Mix Voice With Background Music: Streamer & Podcaster Guide

Mixing your voice with background music is one of those tasks that sounds simple until you actually do it live — and suddenly your voice is buried, the music is overwhelming the chat, or a DMCA bot nukes three hours of content at 2 AM. Getting this right requires understanding a handful of audio principles, the right OBS routing, and a music library that will not get your channel struck.

This guide covers the full picture: target levels, ducking setup, OBS Audio Mixer routing, DMCA-safe music sources, and Discord music considerations. Whether you are running a podcast with lo-fi ambience or a live Twitch stream with a full intro track, the principles are the same.

TL;DR

Voice target: -12 LUFS integrated. Background music target: -20 LUFS or quieter.
Audio ducking (sidechain compression) automatically lowers music when you speak — set it up once and forget it.
In OBS, route music and mic as separate sources, never the same device.
DMCA-safe options: Twitch Soundtrack, Pretzel Rocks, StreamBeats, Lofi Girl — avoid Spotify, Apple Music, and commercial radio.
Discord music bots work well for background vibes during team calls; separate routing keeps levels clean.
Clean vocal audio matters more than fancy music mixing — prioritize noise suppression before worrying about sidechain compression.

Why Voice and Music Clash (And How Levels Fix It)

Human speech sits primarily in the 300 Hz–4 kHz range, with intelligibility cues concentrated between 1–4 kHz. Most music is also dense in that same range — bass, kick drums aside. When both signals hit your stream mix at similar loudness, they fight for the same spectral and volume space. Viewers cannot follow what you are saying, and the experience feels fatiguing.

The solution is not to eliminate one or the other — it is to establish a clear loudness hierarchy before anything else.

The -12 / -20 LUFS Rule

LUFS (Loudness Units relative to Full Scale) is how modern broadcast, streaming platforms, and podcasting standards measure perceived loudness. It is a better metric than peak dB because it accounts for how human ears actually perceive sustained sound.

Voice: -12 LUFS integrated — this is roughly where podcast speech sits for comfortable listening on headphones or speakers, and it aligns with YouTube and Spotify’s normalized playback targets.
Background music: -18 to -20 LUFS — at this level, music is clearly audible as ambience without masking speech.

That 6–8 LUFS gap is perceptible by the human ear as music sitting “underneath” the voice, not competing with it.

To measure LUFS in a stream context: the free plugin Youlean Loudness Meter works in any DAW and as a VST in OBS (via the VST filter on an audio source). Record a short test segment, check the integrated LUFS readout, and adjust faders accordingly. Once calibrated, you can trust your OBS fader positions for future sessions.

OBS Audio Mixer Routing for Voice and Music

OBS Studio has a surprisingly capable audio routing system that most streamers never fully use. Here is a setup that gives you independent control over every element.

Step 1 — Identify Your Sources

In OBS > Settings > Audio, assign your microphone to a dedicated input:

Mic/Aux Input — your microphone (e.g., “Headset Microphone” or your USB condenser)
Desktop Audio — this captures everything playing through your speakers/headphones output

The problem with relying on Desktop Audio alone is that it captures everything: game audio, notification sounds, browser tabs, music — all mixed together. You lose individual control.

Step 2 — Add Your Music Player as a Separate Source

Instead of letting music bleed into Desktop Audio, route it through a dedicated channel:

In OBS, click the + in Sources > Application Audio Capture (Windows) or Audio Input Capture.
Select your music player application (Pretzel, Spotify, a browser tab playing StreamBeats, etc.).
This gives you a separate fader in the OBS Audio Mixer.

Alternatively, use a virtual audio cable (VB-Cable or Voicemeeter) to route your music player’s output to a dedicated virtual device, then capture that device as a separate source in OBS. This approach works even if your music player does not appear in Application Audio Capture.

