Adaptive Fixed Pattern Audio Noise Cancellation for Digital Camera Applications

Attenuating the undesired audio noise generated by lens modules is particular useful for real-time audio/video recording with digital still cameras (DSCs). In this paper, we propose a new approach to audio noise cancellation by adaptively subtracting the coefficients of discrete cosine transform of the noise template from an input signal. The experimental shows that the proposed algorithm effectively deals with the fixed pattern audio noise and reconstructs the original music and sounds.