Chris Marrin had asked about getting access to individual channels. I've
added AudioChannelSplitter and AudioChannelMerger to my proposal document:
http://chromium.googlecode.com/svn/trunk/samples/audio/specification/specification.html
I've also implemented them and checked in the code to the WebKit audio
branch.
These two types of AudioNodes may be used together to implement interesting
types of multi-channel matrix mixing or virtualized down-mixing. I've also
created a "ping-pong delay" demo (not yet checked in) to illustrate other
possible uses.
The more detailed description and diagrams are in the proposal document, but
here's the basic idea:
*AudioChannelSplitter:*
[image: channel-splitter.png]
*AudioChannelMerger:*
*
*
*[image: channel-merger.png]
*
*
*
*
The controversal part may be how I'm proposing that there be a fixed number
of inputs and outputs, which will depend on the highest number of channels
supported (6 channels for 5.1 in this case). But this exact upper limit can
be flexible and could be increased to support arbitrarily numbers of
channels (7.2 and higher). Still, I wanted to avoid cases where the number
of outputs of AudioChannelSplitter would change dynamically depending on
what is connected to it since this could create problems if outputs which
are connected to other things suddenly disappear in the middle of
processing. With a fixed number of outputs, the irrelevant (inactive)
outputs would simply output silence if they're connected to anything at
all...
Chris
*
*
*
*
*
