[theora-dev] Re: YUV to RGB
phil at sanslogic.co.uk
Fri Feb 8 05:34:43 PST 2008
wesley kiriinya wrote:
> I managed to store RGBs in the YUVs then encode the frames (with RGB instead of YUV), but it seems like because of the compression the decoded frames don't look good. I guess the compression would have to be lossless.
> In case you are wondering how the RGB could fit in YUV, I had to adjust the size of the original YUV frames to be bigger in order to accommodate the extra RGB data.
You could try mapping the RGB channels so that they better match what
the encoder expects. Put G where the encoder expects Y, and scale the
R+B channels so they match the size of U+V. This would give you better
results than simply encoding the raw RGB data.
More information about the theora-dev