Australian Hearing’s Telscreen is the “world’s most advanced over the phone hearing test”. It is operated on the Australian Free-call number 1800 826 500 and is not available to mobile phones. This service offers users the ability to check their hearing just by using their phone and about five minutes of their time.
This post might seem a bit of a departure from the norm – but I’m interested in all things tech, and anything attached to a telephone qualifies. Heck, even Steve Jobs and Steve Wozniak were known phone phreaks, and they went on to develop Apple Computer.
Part of what makes this service interesting is the whole concept of telehealth. One of the potential benefits of the NBN was increased or improved telehealth abilities. Telehealth is the provision of a health-related service or information and this is actually an example of telehealth that requires only a telephone. How useful! While it will never be a replacement for a proper hearing check with a qualified hearing specialist, it’s a good “first step”. The fact that it is offered free of charge is a nice touch.
As usual, given that it’s a free call, I couldn’t help myself but investigate and get my hearing checked at the same time. It seems like a pretty nifty service to offer, and is more than meets the eye.
Getting on the Service
Getting onto this service isn’t as easy as it seems. The way you reach the service can and (in my experience) does impact significantly on the ability for the test to work correctly.
The number itself is not available from mobile phones. If you’ve reached the service before, you will be told about this, but if not, you would hear the following message:
The number you have dialled is either not available in your area or not available from a mobile phone. Please check the number before trying again. This is a free call.
I suppose the reason for this is that free calls are never free from mobile phones and the barring of them will help to prevent bill-shock for those who do not know. The other reason is a subtle technical reason – the CODECs and voice processors inside phones and smartphones have the ability to invalidate the test.
I actually gave this a try by using my Samsung Galaxy SIII which has audio processing to “eliminate” background noise. While running the test with the call routed through a G.711a uncompressed VoIP SIP provider, I noticed that any beeps caused the masking noise to be “automatically cancelled” as part of the speech processing logic within the phone. Use of different SIP codecs compromised the beeping in such a way that the test results changed dramatically.
As a result, it became clear that using a quiet land-line connected to a corded (not cordless, as DECT phones have their own CODEC, and others may increase the noise level) would provide the best test result. Using a VoIP SIP provider with G.711a uncompressed codec would also be sufficiently good, although this may depend on your ATA, phone or soft-phone, and connection reliability.
You can actually call this number free via Skype, although it uses G.729 codec which may interfere with the test slightly (although, when I tested it, it seemed G.729 is acceptable).
I have made some recordings of the calls I’ve made to the service in order to analyze just how the system works. These recordings were made by intercepting the packets travelling to and from my VoIP ATA to reassemble the call data using Wireshark. Apologies for the packet losses. This method inserts no additional noise, and results in “absolute” dB values which are not subject to incorrect volume settings, etc. The ATA and the SIP service required configuration to present a land-line number before the call would be connected.
Recordings of the service that I made are presented later on in this post. These recordings are not to be used to test your hearing, and are merely provided as illustrative examples under fair use in this academic analysis and discussion of the service.
The Testing Mechanism
The last time I got a hearing test, I was in Year 3 primary school, and I had visited the school clinic. There, the clinician gave me a set of headphones and told me to nod my head and say yes when I heard a tone. They used three or four frequencies, all at a set loudness, and the test was done in no time. It’s a pretty simple but lousy go/no-go kind of test.
This test operates a little differently, on the principle of Simultaneous Auditory Masking. Hearing loss or damage to the ear can cause changes in the way sounds of different frequency are perceived. It appears that in cases of hearing loss, the ability to distinguish different frequencies (auditory filter) becomes less narrow, and therefore single tones can be masked by adjacent noise or perceived as “one sound”.
This test uses this principle by generating a masking signal …
and asking users to identify the beeps (the tone that sits within the masking signal) …
at different levels.
Depending on how well the user can perceive the signal relative to the masking noise determines if they get a passing grade or not.
The absolute value of the mask in these recordings seems to hover at -80dBFS for frequencies 0Hz – 1.4khz, -65dBFS for frequencies 1.4khz-1.8khz and 2.2khz to 2.6khz. The tone itself is at 2khz and varies from -40dBFS initially to about -66dBFS depending on how the test progresses.
The logic behind the Telscreen system is summarized by my (rather abridged) flow-chart.
The testing itself is pretty straightforward. Test samples are produced and played according to a logic which randomly inserts no-beep samples to ensure that the user is not randomly pressing buttons when they think they can hear a beep. The testing steps the dB level of the beep up and down relative to the previous response to determine where the hearing threshold is. It begins through coarse steps, until it finds a “floor” and then uses fine steps around this to try and determine a threshold and possibly a level of confidence. Once sufficient samples have been accrued, the system delivers the result (one of three) and a choice to find more information by talking to an operator.
The following items are what you would hear (along with recorded audio) if you called into the Telscreen service.
- For English, please press 1
- [In Cantonese] For Cantonese, please press 2
- [In Manderin] For Manderin, please press 3
- [In Arabic] For Arabic, please press 4
- [In Greek] For Greek, please press 5
- [In Italian] For Italian, press press 6
- [In Vietnamese] For Vietnamese, press 7
- [In Spanish] For Spanish, please press 8
- [In Turkish] For Turkish, please press 9
- [In Macedonian] For Macedonian, please press 01
- [In Serbian] For Serbian, please press 02
Your results from Telscreen may vary, depending on whether you are in a quiet or noisy environment. Telscreen is best carried out in a quiet environment.
In this test you must listen carefully for beeping sounds. The beeping sounds are a bit like the warning sound when a truck is reversing. They sound like this.
[Beep, Beep, Beep]
Every time you hear these beeping sounds, press a number once on your telephone keypad. Any number will do. You will also hear a rushing noise while you are waiting for the beeping sounds to come. This is the rushing noise you will hear.
This rushing noise is there to make it harder for you to hear the beeping sounds. Do not press a key when the rushing noise starts. Wait until you hear the beeping sounds. Listen now, to how it will sound, when the beeping sounds come through the rushing noise.
[Mix of rushing noise and beeping sounds]
It is the beeping sounds you should focus on and respond to, not the rushing noise. Press a key only when you hear the beeping sounds. As the test goes on, the beeping sounds will become softer and harder to hear above the rushing noise. Remember, press a key whenever you hear the beeping sounds even if they are very soft.
The test will now start.
Test Phase – numerous samples prepared depending on user responses.
- Beeps -40dB
- Beeps -45dB
- Beeps -48dB
- Beeps -52dB
- Beeps -56dB
- Beeps -59dB
- Beeps -60dB
- Beeps -62dB
- Beeps -64dB
- Beeps -66dB
- No Beeps
Test Phase: Incorrect Response
There was no beeping sound just then. Remember, you should press a key only when you can hear the beeping sounds.
Thank you. Telscreen is now complete. Before I give you the results, please note that this test is just one way to assess your hearing and it is possible that other tests would give different results. If you have any concerns about your hearing, you should contact your doctor who can provide further advice.
Option 1: Result Pass
You have passed Telscreen. Your hearing is within the normal range on this measure. If you would like more information, please visit our website at www.hearing.com.au.
Option 2: Result Edge
Normal hearing falls within a certain range and these results suggest that your hearing may be near the edge of this normal range. This could be for a variety of reasons including the level of background noise around you during the test. Would you like more information? Press 1 for Yes, 2 for No.
Option 3: Result Fail
Your results are outside the normal range, and we recommend a complete hearing assessment. You may be eligible for free services from Australian Hearing – would you like more information? Press 1 for Yes, 2 for No.
Please contact Australian Hearing on 131 797 or talk to your doctor about hearing services in your area.
Thank you for using Australian Hearing’s Telscreen. Goodbye.
In this post, I’ve taken a close look at what the Telscreen service, operated by Australian Hearing, provides. This service is an example of “telehealth” – using technology to deliver health-related information or service. This is the world’s “most advanced over the phone hearing test” as described by Australian Hearing, and its logic seems pretty simple. It operates through testing auditory masking levels through the use of a specially shaped masking noise through which the user’s ability to perceive a beep within a gap of the masking noise itself indicates their relative hearing health. By running this test repeatedly, and with samples containing no such beeps, a level of confidence in the result can be determined (which likely improves the accuracy of the test). Being able to consistently hear the beeps at the level of the masking noise is required to result in a good result.