swscale YUV 10bpc (10-bit per component) -> 8bpc, user controlled dithering

[surfdog2020]

Summary of the bug:
providing arguments into 'sws_dither' never affects the output. It appears that no matter which dithering algorithm you specify in sws_dither (of the 4 that are available), ffmpeg ignores it and uses the same dithering algorithm for colorspace conversion no matter what. MD5s of the output .yuv files confirms this.

How to reproduce:

commands only :
./ffmpeg -v debug -y -i /mnt/storage/ProRes_10bit_422.mov -t 30 -pix_fmt yuv420p -an -threads 36 -vf scale=1920:-1:sws_dither=a_dither mnt/storage/raw_8bit_420_adither.yuv
ffmpeg version 2.6.3-   http://johnvansickle.com/ffmpeg/    Copyright (c) 2000-2015 the FFmpeg developers

./ffmpeg -v debug -y -i /mnt/storage/ProRes_10bit_422.mov -t 30 -pix_fmt yuv420p -an -threads 36 -vf scale=1920:-1:sws_dither=bayer mnt/storage/raw_8bit_420_bayer.yuv
ffmpeg version 2.6.3-   http://johnvansickle.com/ffmpeg/    Copyright (c) 2000-2015 the FFmpeg developers
  built with gcc 4.9.2 (Debian 4.9.2-16)

'''[Jared@transcoder:~/Downloads/ffmpeg-2.6.3-64bit-static]$ md5sum /mnt/storage/raw_8bit_420_bayer.yuv
203cb43ecc97aa16d7203626750a14ad  /mnt/storage/raw_8bit_420_bayer.yuv'''
[Jared@transcoder:~/Downloads/ffmpeg-2.6.3-64bit-static]$ md5sum /mnt/storage/raw_8bit_420_adither.yuv
'''203cb43ecc97aa16d7203626750a14ad  /mnt/storage/raw_8bit_420_adither.yuv'''


