How to Add Background Music in OBS Studio with AI Tools

Why Background Music Changes the Feel of a Stream

Silence is the most underrated problem in live streaming. A viewer who lands on a channel during a slow moment — a menu screen, a queue, a load screen — has nothing to hold onto except the sound of an empty room. Background music fills that gap. It signals that the stream is alive, it sets an emotional register, and it gives the audience a reason to stay during the quiet stretches between big moments.

The catch is that music is also the fastest way to ruin a stream's audio. Add one loud track to the wrong output and you get echo, clipping, copyright strikes, and viewers complaining that they can't hear your voice. The difference between "atmospheric" and "unlistenable" is almost always technical, not creative: source choice, gain staging, ducking, and routing.

This tutorial walks through the whole workflow inside OBS Studio — from preparing files, to adding them as sources, to shaping them with filters so they sit under your voice instead of fighting it. Along the way, we'll look at where AI music generation genuinely helps and where it doesn't, and finish with a troubleshooting section and a repeatable pre-stream checklist. If you already know the basics of scenes and sources, you can skim the setup and jump straight to the mixing and routing sections.

Prepare Your Audio Before You Open OBS

The most common mistake is treating music selection as the first step. It should be the third. Before you drag a single file into a scene, decide what role music plays and how it will be delivered.

Decide on the Role of the Music

Ask three questions:

Is it foreground or background? Foreground music (intro sequences, breaks, highlight reels) can sit loud and carry the moment. Background music is wallpaper and should never be the loudest thing on stream.
Is it continuous or punctuated? Continuous loops suit just-chatting and ambient gameplay. Short stings and transitions suit structured shows, tutorials, and competitive play.
Does it need to be muted on demand? If you plan to mute music during commentary-heavy sections, build that mute into your hotkeys rather than reaching for a mixer slider mid-sentence.

Choose Your Source Type

There are three practical sources of music for a streamer:

Licensed libraries and subscription services. Predictable quality, clear terms, ready-made stems and loops. Good for creators who want zero ambiguity about usage rights.
Original recordings and commissioned tracks. Best fit for a signature sound, but slower to produce and usually more expensive per track.
AI-generated music. Fast, cheap per idea, and excellent for loops, ambient beds, and menu stings. The trade-off is that quality varies wildly, and you still need to understand the terms of the tool you use.

A hybrid approach works best for most channels: AI for ambient beds and transition stings, licensed tracks for intro and outro themes that carry brand identity.

Export Music for Streaming, Not for Hi-Fi

Streaming platforms compress audio aggressively. Tracks that sound rich on studio headphones can turn to mush after the platform's encoder gets hold of them. Export music as high-quality files — WAV, FLAC, or a 320 kbps MP3 — with these habits in mind:

Trim silence at the head and tail. A loop with half a second of dead air will stutter every time it repeats.
Normalize loops to a consistent loudness. If your loop is 6 dB louder than your intro sting, you'll be riding the fader all night.
Use seamless loop points. In most editors, this means cutting at a zero crossing so the waveform meets cleanly. Poorly cut loops produce a click that is very audible on stream.
Keep a mono compatibility check in mind. Some viewers watch on phone speakers, which are effectively mono. If your track depends on wide stereo effects, it can vanish on mobile.

Adding Music Sources in OBS Studio Step by Step

Once your files are ready, adding them to OBS is straightforward — but the source type you choose determines how much control you get.

Pick the Right Source Type

Source	Best for	Trade-off
Media Source	A playlist or single looped file you control precisely	Requires you to manage playback manually or with hotkeys
Application Audio Capture	Music playing from a separate app (browser, player)	Captures everything that app plays, including notifications
Audio Output Capture	Music routed through a specific device	Captures all audio on that device, which can include system sounds
Browser Source	Web-based players and widgets	Heavier on resources and can drift out of sync

For most streamers, a Media Source is the cleanest option. It gives you a dedicated source you can filter, mute, and route independently of everything else.

Create a Dedicated Music Scene or Group

Rather than dropping music into every scene, create a Music scene and place all your audio sources — loops, stings, alert sounds — inside it. Then add that scene as a nested scene or as a scene group in the scenes where you want it. Two advantages:

You tune the mix once and it applies everywhere.
You can enable or disable the whole group with a single hotkey when you need a quiet moment.

Configure the Media Source

Open the Media Source properties and check the following:

Local File rather than a network URL, to avoid buffering hiccups.
Loop enabled for ambient beds, disabled for one-shot stings.
Restart playback when source becomes active — useful if you want the track to start fresh each time you switch scenes.
Close file when inactive — helpful for freeing resources, but it also resets playback position, so decide based on whether continuity matters.

Set the source's volume in the mixer to roughly -20 dB to -16 dB as a starting point, then refine with filters. If you hear nothing, confirm the source isn't muted and that its audio track is actually being recorded or streamed in your output settings.

Mixing: Levels, EQ, Compression, and Ducking

This is where background music goes from "a track playing" to "a professionally mixed bed."

