How to Read Frequency Response Charts: A Beginner's Guide

Frequency response charts show up everywhere in audiophile spaces: on ASR, on Crinacle’s IEM database, in every review on Head-Fi and Resolve Reviews. Three years in, I remember staring at my first graph for the Sennheiser HD600 and having no idea what I was actually looking at. The x-axis, the y-axis, the wiggly lines going up and down, the colored overlays for different units. It looked like science. It was science. But it was also readable, once someone explained the basics.

This guide breaks down how to read a frequency response chart from first principles, with no assumptions about your background. If you want broader context on source chains, headphone categories, and audio fundamentals, the Audiophile Basics hub has you covered.

What a Frequency Response Chart Actually Measures

A frequency response chart shows how loud a device (headphone, speaker, IEM, DAC) reproduces each frequency across the audible range. The horizontal x-axis represents frequency in hertz, running from low frequencies on the left (bass, sub-bass) to high frequencies on the right (treble, air). The vertical y-axis represents amplitude, or volume level, typically measured in decibels (dB).

The line you see on the chart traces how the device performs across that frequency spectrum. A perfectly flat line would mean the device reproduces every frequency at exactly the same relative level, nothing louder, nothing quieter. In practice, no transducer achieves that, and even if one did, “flat” does not always translate to “sounds balanced” to human ears, for reasons related to how we actually perceive sound.

The X-Axis: Frequency Range and What Frequencies Mean

The standard audible range for healthy human hearing runs approximately from 20 Hz to 20,000 Hz (20 kHz). Most charts display this full range, though some zoom in on particular regions.

Here is a rough map of what those frequency regions correspond to perceptually:

• Sub-bass (20, 60 Hz): The deepest rumble in electronic music, pipe organ fundamentals, bass drum body. Hard to hear on most headphones without a seal; felt as much as heard on speakers.
• Bass (60, 250 Hz): Kick drum punch, bass guitar body, the warmth in male vocals and cellos.
• Midrange (250 Hz, 2 kHz): The most critical region for intelligibility. Vocals, guitars, pianos, most instruments live here. Errors in this range are immediately audible.
• Upper midrange (2, 5 kHz): Presence region. A boost here adds clarity and forward attack. A cut makes things sound recessed or veiled.
• Treble (5, 10 kHz): Consonant clarity, cymbal definition, detail and air in strings.
• Air (10, 20 kHz): Spatial extension, shimmer, the sense of “openness” in a recording. Individual sensitivity in this range varies significantly with age.

When I look at the Crinacle graph for a new IEM, the first thing I check is the upper midrange (2, 5 kHz). That region tells me more about perceived tonality than any other part of the chart.

The Y-Axis: Decibels and What the Numbers Mean

The y-axis is measured in decibels. Most charts display a range of roughly 20, 30 dB total. A deviation of 3 dB is perceptible to most listeners. A deviation of 10 dB is dramatic. When you see a headphone’s bass shelf that rises 6, 8 dB from the midrange baseline, that is a meaningful, audible warmth tilt, not a subtle coloration.

One important note: the absolute position of the chart on the y-axis is often arbitrary. Most measurements are normalized at some reference frequency (commonly 1 kHz). What matters is the relative shape of the curve, peaks and dips relative to other regions, not whether the line sits at 0 dB, 70 dB, or 94 dB on the scale.

Target Curves and Why “Flat” Is Not the Goal

This is where a lot of newcomers get confused. Headphones do not aim for a flat frequency response because of how human hearing works in a real room versus in headphones.

When sound reaches your ears from speakers in a room, your outer ear (pinna), head, and torso all shape the sound in predictable ways before it hits your eardrum. That shaping is called the Head-Related Transfer Function (HRTF). In headphones, sound bypasses most of that processing because the driver sits directly against your ear canal. To compensate, headphone makers tune toward target curves that introduce the shaping your pinna and room would have added.

The most widely referenced target is the Harman target curve, developed by Sean Olive and colleagues at Harman International through large-scale listener preference testing. The Harman target has a bass shelf (elevated bass relative to midrange), a presence rise around 3 kHz, and a gradual treble roll-off. It is not the only target, and not everyone prefers it, but it is the baseline most reviewers measure against.

Other targets you will see referenced include the Diffuse Field (DF) target, which aims to replicate an anechoic multichannel speaker presentation, and the Free Field (FF) target, which approximates sound arriving from a single frontal speaker. Crinacle’s “Neutral” target and Resolve’s preferred tuning reference differ subtly from Harman, and those differences matter for how a headphone sounds.

The practical point: a “flat” frequency response on a headphone graph would actually sound thin, bass-light, and unnatural. You are looking for a shape, not a ruler-straight line.

Reading Peaks, Dips, and Shelf Shapes

Once you understand the axes and target curves, interpreting specific features of a graph gets more intuitive.

