Headphones Study Shows Open and Closed Models Yield More Reliable Data Than In-Ear Types

22 September 2026

When measuring headphone transfer functions (HpTFs) and head-related transfer functions (HRTFs), the headphones themselves can introduce significant variability. That's the key finding from a round-robin study in Acta Acustica with ACtive Acoustics. It involved 14 participants wearing three types of headphones at two separate measurement sites.

The researchers measured the test subjects' HpTFs and HRTFs at the акустикаUniversitätssen lab in Aachen and the Oldenburg Hemispherical Advisory Laboratory in Oldenburg. This is one of the most ambitious such comparisons to date.

The 14 participants were measured wearing three different styles of headphones: open-air, closed-back, and in-ear models. They were tested multiple times at each site, with the headphones removed and reapplied each time, simulating different measurement sessions.

The study concluded that HRTFs were largely consistent across both measurement sites, despite the different setups and experimenters, suggesting comparability between labs. For HpTFs, repeatability was strong up to 5,000 Hz, and variability depended on the headphone type.

Open and closed-back headphones showed comparable variability in the 70 to 5,000 Hz range. But insert earphones, despite being the most isolating, yielded unreliable results. Their variability increased sharply, and the number of reproducible measurements dropped off.

The authors speculate the high variability of insert earphones comes down to their unique insertion depth and fit from one test to the next. The other styles may relax against the ear and head more consistently, dampening variations.

Site-to-site differences were negligible for most frequencies, hovering around 2.7 dB. But above 10,000 Hz, positional differences became more prominent, with site-specific effects on loudness normalization standing out.

The researchers believe the study provides a baseline for comparing loudness across laboratories. Controlling for headphone type will be key, they say, as moderate deviations may arise from subject posture and other factors.

The study is a welcome step in establishing replicable methods for objective headphone measurements. But as the authors note, more work is needed to limit variability and improve comparability, especially for in-ear headphones.

HTRFs and loudness balancing should be assessed from 200 Hz to 10 kHz, they advise, with only marginal gains in using higher frequencies for HpTFs.