commands with full output:
[Jared@transcoder:~/Downloads/ffmpeg-2.6.3-64bit-static]$sudo ./ffmpeg -v debug -y -i /mnt/storage/ProRes_10bit_422.mov -t 30 -pix_fmt yuv420p -an -threads 36 -vf scale=1920:-1:sws_dither=a_dither mnt/storage/raw_8bit_420_adither.yuv
ffmpeg version 2.6.3-   http://johnvansickle.com/ffmpeg/    Copyright (c) 2000-2015 the FFmpeg developers
  built with gcc 4.9.2 (Debian 4.9.2-16)
  configuration: --enable-gpl --enable-version3 --disable-shared --disable-debug --enable-runtime-cpudetect --enable-libmp3lame --enable-libx264 --enable-libx265 --enable-libwebp --enable-libspeex --enable-libvorbis --enable-libvpx --enable-libfreetype --enable-fontconfig --enable-libxvid --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-gray --enable-libopenjpeg --enable-libopus --enable-libass --enable-gnutls --enable-libvidstab --enable-libsoxr --cc=gcc-4.9
  libavutil      54. 20.100 / 54. 20.100
  libavcodec     56. 26.100 / 56. 26.100
  libavformat    56. 25.101 / 56. 25.101
  libavdevice    56.  4.100 / 56.  4.100
  libavfilter     5. 11.102 /  5. 11.102
  libswscale      3.  1.101 /  3.  1.101
  libswresample   1.  1.100 /  1.  1.100
  libpostproc    53.  3.100 / 53.  3.100
Splitting the commandline.
Reading option '-v' ... matched as option 'v' (set logging level) with argument 'debug'.
Reading option '-y' ... matched as option 'y' (overwrite output files) with argument '1'.
Reading option '-i' ... matched as input file with argument '/mnt/storage/ProRes_10bit_422.mov'.
Reading option '-t' ... matched as option 't' (record or transcode "duration" seconds of audio/video) with argument '30'.
Reading option '-pix_fmt' ... matched as option 'pix_fmt' (set pixel format) with argument 'yuv420p'.
Reading option '-an' ... matched as option 'an' (disable audio) with argument '1'.
Reading option '-threads' ... matched as AVOption 'threads' with argument '36'.
Reading option '-vf' ... matched as option 'vf' (set video filters) with argument 'scale=1920:-1:sws_dither=a_dither'.
Reading option 'mnt/storage/raw_8bit_420.yuv_adither.yuv' ... matched as output file.
Finished splitting the commandline.
Parsing a group of options: global .
Applying option v (set logging level) with argument debug.
Applying option y (overwrite output files) with argument 1.
Successfully parsed a group of options.
Parsing a group of options: input file /mnt/storage/ProRes_10bit_422.mov.
Successfully parsed a group of options.
Opening an input file: /mnt/storage/ProRes_10bit_422.mov.
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x48e6680] Format mov,mp4,m4a,3gp,3g2,mj2 probed with size=2048 and score=100
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x48e6680] Before avformat_find_stream_info() pos: 14083960832 bytes read:253481 seeks:5
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x48e6680] All info found
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x48e6680] After avformat_find_stream_info() pos: 98224 bytes read:343513 seeks:6 frames:1
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/mnt/storage/ProRes_10bit_422.mov':
  Metadata:
    creation_time   : 2015-03-26 19:59:36
    timecode        : 00:57:50:00
  Duration: 00:13:49.08, start: 0.000000, bitrate: 135899 kb/s
    Stream #0:0(eng), 1, 1/24000: Video: prores (apch / 0x68637061), yuv422p10le(bt709), 1920x1080, 1/24000, 135654 kb/s, SAR 1:1 DAR 16:9, 23.98 fps, 23.98 tbr, 24k tbn, 24k tbc (default)
    Metadata:
      creation_time   : 2015-03-26 19:59:36
      handler_name    : Apple Alias Data Handler
      encoder         : Apple ProRes 422 (HQ)