Peaks and Resonances

A sharp, narrow peak on a frequency response chart usually indicates a resonance. In the upper midrange or treble, these are often called “peaks” or “spikes” and can cause listening fatigue. A 5 dB spike at 8, 10 kHz, for example, is frequently associated with that harsh, piercing treble quality that some headphones are criticized for.

Width matters as much as height. A narrow peak (affecting a small range of frequencies) can be more fatiguing than a broader, shallower rise, even at the same measured height, because it concentrates energy in a very specific tonal region. EQ can address narrow peaks effectively.

Dips and Recessions

A dip below the target curve in the upper midrange (2, 5 kHz) typically produces a “veiled” or “recessed” presentation. Vocals sound distant. Detail feels buried. Some listeners find this relaxed and non-fatiguing; others find it unsatisfying over long sessions. The original HiFiMan Sundara (which I own in the 2020 revision) has a mild dip around 3 kHz compared to Harman that contributes to its slightly relaxed midrange character.

Deep dips are harder to fix with EQ than peaks, particularly if they result from destructive interference in the driver or chamber design. Adding gain at a frequency where the driver is acoustically canceling itself does not always produce clean output.

Bass Shelf vs. Bass Hump

A bass shelf is a gradual, consistent elevation of bass frequencies below roughly 150, 200 Hz. This is what the Harman target prescribes. It sounds like added weight and warmth without muddying the midrange, assuming the transition is smooth.

A bass hump is different: a localized peak in the mid-bass region (roughly 100, 200 Hz), often narrower, that can produce a “boomy” or “muddy” quality. Bass humps bleed into the lower midrange and obscure instrument separation. On a chart, you can spot the difference by how quickly the elevation drops off as it moves toward the midrange.

How to Compare Multiple Frequency Response Plots

Many measurement databases, including ASR and Crinacle’s graph tool, allow overlays so you can compare multiple headphones or IEMs on the same axes. This is extremely useful, but there are important caveats.

Normalization and Reference Points

Before comparing two curves, check how they are normalized. If one curve is normalized at 1 kHz and another at 500 Hz, they will appear offset even if the headphones sound similar. Most serious databases normalize consistently, but it is worth verifying. The shape of the curve relative to its own midrange is what you are comparing.

Measurement Rig Differences

Different labs use different measurement rigs, different ear couplers, and different fixture geometries. A headphone measured on a GRAS 43AG rig at ASR will not look identical to the same headphone measured on an industry-standard IEC 60318-4 coupler. The broad strokes will align, but treble readings above 6, 8 kHz are particularly sensitive to rig variation. This is why I use multiple sources: ASR data for DACs and electronics, Crinacle’s database for IEMs, and Resolve or Headphones.com for full-size headphones.

Unit Variation

Even within the same headphone model, two units from the same production run can measure slightly differently, particularly in the upper treble. Driver matching is better in some price brackets than others. When a reviewer notes a “treble peak at 7 kHz,” it is worth checking whether other measurements of the same unit confirm it, or whether that is specific to their sample.

How Frequency Response Relates to What You Actually Hear

Frequency response explains a lot of what we perceive as sound quality, but it does not explain everything. Distortion, driver speed, soundstage, imaging, and timbre all contribute to perceived audio quality in ways that frequency response alone does not capture.

That said, three years into this hobby, my working rule is: frequency response explains the majority of tonal character differences between headphones. If two headphones have near-identical measured frequency responses and similar distortion profiles, they will likely sound more similar than different. This is a testable, defensible position supported by both psychoacoustic research and the practical experience of blind testing communities.

Where frequency response does not explain differences: very low distortion levels, extremely subtle staging cues, and the rare cases where driver materials and crossover implementations produce genuinely audible textural differences at equivalent tuning. At my experience level, I hold those claims cautiously. I trust the graph first and evaluate from there.

Buying Guide: Portable Audio Sources and Why Measurements Matter

If you are reading frequency response charts, you are also making decisions about source hardware. The chart tells you what a headphone or IEM is capable of; the source determines whether it gets the signal quality it needs. The Audiophile Basics guide at /learn/ covers source chain fundamentals in detail, but here is how portable audio sources map onto what the measurements tell us.

Dedicated DAPs: When a Separate Player Makes Sense

A digital audio player (DAP) is a standalone device designed to play audio files without a phone. The value proposition is dedicated audio hardware, often with better output stages than a smartphone and support for balanced connections. For listeners who want to optimize the signal path to their IEMs or portable headphones, a DAP removes the noise floor and power limitations of a phone’s headphone output.

