This annex provides measurement procedures for verification the presence of audible tones if their presence
is in dispute. Based on the prominence of tones, this procedure also provides recommended levels of
adjustment. The aim of the objective method is to assess prominence of tones in the same way as listeners do
on the average. The method is based on the psychoacoustic concept of critical bands, which are bands defined
so that sound outside a critical band does not contribute significantly to the audibility of tones inside that critical
band.
The method includes procedures for steady and varying tones, narrow-band noise, low-frequency tones, and
the result is a graduated adjustment of 0 dB to 6 dB.
F.2.1. General
The method uses the following steps:
1) Narrow-band frequency analysis (FFT analysis)
2) Determination of prominent tones and their audibility, ΔL
i
, and the uncertainties σ
i
of ΔL
i
3) Grouping the tones inside the common critical bands and recalculating the parameters σ
i
and ΔL
i
4) Calculation of the mean audibility, ΔL, and the extended uncertainty, U, of all spectra measured
5) Calculation of the tone adjustment, spectrum penalty, K
F.2.2. Frequency analysis
A narrow-band A-weighted spectrum is measured by linear averaging for 3 s, that is recommended. If the
shorter averaged time is used, then the number of spectra shall be merged to receive the 3 s averaged
spectrum. Svantek sound measuring instruments do not provide additional spectra merging.
The constant line spacing Δf is equal to 3,125 Hz (the standard requires the line spacing is withing 1,9 – 4 Hz).
The investigation range in Svantek instruments is equal to [50;4575] Hz. The spectrum is measured for the
range of [0;5000] Hz.
It is recommended that the measurement results, including the frequency analyser, be presented in dB re 20
μPa, and that Hanning weighting be used as the window function. The distortions of stochastic signals are
avoided through weighting Hanning function. The Hanning window is used as well as time window to reduce
lateral bands.
NOTE 1. Mean audibility is calculated for the number of spectra that are measured separately.
NOTE 2. The extended uncertainty, U, of the mean audibility, ΔL, with a coverage probability of 90% in a
bilateral confidence interval shall not exceed ±1.5 dB. This is generally the case with evaluation of at least 12
time-staggered narrow-band averaged spectra. If there are less than 12 averaged spectra, then the uncertainty
shall be taken into consideration.
NOTE 3. With the recommended Hanning time window, the effective analysis bandwidth (or the effective noise
bandwidth) is 1.5 times the frequency resolution. The frequency resolution is the distance between the lines in
the spectrum.
To Next Page
Loading...
Need help?
General
