Learning FIR Filter Coefficients from Data for Speech-Music Separation

Abstract

An Finite Impulse Response (FIR) filter is a widely used digital filter technology whose impulse response has a finite duration. An FIR filter is usually favored for many reasons such as easy to design, easy to implement on a variety of system architectures. An FIR filter can be easily designed with a linear phase response and its output is more predictable since it doesn't have feedback components. There are both engineer and mathematical methods for designing an FIR filter so that machine learning doesn't play an important role in the FIR filter design. In this paper, we present an alternative to traditional filter design methods to direct learn the FIR filter coefficients from input data with machine learning algorithm. With the proposed algorithm, we can easily design an FIR filter from the input data mixed with designed all spectrum noise signal. To show the capability of this algorithm, an example application of suppressing background music from speech or vice versa is demonstrated in this paper. Despite that the music and speech have a lot of overlap in their spectrum, the filter designed by our algorithm can successfully suppress music or speech in a mixture of music and speech signals.