ffmpeg help - buffer options, output details?

[JDługosz]

I'm trying to convert a file to something I can work with (dealing with a language barrier between tools)

[E:\Downloads\ffmpeg-20131029-git-f86387b-win64-static\bin]ffmpeg -i E:\scratch\Newyear\00000000.avi
ffmpeg version N-57605-gf86387b Copyright (c) 2000-2013 the FFmpeg developers
  built on Oct 29 2013 18:08:53 with gcc 4.8.2 (GCC)
  configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzli
le-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --
ble-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-libr
eex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-
ack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
  libavutil      52. 48.100 / 52. 48.100
  libavcodec     55. 39.100 / 55. 39.100
  libavformat    55. 19.104 / 55. 19.104
  libavdevice    55.  5.100 / 55.  5.100
  libavfilter     3. 90.100 /  3. 90.100
  libswscale      2.  5.101 /  2.  5.101
  libswresample   0. 17.104 /  0. 17.104
  libpostproc    52.  3.100 / 52.  3.100
Input #0, avi, from 'E:\scratch\Newyear\00000000.avi':
  Metadata:
    TCOD            : 0
  Duration: 00:07:04.26, start: 0.000000, bitrate: 1591842 kb/s
    Stream #0:0: Video: r210 (r210 / 0x30313272), rgb48le(10 bpc), 1920x1080, 23.98 tbr, 23.98 tbn,
[mpeg @ 00000000024d88e0] VBV buffer size not set, muxing may fail
Output #0, mpeg, to 'E:\scratch\Newyear\tryit.mpg':
  Metadata:
    TCOD            : 0
    encoder         : Lavf55.19.104
    Stream #0:0: Video: mpeg1video, yuv420p, 1920x1080, q=2-31, 200 kb/s, 90k tbn, 23.98 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (r210 -> mpeg1video)
Press [q] to stop, [?] for help
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=233950 size=274621
[mpeg @ 00000000024d88e0] packet too large, ignoring buffer limits to mux it
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=233950 size=274621
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=235979 size=274621
[mpeg @ 00000000024d88e0] packet too large, ignoring buffer limits to mux it
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=235979 size=274621
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=238020 size=274621
[mpeg @ 00000000024d88e0] packet too large, ignoring buffer limits to mux it
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=238020 size=274621
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=240061 size=274621
[mpeg @ 00000000024d88e0] packet too large, ignoring buffer limits to mux it
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=240061 size=274621
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=242102 size=274621
[mpeg @ 00000000024d88e0] packet too large, ignoring buffer limits to mux it
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=242102 size=274621
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=244143 size=274621
[mpeg @ 00000000024d88e0] packet too large, ignoring buffer limits to mux it
[mpeg @ 00000000024d88e0] buffer underflow st=0 bufi=244143 size=274621

I can't find anything about 'buffer' in the man page online. But, I think the mpeg details isn't something I'm really caring about—I just used that as a presumably easy test to see if ffmpeg could read and handle the input.

I want to convert this to something that Avidemux can read, and uncompressed. Their docs don't really say much, but I'm guessing that the 10-bit data is what confuses it. So what options do I use to convert this to a "more ordinary" AVI or whatever?

The interesting part above is:
Stream #0:0: Video: r210 (r210 / 0x30313272), rgb48le(10 bpc), 1920x1080, 23.98 tbr, 23.98 tbn,

[poisondeathray]

r210 is uncompressed 10bit RGB

avidemux cannot handle high bit depth formats

What do you need to do in avidemux ? Maybe there are better suited programs

Your options for converting (assuming you want lossless) are 8bit FFV1 YUV 4:2:0 , 8bit x264 YUV 4:2:0 at q=0 . You will degrade the footage (you're going from higher bitdepths to lower bitdepths, and full color to chroma subsampled; but if this was originally your h.264 DSLR footage it's not a problem because it was 8bit 4:2:0 to begin with)

eg

ffmpeg -i input.avi -vcodec ffv1 -pix_fmt yuv420p -acodec copy output.avi

[JDługosz]

What do you need to do in avidemux ? Maybe there are better suited programs

Thanks so much for the reply.
I've played with the x264 options for encoding (for youtube) and appreciate how it gives full control rather than some dumbed-down one-size-fits-all that nonetheless gives warnings from youtube when uploaded.
I also found that one of the noise filters does a good job on the scintillation I get on low-light video from my DSLR.
It's fine for simple trimming and joining when I don't do anything else with editing.

