Stereo Width by Genre: What 152,737 Measured Tracks Show

On the median commercial release, the side channel sits about 10 dB under the middle. Across 152,737 tracks measured by our Sonic Analyzer, the median side-to-mid ratio is 0.30 and the median stereo correlation is 0.83. Classical and score run widest at about 6 dB under, metal about 7, and hip-hop narrowest at about 14 dB under. That gap survives when I compare tracks with the same amount of low end, so it isn't just "bass is mono, bass-heavy genres look narrow." It's how the records are arranged.

The question I get is almost always "how do I make my mix wider," and it usually comes with a stereo widener already on the master. After 20 years, 19 Billboard Top 20 albums and records past 3.3 billion streams, my honest first answer is that most records are less wide than you'd guess. So I pulled the numbers to show what wide actually means, genre by genre.

What the Numbers Mean

The analyzer splits the loudest four seconds of each track (usually the chorus) into a mid channel, what both speakers share, and a side channel, what's different between them. Stereo width here is side level divided by mid level. A correlation meter, the one most DAWs ship with, reads the same thing from another angle. This table translates between them, from the tracks themselves:

Half of everything sits between 0.20 and 0.42. Only 6.5% of tracks go past 0.60, and only 0.53% (817 tracks) read below zero correlation, the zone where parts of the record cancel when they're summed to mono. If your meter is down there in the chorus, that's not a vibe. That's a phase problem.

Your Lane's Lines

Here's every lane with at least 1,000 measured tracks. "Narrow line" is where the narrowest tenth of the lane sits, "wide line" is where the widest tenth starts. If you're outside those two numbers, you're doing something most of your genre doesn't.

The split is clean. Guitar music and anything recorded in a real room sits on the wide side. Groove music built around a centered kick, bass and vocal sits on the narrow side. A hip-hop record at 0.40 is at the wide edge of hip-hop. A metal record at 0.40 is narrower than most metal. Same number, opposite verdict, which is why "make it wider" isn't advice until somebody tells you the genre.

Ambient and classical have the biggest spreads (wide lines at 0.79 and 0.84), and that makes sense to me. Those are the lanes where reverb and real rooms carry the record, and a hall is mostly side information.

It's Not Just the Bass

The obvious objection: bass is almost always centered, so a bass-heavy record will read narrower even if the top is wide. That's partly true. Sorting all 152,737 tracks by how much energy sits below 250 Hz, the median width falls from about 0.37 in the leanest tenth to 0.23 in the heaviest.

But the genre order doesn't come from that. Taking only tracks with 15% to 20% of their energy below 250 Hz, the same low end across the board, hip-hop still measures 0.22 and metal 0.45. Dance sits at 0.28, pop at 0.35, rock at 0.38, classical at 0.45. Almost exactly the order in the big table.

My read is that it's arrangement. Metal and punk are double-tracked guitars panned hard left and right, so the wall of sound literally lives in the sides. Hip-hop puts the vocal, the 808 and the snare in the middle and builds the record around them. Nobody turned a width knob to get either result.

The Mono Dance Track Thing

You hear that club tracks get mixed in mono. In this data, they don't. Only 1.29% of dance and electronic tracks measured close to mono (under 0.05, side more than 26 dB down), lower than the 1.72% across all music. Dance is narrower than pop and rock, but these are stereo records. The lane with the most near-mono tracks is actually hip-hop, at 2.69%.

How I Actually Get Width

I pan left, right and centre, almost nothing in between, and the lead vocal is always centre. My background stacks are five voices: two hard left, two hard right, one in the middle under the lead. That's where most of the width on my records comes from: real performances, different on each side, not a plugin pulling one performance apart.

For mono I don't use a mono button. I put a pair of Avantone Mix Cubes right next to each other, touching, which gives me the mono picture with the depth still there. If the chorus falls apart on those, the width was fake. It matters more than people think, because a lot of first listens happen on a phone speaker, which I wrote about in why your music sounds bad on social media.

Three Checks on Your Own Mix

1. Measure the loudest section, not the intro. Verses and intros are often narrower or wider than the chorus. Our numbers come from the loudest four seconds, so compare like with like.
2. Find your lane's row. If you're past the wide line, check what's getting you there. A widener on the master usually shows up as a correlation reading under 0.5 and a vocal that gets quieter in mono.
3. Fold to mono before you release. If anything important drops out, fix the part, not the master. That's on my is-my-mix-ready-for-mastering checklist for a reason.

The tonal side of the same picture is in the tonal balance by genre study, and loudness is in average LUFS by genre.

Method and Catches

Every number comes from audio captured and measured by the Sonic Analyzer pipeline at Freshly Baked Studios. Stereo measurement was added to the pipeline after it started, so 152,737 of the 354,630 tracks we've analyzed carry it. I dropped 740 rows where the stereo capture failed and the extractor wrote its placeholder (width 0, correlation 1). Width is RMS of the side channel over RMS of the mid channel on the highest-energy 4-second sample of three taken across each track, from the streaming version of the record, so it includes the effect of lossy encoding. Correlation is the Pearson correlation of left and right on the same sample. Genre lanes come from the analyzer's own classification, with regional and case variants merged into the 17 lanes above (130,437 tracks), and anything without a clear lane left out of the lane table but kept in the overall numbers. This is a description of released records, not a rule. Wide or narrow doesn't make a record better, and nothing here says it does.

If you want these numbers on your own file instead of my table, run it through the Sonic Analyzer. It's free. And if the answer is that the mix needs someone else's hands, here's what mixing costs and what mixing and mastering together cost, itemized honestly, alongside what a producer costs if the real fix is the arrangement. Or skip the ranges and get a number calibrated to where you actually are as an artist from the rate calculator.