[Theora-dev] FPGA implementation

[Andrey Filippov]

> Since that I've noticed my FPGA bug - DC predictor supplied two's
complement
> data and the next module expected sign+magnitude.
>

Somewhat better and the source pattern cold be guessed better (that was
diagonal ramp that was jumping back to zero after 255). DC prediction
seems wrong (not surprizingly - there are so many branches) and possibly
color components are shifted half-scale (it is sapposed to be grey scale
image), but ordering of blocks seems working as well as DCT (diagonal line
is precise to a pixel, not just to block).

http://www.elphel.com/downloads/eleventh.ogg
http://www.elphel.com/downloads/eleventh.png