real-time audio synthesis


(Jake Araujo-Simon) #1

Hello! I’d appreciate help with the following:

What’s the most efficient way to use a large array of oscillators for real-time audio synthesis? I’ll need to change the frequency and amplitude of each component almost every frame.

Should I have one dsp per channel, and change the dsp’s frequency and the channel’s volume?

I’m just starting my FMOD research with this - early-stage help would be incredible! Also, performance is important here.
Thanks.


(Brett Paterson) #2

Hi,
If you have a complex model, and want it to be resampled as a group, then use System::playDSP on the root node.

If you want every single node to have a resampler, then you would have to use System::playDSP on every DSP, but then you lose the ability to connect the graph up the way you want it, though you could achieve that with Channel objects, by using ChannelGroup API instead to build a graph, rather than doing it with DSP::addInput etc.

This is just an overview answer, if you have specific questions I can help with that.