Skip to content
Euphona

Free tool

Stereo width analyzer

Drop in a track and read its stereo image: phase correlation, width, balance and what actually happens when it's folded to mono — overall, and in the five bands where width decisions get made.

Nothing is uploaded. The file is decoded and measured by your own browser; this page has no endpoint to send audio to. It works offline once loaded, with no run limits — there's nothing on our side to limit.

Drop an audio file, or choose one. It's read on this device and never sent anywhere.

If the image needs work

Width is a mix decision that mastering has to live with

One contextual next step, offered after your result — never in front of it.

A width problem found here is treated at the mastering stage as far as it can be — Euphona’s AI mastering measures the image per band, applies stereo treatment with the reasons attached, and checks every render as a mono fold-down, because most listeners will hear that version. The free plan runs it on your own track, level-matched against the original.

See how AI mastering works

The numbers

What these numbers mean

Correlation, width, mono loss — three different questions, usually conflated.

Stereo width is how much of a track’s energy differs between the channels; phase correlation is whether that difference adds or cancels; mono compatibility is what survives when a phone speaker, a club system or a Bluetooth pairing sums the two channels into one. A mix can be wide and mono-safe, or narrow and broken — which is why this page measures all three instead of one.

Correlation
+1 means the channels move together (mono); around 0 means they are independent (wide); below 0 means they oppose — and opposing content loses level or vanishes in a mono sum. The worst-window figure matters more than the average: one anti-phase chorus is a problem the whole-file number can hide.
Width in dB
The side signal’s energy against the mid signal’s. −∞ is dual mono; around 0 dB, side and mid carry equal energy. There is no correct value — the number exists so two mixes, or two revisions of one mix, can be compared on something firmer than impression.
Mono sum
The level change, in dB, when the channels are averaged — the fold-down every single-speaker device performs. Uncorrelated width costs about −3 dB and sounds fine; anti-phase content costs far more, and what it costs is selective: whichever instrument was widened that way is the one that disappears.
Why per band
Because width decisions are band decisions. Wide bass falls apart on summing systems; wide highs are usually free. The band table localises a correlation problem to the register it lives in, which is the difference between "something cancels" and knowing what to fix.
Read why low end stays mono

Verification

How to check this tool

The correlation figure is the same normalised cross-correlation every hardware phase meter has shown since the goniometer era; width and mono loss come from the plain mid/side decomposition. All of it is checked on every build against constructed signals: identical channels read +1 and lose nothing in mono, an inverted channel reads −1 and cancels, independent noise reads 0 with the textbook −3 dB mono loss. Play the file into any correlation meter you trust and compare.