I would imagine it wouldn’t take more than a standard stereo mix: there’s still the summing and panning process, but instead of summing and panning channels you’re summing and panning mixes, and then the same DSP found of a standard stereo mix (peq, geq, compressor and delay) gets applied.