Your options for converting (assuming you want lossless) are 8bit FFV1 YUV 4:2:0 , 8bit x264 YUV 4:2:0 at q=0 . You will degrade the footage (you're going from higher bitdepths to lower bitdepths, and full color to chroma subsampled; but if this was originally your h.264 DSLR footage it's not a problem because it was 8bit 4:2:0 to begin with)
eg

ffmpeg -i input.avi -vcodec ffv1 -pix_fmt yuv420p -acodec copy output.avi

Great! Then I'll be back in business. Of course, if there are better ways for subsequent files, I'd love to hear about it.

—John

[poisondeathray]

Of course, if there are better ways for subsequent files, I'd love to hear about it.

Are you referring to batch encoding - yes it's easy to do with ffmpeg

I forgot you should add -vf colormatrix=bt709:bt601 , assuming you want 709 as the previous discussion. The reason is the conversion from RGB to YUV in ffmpeg will use 601 otherwise (if you recall, slight color shifting)

Try it out first on a single file first. If you're happy with the current workflow and you want to batch encode , say a directory of r210 AVI's to 4:2:0 8bit FFV1 AVI's for input into avidemux

for %%a in ("*.avi") do ffmpeg -i "%%a" -vcodec ffv1 -pix_fmt yuv420p -vf colormatrix=bt709:bt601 -acodec copy "%%~na.ffv1.avi"

If you know the name of the denoise filter that you are happy with in avidemux, it might have an equivalent in ffmpeg (then you could do everything in ffmpeg, even encode with libx264) . Or if you want better denoise approaches, avisynth is the way to go

[JDługosz]

r210 is uncompressed 10bit RGB

avidemux cannot handle high bit depth formats
What do you need to do in avidemux ? Maybe there are better suited programs

Hmm, ffmpeg can directly do the x264 encoding with the same details I specified to libx264 via avidemux, if I had any clue as to the way to set those options.

And it does "filters" (http://ffmpeg.org/ffmpeg.html#Filtering) but what filters? Does it have a temporo-spacial noise filter, or does it use filters from some standard plug-in architecture?

[poisondeathray]

yes ffmpeg can use the same encoding options as avidemux using libx264

ffmpeg filters are here
http://ffmpeg.org/ffmpeg-filters.html

hqdn3d is probably a common one used for denoising in ffmpeg (yes it has spatial and temporal components)
http://ffmpeg.org/ffmpeg-filters.html#hqdn3d-1

[JDługosz]

If you know the name of the denoise filter that you are happy with in avidemux, it might have an equivalent in ffmpeg (then you could do everything in ffmpeg, even encode with libx264) . Or if you want better denoise approaches, avisynth is the way to go

I believe it was denoise3dhq.dll with the defaults all doubled.

I'm attracted to avisynth in general, being a software developer and used to expressing myself with computer languages. But I've not gotten it to work on my win-x64 machine. At least, not the way I tried before for frameserving — maybe I should try it again.

I'm happy to use whatever denoise option works, for the tools used. For avidemux I just tried them until I eyeballed what did a good job without messing it up otherwise.

[JDługosz]

yes ffmpeg can use the same encoding options as avidemux using libx264

ffmpeg filters are here
http://ffmpeg.org/ffmpeg-filters.html

hqdn3d is probably a common one used for denoising in ffmpeg (yes it has spatial and temporal components)
http://ffmpeg.org/ffmpeg-filters.html#hqdn3d-1

Great! I'll have to look for more docs in the same parent directory that are not linked from the "man page" document. I thought it was a bit light on details!

