PySceneDetect
PySceneDetect copied to clipboard
Figure out how to properly handle multiple input videos
It might be useful for some cases to be able to concatenate input videos together, although this requires some special handling. This could be done by specifying a new StreamJoiner type that takes a list of paths as well as a type of backend to use, and handles seeking across videos. The constraints for doing so would be that video parameters must match (same framerate/resolution).
The most challenging aspect is likely going to be with the output pipeline, since all commands will need to be updated to handle multiple input paths. How to deal with file naming conventions will be a bit tricky as well.