Free Audio Mixer — Add Background Music to Audio

Worked example

Press "Try a sample". The page switches to "Voice + background music" and adds two files. sample-voice.wav, a synthetic 8-second voice-like clip, is marked "Voice". sample-music-loop.wav, a 4-second tune, becomes "Track 2" with "Volume" at 25% and "Loop" ticked, and under "Trim and fades" it has "Fade in (s)" 2 and "Fade out (s)" 3. "Length of the mix" is on "Same as the voice" and "Lower the music while the voice talks" is ticked, so the page shows "Mix length 0:08 · about 187.5 KB" for MP3. Press "Mix audio". The tune repeats to fill the 8 seconds and dips whenever the voice sounds. The page shows "Your mix is ready." with a player and a "Download" button for sample-voice_mix.mp3. Now choose WAV under "Save as": the mix stays, a note says the settings changed, and the button at the bottom becomes "Mix again".

Which settings to use, and the limits

• Pick how the tracks should play. Above the tracks are three choices. "Mix together", the default, plays every track at once at 100% and runs until the longest one ends: use it to combine songs or lay sound effects over a track. "Voice + background music" marks one track "Voice" (the first file added; press "Make this the voice" on another track to change it), starts every other track at 25%, looped, with a 2-second fade in and a 3-second fade out, and runs the mix as long as the voice. "One after another" plays the tracks in list order. Switching moves only the settings you have not changed yourself.
• Set each track's volume, start time and mute. "Volume" runs from 0 to 200% in steps of 5. "Start at" takes seconds or clock times: 83, 1:23, 1:02:03 or 12.5, and a comma also works as the decimal mark; an empty box means 0, and text that is not a time is ignored. "Mute" leaves a track out without removing it. The play button plays a track on its own, from its trim start to its trim end, so you can find where a sound effect should come in. The ↑ button moves a track up, and a small bar on each track shows where it sits in the mix.
• Trim a track or fade it. Open "Trim and fades" on a track. "Keep from" and "to" take the same times as "Start at"; an empty "to" means the end of the file, and the page shows how much is kept, for example "Keeps 0:03". If the end comes before the start, or the start is past the end of the file, a red note says so and the track is left out until it is fixed. "Fade in (s)" and "Fade out (s)" step by 0.5 seconds. A fade is shortened when the track is shorter than it, and on a looped track the fade out happens at the end of the mix. A trimmed track that loops repeats only the part you kept, up to 3 minutes of it; a longer kept part plays once.
• Choose how long the mix runs. With two or more tracks, "Length of the mix" has two choices. "Until the longest track ends" counts one pass of every track, looped ones included, plus its "Start at" delay; shorter looped tracks repeat to fill it. The other choice is "Same as track 1" when mixing together and "Same as the voice" under a voice: longer tracks are cut off there. One after another, the mix is the tracks back to back. The line "Mix length 0:08 · about 187.5 KB" updates as you change settings. The size is an estimate from fixed bitrates: 192 kbps for MP3, 160 kbps for M4A and about 1,411 kbps for WAV.
• Pick the output format. "Save as" offers MP3, M4A and WAV. MP3 is written at 192 kbps, M4A (AAC) at 160 kbps, and WAV as 16-bit PCM, which is uncompressed and by far the largest. All three are 44.1 kHz stereo, whatever the sources were. The file is named after the first track with "_mix" added, so interview.wav becomes interview_mix.mp3, and the base name is cut to 60 characters. There are no bitrate or sample rate settings to change.
• Know what files it takes. Drop up to 8 audio or video files; the page says video files work too and their sound is used. A video with no sound is skipped. One file on its own can be saved with a new volume, trim or fades: press "Or save just this file with these settings". Each file is checked against your device: it is refused above about 2 GB on a normal computer, 1 GB on a computer reporting 4 GB of memory or less, and 400 MB on a phone, with a caution above 800 MB, 400 MB and 150 MB. Adding the first file starts loading the media engine, about 31 MB, which is cached afterwards. The files are mixed in your browser and not uploaded.
• Play tracks in order with a crossfade. In "One after another" each track starts where the one above it ends, so "Start at" and "Loop" are hidden and each track shows "Starts at" instead. "Crossfade (s)" is 0 by default, which plays the tracks back to back; up to 15 seconds blends each track into the next, and a crossfade never takes more than half of a track. Use ↑ to change the order. For gaps between tracks and a preview of each join, use the audio joiner.
• Let ducking keep the voice clear. In "Voice + background music" with two or more tracks, "Lower the music while the voice talks" is ticked by default. The page says the other tracks dip by about 10 dB whenever the voice makes sound. Music still playing after the voice ends carries on at its normal level. Untick the box for a plain mix. "Even out the loudness" is off by default; it brings the finished mix to −16 LUFS, the usual podcast and streaming level. A limiter is always applied to the final mix, so a track pushed to 200% does not push the output past full scale.

Common problems and how to fix them

It says "None of these files has any sound in it."

Before mixing, the tool probes every track that is not muted and leaves out any without an audio stream, such as a screen recording made with the microphone off. This message appears when all of them were left out. Add a file that plays with sound. If only some tracks are silent they are left out without a warning; if the silent one was the voice, nothing is ducked and "Same as the voice" uses the first remaining track, so check the result.

It says "Could not read how long these files are. Try saving them as MP3 or WAV first."

The mix length must be known before mixing, because looped tracks are cut at a fixed time. Neither your browser nor the media engine could read any track's duration, which can happen with streams that store no length. Save the files as MP3 or WAV in an audio converter, then add them again.

It says "Up to 8 tracks can be mixed. The extra files were left out."

The mixer holds 8 tracks. The message appears when you drop more files than there are free slots, and the files that fit are kept in the order you gave them. Remove a track with the cross button, or mix in two passes: mix one set, download it, press "Start over" and add the result as a single track.

It says "This file is 2.6 GB, which is too large to process in your browser on this device (limit ~2 GB). Try a smaller file or trim a shorter clip first."

The numbers are your file and your device's limit: about 2 GB, 1 GB or 400 MB, as listed above. Only that file is refused; any other files you dropped are still added. If it is a long video, save its sound as an audio file with another tool first, which is far smaller. A caution above the lower size means the mix may be slow or run out of memory.

It says "Something went wrong while processing this file. Try a different output format or a different source file."

This is the shared catch-all message of the media engine, for when it stops without saying why. A file it cannot decode or read gets its own message instead, but no message says which track caused it. Try another "Save as" format, then tick "Mute" on tracks one at a time to find the one that breaks it, and re-save that file as MP3 or WAV. A separate message says when the browser ran out of memory; for that, use shorter files or a desktop computer.

How it compares

This tool combines 1 to 8 tracks into one file in a single pass: played together, as a voice with music lowered under it, or one after another with a crossfade. Each track can be moved, trimmed, delayed, faded, looped and muted. Joining with gaps and a preview of each join is the audio joiner's job, and putting music under a video, which keeps the picture, is the job of adding audio to a video. Compared with a desktop audio editor it has no waveform view, no panning, no equaliser and no per-track effects, and the output is fixed at MP3 192 kbps, M4A 160 kbps or 16-bit WAV. In exchange nothing needs installing.