os33 - video regression ?

[James Cameron]

On Mon, Oct 26, 2009 at 05:37:37PM -0400, Mikus Grinbergs wrote:
> > YouTube video can be encapsulated H.264, On2 VP6, or the older
> > Sorenson Spark. So it's important to try to test the same file.
> 
> Where you are coming from appears to differ from where I am coming
> from.  I am looking at the XO from the point of view of "How usable
> is this?"  That question does not necessarily involve precise
> measurement -- evaluation ought to suffice.

I agree.  Especially as you observed a regression from one version to
another.

You went on to mention formal specifications.  Maintenance of a
sufficiently comprehensive test specification would be a huge effort.
One would also have to select representative videos and ensure they
remain available during the testing without uncontrolled revision by
third-parties.  Controlling for YouTube's transcoding ... which we
don't influence ... would be quite amazing.

I'm curious to know if it was the change in #9538 that caused the
regression.  Could you temporarily change the display depth back to what
it was and see if that restores the original behaviour?

General comment on the typical "have you logged this?" response we tend
to get ...  I've no trouble with hearing from people who have taken the
time to write about their experiences with an XO ... if they don't have
the time to log a ticket and engage in the formal process, that's their
decision.  If other readers here don't attempt to reproduce and create a
ticket, then they mustn't think it is important enough to fix.  ;-)

-- 
James Cameron
http://quozl.linux.org.au/