Broad-dynamic filtering procedure for a recursive digital filter installed in a signal processor (DSP) operating with integers
AssigneeInstitut Franco-Allemand de Recherches de Saint-Louis
InventorsVeronique Zimpfer, Karl Buck
Life of this patent
Estimating the term…
Abstract
The present invention relates to a filtering procedure for recursive digital filters installed in signal processors (DSP) operating with integers. It comprises two calculation procedures coupled with the classical calculation procedure of a recursive filter, the first of which, when calculating the output value of the sample, takes into account the remainders of integer divisions by the scale factor, coming from calculating the output values for the previous samples, and the second of which replaces the default value from dividing the integer by a scale factor by the integer value closest to the result of actual division.
Claims15
A calculation method for producing a recursive digital filter, implemented in a signal processor working with integers, the method comprising:
- calculating an output signal of a sample n based on an algebraic sum of input and output values of signals sampled at a selected point in time and at previous points in time, to which coefficients characteristic of the filter have been assigned;
- applying a chosen scale factor to remainders of integer divisions, the remainders being a result of calculating the output values of the previous samples;
- and
- changing a number obtained from rounding to a default integer value what is obtained from dividing the output value by the scale factor based on a number obtained from rounding to a closest integer to a real-number quotient thereof.
A calculation method for producing a recursive digital filter, implemented in a signal processor working with integers, the method comprising:
- calculating an output signal of a sample n based on an algebraic sum of input and output values of signals sampled at a selected point in time and at previous points in time, to which coefficients characteristic of the filter have been assigned;
- and
- applying a chosen scale factor to remainders of integer divisions, the remainders being a result of calculating the output values of the previous samples.
A calculation method for producing a recursive digital filter, implemented in a signal processor working with integers, the method comprising:
- calculating an output signal of a sample n based on an algebraic sum of input and output values of signals sampled at a selected point in time and at previous points in time, to which coefficients characteristic of the filter have been assigned;
- and
- changing a number obtained from rounding to a default integer value what is obtained from dividing the output value by a scale factor by a number obtained from rounding to a closest integer to a real-number quotient thereof.
Dependencies are inferred from the claim text.