IMHO it is a more straightforward approach, using a double buffer that
is shared between kernel and user space, and events that notify when to
push/pull data to/from the buffer. This means that there is no audio
data transfer using IRPs.
Regards
/R
Den 2015-11-06 kl. 04:04, skrev Paul Titchener:
In terms of additional features and performance, what does the WaveRT port method provide over the Wavecyclic method?
*From:* Cheng-mean Liu (SOCCER) <mailto:soccerl@xxxxxxxxxxxxx>
*Sent:* Thursday, November 5, 2015 6:39 PM
*To:* wdmaudiodev@xxxxxxxxxxxxx <mailto:wdmaudiodev@xxxxxxxxxxxxx>
*Subject:* [wdmaudiodev] Re: simple virtual audio driver
MSVAD should work.
The only downside is : the current MSVAD sample uses Wavecyclic port but some of the Windows Audio Stack enhancement Microsoft did over the last few OS releases only benefit those audio drivers using WaveRT port.
*From:*wdmaudiodev-bounce@xxxxxxxxxxxxx [mailto:wdmaudiodev-bounce@xxxxxxxxxxxxx] *On Behalf Of *Tim Roberts
*Sent:* Thursday, November 5, 2015 6:22 PM
*To:* wdmaudiodev@xxxxxxxxxxxxx
*Subject:* [wdmaudiodev] Re: simple virtual audio driver
Paul Titchener wrote:
I guess one problem with a “simple” driver is there seems to be
differences of opinion in what should be in that simple driver.
Is there any disadvantage in using the MSVAD sample as the
starting point, ie since it was written a while ago is it using
dated approaches that could be done more efficiently or effectively?
Efficiency is irrelevant for the vast majority of the driver -- most of it is only used for setup. It's the data transfer stuff that matters, and that's pretty streamlined.
--
Tim Roberts,timr@xxxxxxxxx <mailto:timr@xxxxxxxxx>
Providenza & Boekelheide, Inc.
