Interpolating filter arrangement with non-rational ratio between the input and the output sampling frequencies

Claim 1

An interpolating time-discrete filter arrangement having a non-rational interpolation factor R for converting a time-discrete input signal formed by a sequence of input samples x(q), q= . . . -2, -1, 0, 1, 2, 3, . . . with which an input sampling frequency f i is associated into a time-discrete output signal formed by a sequence of output samples y(n), n= . . . -2, -1, 0, 1, 2, 3, . . . and with which an output sampling frequency f u is associated which is R-times higher than the input sampling frequency f i , comprising: (a) first means for producing input clock pulses ki(q) occurring at the said input sampling frequency f i ; (b) second means for producing output clock pulses ku(q) occurring at the output sampling frequency f u ; (c) a filter coefficients generator for producing a group of W filter coefficients, the generator comprising: (c1) means to which the input clock pulses and the output clock pulses are applied and which in response to each input clock pulse ki(q) supplies a deviation component d(q) whose magnitude is proportional to the relationship between the time interval T d (q) located between that input clock pulse ki(q) and the immediately subsequent or the immediately preceding output clock pulse and the time interval T u =1/f u between two consecutive output clock pulses; (c2) means for generating the group of W filter coefficients in response to the deviation component d(q) supplied, the filter coefficient having number w being equal to d(d(q),w) and being defined by the expression d(d(q),w)=h(d(q)T.sub.u +wT.sub.u) w=0, 1, 2, . . . W-1 wherein h(v) represents the impulse response of a FIR filter, v is a continuous variable in the interval -∞<v<∞ (d) a signal processing arrangement for generating the output samples y(n) by multiplying input samples by selected filter coefficients and adding together the products thus obtained.