    Stream #0:1(eng), 0, 1/24000: Data: none (tmcd / 0x64636D74), 1/24, 0 kb/s (default)
    Metadata:
      creation_time   : 2015-03-26 19:59:36
      handler_name    : Apple Alias Data Handler
      reel_name       : Unnamed
      timecode        : 00:57:50:00
Successfully opened the file.
Parsing a group of options: output file mnt/storage/raw_8bit_420.yuv_adither.yuv.
Applying option t (record or transcode "duration" seconds of audio/video) with argument 30.
Applying option pix_fmt (set pixel format) with argument yuv420p.
Applying option an (disable audio) with argument 1.
Applying option vf (set video filters) with argument scale=1920:-1:sws_dither=a_dither.
Successfully parsed a group of options.
Opening an output file: mnt/storage/raw_8bit_420.yuv_adither.yuv.
Successfully opened the file.
[Parsed_scale_0 @ 0x48ee3e0] Setting 'w' to value '1920'
[Parsed_scale_0 @ 0x48ee3e0] Setting 'h' to value '-1'
[Parsed_scale_0 @ 0x48ee3e0] Setting 'sws_dither' to value 'a_dither'
[Parsed_scale_0 @ 0x48ee3e0] Setting 'flags' to value '0x4'
[Parsed_scale_0 @ 0x48ee3e0] w:1920 h:-1 flags:'0x4' interl:0
[graph 0 input from stream 0:0 @ 0x48eeae0] Setting 'video_size' to value '1920x1080'
[graph 0 input from stream 0:0 @ 0x48eeae0] Setting 'pix_fmt' to value '74'
[graph 0 input from stream 0:0 @ 0x48eeae0] Setting 'time_base' to value '1/24000'
[graph 0 input from stream 0:0 @ 0x48eeae0] Setting 'pixel_aspect' to value '1/1'
[graph 0 input from stream 0:0 @ 0x48eeae0] Setting 'sws_param' to value 'flags=2'
[graph 0 input from stream 0:0 @ 0x48eeae0] Setting 'frame_rate' to value '24000/1001'
[graph 0 input from stream 0:0 @ 0x48eeae0] w:1920 h:1080 pixfmt:yuv422p10le tb:1/24000 fr:24000/1001 sar:1/1 sws_param:flags=2
[format @ 0x48dbb40] compat: called with args=[yuv420p]
[format @ 0x48dbb40] Setting 'pix_fmts' to value 'yuv420p'
[AVFilterGraph @ 0x48dbe20] query_formats: 5 queried, 4 merged, 0 already done, 0 delayed
[Parsed_scale_0 @ 0x48ee3e0] w:1920 h:1080 fmt:yuv422p10le sar:1/1 -> w:1920 h:1080 fmt:yuv420p sar:1/1 flags:0x4
detected 32 logical cores
Output #0, rawvideo, to 'mnt/storage/raw_8bit_420.yuv_adither.yuv':
  Metadata:
    timecode        : 00:57:50:00
    encoder         : Lavf56.25.101
    Stream #0:0(eng), 0, 1001/24000: Video: rawvideo (I420 / 0x30323449), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 1001/24000, q=2-31, 200 kb/s, 23.98 fps, 23.98 tbn, 23.98 tbc (default)
    Metadata:
      creation_time   : 2015-03-26 19:59:36
      handler_name    : Apple Alias Data Handler
      encoder         : Lavc56.26.100 rawvideo
Stream mapping:
  Stream #0:0 -> #0:0 (prores (native) -> rawvideo (native))
Press [q] to stop, [?] for help
Cliping frame in rate conversion by 0.000008
[output stream 0:0 @ 0x48d1fe0] EOF on sink link output stream 0:0:default.kbits/s
No more output streams to write to, finishing.
frame=  720 fps= 63 q=0.0 Lsize= 2187000kB time=00:00:30.03 bitrate=596600.2kbits/s
video:2187000kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.000000%
Input file #0 (/mnt/storage/ProRes_10bit_422.mov):
  Input stream #0:0 (video): 721 packets read (64913072 bytes); 721 frames decoded;
  Input stream #0:1 (data): 0 packets read (0 bytes);
  Total: 721 packets (64913072 bytes) demuxed
Output file #0 (mnt/storage/raw_8bit_420.yuv_adither.yuv):
  Output stream #0:0 (video): 720 frames encoded; 720 packets muxed (2239488000 bytes);
  Total: 720 packets (2239488000 bytes) muxed
721 frames successfully decoded, 0 decoding errors
[AVIOContext @ 0x48ebca0] Statistics: 0 seeks, 68400 writeouts
[AVIOContext @ 0x48e5e40] Statistics: 65166553 bytes read, 71 seeks





