• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Anyone interested in an at home hearing test?

Would you be interested in doing a test at home to check your hearing thresholds?

  • I'd love to test myself and I'll share the result here, on ASR

    Votes: 30 81.1%
  • I'd like to test myself, but would prefer to keep the result private

    Votes: 6 16.2%
  • Don't want to test myself, my ears are perfect

    Votes: 1 2.7%

  • Total voters
    37

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
This is intended to gauge the interest in a piece of software designed to measure the lower thresholds of audibility across the audio spectrum. I wrote it a couple of years ago for my own use, but had a few folks approach me recently to ask if they could try it. My question is: would you be interested in doing this audiology test yourself, and if yes, would you be also willing to post the result here, on ASR (not required, I just think it would be interesting to share these)?

If there's sufficient interest, I'll start a beta-test thread for the software (called EARFUL). Designed for Windows, but should run under Wine on other O/S.

Since I just dug up the software, I wanted to see how my hearing has changed since the last time I used it, in 2018. I'm well over 50, so some higher frequency loss is expected. Here's the 2018 result (blue), plotted against an ISO audibility threshold curve (red):

1594520945835.png



And here's the result today, 2 years later, actually looks very similar with some minor changes, perhaps due to the listening environment being different:

1594521011332.png
 
Last edited:

JoachimStrobel

Addicted to Fun and Learning
Forum Donor
Joined
Jul 27, 2019
Messages
518
Likes
303
Location
Germany
This is intended to gauge the interest in a piece of software designed to measure the lower thresholds of audibility across the audio spectrum. I wrote it a couple of years ago for my own use, but had a few folks approach me recently to ask if they could try it. My question is: would you be interested in doing this audiology test yourself, and if yes, would you be also willing to post the result here, on ASR (not required, I just think it would be interesting to share these)?

If there's sufficient interest, I'll start a beta-test thread for the software (called EARFUL). Designed for Windows, but should run under Wine on other O/S.

Since I just dug up the software, I wanted to see how my hearing has changed since the last time I used it, in 2018. I'm well over 50, so some higher frequency loss is expected. Here's the 2018 result (blue), plotted against an ISO audibility threshold curve (red):

View attachment 72924


And here's the result today, 2 years later, actually looks very similar with some minor changes, perhaps due to the listening environment being different:

View attachment 72925
I have done a few hearing tests using REW, see
https://www.audiosciencereview.com/forum/index.php?threads/hearing-eq.9632/#post-256552
I would be interested to see how your software compares to my efforts.

I see two issues at the moment:
The hearing test software needs to fit precisely into my current audio path, so that any findings can be used for designing an EQ response. Hence it needs to be HDMI in my case (out of Windows). I would not be interested in having absolut precision, as this would require calibrated headphones, but more like a relative approach that could be quickly used with visiting guests to acquire an own database.
A real improvement over my REW based method would be the use of a frequency range instead of a single sharp one, like 7800—8200 hz. That would overcome the issue that the ear of elderly people might just not be able to hear that sharp frequency that wanders through the room with all its reflections. That can be circumvented by doing many points, but that might not be this same. At the time when I did my test, REW could not produce such a frequencyband noise.

My final thoughts are, that the curve that you use for comparison should be age related. Some people might like the idea comparing their hearing to a young people‘s ISO curve. However, there are a handful publications (see my post), that allow comparison to the relevant age group. That helps to avoid frustration too.

BTW, it is hard to understand why an AVR would not include such a test in their rich portfolio of included stuff. Unless they are afraid of frustrated customers.
 

DChenery

Active Member
Forum Donor
Joined
Feb 6, 2020
Messages
205
Likes
453
Location
Nanoose Bay
Count me in. Always interesting to test drive some new software. As an aside, thanks for Distort. It has been an absolute blast to play with, as well as being quite educational.
 

BDWoody

Chief Cat Herder
Moderator
Forum Donor
Joined
Jan 9, 2019
Messages
6,948
Likes
22,625
Location
Mid-Atlantic, USA. (Maryland)
I'll go for this.
 

LTig

Master Contributor
Forum Donor
Joined
Feb 27, 2019
Messages
5,760
Likes
9,442
Location
Europe
This is a good idea but wouldn't the results depend strongly on the FR of the choosen transducer, and the ambient noise of the room where the test is performed (except when using closed headphones)? How would you compensate for that?
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East

On the surface, similar in many ways to the Digital Audiometer software, except maybe the price; EARFUL will be free.

I recommended to use headphones with the test, although any transducer can be used. The test signal is an interrupted pure tone at any desired number of log-spaced frequencies in the selected frequency range. The user moves the volume control up and down until they can no longer detect the sound. They can go to next or previous frequency and repeat the process and readjust as many times as needed.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
I have done a few hearing tests using REW, see
https://www.audiosciencereview.com/forum/index.php?threads/hearing-eq.9632/#post-256552
I would be interested to see how your software compares to my efforts.

I see two issues at the moment:
The hearing test software needs to fit precisely into my current audio path, so that any findings can be used for designing an EQ response. Hence it needs to be HDMI in my case (out of Windows). I would not be interested in having absolut precision, as this would require calibrated headphones, but more like a relative approach that could be quickly used with visiting guests to acquire an own database.
A real improvement over my REW based method would be the use of a frequency range instead of a single sharp one, like 7800—8200 hz. That would overcome the issue that the ear of elderly people might just not be able to hear that sharp frequency that wanders through the room with all its reflections. That can be circumvented by doing many points, but that might not be this same. At the time when I did my test, REW could not produce such a frequencyband noise.