The FiiO X5 Mark III illustrates both the appeal and the tradeoff well. Dual AK4490 DAC chips and a balanced 2.5mm output represent genuine hardware investment. Verified buyers note that the analog output quality is audible on sensitive IEMs, particularly in low-noise environments. The significant limitation is Android 5.1: current streaming apps, including the versions of Spotify, Tidal, and Qobuz in active development, no longer support that Android version. For local file playback, the hardware case is solid. For streaming-first listeners, the software age is a real constraint.

Current-Generation DAPs With Streaming Support

The FiiO M11 Plus addresses the software gap directly. Android 10 supports current builds of Qobuz, Tidal, and Spotify, and the ESS Sabre ES9068AS chip has measured performance that ASR-style analysis rates favorably. Field reports from Head-Fi’s DAP forum note that the 4.4mm balanced output delivers meaningful power for harder-to-drive portable planar magnetics. The form factor is larger than a typical phone, which some users find inconvenient for commuting.

The honest question with any premium DAP is whether the improvement over a phone plus a quality portable DAC/amp justifies the investment. For users who want a single, phone-free device, the dedicated hardware argument is real. For users already carrying a phone, the comparison gets less clear.

The HiBy R3 Pro Saber shows how far dedicated DAP hardware has come at budget pricing. An ES9219C chip and 4.4mm balanced output in a pocketable form factor represent capabilities that would have been premium-tier a few years ago. Owner reviews consistently flag the small screen and occasionally sluggish touch interface as the meaningful tradeoffs at this price band. For IEM-focused listeners, the audio hardware is well ahead of what the screen and interface suggest.

Portable DAC/Amps: The Phone-Plus Approach

For listeners who prefer keeping their phone as the source, portable DAC/amps provide the dedicated audio hardware without abandoning the smartphone’s screen, connectivity, and app ecosystem.

The iFi xDSD Gryphon adds aptX Adaptive Bluetooth to the portable DAC/amp category, meaning it can operate wirelessly from a compatible phone while maintaining near-lossless audio quality. The physical analog volume dial is a consistent preference signal in owner reviews, particularly among users who find app-based volume controls frustrating. iFi’s XBass and XSpace filters add tunable coloration that some users find useful and others disable entirely. The premium price carries the real-world risk of portable device loss or damage.

The Chord Mojo 2 takes a technically distinct approach: rather than an off-the-shelf DAC chip, it uses Chord’s custom FPGA implementation with their proprietary WTA filter. Measured performance is excellent regardless of the unconventional architecture. The ball-button interface is the most consistent criticism in verified buyer reviews, described as unintuitive and requiring a learning curve that reviewers across Head-Fi and Chord’s own user community have noted. Mojo 1 units available second-hand often represent better value for buyers not specifically interested in the Mojo 2’s DSP enhancements.

True Wireless Earbuds and How Frequency Response Applies to Wireless Audio

Frequency response charts apply equally to true wireless earbuds, though wireless audio introduces codec-related variables that wired measurements do not capture.

The EarFun Free Pro 3 is an instructive case. ASR and dedicated audio review sites have measured its tuning favorably, with a frequency response that tracks closely to the Harman IEM target. Qualcomm aptX Adaptive support at a budget price point means the wireless signal path is not the bottleneck. Verified buyers note that ANC is functional and useful, if not class-leading. The Sony and Bose flagship offerings are meaningfully ahead on noise cancellation performance. For listeners who prioritize tuning accuracy over ANC depth, the measured performance at this price band is notable.

The Sony WF-1000XM5 is the TWS reference point for ANC performance. LDAC support brings the Bluetooth codec closer to lossless quality levels than standard SBC or AAC connections, and the Sony Headphones Connect app provides EQ options that let users adjust from the default tuning. The default frequency response is not Harman-flat (Sony’s consumer tuning tends toward elevated bass and a slightly V-shaped character), but the app’s parametric EQ allows correction. Field reports confirm it as the current benchmark for commuter and travel use.

The Apple AirPods Pro 2nd Generation occupy a specific niche: the best-integrated ANC for Apple ecosystem users. Personalized Spatial Audio, system-level Transparency mode, and tight iOS/macOS integration are genuinely functional features. The measurement story is more complicated. Apple optimizes tuning for its own ecosystem and AAC codec. On Android or non-Apple sources, the AAC codec ceiling limits audio quality below what LDAC or aptX Adaptive devices achieve. For Apple-primary users, the ecosystem integration outweighs the codec limitation. For cross-platform users or Android users, the codec constraint matters.

Frequently Asked Questions

What does a flat frequency response actually sound like?

A perfectly flat frequency response in a headphone would sound unnatural to most listeners because it does not account for how human hearing processes direct versus diffuse sound fields. Headphone tuning requires a bass shelf and upper-midrange presence peak to sound balanced. Flat measurements are generally more useful for evaluating electronics like DACs and amplifiers, where flat means accurate signal reproduction without added coloration.

How much of a frequency response difference is actually audible?

