[PlanetCCRMA] [PD] audio breaking

[Niels Mayer]

On Thu, Mar 4, 2010 at 6:41 AM, Ricardo Lameiro <ricardolameiro at gmail.com>
 wrote:

> I thinj he is using realtime kernel, as he is running jack with the -R flag
> (reatime) otherwise jack wouldn even start.


I run jack -R without a realtime kernel (stock fedora) and have no audio
dropouts with an M-audio delta 66.

IMHO, the more I think about it, the more I find USB/firewire to be a bad
solution for pro-audio, despite the fact that we're being pushed into it as
a solution.. A bad solution, IMHO.

Simply too many layers where things can go wrong. If you're getting firewire
dropouts, consider also that the issue could be coming from your graphics
card hogging the bus.
http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=NVidia+firewire+audio+dropoutis
a good place to start...

IMHO, the device needing the interrupt should generate the interrupt,
directly, for it's own purposes. Interrupt handling is very application
specific. The fact that you've got generic interrupts happening for any kind
of firewire event, with no obvious way to prioritize events (e.g. audio and
video events must take priority over other interrupts and require minimum
throughput guarantees), that to me is the root of the problem.

Firewire audio makes sense electrically (to keep audio separate from
computer noise) it just doesn't make sense computationally.  Electrically,
it is also totally superfluous: WHy not point-to-point gigabit ethernet and
little ethernet D/A and A/D modules and be done with it?

Niels
http://nielsmayer.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ccrma-mail.stanford.edu/pipermail/planetccrma/attachments/20100305/d862989d/attachment.html