XXHighEnd

If your playback device you use for XXHE neither USBDAC nor SPDIF out, try to set the default format to 44100 kHz (or whatever is of your audio files).

It seems to matter in my setup.
I can hear (feel) 44 to 48 conversion when default format is set to 48.
Looks like when playback device is Speakers (not USBDAC or SPDIF) the XXHE is not in exclusive mode.
Anyway it looks like sample rate conversion is taking place as if in shared mode.
Or it may be a fault of the drivers for my sound card.

Andrey

Hi Andrey,

May it help you somewhere, I can assure you that Shared Mode (hence non-Exclusive) has been completely eliminated (since some version maybe 7-8 months ago).

Remember, this is unrelated to what you perceive hence hear. I trust that.
Peter

Peter,
Have you seen these presentations before

http://ampalliance.org/blogs/microsoft/archive/2007/08/14/Microsoft-AMP-Summit-Decks.aspx

I found this link on this blog
http://mark-dot-net.blogspot.com/2008/06/what-up-with-wasapi.html

I am curious if you use "Event Driven Audio" approach.
This approach seems the right one to use, although a driver support is required for WaveRT

Andrey

No I didn't see those presentations, but the software (of this Mark) would not work.
This is kind of logic to me, because the Even Driven feature (at least from SP0) does not work.
Few people asked questions about it back then (end of 2006) and nobody got answers.


Why do you ask ? the 1ms latency is not enough for you ? :)

Actually I feel (and should I think) this is related to the subject of this topic. But I don't see how ... :dntknw:
Do you think I may be doing something wrong ? can improve maybe ?
?

The latest ASIO4ALL implements this pull (event driven) mode for WaveRT drivers.
It looks like it works now.
I even tried it with cPlay->ASIO4All(pull mode)->Realtek WaveRT device.
And sounds very "interesting".

And I am more concerned about the way the cyclic buffer pulling method described in those presentations than about latencies 1ms.

We need to take a closer look at those ppts to discuss it further.

I looked ... :)
By now I kind of could have written them, hahaha

Hi Peter,

Maybe this is not the right place to write about it but I will try:
From this page http://www.guardian.co.uk/technology/2008/jan/31/microsoft.technology:
...Unfortunately many audio hardware vendors have not yet supplied WaveRT drivers for their products. Another weakness of WaveRT is that it only works with internal sound devices, not external devices connected by USB or Firewire....

From the above I understand that if external DACs are used (USB or FW) then there is no WaveRT support for them which then means that it is not possible to achieve maximum audio performance via USB and FW from Vista even with WASAPI deployment. If I understand correctly then WASAPI will fall back to 'compatibility mode' if there will not be WaveRT driver for the audio device.

I am currently using a DAC which is connected to my laptop via dock's SPDIF out. The laptop has IDT audio device with IDT and Microsoft drivers, so I am trying to understand if these folks under my hood are WaveRT compatible?

Any comments Peter?

Thank you

Hey Ivo,

I am trying to understand what you are talking about, but I'm afraid I just don't.
XXHE doesn't make use of WaveRT (well, as far as I know, haha), which btw doesn't mean you can be correct somewhere.

But also : you shouldn't rely too much on any expressions from MS about "new WaveRT" because I can show you a (secrectly hidden) PDF from 1990 which anounces to ditch WaveRT in favor of something new for "Longhorn".

The link you referred to only confirms what I have been saying from off 2006 : nice that new Exclusive Mode WASAPI for Vista, but no existing application like Cakewalk etc.) will be able to use it soon, because everything has to be rewritten.

In the very end the Vista drivers just must be available. Without that, bad luck.
But something falling back to 'compatibility mode' ... I never heard of that, nor can I imagine this would imply non-exclusive mode (because this is really determined by the application (XX), and not by the OS !).

Lastly, don't confuse yourself with the (Dogber) modded Realtek drivers which drivers are -in modded fashion- meant for something like DirectSound in bit perfect mode (which officially can't exist in Vista). These drivers are not meant for WDM (hence WASAPI Exclusive) at all in the first place, but make use of WaveRT (of which I claim is old technology). Look at these two phrases from that article as well :



Notice that for Vista WASAPI WaveRT has been explicitly anounced as "new".
I say it was resurected.


