This is the implementation part. Testing this needs multiple iterations on the try server, and I'll work on that later.

Comment on attachment 745461 [details] [diff] [review] Part 1: implementation Review of attachment 745461 [details] [diff] [review]: ----------------------------------------------------------------- Seems to me that instead of going through mSharedChannels you could work on the mJSChannels directly, and be more efficient and maybe simpler? ::: content/media/webaudio/AudioBuffer.cpp @@ +195,5 @@ > + channels[i] = mSharedChannels->GetData(i); > + } > + > + // Prepare the output channels > + float* downmixBuffer = new float[mLength]; nsAutoArrayPtr. Or else ns(Auto)TArray<float>

(In reply to Robert O'Callahan (:roc) (Mozilla Corporation) from comment #2) > ::: content/media/webaudio/AudioBuffer.cpp > @@ +195,5 @@ > > + channels[i] = mSharedChannels->GetData(i); > > + } > > + > > + // Prepare the output channels > > + float* downmixBuffer = new float[mLength]; > > nsAutoArrayPtr. Or else ns(Auto)TArray<float> I can't use these classes, since AudioChannelsDownMix expects a float** and I can't use it with the rvalue float* accessors of these classes.

https://tbpl.mozilla.org/?tree=Try&rev=f9bef74bc83b https://tbpl.mozilla.org/?tree=Try&rev=dff84f6bff7a (Note that I might do a few try rounds to adjust the fuzziness numbers in the test, but they shouldn't really require their own review rounds.)

https://hg.mozilla.org/integration/mozilla-inbound/rev/3370e2c73cea https://hg.mozilla.org/integration/mozilla-inbound/rev/f958d3fee49f

https://hg.mozilla.org/mozilla-central/rev/3370e2c73cea https://hg.mozilla.org/mozilla-central/rev/f958d3fee49f

Mass moving Web Audio bugs to the Web Audio component. Filter on duckityduck.