Step 3 — Set Fader Levels

With sources separated, open the Advanced Audio Properties (gear icon on any mixer fader > Advanced):

Source	Target Level (dB)	Monitor	Track Assignment
Microphone	-12 LUFS / peaks at -6 dBFS	Monitor Off	Tracks 1 & 2
Background Music	-20 LUFS / peaks at -16 dBFS	Monitor Only	Track 1 only
Game / Desktop Audio	-18 dBFS average	Monitor Off	Tracks 1 & 2

Assigning music to Track 1 only (the stream output) but not to Track 2 (dedicated recording track) means your local recordings will be voice-and-game-only, which is useful if you want to re-edit VODs later.

Step 4 — Filters on the Mic Channel

Add these filters to your microphone source (right-click > Filters):

Noise Suppression — use RNNoise or connect to an external suppressor (see our guide to noise suppression software)
Compressor — Ratio 3:1, Threshold -18 dB, Attack 6ms, Release 60ms. This keeps your voice at a consistent level even when you lean back from the mic.
Limiter — ceiling at -3 dBFS to prevent clipping spikes from hitting the stream.

A well-compressed voice is much easier to balance against music because it does not swing wildly between quiet and loud. This matters more than any EQ trick.

Audio Ducking: Automatic Volume Control That Actually Works

Ducking is the technique where background music automatically lowers in volume whenever your voice is detected, then returns to its original level during silence. Broadcast radio has used this for decades. It is the smoothest way to mix voice with background music for streaming because it removes the need to manually ride the fader.

How Sidechain Compression Works

In a full DAW setup (Reaper, Ableton, Logic), ducking is implemented as sidechain compression:

Your microphone signal is routed as a sidechain input to a compressor inserted on the music channel.
When your mic volume exceeds a threshold (say, -20 dBFS), the compressor reduces the music channel gain by a set ratio (typically 4:1 to 8:1).
Attack time controls how fast the music ducks (10–30ms feels natural). Release time controls how fast it comes back (200–500ms — too fast sounds like pumping, too slow leaves music quiet after you stop talking).
Make-up gain on the compressor is set to zero (you want the ducked level to be lower than unducked).

Ducking in OBS Without a DAW

OBS does not have native sidechain compression, but there are two practical approaches:

Option A — OBS Audio Ducking Plugin

The community plugin obs-audio-ducking adds a filter that uses the microphone signal to duck other sources. Install it via the OBS Plugin Manager, add it as a filter to your music source, and configure:

Threshold: -25 dB (activates when you speak at normal volume)
Duck Amount: -10 to -14 dB (how far the music drops when you talk)
Fade In Time: 200ms
Fade Out Time: 400ms

Option B — Voicemeeter Virtual Mixer

Route audio through Voicemeeter Banana (free) which has a built-in compressor on each bus. Assign your music to Bus B and your mic to Bus A. Use the bus compressor on B with A as the trigger source. OBS then captures Bus B’s processed output as the music channel.

Option C — VoxBooster Virtual Mic + Separate Music

If you are running VoxBooster for voice effects during your stream, your voice is already on a virtual microphone output. You can apply ducking at the VoxBooster routing level and keep the music routed separately in OBS. This separation simplifies the signal chain considerably.

DMCA-Safe Music for Streaming: What You Actually Need to Know

This is the part where streamers get in trouble. Playing music on a live stream or in a recorded VOD without the right license can result in:

Muted VODs — Twitch’s Content ID scanning flags recorded streams after the fact
DMCA takedown notices — requiring action within 24 hours
Channel strikes — three strikes can mean termination on YouTube
Twitch suspensions — repeat violations lead to escalating bans

The safest options, in order of convenience:

Twitch Soundtrack

Twitch Soundtrack is Twitch’s built-in tool that lets you play licensed music during a live broadcast. It routes music directly to your stream without capturing it in your local recording, which means your VODs and clips stay DMCA-clean automatically.

Limitation: Twitch Soundtrack only covers music on Twitch itself. If you also stream to YouTube or export clips to TikTok/Instagram, those exports may not carry the same license coverage. Always check platform-specific terms.

Royalty-Free and Streaming-Licensed Libraries

Library	Cost	Notes
StreamBeats (Harris Heller)	Free	YouTube and Twitch streaming license; download MP3s
Lofi Girl (streaming playlists)	Free	YouTube channel with explicit streaming permission; credit required
Pretzel Rocks	~$8/mo	Large catalogue, Twitch/YouTube/Facebook licensed, desktop app
Epidmic Sound	~$15/mo	Broad licence covers all platforms including TikTok
Monstercat Gold	~$6/mo	Artist revenue share, covers major streaming platforms
Artlist.io	~$17/mo	License is personal-perpetual; covers video sync as well

What to avoid: Spotify, Apple Music, YouTube Music playlists, commercial radio streams, and any music where you do not have written documentation of streaming rights. Even playing music “quietly in the background” does not create a fair-use exemption for live streaming.

The safest habit: keep a folder of StreamBeats MP3s or Lofi Girl stems locally downloaded, and play them from a local media player. No internet dependency, no risk of the source being removed mid-stream.

Voice Clarity Over Music: EQ Considerations

Even at the right LUFS levels, music can mask your voice if both occupy the same frequency range. A small EQ adjustment on the music channel called a presence notch can open up space for speech:

On your music channel’s EQ, apply a gentle notch or dip between 1 kHz and 4 kHz — this is the core of vocal intelligibility.
The notch does not need to be deep: -2 to -4 dB across a broad Q (0.5–1.0) is enough.
Your voice sits in that same range, so by creating a small valley in the music’s frequency content, your voice appears clearer without you needing to raise its volume.

This technique is standard practice in broadcast mixing and is sometimes called “carving space” for the vocal. The effect is subtle when looked at on a spectrum analyzer but significant to listeners.

You can also apply a gentle high-pass filter to your music at around 100–120 Hz if you are speaking over music with heavy bass. This lets the kick drum and bass elements of music remain audible while reducing low-frequency energy that can create a muddy buildup in the mix.

If your background music use case is a Discord server — for gaming sessions, co-streams, or podcast pre-rolls — you have a few options.

Discord Music Bots

Music bots (like Hydra, Rythm-successors, or MEE6’s music module) join your voice channel and play audio directly through Discord’s audio pipeline. For pure background ambience, this is the simplest option.

Limitations of Discord music bots:

Volume control is global; you cannot easily duck bot music when you speak without a bot-side command.
Discord’s audio compression (Opus codec at 64–128 kbps for most servers) reduces music fidelity noticeably.
Music licensed to the bot may not carry rights for you to stream that same bot output to Twitch or YouTube.

For a production setup where you also stream elsewhere, route your music through a virtual audio cable and share that as your “Desktop Audio” in a Discord voice channel. You get:

Independent level control from your mic
Consistent audio quality (you control the source quality)
The ability to run DMCA-safe sources and know exactly what rights you have

Keeping Discord and Stream Audio Consistent

If you stream to Twitch while also on a Discord call, you want both audiences to hear a consistent mix. The cleanest approach:

Route microphone → VoxBooster or your voice processor → Virtual Mic output
In Discord, select the Virtual Mic as your input
In OBS, select the same Virtual Mic as your mic input
Music routes separately to OBS (and optionally to Discord desktop audio sharing)

Both audiences now hear the same processed voice. Music is separate from the voice chain, so Discord noise suppression does not accidentally eat your background music.

For a deeper look at how voice processing integrates with stream routing, see our guide on voice changer for podcasting and the comparison of voice changer vs EQ: when to use each.

Common Mixing Mistakes (And How to Fix Them)

Mistake 1: Setting levels by ear in headphones only

Headphones are louder than your stream sounds to viewers with speakers. Always check your OBS meter (aim for -12 dBFS peaks on voice) and use a LUFS meter to verify integrated loudness. What sounds perfect in cans can sound thin or distant on laptop speakers.

Mistake 2: Using Desktop Audio as a catch-all

Desktop Audio mixes everything together — notifications, game audio, browser alerts, music — and makes individual control impossible. Split your sources.

Mistake 3: No compressor on the microphone

An uncompressed voice swings 10–15 dB between normal speech and excited moments. With background music at -20 LUFS, your quiet moments will be buried and your loud moments will clash. A basic compressor (3:1 ratio, medium attack/release) smooths this out automatically.

