[Enhancement Request] Compare 2 differing files and return common sections/ranges within them about soundfingerprinting HOT 3 CLOSED

Did you look at https://github.com/AddictedCS/soundfingerprinting/wiki/Interpreting-Query-Results ?

Specifically Coverage object, it contains all the required info you've asked about.

from soundfingerprinting.

| Also, is there any way to configure the sensitivity / "fuzziness" of the gap/segment match detection of the coverage?

You can control the sensitivity of the query using ThresholdVotes parameter on QueryConfiguration; the default value is 4, lowering it to 3 will instruct the algorithm to be less precise, increasing recall.

As to the permitted gap - it instructs the algorithm to ignore gaps of a certain length (default value is 2). Lowering it to 0 makes little sense, as it will start reporting too many tiny gaps with less than one fingerprint (~1.86 sec) in length.

Closing the issue as the original question was answered.

from soundfingerprinting.

Thank you so much for your quick response!

I was fooling around with QueryResult to no avail, and completely overlooked the coverage object (in the LCS namespace, no less!).

Currently I am using the coverage.TrackGaps/QueryGaps to determine the bounds of the likely matching segments and then fingerprinting those sections to get the QueryResult with the confidence values. Is this the "proper" way to do that? In my case I have common content to identify and remove from multiple files, so I fingerprint one entire file, query it against one other file to get the gaps, fingerprint the segments between those gaps, and then query the rest of the files against those.

Also, is there any way to configure the sensitivity / "fuzziness" of the gap/segment match detection of the coverage? Is that just done through the permittedgap/stride of the original fingerprint/query? I don't think i really understand what the permitted gap is/does, since I have it set to 0 and gaps between the location (starting time of match) matches in the track and the query are often much greater than that (they do not share the exact same location/timestamp in both files compared). Permittedgap appears to be for accommodating gaps of silence in the fingerprints and/or query result matches (at least that is what I just used it for) - is that right? Also, even when I set the permitted gap in the query, the coverage.permittedgap remained 0 (although it was obviously affected / altered by the query permitted gap) - could that be a bug? There doesn't appear to be any way to alter that value (coverage.permittedGap) in any case.

In any case, it is working excellently for my purposes (though now I need to figure out how to add video fingerprinting into the mix to make it work even better). Many thanks!

from soundfingerprinting.