So, to replace my task with ffmpeg, how do I specify details to the x264 encoder? (Does it use the x264 library?)
Here are the json-format settings from avidemux:

{
	"general" : {
		"params" : "AQ=26",
		"threads" : 99,
		"fast_first_pass" : true
	},
	"level" : 31,
	"vui" : {
		"sar_height" : 1,
		"sar_width" : 1
	},
	"MaxRefFrames" : 3,
	"MinIdr" : 12,
	"MaxIdr" : 120,
	"i_scenecut_threshold" : 40,
	"MaxBFrame" : 2,
	"i_bframe_adaptive" : 2,
	"i_bframe_bias" : 0,
	"i_bframe_pyramid" : 2,
	"b_deblocking_filter" : true,
	"i_deblocking_filter_alphac0" : 0,
	"i_deblocking_filter_beta" : 0,
	"cabac" : true,
	"interlaced" : false,
	"analyze" : {
		"b_8x8" : true,
		"b_i4x4" : true,
		"b_i8x8" : true,
		"b_p8x8" : false,
		"b_p16x16" : true,
		"b_b16x16" : true,
		"weighted_pred" : 0,
		"weighted_bipred" : false,
		"direct_mv_pred" : 3,
		"chroma_offset" : 0,
		"me_method" : 1,
		"mv_range" : 16,
		"subpel_refine" : 7,
		"chroma_me" : true,
		"mixed_references" : true,
		"trellis" : 1,
		"psy_rd" : 0,
		"psy_trellis" : 0,
		"fast_pskip" : true,
		"dct_decimate" : true,
		"noise_reduction" : 0,
		"psy" : true,
		"intra_luma" : 21,
		"inter_luma" : 21
	},
	"ratecontrol" : {
		"rc_method" : 0,
		"qp_constant" : 0,
		"qp_min" : 0,
		"qp_max" : 0,
		"qp_step" : 0,
		"bitrate" : 0,
		"vbv_max_bitrate" : 0,
		"vbv_buffer_size" : 0,
		"vbv_buffer_init" : 0,
		"ip_factor" : 0,
		"pb_factor" : 0,
		"aq_mode" : 1,
		"aq_strength" : 1,
		"mb_tree" : true,
		"lookahead" : 40
	}
}

[JDługosz]

ffmpeg -i input.avi -vcodec ffv1 -pix_fmt yuv420p -acodec copy output.avi

Uh... I get a good frame followed by confetti for a few seconds, then repeat. The screenshot (from VLC) is just a corner of the 1080p picture. Avidemux shows the first frame but turns solid green if I try to play; also loaded instantly which makes me think it didn't really go through it (or maybe AVI doesn't need to scan like other containers?)

General
Complete name                            : E:\scratch\Newyear\tryit.avi
Format                                   : AVI
Format/Info                              : Audio Video Interleave
Format profile                           : OpenDML
File size                                : 9.61 GiB
Duration                                 : 7mn 4s
Overall bit rate                         : 195 Mbps
Writing application                      : Lavf55.19.104

Video
ID                                       : 0
Format                                   : FFV1
Codec ID                                 : FFV1
Duration                                 : 7mn 4s
Bit rate                                 : 195 Mbps
Width                                    : 1 920 pixels
Height                                   : 1 080 pixels
Display aspect ratio                     : 16:9
Frame rate                               : 23.976 fps
Compression mode                         : Lossless
Bits/(Pixel*Frame)                       : 3.915
Stream size                              : 9.61 GiB (100%)

[poisondeathray]

It's working ok here, r210 is decoded and converted to ffv1 ok, and 8bit 420 ffv1 decodes correctly in avidemux 2.6

Test to see if your r210.avi is being decoded correctly - use ffplay

ffplay -i r210.avi

Otherwise try updating your ffmpeg binary, it's a few months old

ffmpeg-20131029-git-f86387b-win64-static

[poisondeathray]

So, to replace my task with ffmpeg, how do I specify details to the x264 encoder? (Does it use the x264 library)

Did you want to pipe ffmpeg to x264.exe, or use ffmpeg libx264 to encode directly ?

