More on Microphones by Michael Williams, (www.williamsmmad.com)   △   ∧   >

1 - INTRODUCTION

The environment of sound recording practice has become more and more sophisticated, especially with the arrival of the digital revolution in the process of recording, transmission and reproduction of sound. Microphone technology on the other hand is being led kicking and screaming into the digital domain – the microphone digital interface being perhaps the only concession to this new digital environment. However great strides have been made in design, so as to improve such characteristics as physical size, frequency response both on and off axis, noise floor and of course dynamic range, which is more than justified by the demands of this new digital medium. Despite the remarkable choice of signal processing equipment available in the present day sound recording studio, the choice of a microphone by the sound recording engineer is still of fundamental importance in determining the final quality of a recording. This choice must be based on a good knowledge of the technical specifications of the microphone as well as a wider operational experience with different sound sources and recording environments.

For many years the frequency response curve has been our first approach in estimating the way in which a microphone will transform the spectral balance of the sound source. In parallel with this scientific measurement, a vast range of subjective vocabulary based on operational experience, has been built up to supplement the limitations of this fundamental measurement. This vocabulary (dry, bright, boomy, resonant, metallic, etc...) has the advantage of seemingly describing in everyday language the basic characteristics of say a particular colouration, but unfortunately suffers from considerable ambiguity in interpretation.

However this only serves to illustrate the fact that, at present, we do not have adequate measurement techniques available to completely characterize this aspect of microphone response. It would seem that the published frequency response measurement can give us only a basic appreciation of the overall tonal balance. But perhaps we are hoping to obtain too much information from only one type of measurement, also our knowledge of the significance of different aspects of this characteristic is perhaps still rather rudimentary. Added to this, it is probable that the perception of overall spectral balance is a combination of many interrelated factors, the on-axis response therefore being only part of the story. On-axis response is certainly of fundamental importance, but the off-axis response and its rendition of early reflections and reverberation also plays an important role in our subjective appreciation of the tonal balance.

Another important contributing factor would seem to be the temporal frequency response. This is often presented in graphical form as a ‘waterfall’ response as shown in Figure 1, i.e. the frequency response measured at periods over the few milliseconds after the microphone has received an acoustic stimulus signal. Some interesting work has been done in the past by Jackie Green (nee Hebrock) in this type of measurement. ^[1][2]

Figure 1 -‘Waterfall’ Response of a Microphone
Reproduced, with thanks to the Author,
from AES preprint 4516 (figure 8 - mic #8)^[1]


This work would seem to suggest that the time domain measurement highlights various resonances in the microphone response, and could go some way to explaining the colouration that we hear in some microphones, which is not obvious from the basic frequency response curve. The purists would say that a detailed analysis of the phase response of the microphone, or even specific peaks in the frequency response curve, should also produce the same type of information. It is certain that individual resonances are quite evident in the time domain as represented by the ‘waterfall’ graph - even to the untrained eye, whereas reading the fine detail in the initial frequency response needs some experience. However effects such as the resonance ‘sliding’ probably due to diffraction effects, around the 400Hz to 1kHz range in the example in Figure 1, can only be highlighted by a time domain response presentation. Let us hope that in the future, it will be possible to integrate this type of measurement into the microphone specification sheet.

However the information that can be ‘read’ from the standard frequency response curve is still a first step in the choice of a suitable microphone for a particular usage, one just has to be conscious of the limitations of this type of measurement for the time being. The low frequency and high frequency roll-off (or otherwise) are quite evident, as well as the overall profile of the response. In hand-held vocal microphones the amount of proximity effect compensation can also be seen clearly from the frequency response characteristic.

However any use of medium frequency range boost to create ‘presence’ must be examined closely, the user should establish his own correlation between the form of the accentuation that has been applied, and their own requirements or subjective appreciation in operational conditions. And again bear in mind that any typical or measured frequency response curve can always hide disagreeable resonances.

It is also important to be sure that we are reading a real response curve of a specific microphone - more often than not, the typical frequency response curve is a hand drawn mean value with, at best, an indication of tolerance. However it is in the fine detail of an individual frequency response curve that some of the more serious problems can be detected.

[1] 1997 – "Psychoacoustic Effects of Ringing in Microphones", by Jackie Hebrock, Kelly Stratham & Chuck Kraft, Audio-Technica, USA, 103rd AES Convention, New York, USA - Preprint 4516

[2] 1998 - "Test Method for the Evaluation of Ringing in Microphones", by Jackie Green & Kelly Statham, Audio-Technica, USA, AES UK Conference - Microphones & Loudspeakers, The Ins & Outs of Audio - Paper MAL-09


Return to 'Contents'(Alt+W) △         
Back to top of page (Alt+C) ∧         
Forward to next chapter (Alt+V)    >