Stereo to Mono Converter

How the Stereo To Mono works

Stereo to mono: what to know first

Converting stereo to mono sounds like it should be a single sum, and most of the time it is, but the interesting cases are the ones where a naive sum goes wrong. The default method here averages the two channels — the sum divided by two — which preserves everything that is common to both and keeps the level where it was. That is the correct choice for the large majority of material: speech recorded on two microphones, a stereo music mix that you need as a single channel for a phone call, a voice note, a podcast bed.

Why the correlation number matters. The readout gives the correlation coefficient between the two channels, from +1 to −1. A value close to +1 means the channels are nearly identical and mono conversion will be essentially free. Around zero means they are largely independent — a wide mix, or two microphones on different sources — and you will lose the sense of space but nothing else. A negative value is the warning sign: it means significant material is out of phase between the channels, and averaging will cancel it, sometimes to near silence. That happens with a miswired cable, a flipped microphone polarity, some cheap "stereo widening" plugins, and a surprising number of old vinyl rips.

When to take one channel instead. If the correlation is negative, or if one channel holds the good take and the other holds a room microphone, an interview guest, or nothing at all, taking a single channel is the honest fix. Left only and Right only copy that channel through untouched, which is also the right move for a file that was recorded mono but saved as dual-mono stereo — the two channels are identical, so there is nothing to average.

The difference method is a diagnostic. Subtracting one channel from the other cancels everything the two channels share and leaves only what differs. On a normal mix that is the reverb, the wide synths and the stereo ambience; on a well-centred vocal it is close to silence. It is a fast way to hear whether a file is genuinely stereo or a fake widened mono, and it is the classic crude vocal-removal trick — crude because it also removes the bass and the kick, which are usually centred too.

Normalising, and when not to. With normalising off, the output is scaled only if the downmix would clip, which can happen when both channels are already near full scale. With it on, the peak is lifted to just under full scale, which is convenient for a single file and a bad idea across a set of files you want to stay level-matched with each other. Downmixing also loses roughly 3 dB of apparent loudness on wide material even when nothing clips, simply because the decorrelated parts no longer add.

What you get out. A 16-bit mono WAV at the source file's own sample rate, so nothing is resampled and no generational loss is introduced beyond the bit depth of the container. That halves the size of a stereo WAV immediately, and it is the format every editor, transcription service and embedded player will accept without complaint.

Going the other way: Mono to Stereo

A mono file has one channel; a stereo file has two. Many tools, players and platforms expect two, and some refuse anything else, so converting is a common chore. The question is what to put in the second channel.

Duplicate. The safest answer is the same audio on both sides. The sound sits dead centre, exactly as a mono file would play through two speakers, and it sums back to mono perfectly. This is what to choose for speech, for a file that has to meet a "stereo only" requirement, or for anything that will later be mixed with other material. The file doubles in size and gains nothing audible, which is precisely the point.

Haas widening. Delaying one channel by a few milliseconds makes the ear perceive width without hearing an echo, an effect described by Helmut Haas in the 1940s. Between about five and thirty milliseconds the two channels fuse into one wide sound; above forty the delay becomes an audible slap. The default of fifteen is a good middle. It works well on guitars, synths, ambience and sound effects. The cost is mono compatibility: when a widened file is summed back to mono, the delayed copy causes comb filtering that thins the sound. Do not use it on a voice that will be played through a single speaker, such as a phone.

Balance. The slider tilts the result left or right by attenuating the opposite channel, using a constant-power curve so the overall loudness stays steady across the range. At the extremes the audio is entirely on one side. Leave it at zero for a centred file.

Already stereo? If a stereo file is dropped in, the page uses the average of its two channels as the mono source and then applies the chosen method, which is a quick way to re-centre or re-widen an existing stereo file.

Output. The exported file is always two channels at the source sample rate.

Related tools