It uses almost the same syntax as x264 with presets and tunings, but there are a few differences

The switches are listed in the ffmpeg help under the libx264 section (ffmpeg -h full) , and any non specified switches can be accessed through -x264-params , or -x264opts

libx264 AVOptions:
  -preset            <string>     E..V.. Set the encoding preset (cf. x264 --fullhelp) (default "medium")
  -tune              <string>     E..V.. Tune the encoding params (cf. x264 --fullhelp)
  -profile           <string>     E..V.. Set profile restrictions (cf. x264 --fullhelp) 
  -fastfirstpass     <int>        E..V.. Use fast settings when encoding first pass (from 0 to 1) (default 1)
  -level             <string>     E..V.. Specify level (as defined by Annex A)
  -passlogfile       <string>     E..V.. Filename for 2 pass stats
  -wpredp            <string>     E..V.. Weighted prediction for P-frames
  -x264opts          <string>     E..V.. x264 options
  -crf               <float>      E..V.. Select the quality for constant quality mode (from -1 to FLT_MAX) (default -1)
  -crf_max           <float>      E..V.. In CRF mode, prevents VBV from lowering quality beyond this point. (from -1 to FLT_MAX) (default -1)
  -qp                <int>        E..V.. Constant quantization parameter rate control method (from -1 to INT_MAX) (default -1)
  -aq-mode           <int>        E..V.. AQ method (from -1 to INT_MAX) (default -1)
     none                         E..V..
     variance                     E..V.. Variance AQ (complexity mask)
     autovariance                 E..V.. Auto-variance AQ (experimental)
  -aq-strength       <float>      E..V.. AQ strength. Reduces blocking and blurring in flat and textured areas. (from -1 to FLT_MAX) (default -1)
  -psy               <int>        E..V.. Use psychovisual optimizations. (from -1 to 1) (default -1)
  -psy-rd            <string>     E..V.. Strength of psychovisual optimization, in <psy-rd>:<psy-trellis> format.
  -rc-lookahead      <int>        E..V.. Number of frames to look ahead for frametype and ratecontrol (from -1 to INT_MAX) (default -1)
  -weightb           <int>        E..V.. Weighted prediction for B-frames. (from -1 to 1) (default -1)
  -weightp           <int>        E..V.. Weighted prediction analysis method. (from -1 to INT_MAX) (default -1)
     none                         E..V..
     simple                       E..V..
     smart                        E..V..
  -ssim              <int>        E..V.. Calculate and print SSIM stats. (from -1 to 1) (default -1)
  -intra-refresh     <int>        E..V.. Use Periodic Intra Refresh instead of IDR frames. (from -1 to 1) (default -1)
  -bluray-compat     <int>        E..V.. Bluray compatibility workarounds. (from -1 to 1) (default -1)
  -b-bias            <int>        E..V.. Influences how often B-frames are used (from INT_MIN to INT_MAX) (default INT_MIN)
  -b-pyramid         <int>        E..V.. Keep some B-frames as references. (from -1 to INT_MAX) (default -1)
     none                         E..V..
     strict                       E..V.. Strictly hierarchical pyramid
     normal                       E..V.. Non-strict (not Blu-ray compatible)
  -mixed-refs        <int>        E..V.. One reference per partition, as opposed to one reference per macroblock (from -1 to 1) (default -1)
  -8x8dct            <int>        E..V.. High profile 8x8 transform. (from -1 to 1) (default -1)
  -fast-pskip        <int>        E..V.. (from -1 to 1) (default -1)
  -aud               <int>        E..V.. Use access unit delimiters. (from -1 to 1) (default -1)
  -mbtree            <int>        E..V.. Use macroblock tree ratecontrol. (from -1 to 1) (default -1)
  -deblock           <string>     E..V.. Loop filter parameters, in <alpha:beta> form.
  -cplxblur          <float>      E..V.. Reduce fluctuations in QP (before curve compression) (from -1 to FLT_MAX) (default -1)
  -partitions        <string>     E..V.. A comma-separated list of partitions to consider. Possible values: p8x8, p4x4, b8x8, i8x8, i4x4, none, all
  -direct-pred       <int>        E..V.. Direct MV prediction mode (from -1 to INT_MAX) (default -1)
     none                         E..V..
     spatial                      E..V..
     temporal                     E..V..
     auto                         E..V..
  -slice-max-size    <int>        E..V.. Limit the size of each slice in bytes (from -1 to INT_MAX) (default -1)
  -stats             <string>     E..V.. Filename for 2 pass stats
  -nal-hrd           <int>        E..V.. Signal HRD information (requires vbv-bufsize; cbr not allowed in .mp4) (from -1 to INT_MAX) (default -1)
     none                         E..V..
     vbr                          E..V..
     cbr                          E..V..
  -x264-params       <string>     E..V.. Override the x264 configuration using a :-separated list of key=value parameters

