Personally, as a template for asynchronous IOCTL integration, I used
https://github.com/microsoft/Windows-driver-samples/tree/master/general/event
This sample provides examples for using both the IRP-Pending method as well
as the Event based method.
- Nathan
On Tue, Apr 28, 2020 at 3:29 AM Giacomo Costantini <
giacomo.costantini3@xxxxxxxxx> wrote:
Hi Tim,
thank you very much for the prompt response.
Is there really any point to this? If you do web searches on this topic,
there are a many technical people who argue that this is a silly thing to
do. There is no spatialization in headphones. It's just two speakers.
One can assume that the applications writers that are PRODUCING the audio
are in a much better position to decide how to reduce the output. It's
better to let them produce 7.1-specific data and 5.1-specific data and
stereo-specific data, because they know what they want.
The intent of our product is to produce adaptive stereo for loudspeakers,
not for headphones. But I suppose the fact of having stereo -> stereo or
5.1 -> stereo is just a matter of configuring the audio filter and handling
the streams in my DSP, so it’s not that relevant for the other questions.
Their driver exposes private ioctls that are used by the application to
access a circular buffer of data. I've done two solutions that use this
same technique.
That starts to make some sense, thanks for the info.
I am not very familiar with IOCTLs on Windows, so if there is any code or
more specific documentation you can share or point me to, that would be of
great help.
Thanks again,
Giacomo
On 27 Apr 2020, at 20:29, Tim Roberts <timr@xxxxxxxxx> wrote:
Giacomo Costantini wrote:
I would like to implement a Virtual Audio Device which processes stereo,
5.1 or 7.1 streams, performs some spatialisation audio processing, and
outputs stereo.
Is there really any point to this? If you do web searches on this topic,
there are a many technical people who argue that this is a silly thing to
do. There is no spatialization in headphones. It's just two speakers.
One can assume that the applications writers that are PRODUCING the audio
are in a much better position to decide how to reduce the output. It's
better to let them produce 7.1-specific data and 5.1-specific data and
stereo-specific data, because they know what they want.
What is not totally clear to me is how to get the output stream from
SYSVAD and re-route it to available stereo speakers.
I understand I will have to write a user application for that, but I would
really appreciate if someone could point me to what specific
API/architecture/technology/code should I use for that.
I have found
https://docs.microsoft.com/de-de/windows/win32/coreaudio/stream-management,
but am I right in thinking that using those APIs I should get the stream
from the loopback pin and reroute it to some speaker device?
If your VAD supports a loopback pin, that may be possible. Otherwise,
you'll end up with what I'll present below.
And they have an app called "WavesLoop Application" which is constantly
running and apparently provides a list of stereo output devices and
performs the operations on the streams in user mode.
What I do not understand is how they manage to get the kernel buffer
stream from the LINEOUT endpoint without apparently having any loopback.
Their driver exposes private ioctls that are used by the application to
access a circular buffer of data. I've done two solutions that use this
same technique.
--
Tim Roberts, timr@xxxxxxxxx
Providenza & Boekelheide, Inc.
