NEW 76394
WebAudio: Optimize AudioBus::createByMixingToMono()
https://bugs.webkit.org/show_bug.cgi?id=76394
Summary WebAudio: Optimize AudioBus::createByMixingToMono()
Jer Noble
Reported 2012-01-16 11:09:38 PST
WebAudio: Optimize AudioBus::createByMixingToMono()
Attachments
Patch (4.40 KB, patch)
2012-01-16 11:17 PST, Jer Noble
no flags
Patch (4.42 KB, patch)
2012-01-16 13:01 PST, Jer Noble
eric.carlson: review+
Incorrect patch, please ignore. (17.90 KB, text/plain)
2012-03-20 13:58 PDT, Jer Noble
no flags
Jer Noble
Comment 1 2012-01-16 11:17:41 PST
Chris Rogers
Comment 2 2012-01-16 11:45:16 PST
Comment on attachment 122664 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=122664&action=review > Source/WebCore/platform/audio/VectorMath.cpp:174 > + *destP = **source1P + *source2P) * *scale; ** should be (*
Chris Rogers
Comment 3 2012-01-16 11:47:01 PST
Comment on attachment 122664 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=122664&action=review > Source/WebCore/platform/audio/VectorMath.cpp:172 > + int n = framesToProcess; I would pull scale into a local here so you don't have to have a pointer de-ref inside the loop float k = *scale; and then use k below
Jer Noble
Comment 4 2012-01-16 13:01:50 PST
Chris Rogers
Comment 5 2012-01-16 13:39:09 PST
Seems fine to me.
Jer Noble
Comment 6 2012-03-20 13:58:34 PDT
Created attachment 132887 [details] Incorrect patch, please ignore.
Jer Noble
Comment 7 2012-03-20 13:59:15 PDT
Comment on attachment 132887 [details] Incorrect patch, please ignore. Uploaded patch to incorrect bug. Please ignore.
Jer Noble
Comment 8 2012-03-27 10:07:37 PDT
(In reply to comment #5) > Seems fine to me. Chris, could you give an official r+ now that you have reviewer powers?
Chris Rogers
Comment 9 2012-04-10 11:58:59 PDT
Hi Jer, I'm confused. Isn't this just an earlier version of: https://bugs.webkit.org/show_bug.cgi?id=74553 which you just landed today?
Jer Noble
Comment 10 2012-04-10 12:40:25 PDT
(In reply to comment #9) > Hi Jer, I'm confused. Isn't this just an earlier version of: > https://bugs.webkit.org/show_bug.cgi?id=74553 > > which you just landed today? No, this optimizes one path of AudioBus to use a vector operation instead of a for-loop. It's orthogonal to the silence patch.
Chris Rogers
Comment 11 2012-04-10 14:52:17 PDT
(In reply to comment #10) > (In reply to comment #9) > > Hi Jer, I'm confused. Isn't this just an earlier version of: > > https://bugs.webkit.org/show_bug.cgi?id=74553 > > > > which you just landed today? > > No, this optimizes one path of AudioBus to use a vector operation instead of a for-loop. It's orthogonal to the silence patch. Sorry, still confused :) The latest patch/attachment is dated 2012-03-20 and seems to be an earlier version of the silence-hint patch. For, example the ChangeLog says: 12011-12-14 Jer Noble <jer.noble@apple.com> WebAudio: propagate a silence hint through the AudioNode graph. https://bugs.webkit.org/show_bug.cgi?id=74553 Do you have a more recent patch I'm not seeing?
Jer Noble
Comment 12 2012-04-10 15:36:35 PDT
(In reply to comment #11) > Do you have a more recent patch I'm not seeing? Ah yes. Webkit-patch upload uploaded to the wrong bug. See comment #7: "(From update of attachment 132887 [details]) Uploaded patch to incorrect bug. Please ignore." I've marked that incorrect patch as obsolete. The patch to which I refer is this one: <https://bugs.webkit.org/attachment.cgi?id=122675&action=review>
Chris Rogers
Comment 13 2012-04-10 17:08:29 PDT
(In reply to comment #12) > (In reply to comment #11) > > Do you have a more recent patch I'm not seeing? > > Ah yes. Webkit-patch upload uploaded to the wrong bug. See comment #7: > > "(From update of attachment 132887 [details] [details]) > Uploaded patch to incorrect bug. Please ignore." > > I've marked that incorrect patch as obsolete. The patch to which I refer is this one: <https://bugs.webkit.org/attachment.cgi?id=122675&action=review> Could you please re-upload it to this bug?
Note You need to log in before you can comment on or make changes to this bug.