Maybe I’m just missinformed, but isn’t the Harman curve derived from the PREFERENCE of a bunch of testers that happens to also agree with the general end user’s preference? Like, it’s basically the lowest common denominator for frequency responce.
When I listen to a Harman-esq tuned headphone (Shuoer Tape for example) it sounds far from neutral to me. To me, it sounds like a lower treble heavy V-shape, and the graphs seem to show that.
I would like to hear what everyone thinks about the Harman target curve and what YOU think is “neutral”
As a disclaimer, I would like to mention that I HATE the sound of the Harman target curve, at least in the IEMs I’ve tried with it (TFZ S2 pro, Tin T4, and Shuoer Tape). The bass tends to be nice, but thats it for me. The mids sound hollow, thin, and sucked out. The upper mids/lower treble are just so shoutty and annoying. Even when it’s lowered from the target curve, it’s still bad because the difference from the mids to the lower treble is like 15dB at least. Then after this mountain of bad, the treble just falls into oblivion after 10k. I think this is at least in part because they just didn’t test for anything above a certain frequency, so they just said fuck it and dropped those frequencies off a cliff (please correct me if I’m wrong about this). It’s just NOT my preferred tuning, which I find funny because it’s supposed to be the most commonly preferred tuning.
I also find it interesting that this is often referred to as ‘neurtal’ when its clearly v-shaped. I think about this in relation to ‘audiophile bass’ where the lows are less emphasized revealing clarity in the mid section. Clearly this target isnt a catch all.
With senses like smell and hearing, it seems language often fails us. I feel like measurements are an attempt to shim that issue a bit. And having a target is just another shim for our lack of language.
In photography there are histogram dos and donts. I feel like this is a very comparable concept. As the histogram of an image leans too far to one side or the other, it starts to get washed out or too dark to read. But sometimes thats exactly the desired outcome. If you look at calibrated monitors they sometimes look like they lack contrast compared to a TV thats in some enhanced mode. They are two different tools for two different jobs. This is why we have headphones for mixing/mastering and headphones that lean a bit one way or the other for enjoyment/preference.
Sorry if I strayed a bit from the core of talking the harman target, just some thoughts came to mind.
I just thought of a better way to describe what I consider neutral. I would say neutral is when no frequencies stand out of the overall sound signature. I say sound signature because things like transient responce and sibilance distortion can make certain frequencies seem louder or quieter than they truly are, so the frequency responce might not be 100% representitive of the actual persived sound.
This is why I don’t consider the Harman target to be “neutral”. The mids and upper treble fall away and the lower treble and bass are more present.
My understanding is that the intent was to match the perceived response of speakers, with headphones or IEMs.
Speakers are designed to measure flat in an anechoic chamber, but listening to headphones isn’t like listening to speakers, so you have to have a none flat response to sound like a flat response.
Personally I’m not even convinced measuring speakers that way is particularly useful, FR curves are one of the least interesting parts of a sound to me.
Im not sure resolve thinks of the harman curve as neutral as much as he uses it to describe a headphones fr
personally I don’t think of it as neutral at all but more of what we like to call “fun” … that said the k371 is supposed to be a good representation of harman and I really enjoy that headphone… as for my neutral I would say its the sundara… some call it neutral bright but for me the treble is just about perfect
That is exactly right. It’s a population-average that is meant to give a good experience to the largest possible number of consumers and be possible to implement with a single (non-personalized) physical design that will be mass-produced. True neutrality requires personalization of the headphone response to each user, like what is done in the Smyth Realiser or like the more work-intensive manual solution from dr. Griesinger: https://www.youtube.com/watch?v=1VaKZr208Dc
None of these fixed-curve solutions like Harman, Olive&Welti or anything based on them like Oratory’s EQ curves, Sonarworks, Toneboosters, nothing like that is truly neutral to most people. Only personalized solutions are truly neutral.
No, that’s a headphone response curve, it should always be “V-shaped” like that because it’s supposed to compensate for various changes that happen when the sound no longer passes around your face and over your pinna but is shot directly into your ears. “Flat” is what the curve should look like when you’re measuring good speakers in an anechoic chamber, but for headphones to sound like that they have to have a different response, more like what you’re seeing on the Harman graph. More discussion here:
Yeah, harman neutral doesn’t sound 100% natural to me either.
Neutral is really misleading – intentionally misleading – if they’re taking random people instead of musicians, who knows what X instrument sounds like, for all their tests, to decide what neutral should be like. I’m thinking it should be called harman conventional or harman consumer curve.
