[MPlayer-users] new inverse telecine filter!

[D Richard Felker III]

On Tue, Apr 29, 2003 at 03:53:06PM -0400, Brian J. Murrell wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]

> On Tue, 2003-04-29 at 15:37, D Richard Felker III wrote:
> > I don't see the file....are you sure you uploaded it?
> 
> I'm sure I uploaded it but for some reason, the name was not as I asked
> in the "put" command(s).  The two files are test.mov and
> mencoder.output.  Strange.
> 
> > BTW, most likely this means the input content is NOT telecined,
> 
> It's most definitely telecined.  I edit it with glav (from mjpeg-tools)
> which allows me to step one frame at a time (something I wish I could do
> with mplayer) and I can most definitely see the 3:2 pull-down.  It
> should also be top-field-first, as I believe lavrec always grabs/stores
> top-field-first.

OK, I've looked at the files. The problem is that you have a very low
quality TV capture with lots of noise, so the filter can't accurately
measure whether to merge frames. Fixing this is difficult, since
making it less sensitive will break handling of complicated telecine
patterns at scene change and such. IMO, the only way vf_ivtc will ever
handle bad noisy input like this is if I find a significantly better
metric for measuring whether fields match...

BTW, a workaround is -vf denoise3d,ivtc (denoise the image first). It
does fairly well with the default values, but you can turn up the
spacial denoising if you find it's still messing up at some parts.

> > but
> > just plain 60 field/sec interlaced.
> 
> Definitely not.

Some segments definitely are, but the beginning is telecine. The parts
meant to look like someone is recording them with a video cam are 60
fps.

> I should have noted in my original message, that I get decent
> (progressive, non-every-4th-repeating frames) results from detc with
> this same file.

Yes, and for this same reason, vf_detc fails on lots of 'clean' input
when the pulldown pattern is tricky or irregular... :(

Rich