Anyone else have over-sensitivity to frequencies around 3.7 kHz, esp. vocals and violins?

I found a long time ago that I was troubled by certain sound frequencies and of course blamed it on my equipment, but after years of experimentation and testing, I've determined that the problem is with my hearing. I've found I'm over-sensitive to frequencies at or around 3.7 kHz and any music with a lot of content at this frequency is unpleasant to listen to, sounding unnaturally bright with ringing and a distorted edge. Vocals and violins are the most likely to trigger it - while no one sings at 3.7 kHz, and no one plays a violin note at that frequency, these sounds seem to have particularly complex harmonics which can often be prominent at or around 3.7 kHz. Even then, it's not all vocals - if people are singing softly it's generally OK but when artists push their voices or use what's called "twang", things can get really bad.

I discovered the other day that the human ear canal, being a tube, has a natural resonant frequency and that this resonance amplifies part of the audio spectrum. That wasn't news to me but what amazed me is that the typical resonant frequency is none other than 3.7 kHz. So I began to wonder if I might not be the only person having a problem with this frequency and, if not, what other people did to mitigate it.

Any takers?

 

No idea. I do suffer from tensor tympanic syndrome, where certain sounds can trigger a spasm. Or sometimes it'll just happen on it's own. I can even voluntarily trigger it in both ears, whereas the involuntary is only my right.

 

Yeah, apparently my hearing has declined in such a way that the range you cite seems to be emphasized. I experimented with a 2db "notch" (smallest available adjustmentat 4kHz on the dbPoweramp equalizer; that smoothed it out fairly well for headphones.

With speakers, it's not so evident, so I go back to "flat" when I use them.

 

One name for it is Hyperacusis. I know it well.

Take these things,
N-acetylcysteine (NAC)
Acetyl-L-carnitine
Alpha-Lipoic acid R-(natural)
B-complex

Also check your blood pressure.

 

Oversensitivity to fr between 3-4 kHz can be ear damage. But this is also a known problem for quite many speakers. It is both related in part to the stereo reproduction of a centered phantom image that is coming from two separated speakers. The centered phantom image is percieved brighter than from a single mono speaker and should be compensated for. But in most part because many speakers have crossover frequencies around 3 kHz. The radiation pattern of such speakers will often have, but not always, a peaking of the energy between 3-4 kHz because the tweeters dispersion is usually much higher than the woofer/midrange. The typical room curve could thus show peaking between 3-4 kHz. Which is not pleasant.

 

I guess it depends on the EQ of the headphones and the speakers. So for me, it's the other way around - my ATC speakers are pretty much flat and so that's where I have the problem, but my Audeze LCD-4 headphones have a big dip in exactly the "right" place and are bliss to listen to.

 

Many thanks but I've done many tests over a very long period of time and ruled out anything like this - for example, by measuring room EQ and, very simply, by cutting out one channel. I'm 100 per cent certain it's my hearing - I wish I could say otherwise.

 

You are not alone my friend! I noticed about 18 months ago, vocals and certain instruments sounded slightly distorted and bright in my right ear. I panicked and thought it was my speaker but alas if I turned around 180 degrees so my right ear was now listening to the left speaker, I heard the same distortion. I double checked with my headphones and the same issue occurs. I went for a hearing test and ear cleaning. No change, they told me my hearing is "normal" for a 63 year old... The good news is I no longer need to spend money on new speakers because the issue will still taint the sound so I am learning to live with it. BTW, I have no issues with my blood pressure.

 

I looked at Stereophiles measurements and the ATCs are far frm flat in in the dispersion plot. The loss of energy in 1-2 kHz and peaking between 3-4 kHz is quite evident. Room EQ does not tell you everything. If there is a mismatch between direct and diffuse sound you can get these effects.

 

Can't get too technical here, but I do hate the sound of flute - not sure if it's over-sensitivity, or the sound is just generally unpleasant to me.

 

Hearing is normally non-linear. Very true the ear canal is a single ended pipe with a resonant frequency. This aids in nature as an acoustic amplifier by perhaps 10 to 12 dB increase (as I seem to recall without checking references) which sharpens hearing acuity and localization. The intelligibility of speech is in the 2kHz to 8kHz range, and most sensitive at 3kHz to 4kHz. The resonant peak varies person to person.

Likewise the ears normally hear unequally, the left and right never exactly the same. This is compensated by the brain and simply being accustomed to our own hearing in live settings... so a good speaker system may closely replicate these live conditions, thus we say a wonderful sounding speaker.

