Merge Audio

Joining audio files without the seams showing

Joining audio is conceptually the simplest edit there is — play this, then play that. What makes it awkward in practice is that the files you want to join were almost never recorded under the same conditions, and audio software is unforgiving about mismatches that the human ear would not even notice.

Why joining sometimes refuses to work

Every audio file carries a set of properties that have to agree before two files can be laid end to end: the sample rate (how many measurements per second — 44,100 for CD audio, 48,000 for most video, 16,000 for a lot of voice recorders), the channel layout (mono or stereo), and the internal sample format.

A voice memo at 16 kHz mono and a song at 44.1 kHz stereo are not directly joinable. Naive tools either refuse outright, or produce a file where the second half plays at the wrong speed or comes out of one ear. Neither is a good outcome, and neither is obvious until you listen to the result.

This tool normalises every input first — resampling to 44.1 kHz stereo — before joining. That is why you can drop in a phone recording, a downloaded MP3 and an exported WAV together and get one coherent file out of it rather than an error.

Whether to leave a gap

The gap control exists because the right answer depends entirely on what you are joining.

• No gap — for a recording that was split mid-sentence, or a continuous mix. Any silence would sound like a fault.
• Half a second to a second — between chapters, between an intro and the main content, between separate takes. Enough to signal a boundary without feeling like a mistake.
• Two seconds or more — between genuinely separate items, such as tracks in a compilation, where the listener should have a moment to register that one thing ended and another started.

A useful check: if the join is meant to be invisible, use no gap. If the listener is supposed to notice, give it a beat. Anything in between tends to read as a glitch.

Order matters more than you think

Files rarely arrive named in the order you want them played. recording_1, recording_10, recording_2 is what an alphabetical sort gives you, and it is almost never right. That is why the list here is explicitly ordered and reorderable rather than sorted for you — the number beside each track is what will happen, not what your file system happened to suggest.

Worth checking the last few seconds of each track before you join. Recorders often capture a second or two of fumbling as the stop button is pressed, and four joined files means four of those moments in the middle of your finished audio. Trimming each one with the audio cutter first takes a minute and is the difference between a recording that sounds edited and one that sounds assembled.

Quality when joining

Joining requires decoding each file and encoding the result, so the output is one compression generation newer than the inputs. At 192 kbps that is not something you will hear on speech or on ordinary listening.

If you are joining lossless files, or the result is going into further editing, pick WAV. Nothing is compressed, the file is large, and every sample survives intact. For anything destined to be listened to rather than worked on, MP3 or M4A is the sensible choice.

Nothing is uploaded

Every file you add is read from your disk into your browser's memory, joined there, and written back to your downloads folder. There is no server involved, which is why there is no account, no queue and no limit on how many files you can add beyond what your device can hold.

For joined audio in particular this tends to matter. The recordings people most often need to stitch back together are interviews, meetings, lectures and personal notes — exactly the material you would least want sitting on an unknown server.