So looking for audio fingerprinting techniques to solve the problem using libraries like, dejavu didn’t not solve the problem, since it results in very similar similarity scores for two different audio tracks of the game. I assume it’s primarly because it’s intended purpose is to detect similar songs and hence I assume they would probably look to reduce noise.

Is there a similar library/technique that could solve the problem for an audio file, such as mine?

To give an overall picture, I’m basically trying to detect if a piece of audio (assume no commentary track just crowd noise), came from game A, if so at what time, if at all this helps.

Thanks.