Audacity multitrack recording: overdubbing and multiple microphones

Overdubbing means playing back existing tracks while recording a new one. It is how you layer a voiceover over music, or record a second part against a first.

Turn it on in Settings under Recording: tick Play other tracks while recording, which is the Overdub option. Without it, Audacity records in silence and you have nothing to perform against.

Wear headphones. If the existing tracks play through speakers, your microphone records them along with your voice, and you end up with the backing bleeding into the new track where it cannot be removed.

Press Record and Audacity creates a new track beneath the existing ones while playing them back. Shift and Record adds to the current track instead, which is occasionally what you want and usually not.

There is always a small delay between Audacity playing a sound and recording your response to it. The result is a new track that sits slightly late against the others, by anything from a few milliseconds to a noticeable fraction of a second.

Audacity has a latency correction setting under Settings and Recording, but the default is a guess rather than a measurement of your machine.

Measure it once: record a short click, then overdub while holding the microphone against the headphones so it captures that click. Zoom in on both tracks, measure the offset, and enter it as the latency correction. It is a ten-minute job that fixes every overdub afterwards.

If you would rather not, the Time Shift tool lets you drag each new track into place by eye, which is fine for speech and inadequate for music.

This is where people hit a wall. Audacity can only record from one device at a time, and two USB microphones are two devices. Plugging both in and expecting two tracks does not work.

The answer is an audio interface with multiple inputs. A two-input interface presents itself to the computer as one device with two channels, and Audacity records both as a stereo track you then split into two mono tracks.

Set the recording channels dropdown to 2 (Stereo), record, then use the track menu and Split Stereo Track to get one voice per track. From there you can level, clean and edit each person independently.

For more than two, you need a larger interface and Audacity set to that channel count. The principle is the same: one device, many channels.

For an interview with someone elsewhere, Audacity has no mechanism for recording them. Capturing the call audio gives you both voices compressed together on one track, at whatever quality the connection managed.

The traditional workaround is a double-ender: both people record themselves locally in Audacity, then one person collects both files and aligns them. The audio quality is excellent because neither track ever went through the internet.

It also depends on your guest successfully operating Audacity, remembering to press record, and sending you a large file afterwards. In practice that is where double-enders fail.

Tools built for remote recording do this automatically, capturing each person locally in their browser and uploading as they go. Hilite is one of them and we make it, so weigh that accordingly.

Name every track as soon as it exists, through the track menu and Name. Four tracks called Audio Track are unworkable within about twenty minutes.

Turn on Tracks and Sync-Lock Tracks before editing. With it on, a cut on one track applies the same timing change to the others, which prevents the gradual drift that ruins a long multi-person edit.

Use the mute and solo buttons constantly while working, and check every one of them before exporting. A forgotten solo is the most common reason an export contains one voice.

Save the project often. Multitrack projects are larger and slower, and losing an hour of alignment work is a bad way to learn this.