Short answer
Before you start
- The correct microphone selected in OBS
- Headphones for monitoring without feedback
- A test recording with normal and loud speech
Bring your setup to a 24/7 AI Streamer Coach
Ask Streamer Coach →Get personalized feedback to fix streaming issues and plan your next content.
Step by step
- 01
Confirm the microphone and gain
Watch this step on YouTube · 0:32 ↗Choose the correct device under Settings > Audio. Check the physical control and Windows input gain before using filters to compensate for a low signal.
Speak from a consistent distance and check the correct side of the microphone. If the meter does not react, fix the input before opening a filter.
What you should see: Normal speech moves the meter without remaining in the red and the correct device receives signal.
The first check is selecting the correct device before adding filters. YouTube · 0:53 ↗ - 02
Turn on monitoring
Watch this step on YouTube · 3:04 ↗In Advanced Audio Properties, choose Monitor and Output. Select your headphones as the monitoring device so you can hear what each filter actually changes.
The tutorial frame shows the monitoring device under Settings > Audio > Advanced. Listen to the same sentence and turn monitoring off if it creates echo or delay.
What you should see: You hear the microphone through the selected headphones without continuous feedback.
The tutorial frame shows where to choose the monitoring device in Audio settings. YouTube · 4:10 ↗ - 03
Apply noise suppression
Watch this step on YouTube · 4:38 ↗Use Nvidia Broadcast if your computer supports it. Otherwise compare Speex and RNNoise. Keep the option that removes the background without clipping words or overloading the CPU.
The tutorial also shows NVIDIA Audio Effects. RNNoise and Speex are separate OBS options. Compare a complete sentence and watch CPU use.
What you should see: Background noise falls without clipped syllables or a robotic voice.
This frame shows the NVIDIA Audio Effects options used in the tutorial. RNNoise and Speex are separate choices. YouTube · 6:10 ↗ - 04
Shape the tone with EQ
Watch this step on YouTube · 6:39 ↗Adjust high, mid, and low bands in small moves. Record the same sentence before and after, then judge with your ears instead of the curve.
Save a recording before the change. Cut one band carefully and avoid using an extreme curve to compensate for poor microphone placement.
What you should see: The new recording sounds clearer than the reference without a harsh change in tone.
- 05
Control low noise and loud peaks
Watch this step on YouTube · 7:40 ↗The expander lowers keyboard and room noise when you are quiet. The compressor brings loud and soft speech closer together for a steadier voice.
Make keyboard noise, whisper, and speak loudly. The expander should lower the room while quiet and the compressor should soften peaks without flattening the whole voice.
What you should see: Room noise falls while quiet and loud speech keeps a natural difference from a whisper.
- 06
Finish with a limiter
Watch this step on YouTube · 10:15 ↗Put the limiter last to catch peaks. Nicole uses -10 dB in her example and checks that normal speech sits between -15 and -10 dB.
The verified frame at 11:10 shows Limiter selected at the end of the chain. Change one control at a time and repeat the same sentence.
What you should see: The limiter sits last and catches peaks during the loud-voice test.
The verified frame shows Limiter selected at the end of the tutorial chain. YouTube · 11:10 ↗
If something does not work
- The voice sounds robotic
- Lower the suppression amount or try another option. Listen to a complete sentence before deciding.
- The keyboard remains audible
- Check position, expander, and gain. Keep a consistent microphone distance before asking the filter to do more.
- The voice reaches red
- Lower gain before increasing compression. Keep the limiter as the final protection.
Test before you go live
- The correct input is selected
- Monitoring plays through headphones
- Noise suppression does not clip words
- Normal speech sits near -15 to -10 dB
- The limiter is last in the chain
Cómo mejorar tu micrófono en OBS con filtros de audio gratis
Watch the full tutorial ↗Common questions
What order should OBS microphone filters use?
Nicole's final chain is noise suppression, EQ, expander, compressor, and limiter. Order matters because each filter receives the output of the one before it.
What level should voice reach in OBS?
In the video example, Nicole adjusts normal speech to sit around -15 to -10 dB and places a limiter at -10 dB.
Should I copy the exact values from the video?
Use them as starting points. The room, distance, and microphone change the result. Decide from a complete recording.
Which filter goes last?
The limiter. OBS describes it as peak protection and recommends placing it at the end of the chain.
