Signal requantization to reduce the word-length of an audio stream introduces distortions. Noise shaping can be applied in combination with a psycho-acoustic model in order to make requantization distortions minimally audible. The psycho-acoustically optimal noise shaping curve depends on the time-varying characteristics of the input signal. Therefore, the noise shaping filter coefficients are to be computed and updated on a regular basis. In this paper, we show that the optimal coefficients can be obtained by solving a simple least squares problem. Thus, an efficient scheme for minimally audible requantization can be proposed in which the noise shaping coefficients are computed from the psycho-acoustic masking curve by solving a linear matrix equation.
