Software to identify duplicated frames in captures?
 

Hi all,

I am wondering if there is any software out there that can detect duplicate frames in an ALREADY CAPTURED video stream (two identical frames one after another) in an AVI file?

The use case would be seeing how often a TBC is duping frames after the fact and how Virtualdub does it if no TBC is used. I am certain that TBCs do duplicate frames, we just don't know when it happens because you can't see a nice report of statistics like Virtualdub shows you.

Would seem like a relatively straightforward thing for software to detect, just never heard of such a thing.

You would be slightly correct, but mostly wrong.

The reason TBCs don't dupe frames is due to the buffering, and it aligns/sync/etc on that buffered data. You only have "dupes" when the data on the tape is essentially missing (corrupted garbled output).

It's not at all the same as what capture cards or capture software does. Same end point, but vastly different route to get there.

I am not aware of such a tool but writing one in a scripting language (shell, PowerShell, python) is relatively easy

Step1: extract all images from the videos, this can be done easily with FFmpeg
Step2: loop through images and check if current = the next image or not
sth like this
https://stackoverflow.com/questions/...ctly-identical

It will be a good exercise to learn how to write ~10 lines of Python with a for loop if you don't know how ;)

But even if you find duplicate frames, what do you gain from it? in drawn movies you can have duplicate frames from the sources, in filmed content, while it is much rarer, you can still have still moments where 2 frames happened to be the same. Out of curiosity, when you identify the duplicate frames, what do you gain? how do you plan to solve it?

Pretty sure there is an avisynth plugin/filter that does that (or perhaps a combination of filters) checking dupes and storing which ones are in a txt file. You'd have to dig doom9's forum or simply ask

Agree that if frames are output erratically they'd be buffered and output smoothly from a TBC.

However, what I'm seeing is a single inserted frame every 2-3 minutes like clockwork depending on the tape.It almost seems like the VCR is playing the tape very slightly slowly (or that the clock on the computer and the VCR don't match exactly) so it needs an extra frame added every few thousand frames to keep it roughly at 29.97 according to the capture PC.

I do have a couple of TBCs, I'm just trying to justify if if makes sense to use it in a situation where there are zero dropped frames without a TBC. If the average frame rate is slightly under 29.97 (say 29.96), That would mean a single frame gets added every 2,996 frames and I think the standalone TBC would do the same thing, it just doesn't get reported to us.

The only real reason of avoiding a TBC in that situation would be potential for added noise from going through a longer video chain.

I'm kind of curious to see if a TBC duplicates those frames exactly as something like virtual dub does, which is why I'm interested in comparing a capture without and with a TBC if there are no dropped frames (only a few inserted ones) to retrospectively see how often it duplicates frames

I would say if you are getting dropped frames the buffering effect of a full frame TBC is definitely needed as this suggests that frames are arriving too erratically for the capture card to grab them all fast enough or within whatever window/variance it will allow for frame arrival and that data is being permanently lost. However, if there are frame inserts only, I am not sure if there'd be a benefit in that situation over just letting the Virtualdub do the frame insertions instead of the TBC.