Telos Zephyr Xstream, User Manual

Page: 123 / 381

USER’S MANUAL
Section 6: AUDIO CODING REFERENCE 111
ear’s sensitivity to signals near the constant tone is greatly reduced. Tones that were previously
audible become “masked” in the presence of “masking tones,” in this case, the one at 300 Hz.
All signals below the upper “threshold of audibility” curve, or Masking Threshold are not
audible, so we can drop them out or quantize them crudely with the least number of bits. Any
noise which results from crude quantization will not be audible if it occurs below the threshold
of masking. The masking depends upon the frequency, the level, and the spectral distribution of
both the masker and the masked sounds.
Masking effects in the frequency domain. A masking signal inhibits audibility of
signals adjacent in frequency and below the threshold.
To benefit from the masking effects, perceptual coders use a filter bank to divide the input audio
into multiple bands for analysis and processing. The maximum masked noise level is calculated
depending upon the spectral content, and the available bits are allocated so as to keep the
quantization noise below the masking threshold at every point in the spectrum.
While coding efficiency increases with more bands and better frequency resolution, the time
domain resolution decreases simultaneously owing to an inevitable side ‐ effect of the band
filtering process. Higher frequency resolution requires a longer time window – which limits the
time resolution. Happily, masking works also in the time domain. A short time before and a
longer time after a tone is switched on and off, other signals below a threshold amplitude level
are not noticeable. Filter banks with higher frequency resolution naturally exploit the ear’s
time ‐ masking properties.