Two main things lead people to hearing aids:
- I can’t understand my spouse, kids, TV across the room
- I can’t understand in a noisy place (office, restaurant)
“Noise” (all unwanted sound) is part of both problems; worse when noise is high.
I’ve done a lot of sound recording. “Noise” is one of the problems. Records and tapes hissss. If the speech/music drops too small, it is hidden in the hiss. Sometimes this becomes the major problem. I was asked to recover some old 78rpm records from a friend’s childhood. Naturally the grooves were full of dust: shsssksssskshssssskTssss. The loud parts (vowels) boomed through; the consonants were buried in hiss. In that case (working with recording and sound files on PC) I was able to use off-line processing, teach it a snip of bare “shsss” and ask it to remove it from the rest of the track. But doing that in real-time and with “babble” noise is much tougher, for people and for aids.
In lecture-hall acoustics we are taught to strive for 25dB signal to noise for very-good intelligibility. At 10dB S/N many lecture-words will be missed even in young ears.
What does (Quick)SIN test?
You hear a primary talker, and 3-4 other talkers much softer. You are asked to repeat what the primary talker says and ignore the others. Then the other talkers get louder. When you can’t get half of what the primary talker says, because the other talkers get in the way, the loudness difference between primary and others is noted as your score.
The SIN tests aim for 50% mistakes because that point is easier to find than say 10% mistakes (which would be preferred in lecture listening).
So if you have 50% mis-hears at 9dB S/N, when lecture-hall experience suggests less than 10% mis-hears at 10dB S/N, it says you could hear better. (However lecture-hall acoustic “noise” is more random, ventilation and chair-creaks; SIN and restaurant “noise” is speech, which is a harder discrimination task.)
Also the SIN 50% level for “normal” folks is 0db to 3dB. The voice you want to hear half of can be very nearly same level as the other voices. A SIN of 9 means the voice you want needs to be twice as loud as others. In a crowd, that can’t happen (we can’t all be twice as loud as everybody else).
Killion (father of SIN and QuickSIN) wrote some on intelligibility in noise and why it does not correlate well with pure loss.
SIN Report: click PDF link top right (open this link in new tab)
The Etymotic SIN tests are in “English” (and perhaps Mid-West American?), so must be adapted for other linguistic populations. This paper (in English!) on SIN for Mandarin starts with a good overview, and rationale for calibrating scores with “normal” people.