DriveSilk

Record a real drive with your phone

The DriveSilk iOS app listens while you drive. It captures tire-related sound levels, vehicle speed, and rough road context — not a studio mic, but a consistent phone sensor you already have. Short loops and long commutes both count.

Tag the setup: tire name + pressure

A single drive is noisy noise. What matters is the setup: which tires, at what pressure, on which car. You label tire name and bar pressure (and optionally vehicle). Near-identical pressures (±0.05 bar) are treated as the same setup so small gauge differences don’t split your data.

Sync automatically — no spreadsheet required

Finished drives sync to the DriveSilk cloud. The public lab can browse setups and compare them. Your phone stays the recorder; the website is where community data becomes comparable science.

Pool many drives into one setup dataset

One rainy Tuesday isn’t enough. DriveSilk merges every drive that shares the same tire + pressure into one dataset. More kilometers → more stable averages and less “I got lucky with that stretch of asphalt.”

Slice the world into comparable situations

Raw “average dB for the whole trip” is unfair: highway vs city, 40 vs 120 km/h. We bin every sample by speed band (0–30, 30–50, … 120+) and road type (city / mixed / highway). Speech-heavy moments are skipped so conversation doesn’t look like tire roar.

Only compare where both setups have data

Setup A and Setup B are compared cell-by-cell — only in situations both have covered. That’s how “quieter by ~1.2 dB on matched highway 90–120” means something, instead of one driver living on motorways and the other only commuting in town.

Read the verdict: quieter setup + pressure & speed tips

The lab reports which setup wins overall on matched situations, the margin in dB, charts of noise vs speed, and plain-language insights — e.g. whether dropping 0.1 bar helped on mixed roads, or which speed band shows the biggest gap.

Same car, different tires — that’s the magic

Suspension, insulation, and wheel size dominate cabin noise. Comparing a soft EV hatch to a sports coupe is mostly noise about the cars. DriveSilk gets powerful when drivers of the same model (or very close) contribute setups: Michelin vs Conti on a Model 3, summer vs all-season on a Golf, 2.3 vs 2.5 bar on your Exact Same Car.

Tag your vehicle carefully. The more Model Y / Octavia / XC40 owners sync, the more useful “community average for this car” becomes.

Why contribute your data

• Personal answer — which of your tire sets is quieter at the speeds you actually use.
• Pressure tuning — small PSI changes with evidence, not forum myths.
• Community baseline — see how your setup sits vs other drivers of the same car.
• Next purchase — when it’s time for new rubber, buy with measured expectation, not brochure claims.

Technical methodology (short version)

1. Align tire dB series with speed and geo samples per timestamp.
2. Drop frames with high speech likelihood (≥ 0.45) when available.
3. Assign each sample to a speed bin × road-type situation.
4. Merge all drives in a setup: concatenate series, then re-bin.
5. For two+ setups, compute per-situation means only where every setup has samples.
6. Aggregate dimensions (mean dB, variability, persistence) and chart noise vs speed.
7. Recommend overall quieter setup and surface pressure / speed insights in plain language.

This is observational field data, not a certified ISO cabin test. Phone placement, windows, and road surface still matter — pooling and situation matching reduce bias; they don’t erase it. Treat results as comparative evidence, not absolute lab truth.