[mythtv-users] Viktor's Bob deint patch for interlaced modelines

[Seth Daniel]

On Sat, Mar 22, 2008 at 04:02:29PM -0400, Michael T. Dean wrote:
[...]
> >   Why?  Well, I discovered
> > that new profiles were created: High Quality, Low Quality, ...  The old
> > profiles still existed also: cpu++, cpu--, etc...  I kept updating cpu++
> > assuming it was the profile being used...I was wrong.  I started seeing
> > a performance degradation in my playback and it turns out this was
> > because the 'High Quality' profile was being used.
> 
> If your profile group specifies no valid profiles, Myth will fall back 
> to using whatever will work (which may, in fact, be similar to the 
> setting in High Quality).


How exactly does a profile groups become 'invalid'?  At one point I had
5 or 6 profile groups (none of which I had created).  Some of them had
similar, or even the same, rules.  How does Myth determine which profile
group to use?  I *thought* that having 'cpu++' selected in the 'Current
Video Playback Profile' drop-down box would do this.  But it didn't.  My
cpu++ profile group had one profile: if rez > 0 0 -> ffpmeg & xvideo.
Using the Bob(2x) deinterlacer.  However Myth decided to use the 'High
Quality' profile group (or, as you suggest, some sort of fallback).  How
did it determine to do this?  I can't believe it would have determined
that my cpu++ profile was invalid.  It wasn't until these new profile
groups were introduced that cpu++ ceased to be used.

-- 
seth /\ sethdaniel.org