Earphones Target Curves - what they measure and limitations

This is a frequently discussed (and debated topic), so we decided to gather some information for those who are completely new to this.

Earphone (and headphone) target curves are reference frequency-response shapes that manufacturers, reviewers, and EQ tools aim for. They represent a preferred or “natural-sounding” tonality when measured on standardized artificial ears/couplers, rather than a purely flat electrical response.

Because earphones and headphones sit on or in the ear, they interact with the pinna, ear canal, and head in ways loudspeakers do not. A raw “flat” measurement at the eardrum therefore does not sound neutral to listeners. Target curves attempt to capture what most people prefer after accounting for those interactions and for typical listening preferences.

Image from Inear Fidelity

The Harman Target Curve

The best-known modern preference target is the Harman target (sometimes called the Harman curve or Harman preference curve). It was developed at Harman International (home of JBL, AKG, etc.) primarily by Dr. Sean Olive, Todd Welti, and colleagues through a series of controlled, double-blind listening tests beginning around 2012–2013 and refined in later papers.

The core idea is straightforward:
Music is mixed and mastered for loudspeakers in rooms. A good pair of headphones should produce a similar perceived tonal balance at the eardrum as a well-designed loudspeaker system playing in a typical domestic or studio listening room.

Harman measured the in-room response of high-quality loudspeakers (flat on-axis anechoic response) in their reference rooms, then translated that to headphone/earphone measurements made on an artificial ear. They further refined the shape through preference tests in which listeners could adjust bass and (in later work) other bands, and rated different tunings of the same transducers.

Key characteristics of the widely referenced versions (Harman 2013/2018 for over-ears; Harman IE 2017/2019 for in-ears) typically include:

• A bass shelf or boost (roughly +3 to +6–8 dB below ~100–150 Hz, with the exact amount and Q varying by version and format). This partly compensates for the lack of room gain that loudspeakers receive and matches average listener preference.
• A relatively neutral midrange with a characteristic “ear-gain” rise in the upper midrange/lower treble (roughly 1–5 kHz region) that approximates the natural amplification of the outer ear.
• Controlled treble that avoids both excessive brightness and dullness, often with a gentle roll-off or specific peaks/dips depending on the measurement fixture and iteration.

There are distinct curves for circumaural (over-ear) headphones and for insert earphones/IEMs because the acoustic coupling and residual ear-canal resonance differ.

Why the Harman Target Works for Many Listeners

1. Empirical preference data
   Large-scale blind tests (hundreds of listeners across age groups, experience levels, and countries in later studies) showed that a majority of people preferred headphones tuned close to the Harman shape over most commercial products of the time. Roughly 64 % of listeners in one major analysis fell into a “Harman lovers” group; smaller groups preferred more or less bass.
2. Psychoacoustic grounding
   It approximates the eardrum response of good loudspeakers in a real room (including room gain at low frequencies and the diffuse-field-like contributions of reflections) while remaining free of the room’s reverberation and crosstalk. This helps break the “circle of confusion” in which recordings and playback systems reinforce each other’s idiosyncrasies.
3. Practical utility
   It gives designers and reviewers a common, research-backed reference. Products that measure close to it tend to sound balanced across a wide range of genres, with good vocal clarity and neither thin nor excessively boomy bass for the average listener. It is far more useful than older abstract targets such as pure free-field or diffuse-field equalization, which often sound too bright or thin when applied to music.
4. Predictive power
   Preference scores derived from how closely a headphone’s measured response matches the target have correlated reasonably well with actual listening preference in multiple studies.

Single-driver IEMs are closer to the ideal “FR tells you almost everything” situation.

Because a single driver operates as an approximately minimum-phase system, its magnitude response, phase response, and time-domain behavior are tightly linked: the frequency response largely determines the group delay and impulse response through the Hilbert transform relationship.

As a result, when you equalize a well-behaved single-driver IEM to a target curve using ordinary minimum-phase filters, you simultaneously correct both the tonal balance and the associated phase/group-delay deviations. This makes frequency-response graphs and target-curve matching far more predictive of the final sound than they are with multi-driver designs, where crossovers and differing driver technologies introduce independent non-minimum-phase effects that the FR plot alone cannot fully reveal.

