or Connect
Head-Fi.org › Forums › Equipment Forums › Sound Science › Everybody has different ears... but HOW DIFFERENT?
New Posts  All Forums:Forum Nav:

Everybody has different ears... but HOW DIFFERENT?

post #1 of 21
Thread Starter 
One of the old saws of audiophilia is "everybody has different ears". It's usually used as an excuse not to have a balanced response curve, because "flat to you might not be flat to me".

I see an obvious problem with this myself, because even if one's ears aren't flat, a flat response is still going to sound natural, because it duplicates how sound is balanced in the real world.

I saw a graph somewhere around here a couple of weeks ago that showed five different individuals' perception of the same sound. All of the individuals had "nomal" hearing and the deviation between them below 10kHz was never more than 4 or 5dB at any point in the response curve. Obviously, the perception of the top octave varied much more, but that isn't a particularly important octave when it comes to listening to music.

Can anyone point to a study that shows the range of normal perception of the core frequencies? Say from 40Hz to 10kHz. Something like the +/- X dB rating on audio equipment frequency response specs.

Curious to find out if that graph I saw was typical.
post #2 of 21
Quote:
Originally Posted by bigshot View Post

One of the old saws of audiophilia is "everybody has different ears". It's usually used as an excuse not to have a balanced response curve, because "flat to you might not be flat to me".

I see an obvious problem with this myself, because even if one's ears aren't flat, a flat response is still going to sound natural, because it duplicates how sound is balanced in the real world.

I saw a graph somewhere around here a couple of weeks ago that showed five different individuals' perception of the same sound. All of the individuals had "nomal" hearing and the deviation between them below 10kHz was never more than 4 or 5dB at any point in the response curve. Obviously, the perception of the top octave varied much more, but that isn't a particularly important octave when it comes to listening to music.

Can anyone point to a study that shows the range of normal perception of the core frequencies? Say from 40Hz to 10kHz. Something like the +/- X dB rating on audio equipment frequency response specs.

Curious to find out if that graph I saw was typical.


Hey bro.

 

First off, the premise may be faulty with this thread, from my perspective, if we are trying to ascertain if a 'neutral sound signature' is preferred by all humans.  Objectivists seem to perpetuate the same cognitive disconnect: that somehow the ears are responsible for the processing and assessment of sound.  This is patently untrue, and in order to be considered true, one must dismiss, ignore, and otherwise deny the processing that goes on within the brain when it comes to assessing sound, in this case music.

 

Secondly, your premise has nothing to do with perception, but instead is fixated on the excitation  of the tonotopic map as sound is received in the cochlea.

 

Am I reading you correctly so far?

post #3 of 21
Isn't a deviation of 4 or 5 dB a whole lot? An increase of 3 dB results in twice the perceived volume... which does mean that there is a huge difference. Also, many people have varying levels of hearing loss, especially at specific frequencies.
post #4 of 21
Thread Starter 
Preference isn't consistent. Accuracy is. The point of a flat response is accuracy... calibrating the sound to the way it occurs in real life. Once that baseline is achieved, then you can always use tone controls like salt and pepper to taste. Besides, the Harmon study already shows that people tend to prefer a neutral response.

But my real interest here is to find out how different our sensitivity to frequencies are. Do I hear with a bass bump and you hear with a lot of treble? What's the degree of difference?
post #5 of 21
I believe your right though. What is neutral is neutral for the majority of people, since it attempts to mimic exactly how an instrument would sound in real life. Of course after mastering, you can never be sure someone didn't tailor the sound the way they liked... not necessarily in a neutral manner.
post #6 of 21
Thread Starter 
Quote:
Originally Posted by ToddTheMetalGod View Post

Isn't a deviation of 4 or 5 dB a whole lot? An increase of 3 dB results in twice the perceived volume... which does mean that there is a huge difference. Also, many people have varying levels of hearing loss, especially at specific frequencies.