[JDługosz]

Test to see if your r210.avi is being decoded correctly -

My earlier test before understanding the output parameters just used the default mpeg output, and that plays. So ffmpeg did read the input and convert to a highly-compressed MPEG file.

Using ffplay on the output files, it freezes after a few seconds while the text keeps showing it's doing something, and clicking does a seek which shows a picture (frozen) when I click in several places. ffplay handled the mpg file, so I know it works in general on my system.

Did you want to pipe ffmpeg to x264.exe, or use ffmpeg libx264 to encode directly ?

I didn't know the former was available. Doing it directly is simpler, but ffmpeg seems to be single threaded (?) so maybe using two processes would be faster. Maybe not, if the input doesn't need any effort but saturates the I/O.

Thanks for your patient help.

[poisondeathray]

mpeg1 is terrible - I'm sure you saw that

1st thing I would try is updating your binary

Using ffplay on the output files, it freezes after a few seconds while the text keeps showing it's doing something, and clicking does a seek which shows a picture (frozen) when I click in several places. ffplay handled the mpg file, so I know it works in general on my system.

Which output files ? From resolve ? Be specific - did you mean the r210 AVI's ? or converted ffv1 AVI's ?

Doing it directly is simpler, but ffmpeg seems to be single threaded (?) so maybe using two processes would be faster.

No, piping adds overhead, and ffmpeg is multithreaded


But if you want to skip avidemux and do everything in batch with ffmpeg (you still have to test -vf filter options for denoising)

I'm not that familiar with avidemux' json interpretation of x264. I'll try to translate it to ffmpeg libx264 syntax but a few Q's:

1) "params" : "AQ=26" - is that AQ strength 2.6 ??!!! It can't be - because later on it says "aq_strength" : 1, Or is that CRF ? AQ means "average quality" ?

2) are you sure you want threads 99? The more threads , the lower the quality. The optimum is logical cores * 1.5

3) it says first pass fast = true, but no bitrate is specified ? Are you doing 2 pass VBR then ? or 1pass CRF, or 2pass CRF ? I'll assume it means 1pass CRF 26

Your avidemux commands would look something like this (no audio in this example -an), assuming you're decoding r210 correctly with that binary

ffmpeg -i input_r210.avi -pix_fmt yuv420p -vf colormatrix=bt709:bt601 -c:v libx264 -profile:v high -level 3.1 -crf 26 -g 120 -bf 2 -b-pyramid 2 -x264-params ref=3:b-adapt=2:sar=1/1 -an x264.mkv

So you would still add -vf some denoise filter

[JDługosz]

Which output files ? From resolve ? Be specific - did you mean the r210 AVI's ? or converted ffv1 AVI's ?

The result of the commands you posted.

