[Webkit-unassigned] [Bug 77224] noteGrainOn is too long

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=77224





--- Comment #3 from Chris Rogers <crogers at google.com>  2012-02-22 14:25:09 PST ---
Sorry, I'll try to explain a bit better.

This is a work-around for the latency in HRTFPanner processing (assuming that there is an AudioPannerNode in the processing chain after AudioBufferSourceNode).

If there is a panner and the extra 512 sample-frames are not generated in AudioBufferSourceNode, then the nodes get torn down immediately when the grain finishes playing.  But, because the HRTFPanner incurs a processing latency, it still has unprocessed sample-data which has not yet been output.  And so the very end of the rendered signal is truncated and can cause glitches.

Ideally, the handling would be more general than just HRTFPanner, and would handle *any* node which incurs processing latency.  We should talk about this offline to consider the best solution.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.