Logo PTB

Directional response pattern of measurement microphones for high-frequency noise measurements


When carrying out noise measurements, it is usually assumed that the directional response pattern of the microphone used can be neglected. This is true when the microphone is clearly smaller than the wavelength of the investigated sound. This usually does not apply in the case of the ultrasound range.

In the audible hearing range, the wavelength amounts to several meters down to a few centimeters (depending on the frequency) whereas in the high-frequency ultrasound range, it only amounts to a few millimeters. Therefore, the wavelength is smaller than the microphones that are typically used, and the higher the frequency of the sound, the less sensitive a microphone becomes to lateral sound incidence. When measuring noise that has a high-frequency sound fraction, the measurement result depends on the precision with which the microphone is directed towards the noise source. If the directional response pattern is known, the result can be corrected. If several noise sources or reflections exist, at least the measurement uncertainty of the result can be determined with the aid of the directional response pattern.

In this context, several microphones of different types and from different manufacturers have been examined at a scanning measuring set-up at PTB (Figure 1). The measuring set-up has been installed in PTB's large anechoic chamber and allows an area in the form of a hemisphere above a test microphone to be scanned with a high angular resolution. For this purpose, a suitable loudspeaker is moved around the microphone that is to be examined, the pulse response is measured for every position, and the directional sensitivity is determined from this.


Figure 1: Microphone set-up for the measurement of the directional response pattern in PTB's large anechoic chamber

It turned out that measurement microphones whose type corresponded to the IEC 61094-4 [1] standard showed very similar directional response patterns, independent of the manufacturer. The differences were mainly due to the ratio of the physical size to the sound wavelength. However, very large deviations were due to the use of protective grids. The protective grids are not standardized and have very different shapes, depending on the manufacturer. Depending on the construction, even the sound that is incident from the front can be falsified.

The results might lead to the conclusion that for measuring noise that has a dominant high-frequency fraction, no protective grids should be used, as the individual geometric design has a strong influence on the measurement result. In the daily measurement routine of an engineer, however, microphones cannot be used without protective grids as the grid is the only protection for the sensitive and expensive sensor against dirt, dust and mechanical impact.

The question which still remains to be answered is how to make high-frequency noise measurements more practicable and reliable in the future.


[1] DIN EN 61094-4: Meßmikrofone - Teil 4: Anforderungen an Gebrauchs-Normalmikrofone (IEC 61094-4:1995); Deutsche Fassung EN 61094-4:1995


Dr. Christoph Kling, FB 1.6, AG 1.63, E-Mail: Opens window for sending emailChristoph.Kling(at)ptb.de