Account for sample-rate in convolution normalization
Tested on many impulse-responses at 44.1KHz, 48KHz, and 96KHz to verify that volume differences are now fixed.
Created attachment 132424 [details] Patch
Comment on attachment 132424 [details] Patch OK. What's the state of testing for this sort of code?
There's a generic convolution test which covers nearly all cases, but we still should write a more specialized test here.
Committed r111100: <http://trac.webkit.org/changeset/111100>