In general, the just detectable threshold between two different volumes in A/B comparison is .5 to 1dB with test tones and around 3dB with music. A 4dB variation would be detectable, but not by much.

Re: hearing loss: Music doesn't contain a lot of audible sound above 13 or 14kHz. Most normal hearing loss due to aging would be right around there too, so I limited my question to core frequencies (40Hz to 10kHz) just to simplify things and focus on normal hearing.
post #7 of 21
Thread Starter 
Quote:
Originally Posted by ToddTheMetalGod View Post

I believe your right though. What is neutral is neutral for the majority of people, since it attempts to mimic exactly how an instrument would sound in real life. Of course after mastering, you can never be sure someone didn't tailor the sound the way they liked... not necessarily in a neutral manner.

But those creative choices were balanced using studio monitors calibrated to a flat response, so we're right back to the flat response again.
post #8 of 21
I think this graph has a reference attached to it:
http://www.head-fi.org/t/324147/strange-frequency-response-graphs#post_4190019
Quote:
Originally Posted by bigshot View Post


Can anyone point to a study that shows the range of normal perception of the core frequencies? Say from 40Hz to 10kHz. Something like the +/- X dB rating on audio equipment frequency response specs.

Curious to find out if that graph I saw was typical.
post #9 of 21
Quote:
Originally Posted by ToddTheMetalGod View Post

Isn't a deviation of 4 or 5 dB a whole lot? An increase of 3 dB results in twice the perceived volume... which does mean that there is a huge difference. Also, many people have varying levels of hearing loss, especially at specific frequencies.

 

Hearing isn't linear and perception is certainly weird, so not really. It's kind of an accepted figure that 10 dB different is roughly perceptually twice as loud.

 

 

 

There are some HRTF measurement databases out there, this for one (CIPIC):

http://interface.idav.ucdavis.edu/sound/hrtf.html

 

Related brief paper:

http://interface.idav.ucdavis.edu/data/doc/CIPIC_HRTF_Database.pdf

 

Response depends on angle of incidence of sound of course. It's hard to get a sense on the vertical scale here, but see here:

 

 

I don't know which paper this is from, but also relevant:

See the gray shaded area for +/- 1 standard deviation. About 32% of people fall outside that range.

post #10 of 21
I found this an interesting light read:
http://www.davidgriesinger.com/binaural_hearing.ppt

Towards the end is a quote:
–“The most immediate observation is that the variation [in sound transmission from the entrance of the ear canal to the eardrum] from subject to subject is rather high…The presence of individual differences has the consequence that for a certain frequency the transmission differs as much as 20dB between subjects.”

Looking at graphs and reading some of these quotes I'm thinking the variation in response could be 15 to 20 db amongst normal individuals.
Edited by GrindingThud - 4/23/14 at 3:22pm
post #11 of 21
Thread Starter 
Quote:
Originally Posted by GrindingThud View Post

I think this graph has a reference attached to it:
http://www.head-fi.org/t/324147/strange-frequency-response-graphs#post_4190019

 

Someone a little further down says that is a measure of bone conduction. Not quite what I was looking for, but thanks. I can imagine that bones would vary a lot, depending on how thick and dense a person's bones are.

post #12 of 21
Thread Starter 
Quote:
Originally Posted by GrindingThud View Post

Looking at graphs and reading some of these quotes I'm thinking the variation in response could be 15 to 20 db amongst normal individuals.

 

I don't have time to dig too deep into this right now, but mikeaj's graph seems to indicate around +/-5dB or so, diverging greatly after 10kHz, which was pretty much what I was expecting.

post #13 of 21
That's a one sigma graph, but yes pretty close. smily_headphones1.gif
Quote:
Originally Posted by bigshot View Post

I don't have time to dig too deep into this right now, but mikeaj's graph seems to indicate around +/-5dB or so, diverging greatly after 10kHz, which was pretty much what I was expecting.

