Hi all,
We are working on beagleboard and want to have a full duplex ALSA based sound processing with DSP run as sound processor. We are trying to run codec engine's example which just copying buffer using memcpy. But we are having some problems. We cannot get the right output using simple codec running on DSP, but when we use simple memcpy on GPP it's fine. We run Angstrom Linux on GPP. We think there is something wrong in GPP-DSP link setting or some overhead in it. Full duplex processing has thight constraint for processing time. Is there any suggestions how we can make this work ?
Thank you.
