Specifications

EAW Smaart 6 Operation Manual Concepts, Glossary, and Bibliography
28
Choosing a Noise Source
The distinction between pink and white noise is not important in Frequency Response
measurements. The transfer function compares the two input signals over a given fre-
quency range point-by-point. This is a concern when looking at a single channel RTA
noise measurement. If it appears flat in narrow band resolution, it is white. If it appears
to be slopping down to the right on narrow band resolution (loss at high frequencies) it
may be pink.
NOTE: White noise is often used to test electronic components but has so much high
frequency energy that it can easily damage loudspeakers if played at high lev-
els. For this reason we DO NOT recommend using white noise as a test signal
for sound system measurement applications.
2.1.7 Data Window Functions
Data window functions are commonly used to reduce truncation errors that arise from
segmenting random signals into FFT-size chunks. The FFT is a circular function that
assumes the signal segment being transformed is infinitely repeating. Therefore, dis-
continuities between the data near the beginning and end of the time domain data series
can result in an excessively high noise component and leakage of data at all frequencies
into neighboring frequency bins when the FFT of the signal is taken.
Data window functions help alleviate this problem by attenuating the samples nearest
the beginning and end of the time record being transformed before the FFT is performed.
Although this technique can dramatically reduce the impact of discontinuities, there are
several trade-offs involved in determining the precise shape of the attenuation curve. A
number of data window functions, with various strengths and weaknesses, have been
developed over the years. For most audio applications, window functions with Gauss-
ian tapers work best. We recommend using the default Hanning window (also called
a Hann window) unless you have a specific reason to use a different type.