Jump to content

Article: Let 1,000 Frequency Responses Bloom

Recommended Posts

7 minutes ago, The Computer Audiophile said:

Thanks so much for your work on this one @Josh Mound. I have much to say, even though I also live in a glass house. But, I'm incredibly short on time and will circle back to this when I have a moment to transribe my thoughts coherently.


I also want to to be the first to say, you do a fantastic job around here, classing up the place with your writing talent. 


Thanks, Chris! Looking forward to hearing your thoughts.

Link to comment
1 hour ago, jhwalker said:

My feelings on this lie somewhere in the middle, I think.


I know exactly which reviewer you're quoting because I frequent his fora daily.   And I think he offers incredible value in the services he offers.


I think he oversteps a bit WRT his comments on the universality of the Harman Curve, BUT ... I also think the Harman Curve offers a "standard" we can (and should) reference as a baseline for comparisons.  I don't think that any headphone that fails to precisely replicate Harman is automatically valueless, BUT I know I, personally, will not like it if if falls too far from from the curve, because Harman is also MY preference among all the variations I've tried.


So, no, people who prefer something else are not necessarily fools (though they may be, or ignorant), BUT I question their judgment when they enthusiastically recommend headphones that are so far from accurate they MUST sound terrible (and they do, the ones I've listened to, at least).  I think headphones which adhere closely to Harman as a baseline generally sound pretty good, and any specific variations may or may not contribute to the special value of that model.

Thank you for your response.


As I mentioned at the end, I don’t have any problem with people enjoying or even touting the Harman Curve. I just don’t want it to become an industry standard. I think the research shows that there are frequency responses that are equally preferred, as well as valid ones that are enjoyed by substantial minorities of listeners, even according to Harman’s research. 

While, as I wrote, I like some IEMs and headphones that tend to follow the curve (or a variant of it), I also usually think the sub-bass is a bit too much and the 3-4 kHz peak is too steep. So, the SoundGuys curve, which is basically Harman but with those traits toned down, might be closer to my ideal. 

That said, I totally agree that there are some IEMs and headphones that stray way too far from any reasonable ballpark definition of “neutral” for me to enjoy them. Those tend to get negative reviews from me. But I also would never say that those products shouldn’t be allowed to exist.

Link to comment
19 hours ago, ecwl said:

I have often wondered if some variations in individual preferences for frequency responses are related to different people preferring to listen at different volumes. In those scenarios, Fletcher-Munson curves/equal loudness contour kicks in. 


But I would still prefer to know the frequency responses of most headphones so I can figure out what frequency response curves I enjoy. That way, I don’t have to try all the headphones that deviate too far from my preferences.


While I 100% agree with the article, I have often seen people use the arguments in the article to say that speakers in-room measurements (or even room acoustics) don’t matter. I think since most headphones have smooth bass response and most rooms don’t, that is a perspective I can’t buy into. Unfortunately, we don’t know what we don’t know. So once again, if people don’t measure their rooms, they don’t really know what they’re hearing.

To be clear, I am most definitely *not* anti-measurements. I just bought a few IEC 711 clone couplers to make my IEM reviews’ measurements more comparable to others’, and I’m setting up a SquigLink site. I measure headphones with my EARS unit, and if I could afford a Gras or similar, I’d use it in a heartbeat. I also measure my listening room with an UMIK-1 and REW and treat it based on those measurements. 


Indeed, I rarely buy a product (especially a transducer) without seeing measurements first, and I value all of the measurements conducted by reviewers, including the pro-Harman reviewer mentioned in this piece. 

So I don’t want anyone to think my skepticism about the accuracy or universality of preference curves for headphones and IEMs is an anti-measurement statement. Far from it. 

Link to comment
2 hours ago, taipan254 said:

Thanks @Josh Mound for your thoughtful piece. I agree: The reviewer in question's assertive promotion of the Harman Curve as the only correct standard, and his dismissal of alternative preferences, is unethical and misleading. 

As my experience with this hobby has grown, I've found myself increasingly drawn to reviews that compare the item in question with something I already own. For instance, in the realm of headphones, I particularly value reviews that juxtapose the model being discussed with the HD6xx. This preference stems from two reasons: firstly, the HD6xx is widely owned and favored, making it a familiar benchmark for many. Secondly, its affordability means that it's accessible to a broad audience, thereby serving as a universal reference point. From there, it's up to the reviewer to make further comparisons. Similarly, when it comes to IEMs, I see the merit in having a Harman-tuned IEM in one's collection. Given the range of affordable options, this provides a baseline for understanding a particular sound signature, enabling more meaningful comparisons with others' experiences.


