11 Jan
2004
11 Jan
'04
7:14 p.m.
On Sun, 2004-01-11 at 14:25, derek holzer wrote:
Actually, it is the Ricoh cardbus controllers which
both RME and myself
would warn against. [But maybe you meant another Derek?] That said, I
*do* have a cardbus HDSP running on a laptop with this controller, but
let's just say it wasn't a picnic getting it this far. I also wonder
what effect this has on my latency, but haven't tried any benchmarks
against the same card with a PCI interface yet. Maybe some weekend when
I have lots of "free time". :-)
Not much - these are PCI type latencies and not audio buffer latencies.
The issue with cardbus controllers, or PCI bridges in general, is that
you are buffering from one PCI bus - the main PC PCI bus - to a second
bus on the other side of the controller. If the memory built into the
Cardbus controller is large enough then generally the whole audio packet
will make it across in one chunk. However, will smaller on-chip buffer
memories it can get broken into pieces and have to be reassembled in the
sound chip. This runs the risk that all the audio data will not be there
when it's needed.
Good to know that my friend's chips are not the ones causing problems,
unless Tim's problems are due to this chip which we do not know yet.
- Mark