SoylentNews is people
SoylentNews
from the "can-you-hear-me-now"-is-too-cliche-of-a dept.
Ever wondered if you've hurt your ears? Too much rock 'n roll, or too many loud race cars? Just discovered this over-the-phone test which is offered free this month in USA, call 866-223-7575.
More info at nationalhearingtest.org:
The critical difference between the NHT and other telephone tests is that the NHT uses digits in noise, rather than pure tones. Previous objections to telephone-administered tests were based on the unreliability of pure-tone tests administered over the telephone. Because the NHT measures an SNR threshold, rather than one based on the absolute level of tones, the NHT can produce a reliable screening measure of hearing, despite the differences in sound levels produced by different telephones.
Scientists at VU University Medical Center in the Netherlands developed the first telephone hearing screenings based on spoken digits in noise. The Dutch National Hearing Test was introduced in 2004 and has since served as a model for similar tests throughout Europe and Australia. The United Kingdom, Denmark, France, Germany, Poland, Switzerland, Spain, and Australia have developed and introduced their own versions of a telephone hearing screening, all using digit sequences presented in noise. The growing popularity of this form of screening test is due in part to its demonstrated validity and reliability. Testing by telephone is also a convenient, inexpensive way to determine whether a person's functional hearing is within the normal range.
(Score: 0) by Anonymous Coward on Sunday May 11 2014, @10:44PM
Now I'll admit that I couldn't be bothered to read TFA, but based on empirical data I'll submit that there are more parameters to audibility then relative sound levels, the frequency response, linear and non linear distortion of the phone in question will differ widely and have a huge impact. Also I kinda also want to call BS on the statement that absolute sound level will not have an effect on how audible the signal is compared to the noise, I mean come on play the same signal at 20 dB and 70 dB, you don't think you have a better chance of hearing the signal buried in the noise with the louder version, even if the SNR remains constant? I hereby declare this test unscientific.
(Score: 3, Informative) by frojack on Monday May 12 2014, @12:00AM
From TFA:
The speech-to-noise ratio (SNR) is adjusted using an adaptive procedure, to seek the threshold for 50% correct identification of the digit sequences.
It seems they are adjusting the signal to noise ratio until you get half wrong.
Meanwhile everyone on a cell phone is adjusting the volume trying to get all correct.
(They tell you not to use a cell phone, and presumably they mean any phone on which you can adjust the volume).
They are ultimately measuring your ability to pick signal from noise.
Most people are trying to simply manage the volume so that they can hear.
At first glance, it doesn't even seem like both parties are working toward the same goal. But perhaps hearing difficulty manifests itself as signal to noise problems for most hard of hearing people.
(I'll admit to not knowing a great deal about how humans actually hear).
No, you are mistaken. I've always had this sig.
(Score: 1) by Ethanol-fueled on Monday May 12 2014, @02:34AM
Depends on your definition of "hear." I called the number just to see what it was like, and it sounded like a reasonably faithful but computer-generated female voice over Brown Noise [wikipedia.org], a lot like how big-city traffic sounds.
The traditional tests that use pure tones and no noise seem being about threshold as a function of frequency, but the test given in the submission is like one of those things that falls into the realm of auditory psychology or neurology rather than the physical condition of the organs specific to hearing.
So, what you said. I think it's kinda arbitrary, though, to use a female voice and a specific noise spectrum -- change it to a male voice and White noise and different people would score better or worse due to their own differing sensitivity of different spectra.
Disclaimer: I'm not a medical professional and am talking out my ass here, but it sounds plausible enough.
(Score: 2) by captain normal on Monday May 12 2014, @04:35AM
That seems what they are measuring. Signal from noise in a very limited bandwidth. About 3000 hertz from a low of ~400 htz to 3400 htz, which is the bandwidth of analogue telephone. They are not measuring range of tone, but ability to discern signal from noise. Which to me is a neurological ability.
"It is easier to fool someone than it is to convince them that they have been fooled" Mark Twain
(Score: 2) by hankwang on Monday May 12 2014, @07:32AM
"It's kinda arbitrary, though, to use a female voice and a specific noise spectrum"
They claim that the test results are highly correlated to those of traditional tests (threshold of hearing over a frequency range). I'm sure they tested a lot of voices and noise spectra in order to find the parameters that produce the highest correlation.
Typically, hearing damage starts in the 2--5 kHz range, where the ear is most sensitive and at the frequencies that are most useful to distinguish consonants and vowels from each other. It's an interesting question why SNR threshold at high SPL would be correlated to loss of absolute threshold. Suppose the signal is S and the threshold of hearing is T. With an added noise N, the overall SNR is S/(T+N). If the test subject's loudspeaker has an unknownn volume gain g, then
SNR=S/(T/g + N).
The quantity they're after is T, which cannot be solved for. So, hearing is a bit more complex than this, if this test still produces valid results.
Avantslash: SoylentNews for mobile [avantslash.org]
(Score: 1) by Anonymous Coward on Monday May 12 2014, @05:19AM
Actually, I suspect it's a concern about how well speech vs. noise is preserved by the lossy, voice-optimized codecs that cell phones use.