judge@gpu.utcs.utoronto.ca (Peter Judge) (06/29/90)
I have in mind an application that creates two or three sounds algorithmically (stuffing them into SoundObjects), and uses the DSP to scale each one separately and then combine them into one sound to output to the DAC. The processing of the previously created sound should occur in as near to realtime as I can make it. I understand that this capability will be offered as a Class in the next version of the OS. In the meantime... I'm not clear on the convention to download more than one sound (each of which are longer than than the DSP hardware buffer) to the DSP and then invoke the appropriate array processing dspwraps to do the work. How can this be coordinated in such a way as to preserve as much realtime performance as possible? Is it even reasonable to expect that, using the sounddriver, I could output these 'massaged' sounds in near realtime? Thanks for your help. Peter Judge (judge@credit.erin.utoronto.ca) -- =============================================== judge@credit.erin.utoronto.ca (Peter Judge) |) judge@gpu.utcs.utoronto.ca ===============================================