A voice changer for streamers and content creators on Windows

What streamers actually need from a voice changer

A voice changer for a stream has a harder job than one for a casual call. It has to sound consistent for hours, survive a game that is already using the graphics card, give you instant control without alt-tabbing, and fit into the way your streaming software routes audio. Most of the frustration people have with voice changers on stream comes from one of these four things: glitches under load, too much lag, awkward controls and a messy audio setup.

Mimicore is designed around those four. It uses your best graphics card by default (or the CPU if you have none) and lets you adjust trained-voice latency and quality with one slider. Controls are hotkeys, so you change voice, mute or start the effect without leaving the game. And the audio is delivered through a standard virtual microphone, so it appears in OBS or any other tool like a normal microphone.

Setting up Mimicore with OBS

First install the virtual microphone from inside the app: Live › Output › Set up virtual mic. Mimicore downloads the free VB-CABLE driver from vb-audio.com, checks its signature and installs it, and you may need to restart Windows once. Afterwards a device named CABLE Output (VB-Audio Virtual Cable) exists on your PC.

In OBS, add an Audio Input Capture source and choose CABLE Output as the device. That source now carries your changed voice. If you also want your friends in Discord to hear it, set CABLE Output as the input device there as well and turn off Discord's own noise suppression and echo cancellation, because they fight a changed voice. In the Mimicore Live page, pick the voice, press Start, and turn on Hear myself with headphones, so you can monitor yourself without the sound leaking back into the microphone.

If you only want the changed voice on stream and not your real voice, make sure your normal microphone is not also added as a source in OBS. Doubling up is the most common reason people hear two voices.

Delay, lip sync and what to do about it

Voice conversion buffers and processes speech, which adds delay. Adjust trained voices with the latency and quality slider and measure the whole microphone-to-output path on your PC. The Live panel gives a local estimate; chat apps and networks add their own delay. Effects need much less processing.

There are two ordinary fixes. The first is to delay your webcam to match: in OBS you can add a Render Delay filter to the video source and set it to roughly the delay of your voice. The second is to avoid the question by using an avatar or a stylised overlay that does not show your real mouth. Many streamers who use character voices do exactly that. Either way, test with a short recording before you go live and adjust the delay until it feels right.

Trained voices have one Latency and quality slider with 15 positions, from about 80 ms to 1.2 s of buffering. Shorter buffering needs more frequent processing with less context, which can reduce voice quality; longer buffering favours smoothness and quality but adds delay you will hear. Stop Live before moving the slider, keep it toward the left for calls and games, and check that processing time stays inside the block on your own PC.