[wdmaudiodev] Re: Looking for Audio specialist

  • From: Clemens Ladisch <clemens@xxxxxxxxxx>
  • To: wdmaudiodev@xxxxxxxxxxxxx
  • Date: Mon, 02 Nov 2009 18:04:00 +0100

Peter Feucht wrote:
> Years ago we developped a system consisting of hardware (which can be
> roughly seen as a wave player) and a software running under WIN XP on a
> laptop.
> All this was working fine so far, but our customers want to switch to more
> modern OS like Vista or WIN7 now and here our WIN software doesn't work any
> longer.
> The software has been developped by a contract IT specialist who is no
> longer available. So I contacted other people here in our region to find out
> where the problem is. The result was, that our software uses DirectSound
> which is no longer available in modern OS (at least, that's what I was
> told).

DirectSound is still available.  Due to the changed audio stack in Vista
and later, it is no longer capable of hardware acceleration, but USB
audio has never supported this anyway, so this should not be a problem.

Your problem might simply be some misconfiguration, or your software
might do something that actually is not compatible with the Vista
drivers.  Can you be more specific than "doesn't work"?

Just a wild guess: Vista doesn't support the DSBCAPS_LOCHARDWARE flag,
so if your software uses it, try replacing it with 0.

> 1. Inside our hardware we are using a TI codec type 2902 which has a USB
> front end, which is connected to the laptop. One problem is, that TI
> supports this IC not very well, means, there are no special drivers
> available, so we have to use generic drivers.

The generic drivers do actually work and are available for practically
every OS you could name (both in contrast to most vendor drivers);
making the codec class compliant is the best support that TI could have
given you.

Your problem seems to be with the APIs on top of the driver.

> 2. We are working with standard audio format (16Bit, Stereo, 44,1KSa) but we
> have to be able to send very short waves (e.g. 50ms long), so there must be
> a possibility to erase data which are already in the audio stack OR the
> stack must be very short [...]
> 5. The audio data is partly generated "on the fly",

So the audio data must be generated in real time (in response to some
event), and you have an upper bound on the latency?  Is that shorter
than 50 ms?

Best regards,

WDMAUDIODEV addresses:
Post message: mailto:wdmaudiodev@xxxxxxxxxxxxx
Subscribe:    mailto:wdmaudiodev-request@xxxxxxxxxxxxx?subject=subscribe
Unsubscribe:  mailto:wdmaudiodev-request@xxxxxxxxxxxxx?subject=unsubscribe
Moderator:    mailto:wdmaudiodev-moderators@xxxxxxxxxxxxx


Other related posts: