Skip to content
HearingTracker
Log in

Why is such imprecision allowed in pure tone audiometry?

12 replies634 viewsLast reply

Link to this topic
Reply
Edited

First, the 5 steps being the norm rather than 2 dB. Maybe that kind of precision is redundant for determination of hearing loss, but in hearing aids it causes an imbalanced frequency response - one threshold being almost inaudible and another very audible.

Then the allowed +/-5 dB error between two properly conducted pure tone audiometries. Assuming they’re even properly conducted… Single Audiometry – A Basis For NIHL Claims Handling Schemes?2

Then the impact of attenuating cups… And “professionals” ignoring you mentioning auditory hallucinations are an issue during the testing and then being mad at me for clicking when there was no sound.

If one test says 30 at 125 Hz and another says 45 dB at 125 Hz for the same ear, which one am I to believe? One says 75 dB at 8000 Hz and another says 60 dB? One says 65 dB at 1000 Hz, another says 70 dB, teenage ones somehow said 75 dB? What’s the value of Real Ear Measurement at this point?

The thresholds in self-performed situ audiometry in Oticon Genie 2 are very stable in comparison, but I don’t know if hearing aids will provide a correct fitting formula, also no 2 dB steps and no uncomfortable level measurement (which wasn’t done at any pure tone audiometry either). The middle frequencies (1000-3000 Hz) are very affected by loudness recruitment for me; VAC+ is too mid heavy, is too quiet, adult is a little better, pediatric DSL finally maximalizes audibility but the peaks in speech are uncomfortable, I’d have to decrease to below even 107 dB which causes very annoying clipping (107 dB MPO clips the obnoxiously loud train door beeping which I don’t hear at all without hearing aids so it shouldn’t even get so loud? the loudest frequency at 80 dB input is amplified by 22 dB to 102 dB and the beeping input shouldn’t even be 80 dB?!). The max compression ratio possible to set is 3.7 which isn’t enough and I wonder if going that high isn’t starting to degrade speech comprehension…

8 people in this conversation7 min readJump to latest

So are you angry at the individual or the test standard that’s been established?

The governing body that establishes the standard based on historical determination and modern best practice. Fitting formulas are a different argument.

Test variability is another issue, there are multiple factors that can cause this, both on your side and the side of the professional. Even something as tame as the residual effects of a head cold or blowing your nose can result in 15-20dB variation in threshold levels.

That being said, in my book, tests relying on headphones rather than using inserts: when inserts have been available for 30 years open the door to several of the questions you are raising about accuracy.

Both, because what’s the point of if pure tone audiometry itself is wrong? And why can’t I do it myself if it’s not for determination of disability, why do I have to guess? I wasn’t sick during testing.

@secondarytrefoil said:

And why can’t I do it myself

Testing bias - if you present a sound to yourself you can ‘tune-in’ to listen for it. You’ll usually yeild a 5-10dB improvement by doing this.

doesn’t just check that the aids are meeting threshold targets, but you’re certainly right that a poor isn’t going to help. However if a 5dB error only results in a 1.5dB difference in gain applied - a 10dB error is only going to give you a 3B error in total output, which isn’t hugely significant unless you’re on really marginal verified gain.

The in-situ process could offer an automated test to present the tones without user selection. Perhaps there’s a reason this isn’t done?

Could one just bump the in-situ by 5dB?

In reply to @Um_bongo

@Um_bongo said:

That being said, in my book, tests relying on headphones rather than using inserts: when inserts have been available for 30 years

I’ve had hearing devices for 15 years now, 7 of these years with a . The only time I had inserts used was when I had my CI evaluation. Every other audiology visit has been with headphones. I might add some of the headphones have been in various stages of disrepair.

Um_bongo are you saying you get a more accurate with inserts?

Usually yes. Especially when there’s significant asymmetry.

