The properties of a microphone are significant for the quality of the
recording and transmission of an acoustic signal. This report presents
possibilities for influencing the linear and the spacial characteristics of
microphones with digital signal processing. The first step for this is the
analog to digital conversion of the microphone signal. With the developed
procedure that uses a combination of a linear and a nonlinear quantisation
the digitalisation of microphone signals with a dynamic range of more than
140 dB is possible without abrupt changes of the signal to noise ratio,
with a level independed signal to noise ratio, without an increasing
quantisation error at decreasing signal levels and without an
overmodulation of the analog to digital converters. The processing of the
linear and the spacial properties of microphones is realised by FIR
filters. For this filters a design process is developed especially for
microphone applications related to the construction of the target transfer
functions, the preprocessing of the microphone transfer functions and the
postprocessing of the calculated FIR coefficients. The calculated filters
can be used for equalisations of the magnitude response combined with an
optimal impulse response of the microphone like the equalisation for
different incidence angles in a direct sound field, the equalisation of the
magnitude in a reverberant sound field or a presence boost. Another design
process is developed for FIR filters that are used for the frequency
dependend processing of the magnitude and the phase response of the
capsules in a TWIN microphone. With such filters a frequency dependend
control of the polar pattern and the directivity of a microphone is
possible. Such microphones can have an equal magnitude response in the
direct and in the reverberant sound field or also a frequency dependend
controlable direct to reverberant ratio. For microphone columns a design
process for application dependend polar patterns using FIR filters is
optimised related to the construction of the target polar pattern and the
pre- and postprocessing during the calculation of the FIR coefficients.
This results in a decreased modulation of the sound level and the magnitude
response when recording extended sound sources.