Hacker News Reader

Comment by r4vik | Hacker News Reader

I don't know if the output is correct but I prompted openai:

  ffmpeg command to convert hdr video to SDR:

  ffmpeg -i video.mp4

and it returned: ffmpeg -i video.mp4 -vf format=pix_fmts=yuv420p -c:v libx264 -crf 18 -preset veryslow -c:a copy output.mp4

  This FFmpeg command will take your HDR video and convert it to an SDR video using the libx264 codec with a CRF value of 18.```

View on HN

This doesn't really fix the color space though so I tried enhanced the prompt

  ffmpeg command to convert hdr video to SDR and fix the colorspace:

and it returned:

  ffmpeg -i input.hdr -pix_fmt yuv420p -filter_complex "colorlevels=entropy_mode=1:levels=srgb:source_range=full:target_range=full:rgb_to_yuv=sigmoidal" -y output.sdr.mp4

I don't really understand this so I asked it to explain:

  how does this fix the colorspace?

It replied:

  The RGB to YUV conversion is done using a "sigmoidal" function, which basically means that it applies a non-linear transformation to the input video. This transformation is designed to "compress" the bright colors in the video, making them more manageable for SDR displays.
  The colorlevels filter is also adjusting the black and white points of the video. This is done to ensure that the video can be displayed correctly on SDR displays, which have a limited color gamut.
  Finally, the target_range and source_range options are used to ensure that the video is encoded in the correct range for SDR displays.

Reply on news.ycombinator.com