The Free App-Based Acceptable Noise Level Hearing Test Everyone Can Use

The Free App-Based Acceptable Noise Level Hearing Test Everyone Can Use

Originally limited to clinicians using an audiometer to vary the levels, the (somewhat-misnamed) “Speech in Noise” test module in the free uHear iPhone app has this very valuable tool; and since it is based on relative, not absolute sound levels, calibration is not an issue. [For this same reason, we do Not Recommend the “Hearing Sensitivity” module in the app: Instead we recommend the Better Hearing Institute’s Across America Hearing Check Challenge, which is based on a questionnaire.]

UPDATE: Although this article is now five years old and needs updated images of the new user graphics in v2.0, everything here still applies — With one notable exception: Although the original version had no lower limit of noise, which is helpful for (Central) Auditory Processing Disorder testing, the newer version has an a lower limit of 20 dB, meaning it can measure Acceptable Noise Levels (ANL’s) up to 20 dB. We have conveyed this to the appropriate people at Unitron.

The Acceptable Noise Level test is a valuable tool both for hearing impaired people as well as for professionals, as it tells us how much distracting noise bothers the person. This now-free test provides both an indication of how much hearing aid noise reduction technology is needed, and more importantly also an overall indication of the success in general with hearing aids.

In fact, the ANL test is such a valuable troubleshooting tool for educational speech pathologists who perform screenings for APD (see the second Bootnote) as well as for hearing aid professionals, we recommend purchasing a $179 iPod Touch 4, if not an iPad, if an iOS device isn’t already owned. [That, and if you’re an SLP you can run all sorts of speech apps; while if you’re an audiologist, you can run the incredibly powerful AudioTools audio analysis app by StudioSixDigital: More on the latter in an upcoming story.]

How to take the ANL:

The way this implementation of the ANL test works is that the patient first sets the volume of the cold running speech on the dial to a “loud but comfortable” level (the MCL), touches the lock button, adjusts the volume of the background babble to the maximum the patient can tolerate, then touches the lock button again:

Unitron uHear screen shots

Left to right: uHear app test selection screen, ANL test setup, speech in quiet, speech in noise.
Note: The ANL test is somewhat improperly called “Speech in Noise”

Type in the initials; and jot down the score:

uHear ANL patient name entry and results screen shots

The patient takes this test 3 times, reporting each score. If there is more than a 2dB difference in the 3 scores, the patient takes the test 4 more times, reporting all 7 scores. If there are just three scores all close together, take the average. However, to get your arms around it when seven scores are measured, calculate the μ (mean) & σ (standard deviation) of the scores: You can use the online standard deviation calculator in Bootnote 3 to quickly get these values.

The 60 second demonstration of the original ANL module is cued to the 2:00 mark, with English open captions:

ANL Test Score Analysis:
  1. If the ANL score is under 7dB (i.e. high ANL), noise doesn’t bother the patient too much, and s/he can get by with less expensive, mid-level technology;
  2. If the score is 13dB or more (i.e. low ANL), noise bothers the patient a whole lot, and they need both the very best noise reduction technology, supplemented with assistive listening technology, such as FM, Mini Mic & such;
  3. If the score is 7 to 12dB, this is where it gets into a funky grey area, as there is really no correlation between how much noise the patient can tolerate and how much it really bothers them, i.e. someone who has an 8dB ANL may be really bothered by noise; and conversely someone who has an 11dB ANL may not be bothered as much. When selecting hearing aids, there are two ways to handle this situation, depending on whether the patient is a first-time user and/or has rejected amplification in the past, using clinical judgment and other factors such as history & lifestyle:
    • If the patient is an experienced user, leads a relatively quiet lifestyle, and is reasonably satisfied with their present hearing aids, then starting with mid-level or even basic hearing aid noise reduction technology should be considered, to save them money by delivering a better value;
    • If the patient is a first-time user, or more importantly has trialed and was dissatisfied with hearing aids, then start with top-tier noise reduction technology, to assure that they have a good experience. What we recommend for first time users is to give them two “everyday” programs, the first with the noise reduction set to the equivalent of the mid-level technology, and the second set to the premium level, instructing them to try both to determine their satisfaction. [Twenty years ago in the early days of laptop computers, when then-new active color displays were horribly expensive, David Pogue recommended placing two otherwise-identical laptops next to each other, with a stack of twenty $100 bills on the one with the passive LCD screen, and ask yourself which looks better.]

