-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Somebody in the thread at some point said: | This filter is useful to reject clicks that are not reliable. We | only care about what happens when we receive DOWN events for the fist time. | If this filter does not reject the first samples then it will change | its internal state to "passed" and the remaining samples | will be sent to the next filter in the chain. | | First we collect N samples, then then we sort them. We discard the borders | of the vector (with a window) and then compute the variance of the remaining | set. If the computed variance is bigger than a threshold, we reject the click. | | The parameters of the filter are: | | * size | * threshold /* maximum allowed variance */ | * window | | I'm using values I tuned by trial an error. I noticed we discard a lot | of bugus clicks. In fact, it seems harder to get them now.
Thanks Nelson I sent these two on stable-tracking. - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkk4TD4ACgkQOjLpvpq7dMpYzgCfd6u2mxXQG3kX8Baa8XUVwjHm W+gAn37QzhOqj2r58BapV25MRIRXU1aI =vEHP -----END PGP SIGNATURE-----
