Headphone Reviewers Can't Be Trusted. Here's Why

You’ve seen it before. One reviewer calls a headphone “endgame,” while another says it’s unlistenable. Your first instinct? One of them must be wrong, or one of them must be a shill.

It’s possible, sure. But far more often, the explanation is much less simple.

Unit Variation is a Real Issue

Two people can review the “same” headphone model, but actually be listening to meaningfully different products.

Manufacturing tolerances—allowances for differences in diaphragm tension/construction/mass, pad variation, assembly differences—can all create measurable and audible differences between units. Sometimes that difference is minor, but sometimes (maybe more than you think) it’s really not.

A clear example is the Meze Strada. When Resolve and Joshua Valour reviewed two separate units with measurements for each on comparable industry-standard rigs, the data showed noticeable differences in performance. One unit was smoother, warmer, and more balanced, while the other exhibited more sibilance and a brighter tilt. 

So of course the reviewers had different impressions of the headphones: they're different headphones!

And unfortunately, even brands known for strong quality control aren’t immune from this. Some designs are easier to manufacture consistently than others, but no headphone is so optimized that this problem ever really goes away entirely.

Additionally, most reviewers aren’t comparing multiple units side by side—they’re evaluating one sample and assuming it represents the whole. What you need to know is: It might not.

Your Head Changes the Sound

This is the big one.

Headphones don’t just sound different because we “hear differently.” While yes, our heads and bodies commit different filtering on incoming sound—different Head-Related Transfer Functions—headphone' don't just sound different because of different HRTFs: they sound different because they physically behave differently on different heads and ears.

That’s not differences in subjective perception—it’s a measurable acoustic quality of headphones.

Take the Hifiman Edition XV. listener loved it enough to call it their favorite over-ear headphone of the year, while DMS found it both too dark, too bright, and overall too strange sounding.

However, when measured with in-ear microphones on these two reviewers, the reasons why they enjoy it differently become absolutely obvious. The frequency response that listener is receiving at their ears is considerably smoother, more extended, and less prone to large peaks and dips, whereas on DMS's head, the response is much more jagged and prone to large magnitude deviations.

Same model. Same general design. Verifiably different sound.

And this effect isn’t limited to “problematic” headphones. In fact, we have reason to believe the Edition XV would be one of the more consistent headphones out there. The issue is not design, but that any headphone when placed on two different heads will simply behave differently.

For example, designs like the Sennheiser HD600 also show dramatic measurable differences when worn by 3 different reviewers—Resolve, listener, and GoldenSound. Even after subtracting each listener’s HRTF (Head-Related Transfer Function), the headphone still varied significantly across heads.

But even with supposedly consistent designs like the Sennheiser HD800S, those dramatic differences still present themselves.

The takeaway? The sound of a headphone not an objective, fixed parameter. It’s interactive. Your anatomy changes the acoustic output of the headphone.

Now if you consider what that means for:

  • Measurements done on standardized rigs

  • EQ profiles based on those measurements

  • Reviews written by literally anyone else that isn't you, the reader

No one in the audience is hearing what the measurement fixture hears, nor are they hearing what the reviewer hears.

Preference Still Matters

Even if we could perfectly control for unit variation and anatomy, preference wouldn’t disappear.

Some people want bass impact and sparkle, while others want surgical precision. There is nothing wrong with different people wanting fundamentally different things from their music.

And to that end, music choice plays a role too. EDM listeners may prioritize low-end weight and treble brilliance. While those focused on acoustic/vocal-focused music may care more about midrange clarity and timbral realism.

Taste matters, perspective matters, and what you listen for in music shapes what you value in a headphone.

Even within the same review team, we all have different taste. Even if we do happen to hear roughly the same thing, we may receive it very differently.

So What Should You Do?

The headphone space is full of confounding variables:

  • Unit variation

  • Anatomical differences (and accompanying headphone behavior differences)

  • Individual preference

It’s a minefield.

The practical answer isn’t to distrust everyone. It’s to broaden your information sources.

Look at multiple reviews.

Look at multiple measurement sets.

Pay attention to trends across units.

Notice how different listeners describe the same traits.

Our new mission of measuring headphones on our actual human heads with in-ear microphones is going to help clarify why impressions diverge, so get ready: the disagreement isn't going anywhere anytime soon. The difference we're hoping to make is not to eliminate disagreement. The hope is that the more data we gather about how headphones behave on different listeners, the more sense these disagreements make.

And ideally, the less confusing, mysterious, or conspiratorial they seem.


This is a companion discussion topic for the original entry at https://headphones.com/blogs/features/headphone-reviewers-cant-be-trusted-heres-why
5 Likes

3 Likes

Well stated. I’ve run into the unit variation problem first had. Pretty annoying when the headphones are north of 2K retail.

Learning how to read the FR graphs as they relate to the end user in-situ is a bit of an acquired skill. I’m still mastering this. :slight_smile:

1 Like

Obviously we must be able EQ our heads so that they conform to an appropriate standard.
First we must establish a committee of experts to develop the appropriate head shape curve.
Thousands of heads must be measured to determine a baseline.
Using both AI and appropriate graphing tools we can represent the ideal head and how various populations and individuals deviate from the ideal.

Unfortunately correction may take time. Medical devices such as existing halos can be used to reshape minor skull aberrations, and injections (we have come far since Carol Doda ) to smooth out problematic curves.

Some heads may require surgery or prosthetics. Users seeking non standard audio responses may need to deform their skulls while the bones and their junctions are still soft, starting in infancy as has been done in some historical cultures.

5 Likes

More information sources isn’t the answer since no quantity of opinions will ever truly answer whether you’ll like the headphone or not. There’s simply too much variability in how a headphone will sound from person to person. The only practical answer is to go try the headphone and decide for yourself if it’s good or not. Try in a retail store or at a show, or by buying only from retailers with consumer friendly return polices such as no hefty restocking fees.

2 Likes

image

1 Like

1 Like

Great article/video. I’d add that reviewers of audio equipment (including headphones) who don’t have their hearing tested regularly up to 20 kHz will produce reviews of unknown validity, despite their best intentions.

Great video - really! Can you describe your in ear mic setup at some point? FC has a short showing it but, it’d be great to be able to replicate it.

2 Likes

They’re open canal mics. It’s a high performance MEMS mic capsule in a resin printed housing that’s mostly open. It fits into the ear in such a way where the outer portion holds it in place while still putting the capsule deeper into the ear canal. This design was originally developed by Blaine but since revised by Oratory1990. I’m not sure it’d be possible to replicate this specific design, but something like it for sure would be doable.

You can probably just convince Konstantin to send you the STL file/capsule spec, Craig :slight_smile: