We don’t tune by ear. We tune by data — but data from humans, not just machines.
The protocol
Each prototype goes through blind A/B testing with 200+ listeners across 6 countries. We send two units with identical appearance but different tuning. Listeners rate them across 12 dimensions without knowing which is which.
Why blind
Humans are terrible at objective evaluation when they can see what they’re judging. A unit that “looks more premium” consistently scores higher in sound quality — even when the tuning is identical.
Blind testing strips away bias. The data tells us what actually sounds better, not what we think should sound better.
What we learned
- Listeners in different countries prefer slightly different bass response (3-5 dB variance)
- High-frequency preference is universal
- ANC quality matters more than raw sound quality for 60% of users