Tutorials Streaming

Livestream Audio That Doesn't Suck

Beginner · ~30 min

Overview

Viewers forgive rough video far more than rough audio. This guide builds a proper audio filter chain directly inside OBS. The same gain staging, gating, and compression concepts from podcast production, applied to a live signal with zero post-production pass available.

What You Need

OBS Studio with your microphone already added as a source
A USB or XLR microphone (built-in laptop mics will always sound rougher, filters or not)
Headphones for monitoring while you set levels

Steps

Set your input gain first

Before adding any filters, get your raw mic level right. Talk at normal volume and watch the Audio Mixer meter. It should peak around −12 dB to −6 dB. Too hot and filters downstream will clip. Too quiet and noise gates will fight you.

Add a Noise Gate

Right-click your mic source → Filters → + → Noise Gate. Set the "Close Threshold" a few dB below your speaking level and the "Open Threshold" a few dB above your room's background noise. This mutes the mic between sentences so keyboard clatter and fan noise don't leak into every pause.

Add Noise Suppression

Add a Noise Suppression filter above the gate, set to RNNoise (the AI-based method) if your CPU allows it. It's noticeably better than the older Speex method at removing fan hum and light room noise without a robotic side effect.

Add a Compressor

Add a Compressor filter with a ratio around 3:1 to 4:1, threshold around −18 dB, attack 5 to 10 ms, release 60 to 100 ms. This narrows the gap between your quiet and loud words so viewers don't need to keep adjusting their volume.

Add a Limiter as a safety net

Add a Limiter filter last in the chain, threshold around −4 dB. This is your insurance policy against a sudden shout or laugh clipping the stream. It should almost never engage in normal speech.

Pro Tips

Filter order matters: Gate → Noise Suppression → Compressor → Limiter. Putting the gate after the compressor lets noise get amplified before it's gated.
Save this exact filter chain: right-click the source → Filters → the gear icon lets you copy filters to another source or scene collection, so you don't rebuild it every stream.
Monitor with headphones during your first few streams, filters can introduce artifacts (gate chatter, pumping compression) that are only obvious while listening live.

Knowledge Base

Live Audio Has No Safety Net

In podcast or video editing you can fix a bad take in post. On a livestream, whatever leaves your mic is what viewers hear, permanently, on VOD too if you're recording. That's why the filter chain needs to be conservative and reliable rather than aggressive: a gate that's too tight cuts off the start of words. A compressor that's too hot causes audible pumping.

Why RNNoise Beats Basic Noise Suppression

OBS's built-in Speex noise suppression is a simple spectral filter, effective on steady hums but prone to artifacts on speech. RNNoise uses a small neural network trained specifically to separate voice from noise, giving noticeably cleaner results at a similar CPU cost on any reasonably modern machine.

Test Before Every Big Stream

Room conditions change. A new fan, a different chair, a laptop that spun up its cooling. Re-check your Noise Gate thresholds and compressor level with a 30-second test clip before any stream where audio quality really matters (an interview, a launch, a sponsored segment).

Where This Fits

This guide covers one specific part of livestreaming. The wider picture (scene and source structure, encoding, bitrate, and audio routing for a reliable stream) is in OBS Studio for Beginners, which frames the discipline as a whole and links out to the detailed guides underneath it, including this one. If you are starting from scratch rather than solving a specific problem, read that first and come back here.

FAQ

Q: Can I use my podcast mic processing chain from Audacity/REAPER for live streaming?
A: Not directly. Those are offline processors. OBS applies its own real-time filters (Gate, Noise Suppression, Compressor, Limiter, EQ) that work the same way conceptually but run live.

Q: My audio sounds robotic after adding noise suppression, why?
A: You're likely over-suppressing. Lower the suppression intensity, or switch from Speex to RNNoise, which handles voice more naturally at similar noise-reduction strength.