[Jared@transcoder:~/Downloads/ffmpeg-2.6.3-64bit-static]$ sudo ./ffmpeg -v debug -y -i /mnt/storage/ProRes_10bit_422.mov -t 30 -pix_fmt yuv420p -an -threads 36 -vf scale=1920:-1:sws_dither=bayer mnt/storage/raw_8bit_420_bayer.yuv
ffmpeg version 2.6.3-   http://johnvansickle.com/ffmpeg/    Copyright (c) 2000-2015 the FFmpeg developers
  built with gcc 4.9.2 (Debian 4.9.2-16)
  configuration: --enable-gpl --enable-version3 --disable-shared --disable-debug --enable-runtime-cpudetect --enable-libmp3lame --enable-libx264 --enable-libx265 --enable-libwebp --enable-libspeex --enable-libvorbis --enable-libvpx --enable-libfreetype --enable-fontconfig --enable-libxvid --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-gray --enable-libopenjpeg --enable-libopus --enable-libass --enable-gnutls --enable-libvidstab --enable-libsoxr --cc=gcc-4.9
  libavutil      54. 20.100 / 54. 20.100
  libavcodec     56. 26.100 / 56. 26.100
  libavformat    56. 25.101 / 56. 25.101
  libavdevice    56.  4.100 / 56.  4.100
  libavfilter     5. 11.102 /  5. 11.102
  libswscale      3.  1.101 /  3.  1.101
  libswresample   1.  1.100 /  1.  1.100
  libpostproc    53.  3.100 / 53.  3.100
Splitting the commandline.
Reading option '-v' ... matched as option 'v' (set logging level) with argument 'debug'.
Reading option '-y' ... matched as option 'y' (overwrite output files) with argument '1'.
Reading option '-i' ... matched as input file with argument '/mnt/storage/ProRes_10bit_422.mov'.
Reading option '-t' ... matched as option 't' (record or transcode "duration" seconds of audio/video) with argument '30'.
Reading option '-pix_fmt' ... matched as option 'pix_fmt' (set pixel format) with argument 'yuv420p'.
Reading option '-an' ... matched as option 'an' (disable audio) with argument '1'.
Reading option '-threads' ... matched as AVOption 'threads' with argument '36'.
Reading option '-vf' ... matched as option 'vf' (set video filters) with argument 'scale=1920:-1:sws_dither=bayer'.
Reading option 'mnt/storage/raw_8bit_420.yuv_bayer.yuv' ... matched as output file.
Finished splitting the commandline.
Parsing a group of options: global .
Applying option v (set logging level) with argument debug.
Applying option y (overwrite output files) with argument 1.
Successfully parsed a group of options.
Parsing a group of options: input file /mnt/storage/ProRes_10bit_422.mov.
Successfully parsed a group of options.
Opening an input file: /mnt/storage/ProRes_10bit_422.mov.
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x37a5680] Format mov,mp4,m4a,3gp,3g2,mj2 probed with size=2048 and score=100
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x37a5680] Before avformat_find_stream_info() pos: 14083960832 bytes read:253481 seeks:5
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x37a5680] All info found
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x37a5680] After avformat_find_stream_info() pos: 98224 bytes read:343513 seeks:6 frames:1
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/mnt/storage/ProRes_10bit_422.mov':
  Metadata:
    creation_time   : 2015-03-26 19:59:36
    timecode        : 00:57:50:00
  Duration: 00:13:49.08, start: 0.000000, bitrate: 135899 kb/s
    Stream #0:0(eng), 1, 1/24000: Video: prores (apch / 0x68637061), yuv422p10le(bt709), 1920x1080, 1/24000, 135654 kb/s, SAR 1:1 DAR 16:9, 23.98 fps, 23.98 tbr, 24k tbn, 24k tbc (default)
    Metadata:
      creation_time   : 2015-03-26 19:59:36
      handler_name    : Apple Alias Data Handler
      encoder         : Apple ProRes 422 (HQ)
    Stream #0:1(eng), 0, 1/24000: Data: none (tmcd / 0x64636D74), 1/24, 0 kb/s (default)
    Metadata:
      creation_time   : 2015-03-26 19:59:36
      handler_name    : Apple Alias Data Handler
      reel_name       : Unnamed
      timecode        : 00:57:50:00
Successfully opened the file.
Parsing a group of options: output file mnt/storage/raw_8bit_420.yuv_bayer.yuv.
Applying option t (record or transcode "duration" seconds of audio/video) with argument 30.
Applying option pix_fmt (set pixel format) with argument yuv420p.
Applying option an (disable audio) with argument 1.
Applying option vf (set video filters) with argument scale=1920:-1:sws_dither=bayer.
Successfully parsed a group of options.
Opening an output file: mnt/storage/raw_8bit_420.yuv_bayer.yuv.
Successfully opened the file.
[Parsed_scale_0 @ 0x37ad3e0] Setting 'w' to value '1920'
[Parsed_scale_0 @ 0x37ad3e0] Setting 'h' to value '-1'
[Parsed_scale_0 @ 0x37ad3e0] Setting 'sws_dither' to value 'bayer'
[Parsed_scale_0 @ 0x37ad3e0] Setting 'flags' to value '0x4'
[Parsed_scale_0 @ 0x37ad3e0] w:1920 h:-1 flags:'0x4' interl:0
[graph 0 input from stream 0:0 @ 0x37adae0] Setting 'video_size' to value '1920x1080'
[graph 0 input from stream 0:0 @ 0x37adae0] Setting 'pix_fmt' to value '74'
[graph 0 input from stream 0:0 @ 0x37adae0] Setting 'time_base' to value '1/24000'
[graph 0 input from stream 0:0 @ 0x37adae0] Setting 'pixel_aspect' to value '1/1'
[graph 0 input from stream 0:0 @ 0x37adae0] Setting 'sws_param' to value 'flags=2'
[graph 0 input from stream 0:0 @ 0x37adae0] Setting 'frame_rate' to value '24000/1001'
[graph 0 input from stream 0:0 @ 0x37adae0] w:1920 h:1080 pixfmt:yuv422p10le tb:1/24000 fr:24000/1001 sar:1/1 sws_param:flags=2
[format @ 0x379ab40] compat: called with args=[yuv420p]
[format @ 0x379ab40] Setting 'pix_fmts' to value 'yuv420p'
[AVFilterGraph @ 0x379ae20] query_formats: 5 queried, 4 merged, 0 already done, 0 delayed
[Parsed_scale_0 @ 0x37ad3e0] w:1920 h:1080 fmt:yuv422p10le sar:1/1 -> w:1920 h:1080 fmt:yuv420p sar:1/1 flags:0x4
detected 32 logical cores
Output #0, rawvideo, to 'mnt/storage/raw_8bit_420.yuv_bayer.yuv':
  Metadata:
    timecode        : 00:57:50:00
    encoder         : Lavf56.25.101
    Stream #0:0(eng), 0, 1001/24000: Video: rawvideo (I420 / 0x30323449), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 1001/24000, q=2-31, 200 kb/s, 23.98 fps, 23.98 tbn, 23.98 tbc (default)
    Metadata:
      creation_time   : 2015-03-26 19:59:36
      handler_name    : Apple Alias Data Handler
      encoder         : Lavc56.26.100 rawvideo
Stream mapping:
  Stream #0:0 -> #0:0 (prores (native) -> rawvideo (native))
Press [q] to stop, [?] for help
Cliping frame in rate conversion by 0.000008
[output stream 0:0 @ 0x3790fe0] EOF on sink link output stream 0:0:default.kbits/s
No more output streams to write to, finishing.
frame=  720 fps= 66 q=0.0 Lsize= 2187000kB time=00:00:30.03 bitrate=596600.2kbits/s
video:2187000kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.000000%
Input file #0 (/mnt/storage/ProRes_10bit_422.mov):
  Input stream #0:0 (video): 721 packets read (64913072 bytes); 721 frames decoded;
  Input stream #0:1 (data): 0 packets read (0 bytes);
  Total: 721 packets (64913072 bytes) demuxed
Output file #0 (mnt/storage/raw_8bit_420.yuv_bayer.yuv):
  Output stream #0:0 (video): 720 frames encoded; 720 packets muxed (2239488000 bytes);
  Total: 720 packets (2239488000 bytes) muxed
721 frames successfully decoded, 0 decoding errors
[AVIOContext @ 0x37aaca0] Statistics: 0 seeks, 68400 writeouts
[AVIOContext @ 0x37a4e40] Statistics: 65166553 bytes read, 71 seeks



