Hi Andy,
What kind of latency figure would be acceptable in your scenario? I don't
think you will ever get it lower than 1 ms on XP.
Maybe, instead of going Windows CE, you could have a look at Linux. There
are some guys there doing good work. For example Ardour seems to be
making
progress. I don't know what kind of latency figures you can expect on
that
platform, but at least you have a lot of options to tweak.
I am not speaking out of my own experience. It's just some things that I
read about it that grabbed my attention.
-Evert
-----Original Message-----
From: wdmaudiodev-bounce@xxxxxxxxxxxxx
[mailto:wdmaudiodev-bounce@xxxxxxxxxxxxx]On Behalf Of Voelkel, Andy
Sent: Friday, March 02, 2007 11:25 PM
To: wdmaudiodev@xxxxxxxxxxxxx
Subject: [wdmaudiodev] Anyone here worked with Windows CE?
Hi all,
Having just gone through another frustrating afternoon trying to reduce
Windows audio latency, I am motivated once again to think of alternatives
for real time audio algorithm development. I have a couple applications
where a minimum 4.5 millisecond latency is just not attractive.
I will probably have to use standalone DSP cards for development in order
to
avoid this problem, but the development tools on such boards just can't
compare to Visual Studio.
I have thought before of building a Windows CE target using a standard
Pentium motherboard, and cross developing from a host Windows XP machine.
I've heard that the Visual Studio tools for this sort of cross
development
are pretty good. I would imagine that Windows CE could be configured to
have
much lower latency than Windows XP.
The problem is that the audio driver model is different, and I am afraid
that finding a Windows CE driver for multichannel audio IO would be
impossible, and that developing a driver myself would be very time
consuming.
Has this idea occurred to anyone else? Is it even feasible? Has anyone
succeeded? Does anyone have opinions on related subjects?
Thanks!
- Andy Voelkel
