WASAPI vs ASIO

The short answer

WASAPI (Windows Audio Session API) is the sound system already built into Windows. Every app that plays or records audio the normal way goes through it. ASIO (Audio Stream Input/Output) is a separate standard, created by Steinberg, that many audio interfaces install alongside their driver to give a program a short, direct path to the hardware.

There is no absolute winner. WASAPI is universal and needs no extra install; ASIO can reach a lower, steadier latency on hardware that supports it well. Which one feels better on your machine comes down to your interface, its drivers, and whether you need to share audio with other apps or hand one app exclusive control.

WASAPI: the path built into Windows

Because WASAPI ships with Windows, it works with everything, no download required. It runs in two modes, and the difference between them is the whole story on latency.

• Shared mode. The default. Windows mixes your app in with every other sound on the system through a shared engine, so a browser, a game, and a call can all make noise at once. That convenience adds a little delay, because everything passes through the system mixer on the way out.
• Exclusive mode. One app takes direct control of the device and bypasses the shared mixer. Nothing else can use that device while it holds the lock, and in return the audio gets a much shorter, tighter path with lower latency.

So WASAPI is not one thing. WASAPI Shared is the friendly, everything-at- once default, and WASAPI Exclusive is the get-out-of-the-way, low-latency option, at the cost of that device being yours alone for the moment.

ASIO: a separate low-latency standard

ASIO takes a different route entirely. Instead of using the Windows audio system, an ASIO driver talks almost straight to the hardware, skipping the operating system's mixing layer. That is why studio interfaces and recording software lean on it: fewer stops between the app and the converters usually means less delay and rock-steady timing.

The catch is that ASIO is not part of Windows. It depends on a driver the hardware maker provides, and, much like WASAPI Exclusive, an ASIO device is typically claimed by a single app at a time. When the hardware ships a good ASIO driver it can be excellent; when it does not, you fall back to WASAPI. For the full picture, see what is ASIO.

Comparing them on latency, exclusivity, and compatibility

Line the three paths up and the tradeoffs are easy to see. The more the signal has to route through the shared Windows mixer, the more latency it picks up; the more directly it reaches the hardware, the lower the delay but the fewer apps can share the device.

• Latency. WASAPI Shared adds the most, since it passes through the system mixer. WASAPI Exclusive and ASIO both cut most of that out and land in the low-latency range. See what is monitoring latency for why that delay matters when you hear yourself.
• Exclusivity. Shared mode lets every app play at once. Exclusive mode and ASIO hand one app sole control of the device, which is what buys the lower latency but blocks sharing.
• Compatibility. WASAPI is on every Windows PC with no install. ASIO only exists if your interface (or a generic ASIO layer) provides a driver, so it is powerful but not guaranteed.

Which one is actually lowest latency?

The honest answer is that it depends on the hardware. On an interface with a well-written ASIO driver, ASIO often reaches the lowest, most consistent latency. But WASAPI Exclusive is genuinely close on modern Windows, and on hardware whose ASIO driver is mediocre, Exclusive mode can match or beat it. The real lever in both cases is your audio buffer size: a smaller buffer means lower latency but more strain on the CPU, and that tuning matters far more than the WASAPI-versus-ASIO label alone.

The practical takeaway: do not assume one badge is faster in the abstract. Try both paths on your own gear at a sensible buffer, listen, and keep whichever stays clean without crackling.

patchd runs over both and locks onto whichever wins

patchd is a Windows virtual audio mixer that runs its real-time engine over both ASIO and WASAPI Exclusive. Rather than making you pick, the engine locks onto the fastest path your setup offers automatically, ASIO first and WASAPI Exclusive otherwise, and reports what it locked to in its master clock readout, so the low-latency route is in use no matter which one your interface prefers.

In practice that keeps the round trip short. Through a bus, the engine adds about 10.7 ms at a 512 buffer, and smaller ASIO buffers take it lower, which is the kind of delay you stop noticing. That is the clean-signal path: your mic runs through a channel's effects rack and leaves on a bus other apps pick up as a microphone, all in real time. (A heavier voice changer using a full AI voice is a different tradeoff, adding roughly 350 ms, which is why you would not reach for it on a live-monitored path.)

How to choose for your setup

Start simple. If you just want everything to work and share sound freely, WASAPI Shared is fine and you may never think about it. The moment latency matters, hearing yourself in a headset, tracking to a beat, running a live chain, move to a low-latency path: WASAPI Exclusive if your hardware is happiest there, ASIO if your interface ships a solid driver.

Then tune the buffer down until you hit the sweet spot between low delay and stable, crackle-free sound. patchd is pre-launch, so if you want a mixer that speaks both ASIO and WASAPI Exclusive and lets you route audio per app in real time, join the waitlist and get notified when it ships.