[Jared@transcoder:~/Downloads/ffmpeg-2.6.3-64bit-static]$ md5sum /mnt/storage/raw_8bit_420_bayer.yuv
203cb43ecc97aa16d7203626750a14ad  /mnt/storage/raw_8bit_420_bayer.yuv
[Jared@transcoder:~/Downloads/ffmpeg-2.6.3-64bit-static]$ md5sum /mnt/storage/raw_8bit_420_adither.yuv
203cb43ecc97aa16d7203626750a14ad  /mnt/storage/raw_8bit_420_adither.yuv

[Carl Eugen Hoyos]

Please test current FFmpeg git head.

[surfdog2020]

Replying to cehoyos:

Please test current FFmpeg git head.

Same thing is still happening , even with the with latest version from git head and the latest version of all libraries.
Here is the output with the new versions numbers of the libraries :
ffmpeg version N-72843-g9c89392 Copyright (c) 2000-2015 the FFmpeg developers
built with gcc 4.6 (Ubuntu/Linaro? 4.6.3-1ubuntu5)
configuration: --prefix=/home/jared/ffmpeg_build --pkg-config-flags=--static --extra-cflags=-I/home/jared/ffmpeg_build/include --extra-ldflags=-L/home/jared/ffmpeg_build/lib --bindir=/home/jared/bin --enable-gpl --enable-libass --enable-libfdk-aac --enable-libfreetype --enable-libmp3lame --enable-libopus --enable-libtheora --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libx265 --enable-nonfree
libavutil 54. 27.100 / 54. 27.100
libavcodec 56. 41.100 / 56. 41.100
libavformat 56. 36.100 / 56. 36.100
libavdevice 56. 4.100 / 56. 4.100
libavfilter 5. 16.101 / 5. 16.101
libswscale 3. 1.101 / 3. 1.101
libswresample 1. 2.100 / 1. 2.100
libpostproc 53. 3.100 / 53. 3.100

