[Webkit-unassigned] [Bug 73545] [GStreamer] FFTFrame implementation

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=73545





--- Comment #28 from Chris Rogers <crogers at google.com>  2012-01-26 12:23:44 PST ---
(From update of attachment 124088)
View in context: https://bugs.webkit.org/attachment.cgi?id=124088&action=review

Philippe, the FFTFrame part looks good overall.  I added one comment there.  It's very encouraging that your code is passing the convolution layout test (with your slight threshold change).  That test really puts your code to work!

> Source/WebCore/platform/audio/gstreamer/FFTFrameGStreamer.cpp:151
> +    VectorMath::vsmul(processedRealData.data(), 1, &scaleFactor, realData, 1, size);

For lines 143:151 I'd recommend either reverting back to your previous approach (not using VectorMath::vsmul()) or calling VectorMath::vsmul() DIRECTLY on m_complexData,
since effectively it's a floating point vector (with alternating REAL/IMAG values).  To operate directly on m_complexData you'd have to use a static_cast<>

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