Limitations of the Harman Target (and Target Curves in General)

No single curve is universal, and the Harman work itself acknowledges important caveats:

• Individual differences dominate. Ear anatomy (pinna shape, ear-canal length and volume), hearing sensitivity, age-related high-frequency loss, and personal taste vary widely. A curve preferred by ~64 % still leaves a substantial minority wanting more bass, less treble, or different midrange balance. Fit and seal (especially critical for IEMs) further change the effective response at the eardrum.
• Measurement-fixture dependence. The original public Harman curves were developed on specific artificial-ear setups (modified GRAS fixtures). Newer, more anatomically accurate heads such as the Brüel & Kjær 5128 produce different raw measurements, so the same physical earphone can look different relative to a target depending on the rig. Translation between fixtures is imperfect, and Harman has not fully published all modern equivalents.
• Preference is not the same as absolute accuracy. The curve incorporates both physical factors (room gain, ear gain) and average listener preference (especially the bass boost). Some critical listeners or specific genres may prefer less bass emphasis or a different treble contour. The IEM versions have been criticized by parts of the enthusiast community for a relative mid-bass dip and elevated 3–8 kHz region that can sound thin or bright to some ears.
• Frequency response is only one dimension. Target curves ignore distortion (harmonic, intermodulation, intermodulation, compression), phase/group-delay behavior, channel matching, spatial cues (or the lack of them), sealing variability, and the effects of active noise cancellation or DSP. Two earphones can match the target equally well yet sound quite different for these other reasons.
• Sample and methodology constraints. Early studies used relatively small numbers of trained listeners and a limited set of products; later work expanded the pool but still relied heavily on Harman employees or specific demographics in some phases. Listener preference can also shift with background noise, listening level (Fletcher-Munson/equal-loudness contours), and music material.
• Not a “correct” or final answer. Other research groups and companies have proposed house curves (SoundGuys, GearLab, Knowles, etc.) that differ in bass shelf height, treble tilt, or midrange contour. Diffuse-field or free-field based targets remain useful for certain technical purposes even if they are less preferred for casual music listening.

What Target Curves Are Good For

• Giving a quick, objective starting point for tonality.
  An IEM that tracks Harman IE 2019, JM-1 / New Meta, or IEF 2025 is statistically more likely to sound balanced to a large percentage of listeners than something wildly off-target.
• Helping you filter out obvious disasters (extreme V-shapes, heavily veiled midrange, piercing treble peaks, etc.).
• Providing a common language when reading reviews or comparing graphs on Squiglink / Headphones.com / Crinacle databases.
• Guiding EQ later if you want to nudge something closer to a preferred shape.

Practical Takeaways

Target curves, and the Harman target in particular, are extremely useful tools: they provide a research-backed, majority-preferred baseline that has improved the average quality of commercial earphones and headphones. Matching or intelligently departing from the Harman shape is a rational starting point for design and EQ.

They are not gospel. The best earphone for any individual is the one whose measured response, after accounting for personal fit and anatomy, aligns with that person’s preference and the music they listen to. Tools such as parametric EQ, measurement databases (with consistent fixtures), and, ultimately, critical listening remain essential. As measurement technology (e.g., better HATS) and preference research evolve, the exact shapes of popular targets will continue to be refined.

A Practical Approach for Buying IEMs

Use target curves as a filter and communication tool, not the final judge:

1. Decide your general preference first (e.g., “I like Harman-ish with a bit more sub-bass,” “I prefer New Meta / slightly tilted DF,” “I want neutral with strong technical performance”).
2. Look at graphs only as a rough screening tool.
3. Prioritize well-regarded reviews from people whose tastes and hearing seem close to yours (and who measure on consistent fixtures).
4. Pay attention to fit, isolation, comfort, and build — these affect daily enjoyment more than a 1–2 dB midrange difference.
5. Whenever possible, try before you buy (or buy from places with good return policies).
6. Accept that EQ is a legitimate and often superior solution. A well-measuring but imperfect IEM + good parametric EQ frequently beats a “perfect-target” IEM that has other flaws.

We will talk about how Frequency Response curves are measured in later articles.