Double pseudo affine projection algorithm for speech enhancement and acoustic noise reduction

In this paper, we consider the speech enhancement problem in a moving car through a blind source separation scheme involving two closely spaced microphones. We propose the use of a new double pseudo affine projection DPAP algorithm to estimate and suppress coherent noise components from speech. The proposed DPAP algorithm is applied to the forward blind source separation FBSS structure and combined with a new whitening scheme of its two inputs. In order to avoid the use of a manual voice activity detector in the new DPAP algorithm, we have adapted then applied the technique proposed in [1] to the FBSS structure and for the new proposed algorithm. The simulation results show that the DPAP algorithm, when controlled by the proposed technique is able to fully cancel the correlated noise components from speech.