FranceBB
15th November 2021, 23:35
Alright guys, the question is the following:
If I have a 4:2:0 Type 2 source (so top left, 0,0 chroma location) and I use:
FFMpeg.exe -i "test.m2ts" -pix_fmt yuv422p
And then save the video, will FFMpeg:
A) Convert from top left to the old MPEG-2 chroma location, hence creating a perfectly fine yv16?
B) Create a non standard yv16 with the chroma sitting in top left, 0,0, which will be read wrongly by any decoder?
C) Upscale the chroma assuming the old MPEG-2 location instead of top left, hence screwing everything up?
I know that you can use:
-vf 'scale=out_color_matrix=bt709:out_h_chr_pos=0:out_v_chr_pos=128'
To go from type 2 to the old MPEG-2 one and also:
-vf scale=out_color_matrix=bt709:out_h_chr_pos=0:out_v_chr_pos=0
To do the opposite and go from the old MPEG-2 to the new Type 2, but the question still remains: what happens if someone just specifies the pix_fmt?
If I have a 4:2:0 Type 2 source (so top left, 0,0 chroma location) and I use:
FFMpeg.exe -i "test.m2ts" -pix_fmt yuv422p
And then save the video, will FFMpeg:
A) Convert from top left to the old MPEG-2 chroma location, hence creating a perfectly fine yv16?
B) Create a non standard yv16 with the chroma sitting in top left, 0,0, which will be read wrongly by any decoder?
C) Upscale the chroma assuming the old MPEG-2 location instead of top left, hence screwing everything up?
I know that you can use:
-vf 'scale=out_color_matrix=bt709:out_h_chr_pos=0:out_v_chr_pos=128'
To go from type 2 to the old MPEG-2 one and also:
-vf scale=out_color_matrix=bt709:out_h_chr_pos=0:out_v_chr_pos=0
To do the opposite and go from the old MPEG-2 to the new Type 2, but the question still remains: what happens if someone just specifies the pix_fmt?