Vocal & Mix Clinic · Lesson 2 of 2

The Squeeze Test: Measure the Phone Gap

Everyone masters for the studio and gets heard on a phone. This diagnostic makes the gap a number — measured on our own loop, with a method you can run on yours in ten minutes.

Time 9 minTools Any DAW + a loudness meterLevel Intermediate

Waveform — the squeezed render (B) of this lesson’s peak-matched pair

Measured — the squeeze’s gain: studio vs phone model

5.68 → 5.01 LU

Same peaks, same loop: squeezing buys 5.68 LU on a full-range system and 5.01 LU through a phone model. The missing 0.67 LU is low-end density the phone can’t reproduce. Measure your own gap before you decide how hard to squeeze.

Step 1

Make the pair: same loop, one squeeze, peaks matched

Take a loop of your own track. Render it straight. Then render it again through the loudness treatment you’re considering — bus compression, saturation, a limiter, whatever your "make it loud" move is — and match the two renders’ peaks exactly. Peak-matching is the whole method’s honesty: with peaks equal, any loudness difference left is real density, not level. The pair below is ours: a two-bar loop at 124 BPM, rendered straight, then squeezed and peak-matched back.

▶ Hear it — our peak-matched pair (A straight, B squeezed)

Step 2

Measure the gap twice: full-range, then phone

Now measure integrated loudness (LUFS-I) on both files. Ours: A reads -12.66 LUFS, B reads -6.98 — the squeeze buys 5.68 LU on a full-range system. Then run both through a phone model — mono sum + a 500 Hz high-pass, which is most of what a phone speaker does to your mix — and measure again: A reads -18.85, B reads -13.84. The gap is now 5.01 LU.

Step 3

Read the result: 0.67 LU of the gain never reaches the phone

On our loop, 0.67 LU of the squeeze’s 5.68 LU advantage evaporates through the phone model — because a share of the added density sits in the low end, and the phone throws the low end away. That doesn’t make the squeeze wrong. It makes it a decision with a price you can now see. The method travels: run it on your own loudest chorus before you print, and choose the squeeze that survives the speaker your listeners actually use.

Sourcing, plainly: this diagnostic circulates from a working engineer’s published test (via @arieldahanaudio). His test’s figures belong to his test and aren’t printed here — every number on this page was measured on our own loop, with the same procedure, in our lab.

Try it in the browser — free, nothing to install

Beat MakerThis grid, loaded for you — a house loop like the one this test runs on on the site sequencer, free packs’ sounds Pattern PlayerThe free MIDI patterns in three voices — the Listen tab

Check your work

Failure check

Common failure: comparing at unmatched peaks

Skip the peak-matching and the "gap" you measure is mostly just level — the louder file wins every comparison and teaches you nothing. Peaks equal first, always. It’s the same rule as every A/B in this Learn section: one variable changed, or the measurement is theatre.

The companion measurements

This diagnostic pairs with True Peak vs Sample Peak — the other half of loudness honesty — and the Mono Fold Test, which our phone model borrows its mono sum from. No product card here: the Clinic sells nothing yet, and a diagnostic you can run for free shouldn’t pretend otherwise.

Keep learning

Vocal & Mix Clinic · Previous

Anatomy of a Hit Vocal Chain

Clinic · 1 of 2 · the documented chain

The chain a hit vocal rides in on — and the doctrine that makes it transferable.

Previous lesson

Low End Foundations

The Mono Fold Test — Measure What the Club Eats

Any DAW · two numbers

Sum to mono and measure what survives — the phone model’s parent test.

Read the how-to