General
Complete name                            : E:\scratch\Newyear\tryit.avi
Format                                   : AVI
Format/Info                              : Audio Video Interleave
Format profile                           : OpenDML
File size                                : 9.61 GiB
Duration                                 : 7mn 4s
Overall bit rate                         : 195 Mbps
Writing application                      : Lavf55.19.104

Video
ID                                       : 0
Format                                   : FFV1
Codec ID                                 : FFV1
Duration                                 : 7mn 4s
Bit rate                                 : 195 Mbps
Width                                    : 1 920 pixels
Height                                   : 1 080 pixels
Display aspect ratio                     : 16:9
Frame rate                               : 23.976 fps
Compression mode                         : Lossless
Bits/(Pixel*Frame)                       : 3.915
Stream size                              : 9.61 GiB (100%)

Interestingly, the "vf" version (the ammended command you posted) gives a larger file.

1) "params" : "AQ=26" - is that AQ strength 2.6 ??!!! It can't be - because later on it says "aq_strength" : 1, Or is that CRF ? AQ means "average quality" ?

This is the options page I worked from.
“ H.264 uses a quantizer scale between 0 and 52. The default quantizer value is 26.”

2) are you sure you want threads 99? The more threads , the lower the quality. The optimum is logical cores * 1.5

That must be a special flag meaning "auto-detect". I have six cores, so 9 would be a good setting.

3) it says first pass fast = true, but no bitrate is specified ? Are you doing 2 pass VBR then ? or 1pass CRF, or 2pass CRF ? I'll assume it means 1pass CRF 26

The UI says "Constant Rate Factor (Single Pass)". I suppose "fast first pass" is ignored if the mode is single pass.

[JDługosz]

ffmpeg -i input_r210.avi -pix_fmt yuv420p -vf colormatrix=bt709:bt601 -c:v libx264 -profile:v high -level 3.1 -crf 26 -g 120 -bf 2 -b-pyramid 2 -x264-params ref=3:b-adapt=2:sar=1/1 -an x264.mkv

So you would still add -vf some denoise filter[/QUOTE]

I gave it a quick spin (I've not upgraded the version yet) and got these warnings:

[libx264 @ 00000000028db4a0] using SAR=1/1
[libx264 @ 00000000028db4a0] frame MB size (120x68) > level limit (3600)
[libx264 @ 00000000028db4a0] DPB size (4 frames, 32640 mbs) > level limit (2 frames, 18000 mbs)
[libx264 @ 00000000028db4a0] MB rate (195644) > level limit (108000)

[poisondeathray]

For the FFV1 export, it can be normal for jerky playback depending on your computer specs - because it's so heavily compressed even more than lagarith . But that earlier screenshot is a decoding error. What version of avidemux are you using ?

Open it in vdub and go frame by frame to see if the frames are all there (you might need ffdshow installed with ffv1 enabled in the vfw configuration)

The ffmpeg -vf colormatrix isn't as high quality conversion as doing it through avisynth directly with ConvertToYV12(matrix="rec709") . Or if you don't care about the colors being precise, you can leave it out

[poisondeathray]

ffmpeg -i input_r210.avi -pix_fmt yuv420p -vf colormatrix=bt709:bt601 -c:v libx264 -profile:v high -level 3.1 -crf 26 -g 120 -bf 2 -b-pyramid 2 -x264-params ref=3:b-adapt=2:sar=1/1 -an x264.mkv

I gave it a quick spin (I've not upgraded the version yet) and got these warnings:

[libx264 @ 00000000028db4a0] using SAR=1/1
[libx264 @ 00000000028db4a0] frame MB size (120x68) > level limit (3600)
[libx264 @ 00000000028db4a0] DPB size (4 frames, 32640 mbs) > level limit (2 frames, 18000 mbs)
[libx264 @ 00000000028db4a0] MB rate (195644) > level limit (108000)

Leave out the -level 3.1 setting in ffmpeg ; it's probably set to "auto" in avidemux, but your sheet said L3.1

[JDługosz]

For the FFV1 export, it can be normal for jerky playback depending on your computer specs - because it's so heavily compressed even more than lagarith . But that earlier screenshot is a decoding error. What version of avidemux are you using ?

