Gyan Doshi
Forum Replies Created
-
I’m fairly sure this isn’t an issue on the decoder side.
FFmpeg will select the ‘best’ pixel format supported by the encoder, which loses the least amount of information relative to the source. Since your input is YUV 4:2:2 and libx264 (default MOV encoder) supports that, that’s what FFmpeg will choose. But most players (and editors) only support 4:2:0 H264, so instead of their decoder registering an error, the s/w will decode only the audio, and display a blank canvas.
The solution is to encode as YUV 4:2:0. Add `-pix_fmt yuv420p` anywhere before the output filename and after the input filename.
-
Keep the shortest=1 in the first overlay filter i.e. [base] [upperleft] overlay=shortest=1 [tmp1]
-
Yeah. The pixel format is identified by the string yuv422p10le which is YUV 4:2:2 (planar, 10 bits per channel stored in little-endian order).
-
The default encoder reports only supporting YUV 4:2:2 10-bit output, so it should be that. Run
ffprobe final.movorffmpeg -i final.movto check. -
Avisynth with a field shift filter should work. See https://compression.ru/video/old_film_recover/field_shift_en.html
-
“-t” will encode for a given amount of seconds starting at the very beginning of the file even if you are using the “-ss” flag
No. ‘t’ represents the duration to encode. So ‘-ss 15 -t 10 -i file’ will work on 10 seconds of the input, starting from the 15th.