Using thre FMOD API, I have a DSP unit which I want to apply to several channels. The normal way to do this would be to group the channels into a channelGroup and, say, add the DSP unit at the tail of the channelGroup.
My question is whether this would have any performance benefit over simply adding each channels output as an input to a DSP mixer unit, then connecting the mixer to the DSP I want to apply to the channels.
If its helpful I can provide more detail on my specific scenario, but provided the performance is not significantly impacted, there is a reason why I would prefer not to use a channelGroup.
There should be almost no performance difference between the two scenarios you describe. A ChannelGroup is a container for a DSP chain, by default that chain only contains one DSP (a Fader). Internally we connect the outputs of each Channel to the fader input, much like what you described for the mixer unit. So the question then becomes is there a performance difference between the mixer and the fader, by default, no, if the fader level hasn’t been changed then no processing is required, same as the mixer DSP.
I would recommend using the ChannelGroup approach for ease of use.