hi iohannes ... hi all ...

you added the changes for a variable dac-size to devel ... cool ... but i somehow doubt that it's possible to reduce the latency to the minimum with that ...

no matter what your audio api is, you will always lose one hardware buffer size

i suppose, you could run the dsp_tick() inside the driver callback to get the optimal latency results, but one would have to rewrite the dsp scheduler ... since the least hardware buffer size i know of is 64, in theory it should work with pd's default dac blocksize ...

tim, who still thinks of an elegant way to rewrite the scheduler