[FFmpeg-user] deprecated pixel format used, make sure you did set range correctly

[Ferdi Scholten]

On 11-06-2023 12:52, Cecil Westerhof via ffmpeg-user wrote:
> Michael Koch <astroelectronic at t-online.de> writes:
>
>> Am 11.06.2023 um 11:33 schrieb Cecil Westerhof via ffmpeg-user:
>>> Ferdi Scholten <ferdi at sttc-nlp.nl> writes:
>>>
>>>> I make a slideshow of my pictures with:
>>>>>        ffmpeg -y -i %04d.jpg -i
>>>>> /home/cecil/Audio/YouTube/Symphony_No_5_by_Beethoven.mp3 -pix_fmt
>>>>> yuv420p -vcodec libx264 -crf 26 -acodec libmp3lame -qscale:a 9 -preset
>>>>> veryfast -vf
>>>>> zoompan=d=3:fps=.333:s=1280x960,framerate=25:interp_start=0:interp_end=255:scene=100
>>>>> huizeDoorn.mp4
>>>>>
>>>>> It looks OK, but I get a lot of the following messages:
>>>>>        deprecated pixel format used, make sure you did set range correctly
>>>>>
>>>>> Am I doing something wrong? Should I be worried?
>>>> No you are not doing anything wrong, the warning is displayed because
>>>> the pixelformat of your input source (the jpeg file) uses a deprecated
>>>> pixelformat with a different range for luma and chroma as your output
>>>> pixelformat, this can lead to differences in color and/or brightness
>>>>
>>>> yuvj422p is the deprecated pixelformat used by your jpg files (range
>>>> 0-255 for luma and chroma) 16 bpp
>>>> yuv420p is the pixelformat of your output. (range 16-235 for luma and
>>>> 16-240 for chroma) 12 bpp
>>>>
>>>> So three conversions need to be made here:
>>>> yuv conversion from 4:2:2 to 4:2:0
>>>> range conversion for y
>>>> range conversion for uv
>>>>
>>>> The warning might disappear (not tested this so not sure about it) if
>>>> you specify the pixelformat of your jpg file in your command as it is
>>>> now being autodetected by ffmpeg and the autodetection might be
>>>> wrong
>>> How can I determine the pixel format of my jpeg files?
>> ffprobe yourfile.jpg
>> or
>> ffmpeg -i yourfile.jpg
> I already tried that and got this:
> ffmpeg version 5.1.git-82278e8 Copyright (c) 2000-2023 the FFmpeg developers
>    built with gcc 10 (Debian 10.2.1-6)
>    configuration: --enable-gpl --enable-version3 --enable-libx264 --enable-libx265 --enable-libmp3lame --enable-libfontconfig --enable-libfreetype --enable-libfribidi
>    libavutil      58.  9.100 / 58.  9.100
>    libavcodec     60. 14.101 / 60. 14.101
>    libavformat    60.  5.100 / 60.  5.100
>    libavdevice    60.  2.100 / 60.  2.100
>    libavfilter     9.  8.101 /  9.  8.101
>    libswscale      7.  2.100 /  7.  2.100
>    libswresample   4. 11.100 /  4. 11.100
>    libpostproc    57.  2.100 / 57.  2.100
> Input #0, image2, from '0001.jpg':
>    Duration: 00:00:00.04, start: 0.000000, bitrate: 53673 kb/s
>    Stream #0:0: Video: mjpeg (Baseline), yuvj422p(pc, bt470bg/unknown/unknown), 1280x960 [SAR 180:180 DAR 4:3], 25 fps, 25 tbr, 25 tbn
>
> But this is I think auto detected and could be wrong. (If I read the
> reply of Ferdi correctly.)
>
Sometimes camera manufacturers diverge from standards, so check with the 
technical documantation of your camera (if available) to make sure! For 
example the layout of the jpeg might adhere to the standard, but the 
manufacturer could have applied different ranges. This can result in 
incorrect range conversions.
But if your output looks okay, the input is likely detected correctly 
and the manufacturer did adhere to the (deprecated) standard.