Decimate while dropping equivilent audio frame

[xbrav]

Hey all,

Stuck with a bit of an interesting scenario here. I have a video that was incorrectly converted from a native PAL source to NTSC. Hopefully I can describe this accurately...

The original reel was 25fps, but was converted to 30 by duplicating every 6th frame, including the audio. This results in a slight doubling of audio, noticeable mostly when there are character vocals.

Decimate drops the frame rate to 23.976, which is expected as they incorrectly scanned the PAL source at too slow of a speed.

Is there any plugin or method to drop the audio applicable to the decimated frame to sync the audio accordingly? Thanks!

[poisondeathray]

Avisynth's SelectRangeEvery() can process video and audio. Instead of "dropping", you are "selecting" or keeping

eg.
SelectRangeEvery(every=6, length=5, offset=0, audio=true)

will pick 0,1,2,3,4, 6,7,8,9,10, 12,13,14,15,16, 18...

Change the frame offset value to start at different part of the cycle

The pattern must be consistent, not adaptive for this to work. eg. If the video was edited after the transfer, the pattern might be interrupted

[jagabo]

If the pattern doesn't change too often you can trim the video into sections that are handled individually then paste them all back together at the end. Unfortunately, previewing that filter with audio doesn't seem to work very well.

And there's another possible problem. If the audio has an advance or delay relative to the video the audio may not be cut at the exact same points in time as the video. So you might be left with some click/pops/etc.

There are runtime filters that may be able to do this too, if the duplicates are exact enough.