[AG-TECH] Re: [ANNOUNCEMENT] H.263 codec beta test

[Lawrence A. Rowe]

Kun -

I have said again and again that the current decoder does not handle
motion vectors (MV). Lloyd had written some code to handle them, but it
didn't work and my quick look at it suggested that there was much to be
done.  Rather than start implementing every one of the MV modes and
special cases in the standard, I wanted to get some test streams so I
could see what features were actually being used.  I can grab uclvic
test streams, but I cannot record polycom or netmeeting streams.  (As an
aside: how do you record NetMeeting streams?  are they RTP-complaint?)
That's why I asked numerous people to give me some short streams that I
can use to analyze coding features used by the standard codecs.

Lastly, did you read the beta release webpage! I mentioned the problem
with MVs and uclvic and the general issue with MVs on that page. Pink
blocks are displayed when it receives a MV so that is what you're
seeing.
	Larry
-- 
Professor Lawrence A. Rowe          Internet: Rowe at BMRC.Berkeley.EDU
Computer Science Division - EECS       Phone: 510-642-5117
University of California, Berkeley       Fax: 510-642-5615
Berkeley, CA 94720-1776            URL: http://bmrc.berkeley.edu/~larry