Active processing for Panner/Convolver/ChannelMerger

[rtoy]

Although it is specified that nodes that are not actively processing produce a single channel of silence, we need to make sure the various nodes alloow for this.

For example, PannerNode, paragraph 3 says:

The output of this node is hard-coded to stereo (2 channels) and cannot be configured.

That's inconsistent with nodes that aren't actively processing. We should update the PannerNode, the ConvolverNode, and the ChannelMergerNode to allow for this.

[rtoy]

A note in 5.3 also says convolvers output stereo all the time. This needs to be updated too.

[rtoy]

For reference, here are the other two places:

• https://webaudio.github.io/web-audio-api/#Convolution-channel-configurations is probably where we want to mention active processing.
• https://webaudio.github.io/web-audio-api/#channelmergernode, second paragraph that says the number of output channels is equal to the number of inputs