Merge audio files

Join or mix

One after another gives you a single file the length of all the parts added up. This is what you want for stitching recorded segments or building a continuous mix.

All at once overlays the files so they play together, as long as the longest one. Use it to put music under a voice recording. Mix a quiet bed under speech by lowering the music first with the volume tool, since this mix does no automatic ducking.

Reorder the files with the arrows before you run, the list order is the play order.

What each mode does underneath

One after another uses FFmpeg's concat filter, which decodes each input and lays the samples end to end. Because everything is decoded first, the files do not have to match: different bitrates, different sample rates, even different source formats join cleanly, and the result is encoded once at the settings you choose. That is slower than a raw stream copy would be, and it is the reason this works on a folder of recordings made on different devices.

All at once uses amix, which sums the inputs into a single track as long as the longest one. Summing is the important word: two signals at full level add up to more than full level, and the result clips. If a mix comes back distorted, that is what happened.

Getting a music bed right

There is no automatic ducking here, so the levels have to be set before mixing rather than after. The reliable sequence is to take the music down to roughly a fifth of the speech, mix, then check.

1. Run the music through change volume at about -14 dB. Under a spoken voice, music usually wants to sit far lower than instinct suggests.
2. Mix the quietened music with the voice here.
3. Listen to the join. If the whole thing sounds squashed or crackly, the sum is clipping: lower the music another few decibels and mix again.
4. Level the finished file to -16 LUFS with the volume tool, which is the podcast standard and what most platforms expect.

Order and gaps

In join mode the list order is the play order, and nothing is inserted between the pieces, so the transition from one to the next is instantaneous. That is usually fine for continuous speech and abrupt for anything musical. To soften it, put a fade on the end of one piece and the start of the next with trim audio before joining. To place a deliberate pause, join a silent file in between; a few seconds of silence trimmed from any recording will do.

Common questions

Further reading

Related tools