It's uncompressed, not heavily compressed. I don't follow.
It's the latest version of avidemux, 2.6.7.

re colormatrix:
The whole reason I have this format mismatch is because I used Davinci Resolve Light to fix the color. I'll try rendering it again in a YUV format if ffmpeg can take it, and that will save the RGB round-trip.

The final mkv from the command you posted does play.
I had previously rendered Resolve to x264 within the program, and I compared the files on the same screen using two instances of VLC. Yours is a little more orange. So maybe the colormatrix was as it wanted it, originally.

[JDługosz]

For the FFV1 export, it can be normal for jerky playback depending on your computer specs

Mercury is a AMD 1100T (six cores) overclocked to 4000Mhz and water cooled. 16GB RAM, OS on SSD and these files on a fast data drive. Readubg 198Mbits/sec should be no problem if the drive isn't doing anything else. I recall it gets something like 150 MBytes/sec.

[poisondeathray]

It's uncompressed, not heavily compressed. I don't follow.

FFV1 is lossless compression, not uncompressed (it's like flac vs. uncompressed pcm wav for audio)

If you wanted uncompressed it would be about 2x the size. You would use -c:v rawvideo instead of -c:v ffv1

re colormatrix:
The whole reason I have this format mismatch is because I used Davinci Resolve Light to fix the color. I'll try rendering it again in a YUV format if ffmpeg can take it, and that will save the RGB round-trip.

There will be a RGB trip if your using things like RGB curves

The final mkv from the command you posted does play.
I had previously rendered Resolve to x264 within the program, and I compared the files on the same screen using two instances of VLC. Yours is a little more orange. So maybe the colormatrix was as it wanted it, originally.

It depends what matrix was used before resolve and what manipulations were done in your workflow. eg. if you used 601 for converting to RGB, then 601 is what you would use to convert back to YUV

[JDługosz]

FFV1 is lossless compression, not uncompressed (it's like flac vs. uncompressed pcm wav for audio)

OK, I see. The rendering from Resolve was uncompressed (though subsampled chroma in some of the trys) but the conversion to "something avidemux can read" was ffv1 which is not just an arrangement of bits and planes that is supported but an actual codec. Maybe I'd have had better luck with lossless H.264 for the same purpose.

Both of the trials to encode using x264 via ffmpeg yesterday gave color shifts. Leaving out the matrix option gave less of a change, but still a bit orange.

Today I re-rendered the file using the AVI/YUV-10bit choice. That is reported as Codec ID v210, YUV 4:2:2 10-bit. (The other choice is 8 bit, otherwise the same.)

When I fed that to ffmpeg, the resulting file matches the color of the H.264 file "rendered" directly from Resolve. (Checking those using the same player, against each other means I leave out any color profile settings in the player and Windows, that may make the result look different than the display in Resolve.)

I don't know if 8b would have been sufficient, but why not? Especially if the denoise filter can work with that, more color resolution when manipulating things around is better.

I'll try adding the noise filter into the mix next. Here is what I did:

