Eli Brandt wrote:
> 
> A third catch for good measure: their clock crystals will drift.
> General solutions to this are expensive, and limited solutions are annoying.

Reviewing some basic facts may save some frustration.

Every DAC has a finite amount of delay while it does the conversion.
This delay time seems to vary with the chip used, design implementaion,
etc. Therefore, using simultaneous digital and analog outputs would
cause the two stereo pairs to be out of sync with each other. 

Analog outputs from two sound cards of different makes/models would
probably have the same problem due to different processing delays in the
DACs. Two identical stereo soundcards would need to be run from a common
clock, as Eli Brandt points out, or phase shift results.

The delays can be less than 1 millisecond up to several ms. With shorter
delays, the problem is with material common to both stereo pairs, i.e.
one sound centered or panned from one channel to the other.

It seems a single integrated solution is in order.

David Boothe

David Boothe wrote:

>Every DAC has a finite amount of delay while it does the conversion.
>This delay time seems to vary with the chip used, design implementaion,
>etc. Therefore, using simultaneous digital and analog outputs would
>cause the two stereo pairs to be out of sync with each other.
>
>Analog outputs from two sound cards of different makes/models would
>probably have the same problem due to different processing delays in the
>DACs. Two identical stereo soundcards would need to be run from a common
>clock, as Eli Brandt points out, or phase shift results.

No promises here for perfect results, as i haven't dealt with quad output,
but Digital Performer does enable you to slightly nudge selected tracks out
of synch. Or it could be nudged in csound before moving into DP or other
commercial software...

(check to see if free ProTools 3.4 still offered)
www.digidesign.com

of course compatibility with your sound card(s) is essential, and this is
what might make it necessary to leave csound in the first place. but such
commercial aps do have their virtues. Anyway, presumably, nudging may
satisfactorily compensate for the above phenomenon *if* analog and digital
outputs may be simultaneously used on a single card. But if two cards are
used, as suggested above: drifting will have to be addressed with a common
clock from MOTU or some other source.

>The delays can be less than 1 millisecond up to several ms. With shorter
>delays, the problem is with material common to both stereo pairs, i.e.
>one sound centered or panned from one channel to the other.

mmm. one milli? in DP, may have to mess with tempo settings and time
feature to trick the software into that kind of resolution without
affecting playback speed, (480 ticks per quarter is limit as midi
reportedly not good for any more than that, and DP which also handles midi,
used that number as a benchmark).

>It seems a single integrated solution is in order.

absolutely.

tolve