HP RM500SL - Manual (Page 105)

07/05/06 © Etymonic Design Incorporated, 41 Byron Ave., Dorchester, ON, Canada N0L 1G0 Page 4
USA 800-265-2093 519-268-3313 FAX 519-268-3256 www.audioscan.com
Figure 3: Microphone location effects in dB re nominal band SPL. Nominal band levels are band
levels for an overall rms level of 65 dB SPL and average vocal effort in a free sound field.
3 Analysis of broad-band signals
Broad-band signals contain energy at many frequencies simultaneously. Such signals are usually analyzed in a
series of narrow frequency bands to produce a spectrum. The auditory system functions on a logarithmic
frequency scale and analyzes broad-band signals in critical bands which approximate 1/3 octave bands (Figure
4). Using 1/3 octave bands for analysis of broad-band signals allows measured levels to be compared more
readily to narrow-band behavioral measures, such as threshold. For this reason, analysis of broad-band signals
in all Real-Ear Measurements (including S-REM in Speechmap) is in 1/3 octave bands. Other analyzers use
narrower analysis bands, sometimes having constant bandwidth. As shown in Figure 4, this can result in
significantly underestimating the audibility (or comfort or discomfort) of a complex signal.
It should be noted that, in all Real-Ear Measurements, 1/3 octave band SPL is displayed (and reported when
saving data to a file) at 1/12 octave frequencies which provides curve smoothing. When calculating overall rms
from these data, the SPL at 225 Hz and subsequent 1/3 octave increments should be used.
Analysis of broad-band signals in all Hearing Instrument Tests is in 1/12 octave bands and the band SPL is
displayed at 1/12 octave frequencies.
3.1 Pink noise signal analysis
All band levels are calculated with an averaging time of 128 ms.
For a linear hearing aid, a pink noise stimulus will result in a response curve that has the same shape as one
obtained using a swept pure tone. However, it must be remembered that, while a swept tone has the same band
SPL as overall SPL, the band SPL for a noise signal is significantly lower than the overall SPL. Consequently, for
a linear aid, output curves obtained using 1/12 octave analysis will be about 18 dB lower than the output curves
obtained using a swept tone at the same overall SPL. As long as the hearing aid is linear, the gain obtained will
be the same for both signals. Figure 5 shows output (A) and gain (B) for a linear hearing aid, obtained using a
swept tone (1) and pink noise (2) with a 60 dB overall SPL.