Closed InoSiX closed 5 years ago
There is a option to set a manual time. It would still require processing the audio, it was added so the user can override the sequence found by ffmpeg and audio analysis. Check the check_db and manually_correct_theme command. There is no way to add a button to all clients, we could add that for the web client using a grease monkey script.
Is there any reason we couldn't skip the processing all together? There is no need for analysis if the user supplies the timestamps.
That’s outside the scope of my project. I have to much content to even consider that. With that being said, it would much simpler to just create a edl and create chapters in the video files.
Some of the automation of this project is really cool. I am not sure how a company like Netflix does it, but I imagine it is not using script like you are doing it here. I imagine they have someone that manually provides timestamps for the beginning and ending of the sequence they wish to skip (intro/outro).
Just thinking out loud here, with the sophisticated work you have already done, perhaps an additional option should be provided: Provide the user the ability to manually provide timestamps, allowing the user to skip on their terms using items/inputs they have provided. It could be stored locally in the folder of the media, and be in simple .txt form. If not providing the user a button to set these times live while watching the episode, at the very least the ability to manually edit this .txt file with the appropriate times.
This would drastically lower the CPU required to compare audio waveforms.
Just a thought.