[mythtv] Denoise3d pixel wrap bug.

[Bruce Markey]

Andrew Mahone wrote:
> I've done some more testing, and the problem disappears when I compile
> with -O1.  If I use -O2 or -O3, even without the extra options I would
> normally use, it breaks again.  If I use -O1 and all of the options
> listed as enabled by -O2...  it still works correctly.  I'm feeling more
> certain that this is compiler bugginess at work.

Andrew, if you're are really itching for something to delve
into, one thing I think would be really useful would be a filter
for the bttv's broken luma range and possibly luma correction
in general. The bttv luma correction is pretty much at the top
of my wishlist for myth.

: bjm at moktoo ; grep 253 /usr/include/linux/videodev2.h
        /* broken BT878 extents (601, luma range 16-253 instead of 16-235) */

Apparently there is a well known bug in the chip design. There
is a filter in tvtime (tvtime.sf.net) that remaps the luma values
and the result clearly fixes image problems with colors and shades
at or near white (see tvtime-0.9.12/src/videocorrection.c).

If you're not interested I might try my hand at this myself.
I imagine I could use invert as a template and build from
there. What would be the best way for me to learn enough
about the frame formats to be able to identify and modify
the luma values?

--  bjm