[MPlayer-dev-eng] New inverse-telecine filter

[D Richard Felker III]

On Thu, Dec 04, 2003 at 11:29:45AM -0600, Zoltan Hidvegi wrote:
> > On Tue, Dec 02, 2003 at 10:10:08AM -0600, Zoltan Hidvegi wrote:
> > > +.IPs luma_only=n
> > > +If n is nonzero, the chroma plane is copied unchanged.  This is useful
> > > +for YV12 sampled TV, which discards one of the chroma fields.
> > 
> > This description is somewhat incorrect. Well-formed interlaced YV12
> > never discards chroma fields. Only broken TV card drivers, which
> > misreport that they support YV12 when in fact they actually don't, do
> > this. A more appropriate description would be "bug workaround for
> > bttv's brain damage". :)
> 
> The problem is that no TV can can properly support NTSC -> YV12
> conversion, because such a conversion would have to know if a frame is
> progressive or interlaced.  For a progressive frame, a simple scaling
> of the chroma plane works, but for interlaced, you have to do
> interlaced scaling. 

TV is always interlaced. :)

> The bug is really in mplayer not supporting 411
> TV capture (or 422, but NTSC is really just 411).  I guess you can do
> uyvy or yuy2 (what is yuy2, I always confuse this stuff), but I did
> not have much luck with that, since it's packed.  bttv supports 411p
> and 422p, but mplayer cannot use that for some reason.

Feel free to send a patch. It's probably just a matter of adding a
case to tvi_v4l.c.

Rich