[linux-audio-user] Re: Stuttering via HDSP, Cardbus, Multiface?

Mark Knecht markknecht at comcast.net
Fri Mar 26 20:18:43 EST 2004


On Fri, 2004-03-26 at 16:08, Tim Blechmann wrote:

> do you still have a configuration that might _break_ the hdsp? e.g.
> booting with acpi enabled? i'm interested in the distortion of the sound
> ...
> what's the output of jackd -R -d alsa (on the broken system)? does it
> complain about anything?
> 

Can we construct some sort of environment to test the problem and gather
real data? 

For instance, if you play a sine wave in some application, and send it
to a specific ADAT (or spdif) output, can you loop that ADAT (or spdif)
output back into the same card and record it? Can you record it on
another spdif enabled recorder and then compare the recordings?

I'm unclear whether the same sine wave, transmitted on every output
produces the exact same problem.

Since cardbus controllers are essentially PCI bridges this means they
have buffer memory that is written by DMA devices. Is this problem
related to buffer memory being over used? 

Or is it purely a timing issue? Is the since wave going through the
controller, but extra data (all zeros) is being inserted? This probably
means an addressing problem as opposed to a data problem.

If you install snd-dummy and run Jack can you show conclusively that
there is no problem back inside the computer? While it's easy to say
this is a card bus problem, what if  it's related to some other part of
the system?

Some carefully constructed experiments could give us a better idea
what's going on.

Just asking...

- Mark



More information about the linux-audio-user mailing list