[theora-dev] yuv strides

Ralph Giles giles at xiph.org
Wed Dec 27 10:50:33 PST 2006

On Wed, Dec 27, 2006 at 05:33:09PM +0000, Philip Heron wrote:

> "the strides may be negative in some configurations"
> I'm curious when that would be the case or what purpose negative strides
> would have?

The libtheora reference implementation uses this to invert the image. 

Theora's native scanline order is bottom-to-top, but the libtheora api 
uses top-to-bottom. If you don't have postprocessing turned on, the
decoder will return a yuv_buffer with a negative stride to avoid doing
an extra copy to invert the row order.

We mention it in the documation to warn people they need to handle that 


More information about the theora-dev mailing list