-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Ooops, the mailing list seems to be down. I CC Pau and Xavi, so they get it anyway.
I did some simple unit testing to freeze the behaviour of the spectral peaks addition/merge before doing further changes. By running such tests i saw what it seems another bug: When both arrays are merged, the last peak of each array is ignored. Is this a bug or is this a feature? Anyone arround requires such a thing? David. El 28/12/13 22:23, David García Garzón ha escrit: > While clearing compilation warnings in CLAM, I detected what looks like > a bug in the SpectralPeakArray addition/merge: we are taking values for > the target phase array from the magnitude array of the origin. The > warning complained about the cached reference to the operand phase > arrays being unused. > > By commenting out the operator, the only code that complains are the > SMSHarmonizer and the SegmentSMSHarmonizer so I guess that's the impact > of the bug or an eventual fix. > > No unit test covers the addition and no functional test covers the > SMSHarmonizer. > > Xavi, my fix proposal is to take the phase as is from the merged arrays > but I dont know if this dumb phase merging has any impact on synthesis. > > David. > > > -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.14 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iF4EAREIAAYFAlK/WrsACgkQxP0Ao7s0Juij9wEAqxi7PmFeHLhWjMmfGVj5hyi5 aeGfNEtS8DjDGo+Jz/sA+QGa5LlkcgBl0eAaT0VyLvIbGw+keEfZedZvic3/Idzj =SkcV -----END PGP SIGNATURE----- _______________________________________________ clam-devel mailing list clam-devel@lists.clam-project.org http://lists.clam-project.org/listinfo.cgi/clam-devel-clam-project.org