[MPlayer-dev-eng] New inverse-telecine filter
vektor at dumbterm.net
Thu Dec 4 23:54:57 CET 2003
Zoltan Hidvegi (mplayer at hzoli.2y.net):
> But it is. Look at that page I've referenced, it writes:
> The MPEG spec makes it clear that while there is a wide latitude for
> the MPEG decoder to choose an upsampling algorithm for 4:2:0 to 4:2:2
> conversion, the decoder must use two different algorithms to be
> correct. One should be used for frames marked "interlaced", and a
> different one should be used for frames marked "progressive."
> So if there are two algorithms for upsampling, there must be two
> algorithms for downsampling as well, the problem is there are no flags
> in broadcast TV to tell you which one to use.
Of course not, because you don't need to. Broadcast TV is sent as
4:2:2. The bttv driver converts this from 4:2:2 to 4:2:0, currently by
dropping every second line of chroma. This happens to be reasonable
(but poor quality - should be filtered!) if the field happens to be
from a progressive frame. But since it might not be, the bttv driver
should use a standard for interlaced chroma, basically, every second
scanline of chroma should be from the next field (again poor, but better
for interlaced content).
You seem to be of the opinion that if the driver were to use
interlaced chroma all the time, that there would be a quality loss on
content that is progressive. I do not see how this follows at all.
Given a series of interlaced 4:2:0 fields, I can convert each to 4:2:2
using a reasonable filter, and now I can do whatever I like with them.
> > Why do you care about 4:2:2 to 4:2:0 anyway though? Let's bring
> > this back to bttv and what it should do. If you ask bttv for 4:2:0,
> > it should _always_ mean that we want interlaced chroma, since this
> > is the only intelligent choice. Furthermore, using the definition
> No, not for progressive frames.
Yes for progressive frames, it does not matter.
> > interlaced chroma from MPEG2 is well defined, with how to go from
> > 4:2:2. This is the correct choice regardless and it works just fine
> > for progressive material: you deinterlace as normal, if you detect
> > 2-2
> No, it does not, interlaced scaling will mess up the chroma of
> progressive frames. And you cannot just deinterlace that, although
> you can probably do some tricks to get back some reasonable pictures
> (like interlaced upscale followed by progressive downscale).
My argument is that these 'tricks' are not tricks at all.
More information about the MPlayer-dev-eng