Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
Splitting videos with an analyzed duration only splits within analyzed area #106
Thanks for your bug report! This definitely seems like an issue, I'll do some further investigation and hopefully have a fix for this included in the next release of PySceneDetect.
Just to help with the analysis, do you see the same behaviour if instead of setting the end time, you set the duration instead? (e.g. replace -e 00:15:00 with -d 00:02:00)
Absolutely! Below is the output for all four scenarios: stream-copy and stream-copy -c in both explicit end and duration variants:
Thanks for such a great tool!
I just pushed a fix for this to master just now (d097275). Feel free to download & install a copy locally to get a patched version up and running until this is available on pip (I'll likely publish v0.5.2 by mid September).
I also updated the relevant unit tests to ensure that this case is specifically checked for now (namely, that all scenes returned by the SceneManager must be in ascending order, and the start timecode must be smaller than the end timecode).
Thanks again @charlesvestal for your well-written and extensive report. All of the included information allowed me to make a much more focused and isolated root cause analysis, as well as updating the unit tests to prevent regressions going forwards. Much appreciated!
Finally had a chance to test this (sorry!).
While i'm now able to succesfully run the command, the analyzed area is treated as the whole area to split. My impression was that the analyzed area is where you would look for splits but the entire video would still be split. Is that incorrect?
i.e. i know there's a break between 13m and 15m, but the video is 30m long, I want to limit the analysis for speed but still have two roughly 15m chunks.