There is an audio file with a soundtrack, SFXs, and voice-overs. I have all of these included elements separately as original audio files.

I need to detect which file is used in what point (timestamps). I know detecting audio patterns is far easier than perfectly removing them and should be practically possible with the help of modern tools(eg. Shazam). Do you know any software/tool can do my job?

Actually I want to remove the voice over from the audio, as per my knowledge and experiments, the current techniques do not perfectly isolate voice from audio, so I am going to reconstruct the audio from the files based on their usage timestamps.


1 Answer 1


It's likely that the final audio file was not just a raw "addition" of the original files, and it was 'mixed' using various techniques (EQ,compression etc.) Therefore it's very unlikely you will be able to reconstruct the exact same mix unless you are willing to spend a lot of time with your ears.

There's no particular software that will do this, other than perhaps the audio synchronization tools in Final Cut Pro or Pluraleyes which might get you started. These tools are usually used to synch audio to video but might help in your case.

  • Then how does Shazam or any audio fingerprinting software work? :) I am just looking for an open software that I can use for the specific task. Commented Dec 1, 2019 at 4:08
  • Audio fingerprinting is a different technique to synchronisation. Shazam/fingerprinting is a different question altogether. All those software tools will do is to tell you what you are listening to, not where it appears in a timeline.
    – Mark
    Commented Dec 1, 2019 at 4:11
  • If you are looking for a software recommendation, probably should be looking on stackoverflow - this is more sound design here.
    – Mark
    Commented Dec 1, 2019 at 4:12
  • Social media platforms such as Youtube, Facebook use the same technique and can tell the exact timestamps of copyrighted audio. Commented Dec 1, 2019 at 4:18
  • 1
    Unfortunately Content ID on Youtube and Facebook is pitifully inaccurate. The number of false positives that assign copyright to an incorrect owner is monstrous, so no, they might think they can, but they can't. Additionally, you will need frame-accurate positioning, which these sorts of proprietary technologies cannot deliver. Hope you find a solution. If you do, please post back here.
    – Mark
    Commented Dec 1, 2019 at 11:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.