stop vs abort segment detection

[v-src]

From what I can tell, the scene detection features let you abort the detection in progress, but any segments that might have been found to that point are lost. It'd be nice if the features let you choose to stop, but pick up the detected-so-far segments rather than only to abort and lose any segments detected so far.

My use case is that in VideoRedo, I often trim off the beginning and end of videos without worrying about segments in between (e.g. recordings of PBS music concert shows that have no commercials within the program, only the beginning and end sponsor credits and station promos that I get rid of). I start scene detection, let it run through the first few minutes, stop it, jump to the last few minutes, do another detection, then review the findings to precisely locate the first and last frames of interesting concert footage, then export only that stuff. It would also be helpful if the progress indicator were not just a percentage, but a dynamic mark on the timeline and maybe even reflected in the thumbnails or video playback.

I suppose that for my use case, I could create a segment from the beginning through that first few minutes, do a detection, then create another segment containing the last few minutes and doing another detection. But as I mentioned in 2324, dealing with heavier weight segments vs scene/book marks can be cumbersome, at least for my video editing style.

[v-src]

Playing with my use case some more in lossless-cut, rather than doing black or silent segment detection, I'm jumping near the time I expect the main program to begin (or end) and using a combination of zooming (16x seems good for a 30 minute program), audio waveform viewing to visually find the silence just before the main program, then doing a split segment at cursor operation to get the begin/end of the main program. I only have two segments to discard, at the beginning and end, keeping the one in the middle for export.

It'd be nice if I could use the automated black or silence detection instead of me looking at the thumbnails or waveform to find where to do the segment split, but for the reduced amount of video editing I've been doing these days, this one-large segment workflow is serviceable with the current feature set. It will still be a bit cumbersome with a multi-segment program, e.g. for an OTA broadcast of a movie full of commercial breaks that I want to record, edit out commercials and watch later uninterrupted. It would still be helpful to start, then stop segment detection rather than aborting and losing findings, and have the progress reflected in the video viewer or thumbnails or timeline.

[mifi]

next nightly build will have streaming detection of segments, meaning they will appear immediately on the timeline once detected, and if you abort, they will not be deleted.

[v-src]

Awesome, hope to try this out soon.