Research from Harman’s listening tests suggests that deviations of roughly 1, 2 dB or less in critical frequency regions are at or below the threshold of reliable detection for most listeners. Differences of 3 dB are perceptible. Differences of 6 dB or more are clearly audible to virtually all listeners. This is why reviewers tend to focus on peaks and dips that exceed 3 dB relative to the reference curve, rather than treating every minor deviation as meaningful.

Can EQ fully correct a bad frequency response?

EQ can address many frequency response issues, particularly peaks and broad tilts. Narrow resonance peaks are often highly correctable with parametric EQ. Deep dips caused by acoustic cancellation within the driver or chamber design are harder to fix, because boosting gain at a frequency where the driver is physically limited does not always produce clean output. EQ works best as a refinement tool on headphones with mostly good measurements rather than as a complete correction for fundamentally problematic tuning.

Why do different measurement databases show different results for the same headphone?

Different measurement rigs use different ear couplers, fixture geometries, and normalization conventions. Treble readings above roughly 6, 8 kHz are particularly sensitive to rig differences because small variations in headphone positioning affect how treble energy couples to the measurement microphone. Broad tonal character (bass weight, midrange balance) tends to be consistent across rigs. Specific treble peak locations and heights should be cross-referenced across multiple databases before treating them as definitive.

Does frequency response explain all the differences between headphones?

Frequency response explains the majority of perceived tonal differences between headphones at any price level, but not all perceptible differences. Distortion, driver speed, soundstage presentation, and imaging all contribute to perceived character in ways not fully captured by a single frequency response measurement. That said, two headphones with nearly identical frequency response curves and similar distortion profiles will likely sound more similar than different in a controlled comparison. Frequency response is the most reliable single predictor of how a headphone will sound.

![learn product image](/images/articles/learn-20.webp)

<script type="application/ld+json">
{
 "@context": "https://schema.org",
 "@type": "FAQPage",
 "mainEntity": [
 {
 "@type": "Question",
 "name": "What does a flat frequency response actually sound like?",
 "acceptedAnswer": {
 "@type": "Answer",
 "text": "A perfectly flat frequency response in a headphone would sound unnatural to most listeners because it does not account for how human hearing processes direct versus diffuse sound fields. Headphone tuning requires a bass shelf and upper-midrange presence peak to sound balanced. Flat measurements are generally more useful for evaluating electronics like DACs and amplifiers, where flat means accurate signal reproduction without added coloration."
 }
 },
 {
 "@type": "Question",
 "name": "How much of a frequency response difference is actually audible?",
 "acceptedAnswer": {
 "@type": "Answer",
 "text": "Research from Harman's listening tests suggests that deviations of roughly 1, 2 dB or less in critical frequency regions are at or below the threshold of reliable detection for most listeners. Differences of 3 dB are perceptible. Differences of 6 dB or more are clearly audible to virtually all listeners. This is why reviewers tend to focus on peaks and dips that exceed 3 dB relative to the reference curve, rather than treating every minor deviation as meaningful."
 }
 },
 {
 "@type": "Question",
 "name": "Can EQ fully correct a bad frequency response?",
 "acceptedAnswer": {
 "@type": "Answer",
 "text": "EQ can address many frequency response issues, particularly peaks and broad tilts. Narrow resonance peaks are often highly correctable with parametric EQ. Deep dips caused by acoustic cancellation within the driver or chamber design are harder to fix, because boosting gain at a frequency where the driver is physically limited does not always produce clean output. EQ works best as a refinement tool on headphones with mostly good measurements rather than as a complete correction for fundamentally problematic tuning."
 }
 },
 {
 "@type": "Question",
 "name": "Why do different measurement databases show different results for the same headphone?",
 "acceptedAnswer": {
 "@type": "Answer",
 "text": "Different measurement rigs use different ear couplers, fixture geometries, and normalization conventions. Treble readings above roughly 6, 8 kHz are particularly sensitive to rig differences because small variations in headphone positioning affect how treble energy couples to the measurement microphone. Broad tonal character (bass weight, midrange balance) tends to be consistent across rigs. Specific treble peak locations and heights should be cross-referenced across multiple databases before treating them as definitive."
 }
 },
 {
 "@type": "Question",
 "name": "Does frequency response explain all the differences between headphones?",
 "acceptedAnswer": {
 "@type": "Answer",
 "text": "Frequency response explains the majority of perceived tonal differences between headphones at any price level, but not all perceptible differences. Distortion, driver speed, soundstage presentation, and imaging all contribute to perceived character in ways not fully captured by a single frequency response measurement. That said, two headphones with nearly identical frequency response curves and similar distortion profiles will likely sound more similar than different in a controlled comparison. Frequency response is the most reliable single predictor of how a headphone will sound."
 }
 }
 ]
}
</script>