Non-linearity of the human ear

[Ian - G4JQT]

A question poised to my class in a sound engineering course: Does the human ear respond to waveform or spectrum?

To those unfamiliar with the concept, imagine two audible sine waves, say 1000 Hz and 2000 Hz. Assuming they are of similar amplitude they will both be heard. As these are harmonically related, assume the waveforms start at the same point in time: i.e. 0 degrees. For every one cycle of 1000 Hz, there will be exactly two cycles of 2000 Hz. They are in phase.

Now, lets delay one of them by say 30 degrees. They are still the same amplitude, they are still harmonically related, they will still both be audible, but if you add the two sine waves together you will get a completely different wave form to when they are in phase.

Both examples have exactly the same spectrum, but the added waveforms are quite different.

From informal tests I have witnessed, it seems some people perceive an indefinable difference and others do not. If there is an audible difference, at what phase difference does this become apparent? Presumably is varies very significantly with frequency.

Regards,

Ian

[John_BS]

Ian:

That's very interesting. Was this a double blind test? I do know that one of the BBC's acoustics gurus( Dudley Harwood) told me that the ear is not primarily concerned by the phase response of a speaker. Phase matching is important for stereo imaging of course (at mid-frequencies).

John

[Synchrodyne]

I am not sure that I have ever fully understood the logic of conventional loudness controls nominally based upon the Fletcher-Munson or similar equal-loudness curves.

I think that this would require that the sound system be “flat” at a given reference sound pressure level (SPL), say 100 dB. At levels progressively below this, the equalization curves would be applied.

On a given amplifier, the volume control setting to achieve reference level would depend upon a whole slew of variables, including the room, the speakers, the electrical output level of the selected source relative to the amplifier input sensitivity, and the modulation level of the specific program being listened too. For example, one might connect two different tuners to an amplifier, and find that different volume control settings were required to equalize their volumes when tuned to the same station. And then different stations would require different volume control settings depending upon their average modulation characteristics and how much compression was applied. Then different sources, such as CD, tape, etc., would likely require different volume control settings to radio.

So to be accurate, loudness compensation would need to have an adjustable reference level that could be reset according to all of the variables noted, none of which are reasonably predictable by the amplifier maker. But it is not done that way. Most loudness controls are fixed, and therefore very much a guesstimate as to what is required in situ. The chances of being anywhere near the applicable Fletcher-Munson curve would be remote. It seems to me that for those who might want this kind of frequency bending, they would have a better chance of getting what was wanted in any given situation from a set of conventional Baxandall tone controls.

Also, what happens when one is listening to a piece of music with wide dynamic range. Let’s say that the loudest peaks are producing 100 dB SPL in-room, and that the volume control setting is accordingly just above that which triggers loudness contouring. Now suppose there is a quieter section, peaking at 60 dB SPL at around 1 kHz. The Fletcher-Munson curves show that at 30 Hz, the corresponding equal-loudness level is around 80 dB. To get this “right” implies that the appropriate curve should be applied. But this in turn means that loudness correction should be applied on a “dynamic” basis, varying according to instantaneous loudness. Or in other words, the amplifier should apply frequency-selective compression according to instantaneous loudness. And that does not seem to be right at all.

I don’t have a problem with the inclusion of loudness controls on amplifiers for those that want them, although I have no use for them myself. But their justification in terms of the Fletcher-Munson curves does look to be flawed, perhaps veering towards “pseudo-science”.

I somehow think that the characteristics of musical instruments, including their dynamic ranges have developed naturally and intuitively around human hearing characteristics, such that those with registers in the contrabass on the one hand and the sopranino on the other generally have lesser dynamic ranges than those in the middle registers. So further frequency correction simply becomes a matter of personal taste, unless there are flaws in the transmission/recording chain that can be corrected in this way.

Cheers,

[colly0410]

Re # 18.

I've just done the test & can hear up to 12 khz, mind you I've got tinnitus as I was in the artillery. BTW I'm 61. Mother-in-law who's 80 can hear up to 8 khz. I'll have to see how Tiddles reacts to it when she comes in..

[Peter.N.]

It would seem by the comments that I am fortunate to still be able to hear 10khz at 75 years although the non linearity of my hearing still gives problems with ordinary conversation if there is a high background noise level.

Listening to music from the Harman/Kardon sound system on my computer it sounds as good as it did in the '50s, I assume that's because I am relatively close to the speakers so I hear the high frequencies before they get absorbed, I do have the treble right up of course. The same with the TV, with top response set at max I can hear it fine, it's just people I have problems with.

Strangely when 625 lines started I was about 30 and yet I couldn't hear the timebase whistle, I could sense it, if I moved my head I could feel by pressure rather than hear the 15khz, so I consider myself fortunate to till have some high frequency hearing.

Peter

[jim_jobe]

I have tinnitus which isn't constant and is of two types:
The first is pulsating, where I can hear my pulse sometimes in my head; sometimes faint and sometimes loud. The second comes on suddenly amd is a piercing whistle, accompanied by deafness. The medics I've seen can't explain these satisfactorily and what I've been speculating is that perhaps there is an AGC mechanism in the auditory pathway to the brain which is causing faint noise in the brain to be amplified/detected when the normal inputs are reduced. I wonder if anyone has done any research on biological amplification in the brain?
Jim

[m0cemdave]

When I was younger I could hear up to around 22 kHz, and would often notice leaky pilot tone (19kHz) from stereo tuners.
As a child I was always very bothered by 405 line whistle, and 625 line whistle was very audible too. I could also detect some species of bats that made very loud squeaks somewhere around 20kHz.

