Multitrack Audio Mixer
Place independent audio files on a timeline and move each track's start time to choose where they overlap. Set track gain, stereo pan, mute and solo. The browser sums decoded samples into stereo; if the combined peak exceeds one, the whole mix is normalized to a 0.98 peak with the pre-normalization over-limit sample count shown. Preview and export a 16-bit WAV and JSON containing source SHA-256 hashes, track settings and output length. Unlike the waveform editor, which joins sections from one source, or the loop finder, which recommends repeating boundaries, this tool blends multiple files in time and stereo space.
Key features
- Show separate waveforms for two to four files and round each track offset to decoded sample frames.
- Set per-track 0–2 gain, -1 to +1 pan, mute and solo.
- Use constant-power mono panning and stereo left/right balance before overlapping sample summation.
- Measure pre-normalization peak and over-limit samples; normalize the full mix to 0.98 peak when needed.
- Preview and export a 16-bit stereo WAV plus track project JSON with source SHA-256 hashes.
How to use
- Choose two to four audio files up to 8 MiB each or load the two-track WAV example.
- Set start time, gain and pan for each track; optionally enable mute or solo.
- Set master gain, render the mix and inspect peak and normalization information.
- Listen, then download the WAV and project JSON.
Use cases
- Overlap narration and a sound effect at different start times.
- Position two melody files left and right for a simple stereo demo.
- Record silent, solo and mute decisions plus pre-normalization peak in a project manifest.
Frequently asked questions
How do gain and pan combine?
For mono, pan moves from left at -1 to right at +1 using cosine/sine constant-power gains. For stereo, balance attenuates the opposite channel linearly. Track and master gain multiply each sample before same-time samples are summed.
What happens when the sum exceeds 0 dBFS?
The mixer first measures the float peak and counts samples with magnitude above one. If any exceed one, it applies one gain factor to the whole result so the peak becomes 0.98 before 16-bit PCM export. This reduces the overall level; it is not a compressor or limiter.
How do silent tracks and solo work?
Silent audio is accepted as a track. If any unmuted solo track exists, only unmuted solo tracks enter the mix. If every track is muted, the output is a silent WAV with the configured timeline length.
How is this different from the waveform editor or loop seam finder?
The waveform editor cuts several sections from one source and joins them sequentially with edge fades. This mixer starts independent files at chosen times and sums their overlapping samples with gain and pan. It does not suggest loop boundaries.
What files and lengths are supported?
Each input may be up to 8 MiB and must be decodable by the browser, such as WAV, MP3, OGG, FLAC or M4A. Each decoded track must be 0.1–30 seconds, mono or stereo, and 8–96 kHz. The output is limited to 90 seconds and eight million frames. Tracks use one AudioContext so they share its decoded sample rate.
Are audio or projects stored on a server?
No. Files are decoded and mixed in your browser. JSON contains names, sizes, SHA-256 hashes, track settings and peak metrics, but no audio samples. Original files are needed to recreate the mix.
Privacy
Source audio stays in browser memory and is not uploaded. JSON includes source names and hashes; review it before sharing. Clearing or leaving the page stops playback and revokes temporary download URLs.
Comments & questions