[C:\Program Files\@AV\ffmpeg\bin]ffmpeg -i E:\scratch\Newyear\00000000.avi -pix_fmt yuv420p -c:v libx264 -profile:v high -crf 26 -g 120 -bf 2 -b-pyramid 2 -x264
-params ref=3:b-adapt=2:sar=1/1 -an E:\scratch\Newyear\yuv10-encode.mkv
ffmpeg version N-60572-gccc48b3 Copyright (c) 2000-2014 the FFmpeg developers
  built on Feb 12 2014 22:09:24 with gcc 4.8.2 (GCC)
  configuration: --disable-static --enable-shared --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --ena
ble-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-lib
modplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinge
r --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis
 --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
  libavutil      52. 63.101 / 52. 63.101
  libavcodec     55. 52.101 / 55. 52.101
  libavformat    55. 32.100 / 55. 32.100
  libavdevice    55.  9.100 / 55.  9.100
  libavfilter     4.  1.102 /  4.  1.102
  libswscale      2.  5.101 /  2.  5.101
  libswresample   0. 17.104 /  0. 17.104
  libpostproc    52.  3.100 / 52.  3.100
Input #0, avi, from 'E:\scratch\Newyear\00000000.avi':
  Metadata:
    TCOD            : 0
  Duration: 00:07:04.26, start: 0.000000, bitrate: 1061492 kb/s
    Stream #0:0: Video: v210 (v210 / 0x30313276), yuv422p10le, 1920x1080, 23.98 tbr, 23.98 tbn, 23.98 tbc
File 'E:\scratch\Newyear\yuv10-encode.mkv' already exists. Overwrite ? [y/N] y
[libx264 @ 00000000002de920] using SAR=1/1
[libx264 @ 00000000002de920] using cpu capabilities: MMX2 SSE2Fast LZCNT
[libx264 @ 00000000002de920] profile High, level 4.0
[libx264 @ 00000000002de920] 264 - core 142 r2389 956c8d8 - H.264/MPEG-4 AVC codec - Copyleft 2003-2014 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=9 lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=2 b_pyramid=2 b_adapt=2 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=120 keyint_min=12 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=26.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, matroska, to 'E:\scratch\Newyear\yuv10-encode.mkv':
  Metadata:
    TCOD            : 0
    encoder         : Lavf55.32.100
    Stream #0:0: Video: h264 (libx264) (H264 / 0x34363248), yuv420p, 1920x1080, q=-1--1, 1k tbn, 23.98 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (v210 -> libx264)
Press [q] to stop, [?] for help
frame=10172 fps= 20 q=-1.0 Lsize=  191009kB time=00:07:04.17 bitrate=3688.9kbits/s
video:190932kB audio:0kB subtitle:0 data:0 global headers:0kB muxing overhead 0.040214%
[libx264 @ 00000000002de920] frame I:86    Avg QP:25.00  size:130775
[libx264 @ 00000000002de920] frame P:3516  Avg QP:27.63  size: 35624
[libx264 @ 00000000002de920] frame B:6570  Avg QP:29.64  size:  8982
[libx264 @ 00000000002de920] consecutive B-frames:  2.0%  4.6% 93.4%
[libx264 @ 00000000002de920] mb I  I16..4: 14.5% 75.7%  9.8%
[libx264 @ 00000000002de920] mb P  I16..4:  1.1%  3.0%  0.2%  P16..4: 64.2%  9.8%  7.0%  0.0%  0.0%    skip:14.6%
[libx264 @ 00000000002de920] mb B  I16..4:  0.1%  0.2%  0.0%  B16..8: 33.9%  1.4%  0.2%  direct: 3.5%  skip:60.7%  L0:43.5% L1:53.2% BI: 3.3%
[libx264 @ 00000000002de920] 8x8 transform intra:72.1% inter:86.5%
[libx264 @ 00000000002de920] coded y,uvDC,uvAC intra: 58.6% 89.7% 60.4% inter: 12.5% 36.4% 2.0%
[libx264 @ 00000000002de920] i16 v,h,dc,p: 23% 22%  9% 46%
[libx264 @ 00000000002de920] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 21% 20% 14%  6%  6%  7%  8%  9% 10%
[libx264 @ 00000000002de920] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 24% 25% 10%  6%  7%  7%  9%  6%  6%
[libx264 @ 00000000002de920] i8c dc,h,v,p: 49% 22% 17% 11%
[libx264 @ 00000000002de920] Weighted P-Frames: Y:1.5% UV:0.5%
[libx264 @ 00000000002de920] ref P L0: 47.2% 13.4% 27.7% 11.7%  0.1%
[libx264 @ 00000000002de920] ref B L0: 63.3% 26.4% 10.3%
[libx264 @ 00000000002de920] ref B L1: 88.1% 11.9%
[libx264 @ 00000000002de920] kb/s:3686.70

Oh, is there an easy way to add the sound in at the same time? I have a 00000000.wav file of the same length (exported from Audacity) that still needs the AAC compression, and then mux into the same mkv.

—John