I've had the idea of applying AI noise reduction and other VST plug-ins to my microphone for some time, but never found a good guide on how to do it. After trying several setups (VoiceMeeter Banana + VB-Audio Virtual Cable, JACK + VB-Audio Virtual Cable, Equalizer APO), all of which failed due to latency, crackling noise, unsupported plugins or other instabilities, I have now found the perfect setup!

Advantages

• Supports any VST plugin as far as I can tell (Equalizer APO only has very limited support)
• Doesn’t rely on abusing the Windows audio API in an unstable way that introduces crackling
• The processed audio can be used anywhere (Discord, Teams, OBS, …)
• Low latency

Disadvantages

• You might need to buy stuff because of hardware requirements

Concept

You need a special audio interface with support for loopback (I use the Arturia MiniFuse 1).

1. Your microphone will go into the VST host Cantabile Lite. The host will apply all the audio effects of your choice and then plays the processed audio back on the loopback speaker.
2. The loopback speaker will be redirected to the loopback microphone by the hardware of the audio interface. This prevents tons of issues you can get if you try to redirect audio with software like VB-Audio Virtual Cable.
3. Now you can just use the loopback microphone like any other microphone in Discord and other programs.

To start the setup you just run Cantabile. That’s it, Cantabile will automatically open your last project file and it can be minimized to tray.

Setting up Cantabile

Select the ASIO driver of your audio interface:

If the loopback is not showing you need to set it up manually here. You can rename your loopback devices in the Windows audio settings in case they are both named identically.