Here's another interesting psychoacoustic phenomenon:
Take a tuning fork (A-440 works well for this) and strike it. Gradually bring it closer to your ear so the volume increases. What happens to the pitch?

[Peter.N.]

My tinnitus is a continuous approx. 10khz whistle that appears to come from the middle of my head rather from one or other ear - life is full of variety.

Peter

[John_BS]

What happens to the pitch?

Worrying! It seems to drop about 2%

[colly0410]

Originally Posted by Peter.N.

My tinnitus is a continuous approx. 10khz whistle that appears to come from the middle of my head rather from one or other ear - life is full of variety.

My tinitus is about 8 khz & in my left ear only, it's always there but varies in loudness. I usually stood on the right of the Abbot SP guns when they went bang, don't think that it's a coincidence it's in my left ear only..

[daviddeakin]

Originally Posted by m0cemdave

When I was younger I could hear up to around 22 kHz, and would often notice leaky pilot tone (19kHz) from stereo tuners.

That is highly unlikely. It is more likely that what you 'hear' is the intermodulation of these tones and their harmonics within the ear.
An interesting result of this in-ear nonlinearity is that if you listen to a 10kHz square wave, it doesn't sound like a pure sine wave, even though all its harmonics are 30kHz and higher!

1. The wanted tone is applied to the human subject at a known sound-power level (raised in stages to eventually generate an audio level v distortion curve).
2. A second pure tone, set to a frequency slightly different to the harmonic to be investigated, is brought up in level until the subject reports a low-frequency beat of maximum amplitude: at this point the second tone is at the same level as the harmonic generated by the ear.
3. The experiment is repeated at several wanted frequencies and for as many harmonics as desired.

Recently I was reading one of M G Scroggie's articles about a wave analyser, in Wireless World. It worked in exactly the same way!

[Jeremy M0RVB]

Crikey are we all going deaf then? I've had a hearing and balance disorder for maybe 16 years now that has left me with very loud tinnitus, now on both sides. On the left I have a pulsating and a constant multi-pitch 'scream', then a constant note on the right, and a hum that moves about and seems to be behind me (it's not, I checked with a mirror!). This leaves holes in speech such that as peoples voices naturally fall and rise during conversation I lose bits which makes hearing all the more hard. IIRC one hearing test reckoned the left was something like 58db down, and it's worse since. So I tend to use headphones or earbuds for the TV, music etc. but I think (not sure) that I now tend to 'fill in' music from memory - I know what the song is and so I guess the bits I don't hear fill in from memory. Most annoying. Look after your ears!

[Humber888]

Re #18.

I've tried the test and can hear up to 10kHz - but well down on 8kHz. So not that great for a 60 year-old who didn't go to many concerts. Then I found the foam on the Beyer headphones was moulting and have had to order new earpads and clean my ears out!

I have been mulling over a different question of how far off pitch does a turntable need to be before 'Mr average' notices, but will post that separately rather than mess up this thread.

Mike

I must be lucky, all I am getting as a low pass filter that goes lower with age, it's about 12-14kHz now (at 54). 4.5kHz bandwidth AM radio is perfectly clear to me, I have some time to go yet, I hope.

[paulsherwin]

Originally Posted by Humber888

Re #18.

I've tried the test and can hear up to 10kHz - but well down on 8kHz. So not that great for a 60 year-old who didn't go to many concerts.

This is pretty typical for a 60 year old with otherwise good hearing - I don't think there's anything to worry about.

I've just done the test and can hear 10kHz but not 12kHz using a pair of Sony MDR-V50 headphones. This is pretty much what I'd expect at 59.

[m0cemdave]

Originally Posted by daviddeakin

That is highly unlikely. It is more likely that what you 'hear' is the intermodulation of these tones and their harmonics within the ear.

I had it measured at the time with pure tones, but I should point out this was when I was about 20 years old. As for the pilot tone, it took quite a bit of investigation to discover why there was a high pitched whistle when my parents' hi-fi was switched to the VHF tuner (an old Whafedale WFM1, remember those?)

[mark pirate]

Good friends of mine run a hearing centre, recently we were chatting about the loss of higher frequency hearing with age. I mentioned that I did a test with my signal generator and a couple of friends who were visiting at the time.
The results were quite interesting, I could just about hear 12.5khz, while my friend who is some five years older than me could hear 14khz.
The youngest amongst us had trouble hearing 10khz!

I had a full hearing test a couple of months back, and when looking at the graph produced was quite shocked at how the high frequency's rolled off
Apparently quite normal for my age (52).
I too tend to have the treble turned up when listening to music these days, but can still plainly hear the line whistle on my 405 line TV's!

Mark

[jamesperrett]

I must admit that I was surprised last night when I found I could no longer hear a 15kHz lineup tone from the Revox - but then I realised that I hadn't noticed the line whistle from the TV for a while either. Fortunately the Revox is still rock solid at 15kHz according to the meters on my recording software.

In my younger days I had no problems hearing the 19kHz FM pilot tone and I remember going to the Foxhills country club near Woking where they had an ultrasonic sensor by the entrance that must have been working around 21kHz. It didn't bother the people working there but I found it extremely annoying whenever I went past.

[fetteler]

Seems to me that the thread has drifted into discussing frequency response rather than non liearity

Steve.