Modartt user forum - Always Use ASIO

Re: Always Use ASIO

null@example.com (EvilDragon) — Fri, 07 Mar 2014 19:40:22 +0000

Reaper's WASAPI works great here, both on my desktop and my laptop.

Re: Always Use ASIO

null@example.com (Skuto) — Fri, 07 Mar 2014 18:19:41 +0000

That just uses the old Win32 Audio APIs, not WASAPI.

I think Modartt has said that you can use a VST host with WASAPI support, but obviously this is more burdensome than native support, and for example I found Reapers WASAPI support to be spotty.

Re: Always Use ASIO

null@example.com (EvilDragon) — Fri, 07 Mar 2014 17:07:31 +0000

Pianoteq 4 already supports WASAPI, I think... select Windows Audio in "Audio device type".

Re: Always Use ASIO

null@example.com (Skuto) — Fri, 07 Mar 2014 16:50:20 +0000

What you're saying is right EvilDragon, but note than it my test these additional latencies were constant, so it's also completely irrelevant. The displayed DirectSound latency is simply a lie.

I wonder if Pianoteq 5 will support WASAPI. The latency there has improved in recent Windows version (8.1 even improves over 8.0) and MS seems determined to make it usable for music. ASIO works but it's not perfect either.

Re: Always Use ASIO

null@example.com (Luc Henrion) — Fri, 07 Mar 2014 16:08:46 +0000

but the conclusion is the same: always use ASIO :-)

Re: Always Use ASIO

null@example.com (EvilDragon) — Fri, 07 Mar 2014 14:48:16 +0000

http://www.presonus.com/community/Learn...io-Latency

Audio drivers can never report the total round-trip latency, no matter if it's ASIO or DS or whatever. There are other factors that need to be accounted for, such as USB bus latency and AD/DA conversion time.

Always Use ASIO

null@example.com (Skuto) — Fri, 07 Mar 2014 12:57:32 +0000

I spent about half a day tracking down the cause of a latency issue with a new setup here.

To make a long story short: the buffer size shown when DirectSound is selected as audio output is a lie. The driver and device can and do add much more hidden buffering beyond that.

Here's a recording with 3 parts, playing a keyboard with Pianoteq linked up through a fast MIDI cable.

The first is DirectSound with 192 samples@48kHz buffer. The second is ASIO4ALL with the same buffer size. The last is ASIO4ALL with 64 samples.

https://mega.co.nz/#!6xwzARTA!13eWGJAAc...jzm5Gjf1h4

The difference between 192 and 64 isn't audible to me. The difference between DirectSound and ASIO on the other hand, is gigantic and way more than the claimed 4ms. This happens even with all DirectSound processing disabled and everything forced to 48kHz