Even with symmetrical losses, the variation in pinna/concha size, canal volume, pressure from cans, leakage, transcranial vibration and pad surface condition: all can lead to errors beyond the knowledge of the Audiologist. Each time you place the cans differently on a different head you roll the dice.

I’m not saying inserts are perfect, but they certainly cut down the variables.

I got this from Google

why are 5 db steps being the norm rather than 2 db

In audio engineering and acoustics, 5 steps are often preferred as a “practical” norm because they represent a clearly noticeable change in loudness to the average listener. While smaller steps like 2 dB or even 1 dB are used for precision, 5 dB provides a more efficient balance between perceivable impact and user control.

Perceptual Reasons
The human ear does not perceive sound intensity linearly; it follows a logarithmic scale.

1 dB: Typically the smallest change a human can detect under ideal conditions, often considered imperceptible in a general environment.
3 dB: Represents a doubling of power but is only barely noticeable to the ear.
5 dB: Often cited as the threshold for a clearly discernible or significant change for most listeners.
10 dB: Perceived as twice as loud (a rule of thumb).

Practical and Efficiency Reasons
Decision Speed: Using 5 dB steps allows engineers and users to make rapid, significant adjustments without getting bogged down in minor “micro-adjustments” that the listener might not even hear.
Equipment Design: Many physical controls (like stepped attenuators or EQ knobs) use 5 dB increments to provide a wide range of control without requiring an excessive number of physical “clicks” or “detents”.
Hearing Safety: In industrial standards like the 5-decibel rule, for every 5 dB increase in noise, the safe exposure time is cut in half (e.g., 90 dB for 8 hours vs. 95 dB for 4 hours). This makes 5 dB a vital benchmark for safety assessments.

When 2 dB (or smaller) is used instead
While 5 dB is the “norm” for broad adjustments, smaller steps are standard in specific scenarios:
Mastering: Engineers often work in 0.5 dB or 1 dB increments because they are making final, surgical adjustments to an entire mix.
Audiometry: Hearing tests may use 2 dB steps to find a more precise “threshold of hearing,” though studies show 5 dB steps are often just as reliable for general testing.
Fine Mixing: High-end faders on mixing consoles or within DAWs allow for infinite or 0.1 dB resolution to perfectly balance competitive elements like a vocal against a snare drum.

In reply to @Um_bongo

@Um_bongo said:

Usually yes. Especially when there’s significant asymmetry.

Well, that describes me. And yes, for years now, the good folks at Kaiser have used inserts rather than headphones for testing. Add to that: tinnitus in my right ear. Pretty extreme, and I notice it most of all in the quiet of a testing room. Which doubtless leads to a fair amount of false positives for the beeping sound. No matter, the audis are used to this. I’ve had massive surgeries done on my ears and the canals–and middle ear–structures vary widely from one ear to the other. So…apparently it’s been determined that I require inserts for testing. Fine with me! p.s. Kaiser always tests for bone conduction as well. I have trad aids, but also the Osia baha on my right ear, which I use instead of BTE’s for everyday. It’s surprising how much bone conduction of sound to the opposite ear there is.

I’m not sure many people run with both calibrated at the same time in the private sector. I could see why you would do it in a diagnostic setting though.

I run them because they give more accurate results. I don’t particularly see any advantage in running cans. Especially if you’re doing , you want reliable threshold results.

They use on ear headphones at the NHS hospital audiology department I go to. Think they are a rubber material. Maybe it’s cheaper.

There is some research out there about how insert earphones and supra-aural earphones will yield different low frequency thresholds in cases of eardrum perforations i.e. low frequencies will be better with supra-aural earphones in ears with perforations. However, high frequencies will tend to be better with insert earphones (I am less familiar with research on that topic). I have heard some examiners discuss using supra-aurals for frequencies up to a certain point and then switching to inserts for frequencies above that. I think the most important part is being consistent, testing the same way every time so your data ends up being a truer reflection of hearing sensitivity/changes rather than an amalgamation of several different variables.

Related discussions