- Voicemeeter Banana allows you to separate and control microphone, games, music and chat with A/B buses.
- Configuring 48 kHz and WDM drivers improves stability and reduces live latency.
- Integration with OBS/Streamlabs via B1/B2 and multi-tracking facilitates VOD without music.
¿How to set up Voicemeeter Banana for streaming? If you use Twitch or YouTube and want clear, flexible, and professional audio, Voicemeeter Banana is a true Swiss Army knife. It lets you separate sources (microphone, game, music, chat), apply compression and EQ, and send separate mixes to your live stream and your headset. Still, many guides fall short or overlook key steps, which is why common questions arise when getting started.
The community has tried to fill that gap with videos and collaborative contributionsBut here you'll find a thorough walkthrough of Spain in Spanish to get everything set up with meaning: from installation and drivers to routing, effects, and OBS/Streamlabs integration. The idea is that you end up with a stable setup, with low latency and complete control over what your audience hears and what you hear.
Requirements, installation and basic settings in Windows
Before touching anything on the mixer, it is a good idea to prepare the system. Download Voicemeeter Banana from the VB-Audio website and also install the associated virtual drivers (Voicemeeter VAIO and AUX are installed with Banana itself). Restart the PC afterward; many strange errors (such as problems with the process audiodg.exe) disappear with a simple reboot after installation.
Go to Windows Sound and define a common base. Set to 48 kHz (48000 Hz) as the sample rate for microphones and main outputs; it's a convenient standard for streaming and reduces desync issues with OBS. If you already have projects or devices at 44.1 kHz, unify everything anyway to avoid constant resampling.
In Advanced Device Properties, check two boxes: exclusive mode (sometimes it's a good idea to enable it to avoid conflicts with apps) and default quality (choose 16 or 24 bits depending on your hardware, with 48 kHz). The important thing is that all critical elements (microphone, VAIO VoiceMeeter/AUX, your audio interface if you have one) share the same format.
Open Voicemeeter Banana. In the top right area, you'll see A1, A2, and A3. Select your main physical output in A1 (your headset or interface). Choose the WDM-priority driver, which offers lower latency. MME is more compatible but slower; KS is very straightforward but can be finicky depending on your hardware. Start with WDM and only change if you notice crackling or instability.
On the input strips, Banana offers three hardware channels (Hardware Input 1-3) and two virtual channels (Voicemeeter VAIO and Voicemeeter AUX). Assign your physical microphone to Hardware Input 1 with its WDM driver whenever possible. This way, you'll have independent control over EQ, compression, and bus output.
The A1/A2/A3/B1/B2 buttons under each strip are the sending matrix. Think of A as outputs to your ears (physical A1-A3) and B as virtual outputs to programs (B1 and B2). A practical rule of thumb: the microphone goes to B1 (for OBS), and to A1 if you want to hear yourself monitored. If you don't like hearing yourself, turn off A1 on the microphone to prevent feedback.
By default, Windows may continue to send system audio to your direct headphones, bypassing Voice meeter. Change the default output device in Sound to "Voicemeeter Input (VB-Audio Voicemeeter VAIO)". This will input all PC audio into the VAIO virtual strip, which you can control from Banana. Optionally, use AUX to separate music or chat.
In Windows Advanced Sound Settings (volume per app), route specific apps to VAIO or AUXFor example: games to VAIO and Spotify to AUX. This way, you can mute or lower your music without affecting the game or chat audio during the stream.
Once you have assembled the basic circuit, open the Menu in Voicemeeter and check the buffer if you notice any micro-cuts. Increase the buffer size from 256 to 384/512 If your CPU is tight or your interface is capricious. The smaller the buffer, the lower the latency and the greater the demand on your computer; find the stable point for your situation.
Tickets, buses and routing for live performances
Organizing your routes well is what will give you real control. Design your mixes with two destinations in mind: what your audience hears (B1/B2 to OBS) and what you hear (A1/A2/A3). This allows you, for example, to raise your voice a little through headphones without altering the output.
For a typical setup: Mic on Hardware Input 1, Games on VAIO, Music on AUX, Chat (Discord) also on AUX or another channel if you split it. Send the microphone to B1 (for OBS to capture) to A1 only if you want monitoring. Games to B1 and A1, music to B1 and A1, and chat to B1/A1. If you record VODs and don't want claims for music, you can send the music to B2 and not capture B2 on the VOD track.
Voicemeeter Banana incorporates a compressor, gate, limiter and the Intellipan module per channel. The gate helps reduce background noise cutting the microphone below a certain threshold. Reasonable starting values: Gate between -40 and -35 dB, being careful not to swallow syllables. Adjust by speaking softly and normally until it only closes when you're truly silent.
The compressor controls peaks and makes your voice more consistent. Start with a 3:1 or 4:1 ratio, threshold around -12 to -18 dB depending on your mic and gain, and compensate with Makeup to recover level. Avoid over-compressing to avoid sounding squashed or noisy.
Intellipan is a fast character shaping wheel: you can add brightness or warmth without going into the detailed EQ. Use it sparingly; if you need fine control, activate the strip's 6-band EQ and apply a low-cut between 70 and 90 Hz, a slight presence boost between 2.5–4 kHz (+1 to +3 dB), and, if your mic is shrill, a soft cut between 6–8 kHz.
In the MASTER section, set output limiters/comps if necessary. A soft limiter on B1 You can prevent clips if you scream or the game crashes. Remember that OBS can also have filters; decide whether you prefer to handle them in Voicemeeter, in OBS, or share the work (but don't duplicate aggressive compressors in both places).