Establish Target Loudness for Each Element

A simple hierarchy works for almost every stream:

Voice: the loudest element, peaking around -6 dB to -3 dB.
Game or content audio: roughly -12 dB to -8 dB.
Music: -24 dB to -18 dB under speech, louder only during dedicated music moments.
Alerts and notifications: brief peaks, but never loud enough to make viewers flinch.

Mixing by ear in a quiet room is unreliable. Watch the meters in OBS's audio mixer and use the color indicators: green is healthy, yellow is getting hot, red is clipping.

Sidechain Compression for Automatic Ducking

Ducking means the music automatically gets quieter whenever you speak. OBS doesn't have a native sidechain input on the compressor filter, but you can approximate it in two ways:

Approach one — manual hotkey ducking. Assign a hotkey that lowers the music source's volume while pressed or toggles a "music down" filter. Simple and predictable.

Approach two — external mixer or audio plugin chain. Route your microphone and music through a virtual audio cable into a small mixing application, apply real sidechain compression there, and send the result back into OBS as a single source. This gives you genuine ducking and a cleaner overall mix, at the cost of extra setup and a bit more latency to manage.

If you use a dedicated mixer or audio interface with onboard ducking, that's often the easiest path of all — the hardware does the work before the signal ever reaches OBS.

EQ and Filter Choices That Actually Help

Three filters do most of the heavy lifting on a music source:

High-pass filter at 80–120 Hz. Removes rumble that competes with the low end of your voice and the game. Music often sounds clearer after this, even though you're removing information.
Compressor, gentle settings. A ratio around 2:1 with a slow attack and moderate release smooths out loud and quiet passages so you're not constantly adjusting levels.
Limiter. A safety net that catches sudden peaks from a track that was mastered far louder than you expected.

On the microphone chain, do the mirror image: a high-pass filter, a compressor, and a noise gate. The goal is to make the two signals occupy different spaces so they don't compete.

Mind the Masking Problem

If your music sits in the same frequency range as your voice — roughly 300 Hz to 3 kHz — it will make you sound muddy no matter how low you set the volume. A small dip in the music's EQ around 1–3 kHz (about 2–4 dB) opens up space for speech without making the track sound thin. This single move often does more for intelligibility than any amount of volume fiddling.

Scene Routing and Advanced Setups

Once the basics work, routing is where you gain control and lose headaches.

Use Nested Scenes for Consistency

Build a Music scene that contains only audio sources. Add it as a nested scene in your gameplay, just-chatting, and break scenes. Now a single change to the music scene propagates everywhere.

Separate Music from Game and Chat Audio

If you stream with friends or use voice chat, keep music on its own audio track so your recording contains music but your collaborators' audio doesn't get polluted. In OBS's output settings, you can assign sources to up to six audio tracks for recording — a huge advantage during editing. Music on track 2, voice on track 1, game on track 3 is a common layout.

Handle Alerts and Overlays

Alert sounds and music occupy the same sonic space. If an alert fires while a loud passage of music plays, the two stack and clip. Either duck the music slightly whenever an alert is active, or keep alert sounds short, bright, and mixed a few decibels below the music bed.

Control Everything with Hotkeys

Set hotkeys for the things you'll reach for mid-stream:

Mute/unmute music
Lower music volume temporarily
Skip to the next track
Stop all audio except your microphone

The last one is your emergency brake. If a track turns out to be problematic — wrong mood, unexpected lyrics, a copyright flag — one keypress should silence it without disrupting the rest of your stream.

Using AI Tools in the Music Workflow

AI music generation has become genuinely useful for streamers, but the useful part is narrower than the marketing suggests. Here's where it earns its place.

Generating Ambient Beds and Loops

AI is strongest at producing long, unobtrusive, repeatable textures: lo-fi beds, synth pads, ambient drones, and simple rhythmic loops. These are exactly the tracks that work best as background music because they don't demand attention. Prompt for mood, tempo, and instrumentation, generate several variations, and keep the one that doesn't fight your voice.

Creating Transition Stings and Branding

Short stings — three to five seconds — are ideal AI territory. Generate a few, cut them tightly, and normalize them to the same loudness as the rest of your audio package. A consistent set of stings can make a small channel feel surprisingly polished.

Speeding Up Selection

One underrated use of AI is search. Instead of listening through a library, describe what you want in plain language, generate five candidates, and audition them in seconds. Even if you end up licensing a human-made track, the AI pass helps you articulate what you're actually looking for — tempo range, instrumentation, energy level.

Where AI Falls Short

AI struggles with anything that needs a strong melodic identity: branded intro themes, memorable outros, anything a viewer should recognize instantly. It also tends to produce tracks that are subtly over-compressed or that loop poorly. Treat AI output as raw material, not a finished product. Trim loop points, check for level consistency, and clean up with a quick pass in an audio editor before it ever reaches OBS.

