On Thu, 16 Feb 2006 17:45:56 -0500 Paul Davis <paul@xxxxxxxxxxxxxxxxxxxxx> wrote: > On Thu, 2006-02-16 at 22:14 +0100, Arnold Krille wrote: > > I just want to add another trick I learned from my hardware-dealer who > > syncs ISDN-cards: Maybe its not enough to just couple their clocks, > > they also change the driver to mask the IRQ's from the slave cards and > > just react on the master-IRQ and then the driver automaticly > > proccesses the data on all cards. > > thats a very cool idea. i wish i had the time to add that to the RME > hammerfall/hdsp/madi drivers because they can all have IRQ's disabled or > enabled. you could then load the driver with an option to disable IRQs > for all but the first card. nice. One would need to ensure that the start/stop time of the cards is right in sync in the first place no? After that, if properly sample synced everything is fine. I don't know how the hw works, but i imagine it has a "pointer" into a sample buffer where it puts each new sampled value (and another buffer with another "pointer" where it reads from, etc pp). These "pointers" need to have an offset as small as possible between two cards, ideally it would be 0 (So that in a sense, if both cards still generated interrupts, they would happen at the exact same moment). I always wondered about this part of ALSA. How do you start/stop soundcards at the exact same moment? Does stuff like the multi or whatwasitsname ALSA pcm plugin care about this? Flo -- Palimm Palimm! http://tapas.affenbande.org
