Building Acoustics

Oscillating systems 3

MLS (maximum length sequences) which are binary sequences repeating themselves
exactly. Measurements may then be performed where the useful signal, i.e. the MLS
signal, has an amplitude not much higher or even lower than other disturbing signals
(background noise). This is possible when performing a synchronous averaging over
several sequences. As one may understand this presupposes that the tested system is time
invariant.
A further development in the measuring technique is by using so-called chirp
signals or swept sine signals. As the latter name implies, these are transient signals where
the frequency sweeps from a starting frequency to a final frequency during the
measurement period. The application of such signals in system testing is at least 30 years
old but due to modern digital technology it is now in widespread use. We shall deal with
such test signals in section 1.5. A recent international standard, ISO 18233, is devoted to
the application of such new measurement technique in building acoustics.

1.3 Methods for signal analysis

The methods or techniques that are used in practical signal analysis may be divided into
three main groups as follows: 1) signal amplitude analysis, 2) time domain analysis and
3) frequency domain analysis (spectral analysis). Figure 1.2 gives an overview of some
concepts associated with these main groups.
For many purposes the only information needed is the absolute value of the
oscillation, normally specified by the RMS-value (root-mean-square value) or the peak
value. The reason is that these values may be specified as a legal limit value, e.g. the
maximum vibration amplitude for a certain type of machine, the maximum A- or C-
weighted sound pressure level in a living environment or in a work space. These types of
analysis may normally be carried out using simple and low-cost instrumentation.

Signal analysis

Magnitude

analysis

Time domain

analysis

RMS-value

Max./min. values

Variance, probability

density etc.

Time history

Impulse response

Correlation

Covariance

Spectrum

Frequency response

Coherence

Cepstrum

Frequency domain

analysis

Fourier transform

Figure 1.2 Signal analysis types and methods.

More advanced equipment is needed for accurate analysis in the time or frequency
domain, in particular when information from many parts of a system is needed and/or
when the task is to map the relationship between data from the various parts. Time
domain analysis as cross correlation or the equivalent frequency analysis, i.e. transfer