Abstract:A new method for speech denoising is presented which uses a new threshold different from DJ's thresholds. Preliminary computer simulations were performed on speech signals and resulted in better performance than that by using DJ's thresholds.