JorenSix
commented
2 years ago Thanks for the feature request / notification of a limitation.
In these cases it can indeed be of interest to graph the time of matching fingerprints. Programatically these can be accessed in the code matching fingerprints. Have a look there if you want to implement this.
It is a rather fringe use-case so in order to keep code and result reporting simple Panako, this functionality is not planned.
I have been exploring using audio fingerprinting to determine sync offsets in media files. So far Panako has been extremely accurate in certain cases. If I have 2 files that are identical, but one has 160ms of silence at the beginning of the file, Panako reports this as a 160ms offset. Perfect! If I have a second set of files the the same 160ms offset at the beginning and an edit in the middle that adds 250ms of silence, bringing the total offset by the end to 410ms, Panako reports this as a single offset value of 410ms.
If Panako could report a more verbose analysis, I could potentially use ffmpeg to realign the test file to the reference.
A real world example would be the reference file as an audio stream for episodic content and the test file being Dubbed in spanish. This would contain the same music and effects and should match a fingerprint for the music.