Furthermore, I believe that frequency response is not the sole determinant of quality. Factors such as detail / clarity, damping / decay, and coherence/imaging—attributes particularly noticeable in the variance between single and multi-driver IEMs—cannot be fully encapsulated by frequency response charts alone.


While that reviewer's website can be a source of frustration for me, I find myself consulting it occasionally for its measurements, which can be informative. I'm thankful for the presence of other resources (SBAF, @GoldenOne, Danny Ritchie) in the community that also conduct equipment measurements. The less time I need to spend navigating that particular site, the better!

PS - is there a reason we're not naming the reviewer? I feel like we're talking about Voldemort and the Death Eaters here - he who shall not be named! A fun project could be assigning Harry Potter characters to folks in the Audio Industry =)

I completely agree about common model benchmarks being valuable. I include the HD6XX in almost all of my headphone reviews for that reason.


The reason for not naming the reviewer is that I want to talk about the broader issues at stake, rather than get bogged down in personalities. After some thought, it seemed likely that foregrounding the reviewer would simply devolve into a debate about that reviewer, which isn’t what I wanted this article to be about. 

Link to comment
On 3/24/2024 at 8:19 PM, mitchco said:

Love your article Josh! Some comments if you will.


As @ecwl has pointed out, the equal loudness contour comes into play as folks listen at various sound pressure levels which impacts how much bass and to a certain degree, treble one is perceiving. In other words, tonal balance is affected by how loud or quiet one listens. For those that think the Harman curve is too much bass, it makes me wonder how loud folks are listening...


I have yet to see any measurement rig that can measure over the ear headphones beyond 6 to 7 kHz due to the measurement being swamped with internal reflections between the headphone and ear due to the (short) wavelengths involved.


Related to the above paragraph, I also wonder how accurate the measurements are wrt headphone positioning on the measurement rig. I don't know about others, but there is "sweet spot" when positioning headphones on ones' head and it takes some maneuvering to hit it precisely. This can have an impact not only tonal response but also soundstage.


As a studio mixer, perhaps the requirements are slightly different as we are trying to make tonal eq and level decisions based on what we hear. So if the speakers/room or headphones are not neutral or colored in anyway, then we are making decisions that won't "translate" across a wide range of speakers and headphones.


Relative to the above, I chose a different approach to measuring and eq'ing headphones. As one can see by looking at the individual headphones measurements, it is most certainly the Wild West when it comes to headphone frequency response. There have been headphones I have measured, that costs thousands of dollars but the frequency response is so bad, they can't even be eq'd. I would call them "broken" yet I have seen others verify my measurements and but seem to enjoy the sound. Sometimes one has to realize that statistically, 4% of the population have "tin ears." I wonder how many audio reviewers this applies to...


Here are a couple of convolution filters I developed for the AKG K371 and Sennheiser HD560S, both popular inexpensive headphones, but I would argue sound more "neutral" with the convolution filters. What I mean by neutral is that no one frequency or range of frequencies sound less of more than any other range of frequencies. Please be sure to level match if comparing to no filter ;-) Enjoy!


Keep up the great writings Josh!


Sennheiser_HD560S_neutral_filterset.zip AKG_K371_neutral_filterset.zip


The point about volume and equal-loudness contours is a good one. At least in this Harman study, the average playback level was set to 78 dB (b-weighted). I don't know how the level in the Harman study was measured or what that means in the context of dynamic music (especially since they were using dynamic songs like Steely Dan's "Cousin Dupree" as material). But based on my attempts, that seems low.


I fired up "Cousin Dupree," put in my Moondropo Kato IEMs, brought the volume to a "normal" (for me) listening level, and measuriung it with both my 711 coupler and Dayton USB-C mini-microphone. I calibrated the former with an external calibrator, and the latter is calibrated by Dayton. On both, I was in the mid-to-upper 80s. When I lowered the level to 78 dB, it seemed too soft for "normal" listening, and I don't have any hearing loss (at least in the range tested by audiologists and the Etymotic home test system).


Now, this could all be chalked up to my measurement process and the study's being totally incompatable. (Indeed, that seems more likely than them using "my" 78 dB.) But I'd love to actually hear how loud the music is in their IEM and headphone tests.


EDIT: See below. I made a dumb mistake. The Harman study’s listening level seems great to me.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Create New...