A Practical AI Workflow

Generate five to ten candidates per mood you need (calm, energetic, tense, playful).
Import into an audio editor and trim to clean loop points.
Normalize every track to the same integrated loudness.
Export as WAV or high-bitrate MP3 into a dedicated music folder.
Load the folder into a Media Source playlist or a local player, and control it from OBS.

Legal and Platform Considerations

No technical setup matters if your stream gets muted or flagged. A few principles keep you safe.

Understand Your Rights to the Music

Every track you play on stream needs a usage right that covers live broadcast and, ideally, recorded archives. "Royalty-free" is a marketing term, not a legal one — it means no ongoing payments, not no conditions. Always read the actual terms: some licenses exclude monetized streams, some exclude video-on-demand archives, and some require attribution.

Keep Records

Maintain a simple spreadsheet or document listing each track, its source, the license type, and a link to the terms. When a platform asks a question about a muted section six months later, having that record is the difference between a quick fix and a lost archive.

Match the Music to the Platform

Music that is perfectly fine on one platform may be auto-flagged on another. Content ID-style systems are automated and frequently wrong, and appeals take time. Where possible, prioritize tracks from libraries that explicitly state platform compatibility, and keep a "safe set" of tracks you know work everywhere for high-stakes streams.

Think About Your Archive

Live performance rights and recording rights are separate things. If you plan to upload highlights, confirm that your license covers permanent distribution, not just the live broadcast. If it doesn't, you can still stream — just mute or replace the music during your editing pass.

Troubleshooting Common Music Problems

Music plays on stream but the recording is silent. Check your audio track assignments in the output settings. The source may be routed to a track that the recorder isn't capturing.

Everything sounds distorted when the music comes in. You're clipping on the sum. Reduce the music source by 3–6 dB and check the master meter. If the mix still peaks, apply a limiter to the master output.

Viewers say the music is too loud, but it sounds quiet to you. Headphones with strong bass response flatter low frequencies. Check your mix on laptop speakers and a phone before trusting your ears.

The loop clicks every time it repeats. The cut point isn't at a zero crossing, or the file has a stray sample at the start. Fix in an editor rather than trying to hide it with a fade.

Music keeps playing after you switch scenes. The source is in a scene that's still active, or it's set to "close file when inactive" with a nested scene that never deactivates. Check your scene structure before touching the mixer.

Latency between the music and the game audio. Music playback latency usually comes from processing. Move music to a source with fewer filters, or pre-render the loop with effects already applied.

Audio drifts out of sync over a long stream. Long-running browser sources and some media players can drift. Restarting the source between segments, or using a dedicated audio player outside OBS, solves it.

A Repeatable Pre-Stream Checklist

Run this ten minutes before going live and most audio problems never make it to air.

Confirm your music folder is loaded and the playlist is playing in the right source.
Spot-check levels: voice peaks around -6 dB, music sitting 6–10 dB below it.
Test one full loop for clicks or dead air.
Verify hotkeys for music mute and emergency silence.
Confirm track assignments for recording versus streaming.
Do a five-second local recording and listen back with headphones.
Check the platform's audio meter during that recording.
Have a backup playlist of known-safe tracks ready.

Frequently Asked Questions

Can I add background music to OBS with just one source?
Yes. A single Media Source with a playlist handles most needs. Filters, volume, and hotkeys all apply to that one source, which keeps your scene collection simple.

Is it better to play music from a browser or from a local file?
Local files. Browser sources consume more resources, can drift out of sync, and may pick up unwanted notification sounds. Local media sources are more predictable and easier to route.

How loud should background music be under speech?
Aim for roughly 6–10 dB below your voice, with music peaking around -20 dB to -18 dB when you're talking. During dedicated music moments, you can push it to -12 dB or higher.

Do I need a separate audio interface for ducking?
No, but it helps. You can approximate ducking with hotkeys or an external mixing application. A hardware mixer or interface with onboard ducking simply makes it automatic and consistent.

Can AI-generated music be used on a monetized stream?
That depends entirely on the terms of the tool you used. Read the license carefully, check whether commercial and broadcast use are permitted, and keep documentation of the specific terms that applied when you generated the track.

Why does my mix sound fine live but bad in the recording?
Recordings bypass platform-side processing. What sounds smooth after a live encoder's loudness normalization can sound unbalanced in a raw recording. Monitor your local recording file, not just the live output.

How often should I rotate music tracks?
Every couple of weeks for regular viewers, more often for daily streams. Repetition is comfortable for a few sessions and grating by the tenth. Build a playlist of at least ten to fifteen loops and shuffle it.

What's the fastest way to fix muddy-sounding music under my voice?
Apply a high-pass filter at around 100 Hz and a gentle 2–4 dB dip between 1 kHz and 3 kHz on the music source. That single EQ move clears space for speech better than lowering the volume.

Should music be on its own audio track?
If you record for later editing, yes. Separate tracks let you adjust or remove music in post without affecting voice or game audio. It costs nothing and saves hours later.