>which shares the central ice1712 chip with the delta series. there were >spurious strange discontinuities in the recorded signal then, but i >haven't found out what caused them - might be they were unnoticed xruns.
i think these might have been fixed by some recent changes to the audioengine code (changes that originated in JACK). >iirc there were interrupts with a very small number of frames available >on one stream, and >= full frames / cycle for the other, which should >not happen at all with linked ALSA-streams (duplex setup) on this chip. >however for one i haven't run ardour or jack in a while now, for another >ALSA has evolved, and for yet another i have just seen the card perform >flawlessly at 16 frames / cycle with different code. 16 frames per cycle! yow! this h/w can really do that? --p
