #tb 0: 1/1 #media_type 0: video #codec_id 0: rawvideo #dimensions 0: 320x240 #sar 0: 1/1 #stream#, dts,pts, duration, size, hash Output #0, framemd5, to 'pipe:': Metadata: encoder : Lavf57.72.101 Stream #0:0: Video: rawvideo (Y3[10][10] / 0xA0A3359), yuv422p10le, 320x240 [SAR 1:1 DAR 4:3], q=2-31, 1536 kb/s, 1 fps, 1 tbn, 1 tbc Metadata: encoder : Lavc57.95.100 rawvideo 0, 0, 0,1, 307200, b8ef3d6c4fc96d70fafbf20afd8f9cdf 0, 1, 1,1, 307200, dada36d572dfe9b75b799b207e54590b 0, 2, 2,1, 307200, f6b2bd8a2cafbe7261bd97cd41930ad3 0, 3, 3,1, 307200, 5efbf76a4e8af59271de6a3e8cfb63f4 0, 4, 4,1, 307200, f2289ce67d1f24230cee1219dc2557a5 frame=5 fps=0.0 q=-0.0 Lsize= 1kB time=00:00:05.00 bitrate= 1.0kbits/s speed= 754x video:1500kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown ffmpeg -f lavfi -i testsrc2=r=1:d=5 -filter_complex format=yuv422p10le -f framemd5 - ffmpeg version N-47211-g164e277326 Copyright (c) 2000-2017 the FFmpeg developers built with Apple LLVM version 8.1.0 (clang-802.0.38) configuration: --prefix=/usr/local/Cellar/ffmpeg/HEAD-164e277 --enable-shared --enable-pthreads --enable-gpl --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-ffplay --enable-libfreetype --enable-libmp3lame --enable-libopus --enable-libtesseract --enable-libvpx --enable-libx264 --enable-libxvid --enable-opencl --disable-lzma --enable-vda libavutil 55. 62.100 / 55. 62.100 libavcodec 57. 95.100 / 57. 95.100 libavformat57. 72.101 / 57. 72.101 libavdevice57. 7.100 / 57. 7.100 libavfilter 6. 88.100 / 6. 88.100 libavresample 3. 6. 0 / 3. 6. 0 libswscale 4. 7.101 / 4. 7.101 libswresample 2. 8.100 / 2. 8.100 libpostproc54. 6.100 / 54. 6.100 [opencl @ 0x1026d6d68] Same kernel code has been registered Last message repeated 1 times Input #0, lavfi, from 'testsrc2=r=1:d=5': Duration: N/A, start: 0.00, bitrate: N/A Stream #0:0: Video: rawvideo (I420 / 0x30323449), yuv420p, 320x240 [SAR 1:1 DAR 4:3], 1 fps, 1 tbr, 1 tbn, 1 tbc Stream mapping: Stream #0:0 (rawvideo) -> format format -> Stream #0:0 (rawvideo) Press [q] to stop, [?] for help #format: frame checksums #version: 2 #hash: MD5 #software: Lavf57.72.101 #tb 0: 1/1 #media_type 0: video #codec_id 0: rawvideo #dimensions 0: 320x240 #sar 0: 1/1 #stream#, dts,pts, duration, size, hash Output #0, framemd5, to 'pipe:': Metadata: encoder : Lavf57.72.101