High frequency limit

From Wikipedia, the free encyclopedia

The high frequency limit of hearing is the upper extent to which a particular animal can perceive sound.

Perhaps the most commonly known aspect of the psychoacoustic model is that humans cannot hear frequencies above and below certain thresholds; in fact, most humans can only hear frequencies between 20 Hz and 20,000 Hz (E0 to D#10 in Scientific pitch notation). High frequencies have high pitches and lower frequencies have low pitches. So-called "silent" dog whistles exploit this phenomenon by producing sounds at frequencies higher than those audible to humans but well within the range of a dog's hearing. Likewise, when compressing a digital signal, an acoustic engineer can safely assume that any frequency beyond approximately 20 kHz will not have any effect on the perceived sound of the finished product, and thus use a high-shelving filter to cut everything outside this range. The sound can then be sampled at the standard CD sample rate of 44.1 kHz (or 48 kHz in DAT), set somewhat higher than the calculated Nyquist-Shannon rate of 40 kHz to allow for the cut-off slope of a reasonable band-pass filter.

When additional compression of sound is required, higher frequencies are usually cut-off first, because regular adult's hearing in those areas is often even less than 20 kHz. This is because of loss of hearing happening either due to hearing damage through e.g. listening to loud music or due to hearing loss that happens naturally while one grows old. For instance, the commonly used MP3 coding often cuts sounds above 18 kHz, or when compressing as high as 128 kbit/s, at 16 kHz[1].