I agree with @Thomas_A , the crossover point of many speakers lie within the 3kHz range. The crossover point may sum the drivers somewhat, resulting in a peak, and then multiply times 2 for two speakers. The phase relationship of two drivers can be particularly annoying, even painful as the sound compression waves "fight" each other, and can cause other odd distortion harmonics or ringing. (in such case a poorly designed speaker) When more out of phase than in , the effect is like the sound originates inside the ear, and certainly we don't want that!

 

It is one of the symptoms of my cochlear hydrops (right ear only), which comes and goes. Fortunately hasn't been a prominent symptom for about a year now, but very annoying when it crops up.

 

I think the trouble with hearing tests is they only test whether hearing is present at various frequencies - they don't test for unevenness in hearing levels at different frequencies nor for distortion in the perceived sound. Not in a simple hearing test, anyway - there are probably more sophisticated tests available at a price.

 

Just to say, that Stereophile review of the ATC SCM 11 is very old and for a previous model. I do appreciate your help and I'm happy to accept that the problem you describe can exist but as I say, I still get this effect when listening to mono sound from a single speaker so it's not the explanation in this particular case - it really is my hearing.

I find one way of solving the problem, though bizarre, is to put foam earplugs part-way into the ear canal - not so far as to block the sound but far enough to kill the natural resonance and the peak in the frequency response that results from it. Another way is to put an equaliser in the chain and insert a very large notch at 3.7 kHz but it's hard to find a setting that sounds natural. Another way, as bizarre as the first, is to wear my Audeze LCD2 or LCD4 headphones unconnected to anything while listening to music coming through my main speakers. The LCD headphones are open backed so I don't lose much sound and they seem to filter out all the horrible stuff around 3.7 kHz leaving a very enjoyable sound even with the worst recordings. The problem is that it's crazy.

 

I think there are many possibilities and I wouldn't be too quick to blame your ears. A lot of music is poorly recorded and mastered, which tends to result in fatiguing highs and a forward midband. Then there's the propensity for speaker designers to seek flat measurements. The problem is our ears, even at their best, do not function like a microphone. Hearing sensitivity does plateau near 4kHz for most people. However, it seems to me that speaker crossovers near these frequencies usually result in dips, not peaks.

There's a phone app called Neutralizer that makes this sensitivity quite evident. It plays test tones at various frequencies (including 4kHz) but at the same SPL. Use this app on anyone and probably 99% will identify 2 and 4kHz as the loudest tones.

I've found my state of fatigue/energy prior to listening plays a significant role in how I perceive brightness.

 

My wife can't stand listening to Stevie Nicks and other singers in that range due to a perforated eardrum caused by measles when she was a child.

 

It's certainly true that the quality of the recording makes a big difference. Many recordings are fine for me, especially instrumentals with not too many instruments playing at the same time - jazz, for example.

 

YES! Stevie Nicks is one my "problem" vocalists! I think it's because she has a nasal intonation. I ought to make a list - Bonnie Raitt is another, which is great pity because I'm a huge fan. Lots of male artists, too - David Byrne is a big problem for some reason. Not deep voices, though.

 

Are you sure she just doesn't have good taste?

 

Could well be that!

 

Yes, you are absolutely right, it is common practice to design a slight dip at the crossover point. The crossover region will always be the speaker's weakness. So, by allowing slight attenuation at the crossover point, the result can be more pleasant overall system performance.

I'm a speaker builder/ hobbyist.. some observations:
This is not to negate very acute hearing sensitivity in the 3kHz to 4kHz range, so even a slight dip in speaker response may result in annoyance. (worse of course when the crossover point has a peak) As a speaker builder, I have observed some "weird" things in multiple driver behavior. Resonant peaks can be induced, even with a measured dip at the crossover point.

Let's say we have a two way speaker, and crosses at 2.5kHz, a common point for two ways... The woofer is still "on" at frequencies above 2.5kHz, and attenuates in level with increase of frequency. So, the woofer will be rolling off at 4kHz, perhaps -8dB. (given a 12 dB/oct rolloff) This is still enough energy to induce odd harmonic artifacts. (ringing and buzzing) Even when time aligned, the slope introduces phase angle, so the phase alignment will be off to some degree, and interference/ degradation of the quality of the tweeter. The phase/ time alignment difference is further aggravated by vertical off-axis of the listener. (sometimes beneficial if the two drivers "line up" on vert off-axis) In fact, I've tested this well beyond the crossover points of multiple drivers... very important the design considers the slopes, and the alignment of drivers outside their effective operating ranges. I have designed crossovers that address this type of distortion, and the results have been very good. (if only I would patent and build the these circuits?.. Back in the 80's, I designed the current KEF coincident source driver before the folks at KEF did, have my drawings, never patented it, which now KEF holds international patents. I never DO what I know)