[Hendrik]

Try to pass the sws flag to make it print info about the conversion parameters used, I don't remember the syntax off-hand .. but if it says that its using a "specialized" converter, then it won't execute the generic dithering pass. You can usually force the generic path by passing the accurate rounding and/or full internal chroma flags.

[surfdog2020]

I'm guessing that you wanted me to add '-sws_flags print_info' to my command line. If that is the case, I see no mention of a "specialized" converter being used. I only see the following get added to my command's output:

SwScaler: reducing / aligning filtersize 1 -> 4

Last message repeated 1 times

SwScaler: reducing / aligning filtersize 1 -> 1
SwScaler: reducing / aligning filtersize 9 -> 8
[swscaler @ 0x43a9060] bicubic scaler, from yuv422p10le to yuv420p using MMXEXT
[swscaler @ 0x43a9060] 1920x1080 -> 1920x1080
[swscaler @ 0x43a9060] lum srcW=1920 srcH=1080 dstW=1920 dstH=1080 xInc=65536 yInc=65536
[swscaler @ 0x43a9060] chr srcW=960 srcH=1080 dstW=960 dstH=540 xInc=65536 yInc=131072

Can you also clarify how one can pass in the "accurate rounding" of the internal chroma flags?

[Hendrik]

Try with "-sws_flags +accurate_rnd+full_chroma_inp+full_chroma_int", that should hopefully make it perform a high quality conversion with dithering. The last option is probably not needed for your use-case, IIRC its only needed when you upsample chroma, not downsample.

[surfdog2020]

I'm looking to be able to specify the dithering algorithm used on the 10-bit to 8-bit conversion. I tried:
./ffmpeg -v debug -sws_flags +accurate_rnd+full_chroma_inp+full_chroma_int -y -i /mnt/storage/ProRes_10bit_422.mov -sws_dither a_dither -t 30 -pix_fmt yuv420p -an -threads 36 -vf scale=1920:-1:sws_dither=a_dither /mnt/storage/raw_8bit_420_adither.yuv

./ffmpeg -v debug -sws_flags +accurate_rnd+full_chroma_inp+full_chroma_int -y -i /mnt/storage/ProRes_10bit_422.mov -sws_dither bayer -t 30 -pix_fmt yuv420p -an -threads 36 -vf scale=1920:-1:sws_dither=bayer /mnt/storage/raw_8bit_420_bayer.yuv

Both output files are still equal to each other, have the same md5 value.

[Michael Niedermayer]

I dont think this case is supported currently but it would be nice to add support for

[surfdog2020]

Replying to michael:

I dont think this case is supported currently but it would be nice to add support for

The documentation says that this should be already supported, see: https://ffmpeg.org/ffmpeg-all.html#Scaler-Options, this was my reasoning in marking it as a defect instead of enhancement.

Regardless of the label, I agree that this is functionality that would be beneficial to have, as some folks prefer one way of dithering over another.

[gdgsdg123]

Build from: ​https://zeranoe.com/builds/win64/static/win64/static/ffmpeg-20190416-036b4b0-win64-static.zip


It's still the case in recent version.



What if one preferred no dither? (for the sake of compressibility... or just because banding/discoloration is preferred over noise)







Suggested workaround (for no dither):
YUV 10bpc -> RGB24 -> YUV 8bpc


And don't forget ​this...



Or use zscale?..

[Balling]

That is no longer the case with -vf scale=sws_dither=0 or -scale=sws_dither=none, both produce same results different from auto dither.

Please test! Please note that there still appears some dither from 10 bit to 8 bit that produces a pair of colors instead of one.