If you're using music, it's a good idea to enable ducking so it quiets down when you talk. Voicemeeter doesn't have a classic sidechain, but You can simulate it with the music channel compressor. feeding your voice through the K knob and routing it to buses, or ducking in OBS with the sidechain compressor filter pointed at the microphone. The important thing: always keep your voice first.
A useful trick is to clearly separate the mix that goes to the stream from the one you hear. If the return distracts youRemove the microphone from A1 but keep it on B1. If your latency allows it and you want to hear yourself, leave it on A1 very softly. Remember that singing or voiceovers with high latency can be uncomfortable; if you notice an echo, it's best to turn off that feedback.
Finally, document your routing. Create a small outline of what goes to B1/B2 and A1It will save you headaches when you add new fonts or change something in Windows and everything seems to move around.
- Quick recommendation: WDM on inputs/outputs, 48 kHz, music on AUX, games on VAIO, mic to Hardware 1, stream through B1.
- Target levels: peak voice at -6 dBFS, overall mix max -3 dBFS, nothing in red.
- Subtle EQ: Bass cut and a touch of presence; less is more.
OBS/Streamlabs integration, testing, and troubleshooting
Open OBS (or Streamlabs Desktop) and go to Settings > Audio. Set the microphone device to "Voicemeeter Output (B1)" If you've decided to use B1 as your main mic mix, if you want two separate buses (for example, one with music and one without), add a second device with "Voicemeeter AUX Output (B2)" and divide your Voicemeeter sources accordingly.
Another option is to not use the global “Mic/Aux” and instead add "Audio Capture Device" by source in the scene (one pointing to B1 and one to B2). This allows you to mute or record tracks separately with more control from the OBS mixer.
In Settings > OBS Output, turn on Multiple Audio Tracks for recording. Assign your voice and essential effects to track 1, gameplay on 2, music on 3, chat on 4, etc. For Twitch, use the “VOD Track” feature and leave out the music if you want to avoid claims: send the music only to B2 and do not include B2 in the VOD track.
Test the synchronization. If you notice that the audio is ahead or behind the image, in OBS you can adjust the sync delay in ms in the advanced properties of the audio device. Start with 0 ms; if your capture device or network introduces video lag, add 100–200 ms as needed to match your lips and voice.
When you hear clicks or cuts, check your buffers. Increase the buffer in Voicemeeter and/or your interface and close applications that block audio in exclusive mode (DAWs, WebRTC-enabled browsers, etc.). Switching from WDM to MME can provide additional stability, at the cost of latency.
A key step is to lock in levels before going live. Speak like you're on stream, play a game, put on some background music, and check the meters. Make sure your voice never clips and the game doesn't cover up the voiceover. If everything sounds low in the stream, increase the output gain on B1 or the corresponding channel fader, not the headphone fader.
To quickly manage your live mix, use hotkeys. Voicemeeter MacroButtons It lets you create buttons to mute music, activate chat push-to-talk, or toggle EQ profiles. Assign convenient hotkeys and add them to your stream deck if you have one.
If you're using Discord, select "Voicemeeter AUX Input" as your speaker in Voice Settings to send it to the AUX channel and separate it from the game (if you're experiencing dropouts or crashes, see how to fix them). Disable aggressive noise suppression if it generates artifacts, since your own gate/compressor in Banana can do it better. Also, check that Discord doesn't change your default device when you restart.
Regarding music, services like Spotify can be moved from device to device after updates. Check the routing in the volume mixer per application Windows whenever you notice that the music is not responding to the AUX fader (for example Spotify). It is common after reboots or after connecting/disconnecting USB headphones.
Good maintenance practices: save presets. In the Voicemeeter menu, export your settings to a file to return to a stable state if something goes wrong. Create one copy with and without music, another for interviews (dual mic), and another for competitive gaming (less latency, fewer effects).
- Buffers and latency: set to 256–384 samples for live; turn up if you hear pops.
- Consistent formatting: all at 48 kHz where possible to avoid drift.
- Ducking the microphone over music: Do it in OBS with sidechain compressor if it is more comfortable for you.
What if you're using a console capture card? Assign its audio to a separate channel in OBS, or route it to a Voicemeeter strip If it's coming in through your PC's sound card, keep the total latency low so you don't notice any delay between what you see on the capture card monitor and what you hear.
When you want to take your sound to the next level, play with the master EQ or subtle effects. Small adjustments give big results: a +1–2 dB high-shelf over 8–10 kHz for air, a 200–300 Hz cut if your room adds snare drum, and a soft limiter to protect the audience from unexpected peaks.
If at any time Windows “loses” the devices, reinstall the VB-Audio drivers and reboot. Avoid changing default devices during a stream; perform these tests off-air. It's also a good idea to disable Windows audio enhancements and manufacturer utilities that interfere with the stream.
Finally, spend 10 minutes setting up a test scene in OBS with a local recording. Record 1–2 minutes of talking, with game and music, and check with headphones. Look at the meters, see if there's any compressor pumping, and adjust until you're happy. Going live with confidence in the mix is half the job done.
With this Voicemeeter Banana setup You'll have clear paths, useful effects, and fine control over what your audience hears and what you hear. The key is to unify the format at 48 kHz, choose WDM for low latency, separate sources in VAIO/AUX, use B1/B2 wisely, and spend time testing levels. Once you get it right, your stream will sound cleaner, more balanced, and more professional, and you can focus on creating content without fighting over the audio.
Passionate about technology since he was little. I love being up to date in the sector and, above all, communicating it. That is why I have been dedicated to communication on technology and video game websites for many years. You can find me writing about Android, Windows, MacOS, iOS, Nintendo or any other related topic that comes to mind.