Cannot get native DSD on Kernel Streaming mode

Great job !
But I have a Nuprime IDA16 Amp and even when I set “Native DSD ” on Kernel streaming Mode, I can’t get DSD on my amp, only 176.4 kHz.
Like 3.5, I have to use ASIO mode to get DSD to work.

To get native DSD playback you need to use ASIO. In KS you will be able to play DSD track in DoP.

1 Like

Hi, @Damien3 Damien!

Is there is any audio improvement using ks instead of asio?



Hi @antoine and colleagues, is it possible to have full native DSD in Kernel streaming in any shape or form, or KS only accepts DoP forever and that’s it?

Also I don’t understand what’s the problem. Is DoP such bad in terms of audio quality, things that can be percieved by a normal ear? Anyway, for me with actual equipement ASIO 24 bit sounds best, just curious.

DoP (DSD over PCM) = FULL DSD.
The only difference is that it is transported in a PCM container. When it reaches the DAC, the DAC recognizes it is DSD disguised in a PCM and will play it as native DSD.

It goes too far to technically explain this all in this forum, but suffice to say that DoP’s endresult is exactly the same as full DSD. Here is a link to a Youtube video that explains how it works.

DoP has to be used with WASAPI and Kernel streaming. That is not a limitation of Audirvana but a technical limitation of the computer. Mac and Windows both simply do not understand native DSD and both have to be ‘tricked’ into playing DSD. There are of course differences in how this is done on a Mac or on Windows, but that is a subject that goes too far in this post.

Again in short (I have the habit of repeating myself :grin:):
There is no difference between DoP and full DSD (except in the way they are transported). The endresult with both is exactly the same: full DSD. Lots of people have the misconception that DoP is a conversion to PCM and degrades the quality but that is simply not the case.

If you want to know more technical details: Google is your friend :grinning:.

1 Like

Hi @EJ51,
If you activate DoP (DSD over PCM) in the Audirvana settings when using Wasapi or Kernel streaming mode you are fine.

Maybe I am a bit superfluous here (because most of it I already explained in my post above this one). Windows can not recognize native DSD when in Wasapi and Kernel streaming mode. That can only be bypassed via an ASIO driver.
Audirvana can not do anything about that, it is a limitation of the OS. (No need to buy a Mac instead, because MacOS has exactly the same limitation :grin:).
DoP is a trick to fool Windows that it plays PCM, but the endresult to the DAC is still native DSD.
In short DoP is basically native DSD disguised as PCM to fool the computer, but the endresult is also native DSD and sounds exactly the same as native DSD.

So, rest assured: Nobody can here the difference between DoP and ‘native’ DSD because both have native DSD as endresult (only the way of transporting the data differs).


Thanks @AndyLubke, learned something new today. Sometimes I’m too lazy to Google :blush:

Thank you Andy, much appreciated!

But what I do notice is that, although my DAC recognizes the DSD stream, Audirvana doesn’t send it at DSD512 (max spec of my DAC), but rather DSD256, which indicates me that DSD over DoP loses resolution; right?

If you on a mac… mac cant do dsd512

Are you playing DSD512 files or is Audirvana converting the stream to DSD?

I believe that DoP has a maximum of DSD256 (on Mac and Windows). If you want to do DSD512 you need an ASIO driver (in Windows). On a Mac you can not do DSD512 at all (as far as I am aware, I am a Windows user myself).

Damien is in the list of contributors.
He will know roughly how it works.

welcome to the lounge and take a pint.

@Damien, @Antoine (and Audirvana colleagues), will it be possible for AS to do DoP at DSD512? I can see from the documentation above that it is possible, just not implemented in Audirvana yet.

The primary objective would be to stretch r8brain upsampling to beyond DSD256 - indeed current implementation of DoP does impact max possible quality as the DAC can play DSD512 but upsampling is limited to DSD512 with KS.

Thank you all for the info sharing; it helps us all grow. If you need an extra pair of hands to help with development or beta-test, I’m available.

Cheers, Ed

Do you have a DSD 1024 device? because it’s the only way to do this.

Hi @Antoine, my device maxes at DSD512, PCM 768 KHz. Isn’t it enough?

Thanks for the quick reply!

No it’s not, in WASAPI or Kernel Streaming, you can’t send raw DSD streams. That’s why the DSD over PCM standard has been developed: to send raw DSD in a fake PCM stream.
To identify it as such, there are 8 bit markers in it.
And the presence of these markers doubles the need of bandwidth, thus explaining why on a mac only half of the max DSD rate is available.

That’s why if you have a DSD 1024 device, you should be able to do DSD 512 in KS.


I now fully understand! Thank you for this!

Hey @Antoine, do you think that in the future DSD could be natively supported (say in Win11 and/or a new USB4 protocol, thus eliminating the need for DoP in KS?

ASIO supports DSD.
DSD is a niche product. I would not expect it to be included in new protocols.