Right. I hope you are as confused by my text as I am with yours. :)
But maybe now you can shed a better light on what you mean ? (and I *don't* claim I know it all !)

Peter,

OK, I did a little more investigation on Vista audio and on my particular setup and:
FINDING#1:
I determined that my laptop's internal audio HW (made by IDT) does not rely on Vista's UAA HDaudio.sys driver, but rather than on its own custom WaveRT port driver. What I do not currently understand is which approach (relying on Vista's UAA driver OR design and use of a custom WaveRT port driver via PortCIs.sys) is better in terms of the latency and audio performance. Let us assume there are 2 soundcards in one PC from which one uses HDaudio.sys driver as provided by Vista and another one uses custom design WaveRT driver based on PortCIs.sys driver. So the question is: Which route for the audio signal (let us take digital one) from the application down to the output jack is better?

FINDING#2:
I know that XXHE can use WASAPI (only in Vista) and I use it by choosing Engine#3. So, what I do not currently understand:
 - How is WASAPI connected with audio HW drivers? If my audio HW utilizes a custom WaveRT driver, does it mean XXHE will actually play music via it by the help of WASAPI? Or is WASAPI something that is completely working on its own? You mentioned that XXHE does not make use of WaveRT, but I think that this may be wrong in my setup here. Maybe the attached picture will help.

Thanks and Happy New Year! :)

Ivo

Seems like I did found something on FINDING#1 here:
http://msdn.microsoft.com/en-us/library/ee220332.aspx

Cheers,
Ivo

Hi Ivo,

By now I dare say you may just be right; yesterday I prepared some 5-6 links for you, similar to your last. I just lost them because the PC got stalled, but since you look at the same matters already, I think you will have found those too.
The problem is, these pages are from the driver development perspective, and to me it doesn't say (or at least I can't find the proof of it) that this one way or the other can be influenced by the "user level" which is what WASAPI has been made for. Thus, all that stuff I don't see, and really the only thing I can see is the devices as you see it in the sound devices list, and WASAPI talks to those. What happens internall, I don't know.

Funnily I automatcially ended up by those links at getting the "device" selection to work for Kernel Streaming. This, because that higher level (the devices in the device list) are not valid for Kernel Streaming, and I really have to dive deeper for that. I can't say yet whether this enables the selection of just what you may want, but at least indirectly (by means of "filters") I think this will be the case.

Don't start to be happy yet, because thus far I could not find a way to select those filters *and* play them the way I have set it up right now. Otoh, when I won't succeed on that I won't succeed at all, because otherwise I can't let you select "devices" in the first place.
Note : I can play with KS by pure accident because the code at hand for that selects the first device, which in my system (concidentally) is the right one.

This all takes a lot of time and knowledge I don't own yet ...
:no:
Peter


PS: But keep in mind what I said elsewhere about my objective and KS : allow "you" to select all the detail there is internally which is about Filter Graphs. This is exactly your topic ...

A key page is this one : http://www.microsoft.com/whdc/device/audio/wavertport.mspx
(also look at the Word document at the right).

By now I had already forgotten about this, but note this is still all transparent to me. I am (WASAPI is) just talking to the other side of the interface ...
This doesn't mean though I could go more directly (as it seems), especially seeing that (indeed) under the hood KS parts are being used.

Hi Peter,

Thanks for answers.
I am starting to think that either way be it HDaudio.sys or PortCIs.sys being talked to by WASAPI produces the same effect in terms of latencies and distance from app to hw. Therefore I will stop for now looking more deeply as it is more or less clear to me.
The only thing that interests me is if Windows 7 is somehow better than Vista regarding what we talked about here?

Regards,
Ivo

Hi Ivo,

I am not sure what to think of it, but things can be looked at a kind oif other way around :

The modern WaveRT drivers (like used in HDAudio !) don't expose anymore to Kernel Streaming (or they are not supposed to anyway). So, I am not sure what this tells, apart from devices like HDAudio not being able to be used with Kernel Streaming.
I know this is not your subject, but generally this is important (well, for me it is and in the end for you, the users).

Ok, we'll let this rest for now.
Thank you.
Peter