Loading tool…
Putting narration over a music bed or balancing a few stems shouldn’t need a full DAW. Add your files, and each one becomes a track with its own volume, pan, mute, solo and start time. Play everything in sync to check the balance, then bounce it down to one file.
Processing locally on your device
Overview
A mixer adds several signals together. Here each track is an AudioBufferSourceNode routed through its own gain (volume) and StereoPannerNode (left–right position) into a master bus, all inside one Web Audio graph. Because every track lives in the same graph and is scheduled against the same clock, playback is sample-synchronized, and volume, pan, mute and solo changes are heard immediately without restarting. Moving a start time or adding a track rebuilds the graph at the current position.
Summing adds levels, so mixes get loud quickly. Two tracks that each peak at −6 dBFS can reach about 0 dBFS when their peaks line up. The master section shows a live level meter during playback and turns red at full scale. The master limiter, on by default, catches peaks above −1 dB so the export doesn’t clip, but the cleanest mixes come from lowering track volumes until the limiter rarely has work to do. For voice over music, a common starting point is to keep the music 15–20 dB below the voice while someone is speaking.
Panning places a source in the stereo field. Web Audio’s stereo panner uses an equal-power law, so a mono sound moved from center to hard left keeps roughly the same perceived loudness. Spreading supporting parts, for example two guitars at 30% left and right, leaves room for a vocal in the center. Each track also has a start offset, set with a slider or typed to the hundredth of a second, so you can delay music until after an intro or line up separately recorded guests.
When you export, the same graph is rebuilt in an OfflineAudioContext, which renders the mix much faster than real time, so what you heard is what you get. The mixdown is stereo at your files’ sample rate. WAV is written instantly in JavaScript, and MP3, M4A, OGG, Opus and FLAC are encoded with FFmpeg in WebAssembly after a one-time 31 MB download. Each decoded track uses roughly 21–23 MB of memory per stereo minute, so phones handle fewer long tracks than desktops.
| Mix | Main element | Supporting element | Pan |
|---|---|---|---|
| Voice over music | Voice at 0 dB | Music about −15 to −20 dB | Both center |
| Two-person podcast | Both voices at 0 dB | Room tone or music bed −25 dB | Voices 10–20% apart |
| Band demo | Vocal at 0 dB | Guitars −4 to −8 dB | Guitars 30–60% L/R |
| Sound effects on a scene | Dialogue at 0 dB | Effects −6 to −12 dB | Effects placed to match the picture |
Drop two or more audio files onto the mixer; each becomes a track.
Set each track’s start time so the parts line up on the timeline.
Balance the tracks with the volume and pan sliders, using M (mute) and S (solo) to check parts.
Press Play mix and watch the master level meter; lower volumes if it turns red.
Choose a format and click Export mixdown.
Use cases
Put narration on one track and a music bed on another, delay the voice by a couple of seconds for an intro, and keep the music low enough that every word is clear.
Line up separately recorded host and guest files with start offsets so they play as one conversation, and pan them slightly apart for a sense of space.
Load drums, bass, guitar and vocal recordings, balance them with volume and pan, and bounce a quick demo to share with the band.
Mute or solo stems to hear one part clearly, then export a mix without your own instrument to play along with.
Pro tips
Answers to common questions about the multi-track mixer.
There’s no fixed limit. Each decoded track uses roughly 21–23 MB of memory per stereo minute, so a desktop can handle many long tracks while a phone may manage fewer.
Signals add together. When peaks from different tracks coincide, the sum can exceed 0 dBFS even if each track is well below it. Lower the track volumes, or leave the master limiter on to catch the occasional peak.
Yes. The mixdown uses the same Web Audio graph as playback, with the same volumes, pans, offsets, mute and solo states and master limiter, rendered offline instead of in real time.
Not inside the mixer, which focuses on level, pan and timing. Process a stem with the separate equalizer, reverb or noise-reduction tools first, then add the result as a track.
They are silent in the mixdown, exactly as you hear them. If any track is soloed, only soloed tracks are included.
No. ToolEnhancer processes files on your own device inside the browser. They are never sent to, stored on or seen by our servers.
Yes. The multi-track mixer is free to use with no sign-up, no watermarks and no daily limits.