Show similarities of 2 audio files about soundfingerprinting

[AddictedCS]

This issue can be solved using dynamic programming techniques. They fit exactly for the problem described above.

[lvaleriu]

I'm working on this right know. I'll probably need to modify some interface methods return values in order to have some localisation informations about the fingerprinting data.

[AddictedCS]

Please do it on a separate branch.

Also, could you please elaborate on what technique will you apply. I was previously thinking about applying dynamic programming techniques. The only thing is that sub-fingerprints will have to have a sequence number which is currently not the case as I am using parallel procedures and resulting list of sub-fingerprints is not ordered.

Details about a potential solution can be found here.

[lvaleriu]

Ok, I'll do this on a separate branch. I've already modified the sub
fingerprint data in order to have a time position in the sound file ( I
think that I need more than simply one index ). As for the technique to get
2 sounds similarities i was thinking indeed on using the LCS technique as
described on the Wikipedia page.

Le jeudi 30 janvier 2014, Ciumac Sergiu [email protected] a écrit :

Please do it on a separate branch.

Also, could you please elaborate on what technique will you apply. I was
previously thinking about applying dynamic programming techniques. The only
thing is that sub-fingerprints will have to have a sequence number which is
currently not the case as I am using parallel procedures and resulting list
of sub-fingerprints is not ordered.

Details about a potential solution can be found herehttp://en.wikipedia.org/wiki/Longest_common_subsequence_problem
.

Reply to this email directly or view it on GitHubhttps://github.com//issues/18#issuecomment-33668312
.

[AddictedCS]

In case you need the right sequence number you will have to modify HashFingerprints method, from Parallel.For to a regular foreach in this class.

[AddictedCS]

First draft available starting from v > 2.0.0