It’s like THD measurements for DACs and amps. It can be seen as a way to know if something went really, really went wrong.
IMO, all ears are different – literally – all pinnas and conchas etc. So you got to do your own curve.
Resolve drives me crazy by having understood about 80% of the concepts that go into FR and measurements … but mixing the 20% that he’s got wrong at any given time. In this case he’s using “Harman” to mean his personal amalgam of the 2013 Harman curve’s bass with the 2018’s mids & treble. So it’s not quite as radical a departure from some mythical over-ear headphone neutral as the full 2018 Harman curve.
If you want to hear what actual FR neutral sounds like on a headphone, it’s simple (grin). Listen to your headphones using the Online Tone Generator or your favourite sine sweep tool. Set the volume to quiet conversation but not whisper quiet. Move the slider slowly from 20 Hz to however far into the stratosphere your hearing reaches. Note the frequency ranges of any meaningful departures – either dips or rises. Then EQ them out. Now use that EQ to listen to a variety of your usual music at your usual loudness levels.
It takes a bit of practice to learn to hear very gradual rises or falls in loudness. And also to distinguish what’s actual loudness vs the psychoacoustic softness of low frequencies and the psychoacoustic piercing-ness of high frequencies.
And of course (this keeps on needing to be repeated) neither the three flavours of Harman nor the multiple flavours of diffuse field nor even your personal neutral is in any way the “correct” tuning for a headphone. The correct tuning is the one that rocks your boat. period. full stop.
I guarantee neither Sean Olive nor Todd Welti has ever described the Harman FR curve as either Neutral nor Correct in some Absolute Truth sense. They know very well it’s the statistical average consumer preference tuning. They even assign this a number: 71% likely to be at least acceptable to any given consumer.
Yes neutral is relative, however there should always be some degree of elevation past 1k because of pinna gain, even more so for in ears. Other than that it’s really a matter of taste, I do like the idea of having harman as a neutral standard because it gives a objective understanding of what people mean when they say neutral, otherwise it would be so much harder to convey how a headphone sounds if we don’t have a baseline.
I for one don’t think the 2013 over ear harman response is perfect but it is a pretty good safe tuning that most would not find offensive, the in ear harman response is less ideal imo.
For in ears it’s much more elevated because the pinna is bypassed entirely, for over ears, I’m actually not completely sure, but a strong suspicion of mine is that most measurement rigs, even the cheapest minidsp ears has plastic ears outside the microphone, these fake ears probably simulate the effect pinna pain resulting in the final measurement containing that elevation which the brain corrects for. The harman target was developed with a measurement rig that had “ears” thus having that elevation too.
The key concept here is that all headphones, in-ear, on-ear and over-ear, are directly coupled to the ear. There’s no open air environment between the driver and the ear drum as there is with both loudspeakers and live sounds.
Direct coupling cancels out the natural amplification from the outer ear’s shape. Therefore, the amplification of frequencies starting in the midrange has to be artificially engineered into the headphone or it doesn’t happen.
Oh thank God. I’ve listened to the AKG TWS N400 that are specifically tuned to the harman curve and I can’t listen to them, they drive me mad. To my ears they sound exactly like you described - I was head nodding while reading your post. Shouty, but also thin and metallic. Gross.
I honestly think the harman curve is one of the worst things to happen in audio. Maybe it’s just me hating v-shape (where are the mids??), but just irks me when someone else has almost arbitrarily decided what sound is “good”.
Blame audiophiles and audio reviewers who have popularized the Harman curve among laypeople without explaining it properly.
The Harman target curve is what frequency profile the sound should have by the time it reaches your eardrum in order to sound to you like good speakers in a good room. You can’t EQ the response at your eardrum without measuring it at your eardrum. Everyone who is “EQ-ing to Harman” without using deep-seated in-ear microphones (or a professional head simulator with eardrum microphones) to measure the results and see what to adjust is only kidding themselves. Then these people get horrible results and go around blaming the Harman curve.
If you’re doing it right, it should at least sound decent, it shouldn’t sound completely wrong.
That’s why Oratory1990 can create very good Harman-based EQ curves and you can’t: he uses a GRAS measuring rig, which is one of those types of products they don’t even publish the price for, you have to ask for a quote. Also why people ask to send Oratory their headphones to get a custom Harman-based EQ - he can put your headphones on his GRAS rig and do this right.