My final thoughts are, that the curve that you use for comparison should be age related. Some people might like the idea comparing their hearing to a young people‘s ISO curve. However, there are a handful publications (see my post), that allow comparison to the relevant age group. That helps to avoid frustration too.

BTW, it is hard to understand why an AVR would not include such a test in their rich portfolio of included stuff. Unless they are afraid of frustrated customers.

Interesting point about the range of frequencies, something I can add as an option.

The output is using standard windows audio, using WASAPI, ASIO or Direct Sound.

You can load any curve you like for comparison, including your own previous measurements. The ISO curve is just one that I had on hand.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
This is a good idea but wouldn't the results depend strongly on the FR of the choosen transducer, and the ambient noise of the room where the test is performed (except when using closed headphones)? How would you compensate for that?

Calibration can be applied to the result, for example an REW calibration file. Ambient noise level meter can be added, although that will add to the complexity of the test by requiring a mic and additional calibration.
 

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
I think I'd do this, why not, I'd do it on my closed back NAD HP50 headphones that extend all the way into the low bass and they're EQ'd to the Harman curve by Oratory1990 so that should be pretty valid, although I have a feeling they would suck at high frequency because they're not high frequency rich compared to my AKG K702 and Oratory1990 EQ procedure only calibrates up the 10kHz. I'd give it a try though, would have to make sure my long Covid hair is not obstructing my headphones! :p

EDIT: actually I could do it on my AKG K702 too to see how the results differed.

EDIT#2: I'd probably look to do the listening tests at my normal music listening volume level (which would be roughly controlled by the position of the volume knob on my headphone amplifier, lol. Obviously I understand that I'll be digitally altering the volume using your software during the test, but I guess my normal listening volume would be the maximum volume I would experience....afterall I don't want to wreck my ears doing your hearing test, and I'm not aware of any guidelines on what volume we should be doing this test at.
 
Last edited:

Hayabusa

Addicted to Fun and Learning
Joined
Oct 12, 2019
Messages
787
Likes
518
Location
Abu Dhabi
I am interested, but how would you calibrate the absolute level?
 

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
I am interested, but how would you calibrate the absolute level?
Yes, that was what I was insinuating in my edits. Also the effect of the EQ and/or natural frequency curve of the speaker/headphone; however, if you capture enough data (test enough people) then these irregularities would eventually equal out as long as there were no pathological patterns in widespread 'mis-calibration' of speakers/headphones.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
I am interested, but how would you calibrate the absolute level?

Without calibration this will give a relative difference between frequencies, which is probably the more interesting part of the test. To compare two different captures by different people you’d need to index them to, say, the lowest amplitude in the two measurements or to a specific frequency.

With speakers an SPL meter can be used to measure an absolute level. With headphones, this will be more involved.
 

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
Without calibration this will give a relative difference between frequencies, which is probably the more interesting part of the test. To compare two different captures by different people you’d need to index them to, say, the lowest amplitude in the two measurements or to a specific frequency.

With speakers an SPL meter can be used to measure an absolute level. With headphones, this will be more involved.
I can visualise that (and I think that's valid).

(You'll lose the bottom and top ends of the frequencies though for people who have run at lower volumes, their data points won't exist there because they won't hear them, but you'll get correlation around the more audible frequencies).
 
Last edited:
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
Obviously I understand that I'll be digitally altering the volume using your software during the test, but I guess my normal listening volume would be the maximum volume I would experience....afterall I don't want to wreck my ears doing your hearing test, and I'm not aware of any guidelines on what volume we should be doing this test at.

Correct. The maximum volume from EARFUL is 0dBFS into the selected Windows driver. This will never exceed the volume set by Windows volume controls, as well as, by the volume control of the preamp in the audio chain. That said, I’ll certainly add some warning before you use it to set the levels appropriately! Maybe I’ll also add a setting to limit the maximum level to something below 0dBFS, for those who want to be extra safe.
 

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
Correct. The maximum volume from EARFUL is 0dBFS into the selected Windows driver. This will never exceed the volume set by Windows volume controls, as well as, by the volume control of the preamp in the audio chain. That said, I’ll certainly add some warning before you use it to set the levels appropriately! Maybe I’ll also add a setting to limit the maximum level to something below 0dBFS, for those who want to be extra safe.
I think you'd be alright with 0dBFS and a warning, because a lot music approaches 0dBFS although pure tones do sound a little louder, maybe something like operating at -2dBFS would be safe enough although I'm sure 0dBFS is safe & a warning.....you could ask them to check/calibrate (comfortable) volumes at 1kHz or something so they don't end up trying stupidly high volumes at the low & high end of the frequency spectrum (in a bid to hear the tone).
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
Thanks everyone for voting! Appears to be enough interest, so I'll spend some time cleaning up and testing EARFUL and post a new thread for those willing to try it out. Keep the feedback coming, if you have some other ideas on features/improvements or potential problems I might want to address in this software.
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,825
Could you build in more of the equal loudness curves and allow the user to shift their result up and down? That may help explain why it may be flatter, since users will likely be listening closer to the middle of the dynamic range of hearing than the extremes.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,630
Likes
10,205
Location
North-East
Could you build in more of the equal loudness curves and allow the user to shift their result up and down? That may help explain why it may be flatter, since users will likely be listening closer to the middle of the dynamic range of hearing than the extremes.

The design already allows any curve (defined in any text file) to be loaded for comparison. I just used an ISO threshold equal-loudness curve for my own testing as I was interested in how my hearing compared to an average. In theory, you can load other loudness curves, measurements from others, your own last year's measurement, whatever else you'd like.

Moving the curves up and down is a good idea, I'll definitely add it to the feature list.
 
Top Bottom