2

There are several online hearing frequency tests. Here is one example with over 10 million views: https://www.youtube.com/watch?v=H-iCZElJ8m0

The idea is that the recorded audio consists of a pure sine wave with frequency ranging from some small number up to tens of thousands of Hertz. By stopping the video when you can't hear it any more, you find out how high a frequency you can hear.

It occurs to me there are many problems with this. Even if the sound was recorded with CD quality of 44.1K sample/seconds, it would have trouble representing a sine wave in the range > 10,000 Hz. Then you take that signal and pass it through MP3 compression. I imagine that the actual recorded signal at this point is barely more than a crackling sound.

Then of course there's the limitations of my device's sound card, and the limitations of my speakers.

Are these able to test anyone's hearing above? Is it even possible to perform such a test on a consumer device?

Improve this question
1
  • 1
    ncbi.nlm.nih.gov/books/NBK53869 suggests that standard hearing tests only go up to about 8000 Hz, and that criteria for hearing loss are only based on hearing ability in the range below 4000 Hz. So the range above 10 kHz may not be diagnostically meaningful anyway.
    – Nate Eldredge
    Commented Jul 26, 2020 at 17:34

2 Answers 2

Reset to default
4

I have downloaded audio using youtube-dl in the best quality(opus @160k (48000Hz)) from the linked video and opened it in Audacity. The spectrogram displays that the file contains the claimed frequencies.

enter image description here

The concerns regarding troubles in representing sine waves are unwarranted due to Nyquist–Shannon sampling theorem, which proves it is possible to accurately represent frequencies up to half of the sampling frequency. In the case of a 48kHz file, it means the maximum sound frequency of 24kHz.

A quick check of cheap speakers shows that they should be capable of playing 17kHz frequency. The more expensive model claims the possibility of playing up to 20kHz.

The requirement for precise results(which is necessary for medical diagnosis) is a controlled environment and calibrated hardware, so the test in question is purely indicative. (Credits to Mast)

In conclusion, by performing the test one should be able to determine one's perceivable frequency range. However, the results may not be accurate due to noise present around the tested frequency or lack of hardware calibration.

Improve this answer
3
  • 2
    "I cannot address the part of the question about the medical validity of the results." Quite simply put: if you really want to know, you need a test in a controlled environment with calibrated hardware. You don't get that at home. Online tests are purely indicative, after taking into account the hardware limitations as described in the answer.
    – Mast
    Commented Jul 26, 2020 at 20:02
  • 2
    Welcome to Skeptics! This answer has a lot of assumptions, which means its conclusions are rather unsafe.
    – Oddthinking ♦
    Commented Jul 26, 2020 at 21:27
  • Main problem with this test is hardware calibration. First one is sensitivity of used headphones (in dB/mV) which vary between each headphones and second is their frequency range. Manufacturer might say they work up to 20 kHz, but question remains how loud. For manufacturer -30 dB might be good enough at 20 kHz.
    – Rok Tavčar
    Commented Jul 27, 2020 at 11:43
1

The question is complicated.

First, we should be clear that audiologists conduct more elaborate tests than just "pure tone audiometry". Wikipedia provides a good summary, including examining the ear, and conduction more functional tests like the Words-in-Noise Test.

If we look at the hearing tests conducted by YouTube memes, Miles Kitaro tweeted an analysis of a tweet of a YouTube test:

due to video compression, the audio in this video past ~15,500 hz is absent.

(I checked Kitaro's bio, and there doesn't appear to be any relevant evidence of expertise in audio analysis, so this should be treated as "some random person on the Internet claims it.")

The next step up after a video with a fixed range is a mobile app designed specifically to test hearing.

In 2016, there was a review of the evidence, Validated Smartphone-Based Apps for Ear and Hearing Assessments: A Review.

App store search queries returned 30 apps that could be used for ear and hearing assessments, the majority of which are for performing audiometry. The literature search identified 11 eligible validity studies that examined 6 different apps.

They concluded:

Very few of the available apps have been validated in peer-reviewed studies. Of the apps that have been validated, further independent research is required to fully understand their accuracy at detecting ear and hearing conditions.

So, if even the dedicated apps aren't to be trusted yet, the YouTube variants should be treated with skepticism.

(To be fair, yet another peer-reviewed analysis of one smartphone app, that was dated after this review that was quite positive:

Within the adult sample, 94.4% of thresholds obtained through smartphone and conventional audiometry corresponded within 10 dB or less. There was no significant difference between smartphone (6.75-min average, SD = 1.5) and conventional audiometry test duration (6.65-min average, SD = 2.5). Within the adolescent sample, 84.7% of thresholds obtained at 0.5, 2, and 4 kHz with hearTest and conventional audiometry corresponded within ≤5 dB. At 1 kHz, 79.3% of the thresholds differed by ≤10 dB. There was a significant difference (p < 0.01) between smartphone (7.09 min, SD = 1.2) and conventional audiometry test duration (3.23 min, SD = 0.6).

They concluded that, combined with calibrated headphones, this could be used to conduct "air conduction hearing threshold" tests.

The conclusion seems to be: There is potential for relatively inexpensive equipment to be used to perform one sort of testing, but simply relying on typical equipment deployments and YouTube and Twitter's video compression algorithms cannot be relied upon.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .