WASAPI Shared mode - buffer size

Hi,
While I’m using an external audio interface for gigs without problems, I would find it really convenient if I could use the integrated realtek hd (HP EliteBook) output with WASAPI shared mode driver at home when preparing setlists / practicing / transcribing.
However I can’t seem to get an adequately small buffer size - no matter what I select as buffer size, what C3 gets is 528 samples as shown in the profiler.

From what I read here: https://docs.microsoft.com/en-us/windows-hardware/drivers/audio/low-latency-audio
" * Allow an application to discover the range of buffer sizes (i.e. periodicity values) that are supported by the audio driver of a given audio device. This makes it possible for an application to choose between the default buffer size (10ms) or a small buffer (<10ms) when opening a stream in shared mode. If an application does not specify a buffer size, then it will use the default buffer size."

It seems like smaller buffer should be possible (but maybe I’m over-simplifying). Any idea whether or not (and how?) I could get at least to 256 samples (or double 128)? Maybe with a different driver - not the realtek one? I tried with the Microsoft one but it was even worse.

What bothers me most - I could get low latency in my 2011 MacBook Pro running Boot Camp but can’t get it on the new EliteBook 840 G6…

Thanks and best regards,
Matej

Voicemeeter Banana has the ability to set buffer sizes, but it is WDM/KS, not WASAPI. Still, using that as a virtual ASIO might be fruitful. Plus, you would get both audio outputs AND inputs. *If you get a “robot voice” output, try turning Cantabile’s “power” switch off and then on again so the two programs can attempt again to sync up.

https://www.vb-audio.com/Voicemeeter/banana.htm

I made a video on setting the buffers in Voicemeeter a few years ago. Here it is:

Terry

Hi @matejb

Cantabile doesn’t support the new IAudioClient3 interfaces described by the linked article. Not sure how much is involved in adding support for that. I’ll look into it.

Brad

1 Like

Terry thanks for letting me know about the Voicemeeter.
However it is even worse - even with very high buffers there is always distortion or crackling no matter what I choose (WDM, KS, MME, …).
It seems like the Realtek driver is really messed up. I don’t even get the normal WDM choice in Cantabile, only WASAPI.
Thanks,
Matej

@brad thanks very much for your will to look into it. Not being able to simultaneously run cantabile with ~10ms latency and listen to music “to-transcribe” in 2019 is really an unexpected situation coming from mac background… don’t understand how Realtek can possibly exist with this bad drivers…
Regards,
Matej

That is why cheap ASIO USB interfaces have such a wide customer base. The newest generation of the Focusrite 2i2 is possibly your next stop. The difference is like night and day. (It really is in this case!)

Terry

Before I take a closer look at this I’m going to do some research to see if there’s now a way to do wasapi audio input/output streaming in a synchronized manner so I can enable WASAPI input in Cantabile. Then, I’ll dive in and see if I can fix both issues at the same time.

Terry thanks. Well it might be the only solution for now but it’s not a money issue really - I do have a spare audio interface - I’d just like to avoid complicating setup and cabling when it should be much simpler. I’m used to having macs that could do things like this since 2007.

Brad, thanks. Looking forward to a solution.
regards,
Matej

1 Like

I think those darn Macs spoiled you!

The rest of us (PC users, that is) take it for granted that we will require the few extra cables and a USB cord at performances. So much so that it doesn’t even register as an inconvenience. Windows audio has ALWAYS been crap, though they are slowly (too slowly, IMHO) improving it.

Terry