Edited by GrindingThud - 4/23/14 at 6:33pm
post #14 of 21
Quote:
Originally Posted by ToddTheMetalGod View Post

I believe your right though. What is neutral is neutral for the majority of people, since it attempts to mimic exactly how an instrument would sound in real life. Of course after mastering, you can never be sure someone didn't tailor the sound the way they liked... not necessarily in a neutral manner.

Actually this is a good point.  Even though our ears have various FR, that FR is being used all the time, so therefore, it is what is heard, and that's reality.  If the audio system can represent reality of the sound, I don't see how it can differ from reality to each individual's unique hearing.  

 

So, the variations to people's hearing should not matter.  What I believe matters is if the system is providing what is represented in reality.  If one's ear has high frequency dip, and another low.  For real world sounds, they are hearing how their ear is meaning for them to hear.  If the audio system presents the sound the same as reality, the listener's ear variance should not factor into it since it is how they are meant to hear.

 

This is in regards to accurate audio system.  As for the listener to perceive neutral sound, the system has to accomodate for each unique hearing ability.  That brings the question of what is neutral to people, it seems relative.  It's easy to show it on electrical system with graphs, but for in regards to perception, there will be variances to people's perception.


Edited by SilverEars - 4/23/14 at 6:53pm
post #15 of 21
Just to quickly interject on the confusion of 3 dB/ 6 dB/ 10 dB/ etc. (and to hopefully not make the confusion any worse smily_headphones1.gif ), let's remind ourselves about decibles and how they relate the amplitude of signals.

dB = decibel = 1/10 of a Bel
Bel = Log[ value / reference ] (Here, 'Log' means Logarithm base 10.)

Therefore, the value in dB of some signal is given in relationship to a reference as 10* Log [ signal/reference ]

A Bel is a unitless measure of amplitude on a logarithmic scale. It is given in terms of a reference value. Not all decibel measures are equal, due to differences in the reference value.

For example, the amplitude of an electrical signal may be given in dB amplitude ("dBu") or dB power (simply "dB"). This is a frequent source of confusion, because folks often think in terms of the voltage amplitude of a signal; however, folks also often think in terms of the power of sound. The problem arises because the voltage sent to audio transducers is typically related to the sound pressure; however, the power of the electrical signal and the sound wave are proportional to the voltage-squared and the pressure-squared, respectively. One can see how this leads to a factor of two difference between the dBu and dB ratings for the same signal.

Lets say the voltage amplitude is V and is being driven across a load with impedance R. Furthermore, lets use a reference voltage amplitude of V_ref.


(Recall P= V^2/R)

Similarly, V and V_ref could be a pressure amplitude and reference, respectively. R would be the acoustic impedance of the the medium through which the sound is traveling, and P would be the power-per-unit-area.

dBu = 10 * Log[ V / V_ref ]
dB = 10 * Log[ (V^2/ R) / ((V_ref)^2/R) ]
= 10 * Log[ ( V / V_ref )^2 ]
= 20 * Log[ V / V_ref ]
= 2*dBu

Hence, a 1 dBu change (amplitude) corresponds to a 2 dB change (power).

When folks talk about a 10 dB change corresponding to a doubling in perceived volume, they're talking about 10 dB power. Here, the amplitude change is 5 dBu.

As far as base 10 is concerned:
- 10 dB corresponds to a change in power by a factor of 10
- ~3.01 dB corresponds to a change in power by a factor of 2 and a change in amplitude by a factor of sqrt(2)~1.41
- ~6.02 dB corresponds to a change in power by a factor of 4 and a change in amplitude by a factor of 2


Hope this helps clarify how dB differences relate to amplitude and power.

Cheers

(Naturally, one might have googled this to find the wiki page on the decibel scale linked here for convenience)
EDIT: errors fixed, thanks bigshot & wakibaki
Edited by ab initio - 4/24/14 at 7:28pm
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Everybody has different ears... but HOW DIFFERENT?