# Spectrogram

Old sonagraph from the 1980s

A spectrogram is the graphic representation of the time course of the frequency spectrum of a signal . Spectrograms are widely used, for example for the purpose of analysis of sound signals , the signal processing and image processing or display of trace elements in the neutron activation analysis .

A spectrogram represents the composition of a signal (e.g. a sound or spoken language) from individual frequencies , usually over time. In this case, a spectrogram is, more precisely, a time-variant representation of the frequency distribution using the short-term Fourier transform . This results in a complex-valued function that depends on the time t and the frequency . The spectrogram is a representation that plots the values in a time-frequency diagram (for example color-coded). ${\ displaystyle f (t, \ omega)}$${\ displaystyle \ omega}$${\ displaystyle | f (t, \ omega) | ^ {2}}$

The interpretation is that it indicates the frequency component of the signal at time t . ${\ displaystyle | f (t, \ omega) | ^ {2}}$

## Sonagram

Sonagram of the Japanese word minato . The sound properties of the individual speech sounds can be recognized over time.

A sonagram (from Latin sonare ' to sound' and from ancient Greek γράφειν grafeïn , German ' to write' ) is a sound spectrogram. Originally, the sonagram and sonagraph were trade names : The sonagram was the representation of rapidly changing sound processes in rectangular coordinates (time and frequency axes) supplied by a sonagraph. The strength of the frequency components was made visible by blackening the recording paper. Today the term Sonagraf (or Sonagraph ) is used regardless of the make of the device. Nowadays, sound analysis is only carried out digitally, mostly with the help of computer programs.

The sonagram is also known as a sonogram . However, a sonogram is usually an image generated by ultrasound , see sonography .

Sonagrams are used, among other things, to analyze speech signals, e.g. B. to determine speech sounds ( phonetics ). Vowels can be recognized in the sonagram by characteristic patterns, the so-called formants . Consonants can be recognized, for example, by the course of the black coloration (depending on the intensity), by the existing or non-existing voice bar (a black bar in the lower area that shows the voicing ) and by various transitions.

## Campbell diagram

In vehicle and machine acoustics, there is often a particular interest in the relationship between the spectrum and the engine speed . The spectral course of a high rev is therefore not shown as a function of the measurement time, but as a function of the speed. The term Campbell diagram is used for this form of the spectrogram .

## literature

• Uwe Kiencke , Michael Schwarz, Thomas Weickert: Signal processing. Time-frequency analysis and estimation methods . Oldenbourg, Munich 2008, ISBN 978-3-486-58668-8 .