Mistake 4: Music with vocals

Instrumental background music is much easier to mix under a speaking voice than music with vocals. Two vocal streams fight for listener attention in a way that two separate instruments do not. Stick to instrumentals, lo-fi, ambient, or orchestral tracks for live commentary situations.

Mistake 5: Ignoring intro/outro loudness

Intro music at full loudness (-10 LUFS) that abruptly fades when you start talking is jarring. Fade the music down over 2–4 seconds as you start speaking, and give it a second to duck before your voice reaches full volume. Most DAWs and playlist tools support volume automation; even OBS can do this manually with a slow fader move.

Quick-Start Checklist

Before going live with background music, verify:

Microphone peaks at -6 dBFS or lower in the OBS meter
Background music source is separate from Desktop Audio
Music fader set to approximately -8 to -10 dB below the voice fader
LUFS meter confirms voice at -12 LUFS integrated (test with 30-second recording)
Music source confirmed DMCA-safe (have written documentation or use Twitch Soundtrack)
Compressor filter applied to microphone channel
Ducking enabled (plugin or manual fader automation plan)
Tested on stream preview with headphones off for 60 seconds

Frequently Asked Questions

What volume should my voice be relative to background music?

Target your voice at around -12 LUFS (integrated) and your background music at -20 LUFS or lower. That 8 dB gap keeps speech intelligible over music without making the music feel absent. In OBS, use the Audio Mixer faders to set these levels before going live.

What is audio ducking and how does it work for streaming?

Ducking automatically lowers background music volume whenever your voice is detected, then brings it back up during silence. In a DAW this is done via sidechain compression — your mic signal triggers a compressor on the music channel. In OBS, plugins like obs-audio-ducking or a virtual audio cable chain can approximate the same result.

Is background music on Twitch streams allowed?

Only if you use licensed music through Twitch Soundtrack, fully royalty-free tracks from libraries like Pretzel Rocks or StreamBeats, or Creative Commons music from Lofi Girl. Playing commercial music without a license risks DMCA mutes or account strikes. Spotify and Apple Music are not licensed for live streaming.

Can I play music through Discord for my stream team?

Yes, using a Discord music bot or by sharing your desktop audio in a voice channel. For production-quality mixing, route your music through a virtual audio device and send that channel to both OBS and Discord so levels are consistent everywhere.

How do I route background music in OBS without it going into my mic?

Add your music player as a separate audio source — Application Audio Capture or a virtual cable device. Keep your microphone on a dedicated input source. In OBS’s Audio Mixer, use Advanced Audio Properties to assign each source to specific output tracks independently.

What DMCA-safe music libraries do streamers use?

The most popular are Pretzel Rocks (subscription), StreamBeats by Harris Heller (free), Lofi Girl’s streaming-safe playlists (free), Monstercat Gold (subscription), and Epidmic Sound (subscription). All provide written streaming licenses you can document if a DMCA claim is filed.

Does noise suppression affect background music in a stream?

A voice-targeted noise suppressor is trained to remove non-voice noise from a microphone channel — it does not touch a separately routed music source. As long as your music is on its own audio track and not feeding through the mic, suppression will not alter it.

Conclusion

Mixing your voice with background music for streaming or podcasting comes down to three fundamentals: the right loudness targets (voice at -12 LUFS, music at -20 LUFS), proper OBS routing with separate sources for every audio element, and music that you are actually licensed to use.

Ducking handles the dynamic relationship between voice and music automatically once configured, which means you can focus on what you are saying instead of manually riding faders mid-stream. On the music side, Twitch Soundtrack, StreamBeats, and Pretzel Rocks cover most use cases without DMCA risk.

The rest — presence notches, multi-track recording, Discord bot routing — are refinements you layer on once the basics are solid. Start with clean voice audio and a properly calibrated mix; everything else follows from there.

If you want to add voice effects or real-time voice processing on top of your mix, VoxBooster integrates cleanly into this routing setup — it creates a standard virtual microphone output that OBS and Discord both recognize, with built-in noise suppression so your voice stays clean even when the background music level creeps up. The 3-day free trial lets you test it against your actual hardware before committing.