IVTC Deinterlaced Film

[jmac698]

I have film material that was pulled down, then deinterlaced. As a result, every 5th frame is an exact dup and the deinterlaced footage has shimmered in areas of detail, due to the poor deinterlacing.
If the deinterlacer didn't touch each even(odd) field, I should have the exact fields still present, therefore I can IVTC it again.
The question is, how to adapt to find where I'm at in the pulldown pattern. I could end up IVTCing on the interpolated fields and it would still look like crap.

[Didée]

No, you cannot "IVTC it again". Because of the deinterlacing performed, there's nothing to field-match anymore. What the deinterlacer has discarded is gone forever. And basically it's not even possible to tell where the deinterlacer has deinterlaced and where not. ;)

The easy step is to Decimate it from (now) 30p to 24p: TDecimate() and be done.

Options after that step:

- if the deinterlacer was particularly poor, you might consider to re-create the odd-parity fields with e.g. NNEDI3.

- the shimmering possibly can be reduced by MDegrain1/2/3. It's probably needed to use a custom "search" clip that also reduces the shimmering - else, you'll likely get too big SADs in the affected areas, causing MDegrain to do nothing there.

[2Bdecided]

You could post a sample.

qtgmc can help with some of these kinds of shimmer - either the repair modes, or the full de-interlacing. Probably won't be perfect though. decimate after (or before might be better, depending on which mode you try).

Cheers,
David.