The ANL test has been clinically validated (see the References in the Bootnotes); and this iPod-based version is a very good (and free) implementation of it.

Post your ANL scores:

Go ahead, take the ANL test; and then post the results along with your hearing loss & type (if applicable) in the comment section of this article: You can leave it with your name, or leave it anonymously.


1) We use this $10 GLS Audio 6ft wye-cable splitter cord to connect our iPod to our audiometers, which has a ⅛” TRS stereo plug on one end & two ¼” TS mono plugs on the other;

2) Calculating the mean & standard deviation when there is a wide variation in ANL scores is something we stumbled across, as something unusual we’ve seen when (central) auditory processing disorder (APD) along with autism spectrum disorder is in play are astronomical and wildly varying ANL figures: One 21 year old lady has a μ of 39.6dB & σ of 6.8dB, while another 16 year old lady has a μ = 66.8dB & σ = 7.3dB. It raised our eyebrow when we saw these ANL scores; and when we bounced this off of a renowned professor who is an expert in APD to find out if there is any research into this, she responded as follows (links added)

I LOVE your question as it’s the question dying to be answered. I use the BKB-SIN [Bamford-Kowal-Bench speech-in-noise – Ed.] in many kids that I see with APD and although there is no published data/research, we see some astronomical results for that test (which is obviously different from the ANL but same concept). I wish I had research we could refer to for the ANL or any other well normed speech in noise test.

Ironically, I’ve been messing around with the idea of using the AZBio [speech-in-noise test – Ed.] as a test to look at APD… One of my colleagues does many [cochlear] implant evals and she leaves tools lying around the booth – I get dangerous with those tools.

I find it particularly interesting that both of these young woman are on the autism spectrum. Seems to me that this may support the inconsistency in listening skills that seem to result in some of the listening complaints of people on the spectrum.

Sorry not to be more insightful… I find it all very interesting and may look around at some more speech and noise with APD issues… I think this is a dog that hunts with a tool that is sensitive enough.

3) The population standard deviation measures the variability of data in a population; and is usually an unknown constant. The standard deviation σ & variance σ² of the population are given by:

population standard deviation and variance formulas

σ = population standard deviation
σ² = population variance
x1, …,xN = the population data set
μ = mean of the population data set
N = size of the population data set

If you want to delve a bit further into analyzing ANL scores, the sample standard deviation is an estimate, based on a sample, of a population standard deviation. For your convenience, the population standard deviation s and variance of the sample are given by:

sample standard deviation and variance formulas

s = sample standard deviation
= sample variance
x1, …,xN = the sample data set
= mean value of the sample data set
N = size of the sample data set

You can also use this online calculator to determine μ, σ, s, &

4) References with links:

5) Permalink to this article: Short link:

Comment problems: Read before posting
It’s been brought to our attention from several of our readers that they were having their comments rejected by the Akismet plug-in for WordPress as spam. This is unacceptable to us; and we are soliciting suggestions for a replacement. Unfortunately, we have to use something to screen for spam, as we were receiving over 100 spam comments per day at its’ peak. In the interim, to save retyping, we recommend selecting & copying all of your text to the clipboard: If your comment is accidentally rejected, simply paste it into an e-mail message, put “Rejected Comment” in the subject line, and send it to us at Dan@Snip.Net and we’ll manually post it for you~

Print Friendly

← Made for iPhone hearing aids & CI's: Android catches up with Bluetooth 4.0 Low Energy support LOCKED OUT! The Hearing Blog Editor was ordered out of the HLAA Convention! →

About the author

Dan Schwartz

Electrical Engineer, via Georgia Tech

Leave A Reply

Time limit is exhausted. Please reload CAPTCHA.

%d bloggers like this: