The human voice, like any sound produced by thrumming a stretched string, has a fundamental frequency. For voice, the centre of that frequency lies mostly below 300Hz depending on the speaker"s . Information is conveyed through simultaneous higher-frequency overtones(泛音)and additional components that can stretch up to 20,000 Hz(20kHz). Modern hearing are able to distinguish only a small part of that range, typically between 300Hz and 6kHz, reducing noise and amplifying those frequencies where the wearer"s hearing is the weakest. But differentiating elements of many common parts of speech occur in higher frequencies. This is the result both of harmonics(和声)that ripple out from the main tone, and from non-voiced elements used to utter consonants(辅音), which employ the tongue, teeth, cheeks and lips. Take the words "sailing" and "failing". Cut off the higher frequencies and the two are indistinguishable. The problem is compounded on telephone calls, which do not transmit frequencies below 300Hz or above 3.3kHz. People with hearing experience this problem constantly, says Brian Moore of the University of Cambridge. Typical hearing loss tends to be most acute at frequencies above 10kHz, which contain quieter sounds but where speech can still include important cues. Older hearing cut off at no higher than 6kHz, but much modern equipment stretches this range to 8-10kHz. However, a problem remains, Dr Moore says, because bespoke hearing-aid calibrations for individual users, called "fittings" , do not properly boost the gain of these higher frequencies. So Dr Moore and his colleagues have come up with a better method. Their approach can be applied to many existing devices, and is also being built into some newer ones. A key step in any fitting involves testing an individual"s ability to hear sounds in different frequency bands. Each hearing loss is unique, and for most users a standard profile would be too loud in some ranges and too soft in others. But current tests pay scant attention to the higher frequencies that a device"s tiny speaker can produce, regardless of whether the user needs a boost. Dr Moore"s new test, known as CAM2, which is both a set of specifications and an implementation in software, extends and modifies fittings to include frequencies as high as 10kHz. When the results are used to calibrate a modern hearing aid, the result is greater intelligibility(可懂度)of speech compared with existing alternatives. CAM2 also improves the experience of listening to music, which makes greater use of higher frequencies than speech does. The text mainly focuses on modern hearing aid"s ______.