[Ffmpeg-devel] Input image format for encoding...

[Mat]

Michel Bardiaux wrote:

> A note of caution: YUV is rather vague. 

Yep.
Many standards.

> What you need is YUV420P which is: HEIGHT lines of WIDTH bytes of 
> luminance, followed by HEIGHT/2 lines of WIDTH/2 bytes of chrominance 
> (I can never remember if its Cr or Cb), followed by the same for the 
> other chrominance.

4:2:0 planar - 2 bytes per pixel

> If your grabber doesnt support that format, you have to use another 
> one then use img_convert.

Yep.
Like in  output_example.c

But my grabber seems to like YUV too :)