Apparatus for converting a signal of a first sample rate into a signal of a second sample rate

DETAILED DESCRIPTION It is an object of the present invention to provide equipment by which signals of different sample rates can be converted so as to provide a common sample rate at which all can be processed.
Briefly, the sample values of a signal are supplied to a digital filter of which the coefficients are provided in a manner dependent upon the ratio of the sample rates and the sample value are read out of the digital filter at the sample rate into which the input samples were to be converted.
The filter is preferably a FIR filter.
The coefficients are formed according to a function of which the Fourier transform satisfies a specified set of conditions further set forth below.
The apparatus of the invention has the advantage that the sample rates of the individual signal sources do not need to have any particular number ratio to each other or to the unitary sample rate to which they are being converted.
In addition outstanding signal quality and an excellent margin against disturbance is obtainable.
In an important embodiment of the invention the signal of a first sampling rate is supplied through a device for oversampling.
In that way a particularly favorable implementation of the apparatus of the invention is possible by means of a digital signal processor.
In that case the timing pulses of the first and second sampling rate are so intercoupled that they then have the same error tolerance.
In the apparatus of the invention, a mathematical procedure is used for calculating the sample values for the second sample rate.
This procedure is known under the designation of "generalized sample sequences.
" For this purpose the calculation begins from a Fourier transform of a function g that is defined by ##EQU1## and in which v is a real number and i = .
sqroot.
-1.
If .
PHI.
is a time bounded function with the time boundary T .
gtoreq.
O, which means .
PHI.
(t) = 0 for .
vertline.
t.
vertline.
.
gtoreq.
T, and the following holds for the Fourier-transformed .
PHI.
: ##EQU2## wherein D.
sup.
j is the jth derivative,