What changed
Shipped in this update
- Harmonic complexity, stereo width, and 3D spatial scoring rebuilt — the old formulas were mathematically capped.
- Median score rises about 13 points; validated on commercial masters and real user mixes.
- Every analysis now carries a scoring version; old analyses render against their original thresholds.
- A percentile line shows where your track sits among all tracks analyzed.
- Score colors and verdicts recalibrated to the real production score distribution.
- Dashboard trends and song timelines never compare across scoring versions.
The uncomfortable number
For three years, Mix Analyzer told you a score of 80 meant "excellent." Here is the number behind that promise: across 14,371 analyses in the last six months, the best score anyone ever reached was 77.5. The median was 61.9. Nobody — including commercial masters — could touch 80.
That wasn't your mixing. It was our math.
What was broken
We audited the scoring engine component by component and found three parts that could not score high for any real audio:
- Harmonic complexity (13% of your score) used a formula whose maximum value for real music sat around 40 out of 100. Every mix on the platform lost 8–11 points here before you touched a fader.
- 3D spatial analysis scored height and depth on a scale where typical stereo mixes landed between 7 and 11 out of 100.
- Stereo width only awarded full marks at one exact mid/side ratio — a knife edge no mastered track sits on.
Add those up and the theoretical ceiling landed almost exactly where the data said it did: 77.5.
What changes for you
Your score now comes with context. Under the score you'll see where your track sits among all tracks analyzed on the platform. "Top 5% of all tracks analyzed" tells you more than any absolute number. A 71.9 under the old system was better than 95% of everything we had ever analyzed — and the UI told that producer to keep chasing 80. That stops now.
Colors and labels match reality. Green used to start at 80, so nobody ever saw green. The color scale, the verdict labels, and the dashboard messages are now calibrated to the real score distribution.
Scoring v2 fixes the broken components. Harmonic complexity uses a proper entropy measure, stereo width rewards a healthy range instead of a single point, and the 3D spatial scores use maps derived from real stereo material. We validated the new engine on commercial masters and real user mixes: the median score rises about 13 points. A mix that scored 62 last week scores around 75 now — same mix, fairer ruler.
Your old analyses keep their meaning
Every analysis is stamped with a scoring version. Your history renders with the thresholds it was scored under, and your dashboard trend never compares old-formula scores against new-formula scores — so the change can't show up as fake improvement. The same goes for the new song timelines: the progress delta is always computed within one scoring scale.
When you want your best tracks re-scored under v2, just re-analyze them.
Why we're telling you this
A scoring tool is only useful if you can trust it. We could have quietly raised everyone's numbers and called it an upgrade. Instead you get the audit: what was broken, what it cost you, and how the new scale is anchored. The v2 thresholds are marked provisional in our own code until two weeks of production data confirm them — and when we re-baseline, we'll do it in the open, the same way.
Questions about your score history? Write to hello@mixanalytic.com.