Support fast-copy mode with FFmpeg instead of only Mkvmerge

[raj6996]

Bug/Issue Description:
Mkvmerge works great, but output files are .mkv instead of mp4.
As Vegas software is not works on mkv. so I need to re-encode to mp4 from mkv,
I replace ffmpeg.exe with mkvmerge on scenedetect, but it give me this.

Error: Invalid value for split-video: mkvmerge is required for split-video -c/--copy.
Install the above tool to enable video splitting support.
Is there anyway to copy option with ffmpeg, I don't need to re-encode!

Computing Environment:

• OS: 64bit Win10 Pro
• Python Version: 64bit 3.7.0
• OpenCV Version: Latest

[Breakthrough]

This is a good idea, and I definitely want it implemented - I'm just not sure the best way to resolve the discrepancy of when you have both tools installed, which one should be used. I could allow specifying the file extension for the video name and using that, but that feels inelegant.

I think the best way to solve this is to have a settings file of some kind, and maybe have a command line option to set which program will be used (e.g. scenedetect set copy_program ffmpeg).

Another option is to always default to ffmpeg, and have another command line switch to select a particular program. Not sure which is the best option yet.

EDIT: This is now possible by setting the -a/--override-args option with the split-video command, e.g.:

split-video -a "-c:v copy -c:a aac"

[raj6996]

scenedetect -i "G:\M\M\Mno.mp4" -o "G:\MGPM\2" detect-content split-video --copy
Original file was .mkv, then I converted it to MP4 with copy option of ffmpeg. Then I ran upper command, I removed mkvmerge.exe file from scenedetect folder, and inserted latest ffmpeg.exe into that folder. Finally i got that error. Hope take a look into it.
By the way. I converted all splitted mkv files to mp4 with batch process. more improved and better and faster detection than previous version beta version. Thanks

[anujtyagi2802]

scenedetect -i video.mp4 -s video.stats.csv detect-content list-scenes -n gives syntax error

[Breakthrough]

Hi @anujtyagi2802;

Would you be able to submit a new issue report for that, or is it related to #77? Could you share the console output as well as the debug output to help investigate the error you're having?

Thank you!

[anujtyagi2802]

[Breakthrough]

Hi @anujtyagi2802;

See this page in the manual for an example on how to use the Python API. What you typed there is for the command line interface (CLI), not for a Python script or interactive environment. For that you need to use the SceneDetect API directly.

Thanks.

[Breakthrough]

This is now possible by setting the -a/--override-args option with the split-video command, e.g.:

split-video -a "-c:v copy -c:a flac"

[Breakthrough]

In v1.0, -c/--copy will now use ffmpeg, and added a separate option -m/--mkvmerge to use mkvmerge for splitting instead. This means that if ffmpeg is no longer available, the -c/--copy option will result in an error. Therefore any tools/scripts relying on the previous behaviour will need to replace -c with -m.

[Breakthrough]

This is now available in the v0.6 beta.