Currently we just index in the frameDir. This might cause issues if the files are sorted differently on different hosts.
We should count how many frames there are and create a full frame name with passed 0 at the start, just like ffmpeg names them.
This way we can uniquely identify the files / frames and don't rely on the pre-processing being 100% done
Currently we just index in the frameDir. This might cause issues if the files are sorted differently on different hosts. We should count how many frames there are and create a full frame name with passed 0 at the start, just like ffmpeg names them. This way we can uniquely identify the files / frames and don't rely on the pre-processing being 100% done