Laguerre-Based Linear Prediction Using Perceptual Biasing

Warped or Laguerre-based linear prediction can be utilized in audio coding in such a way that it accounts for the spectral masking effects. This requires running a psycho-acoustic model to calculate the masking curve; and then the corresponding prediction parameters are derived from the masking curve. As an alternative, we propose to bias the linear prediction solution in a perceptually relevant way using only simple modifications of the coefficients defining the normal equations for a least- squares error. This implies that the complexity is substantially reduced. The approach was implemented in a Laguerre-based linear prediction scheme and has been compared to a system where the linear prediction is based on a psycho-acoustic model. The results clearly demonstrate the viability of the method. The proposed perceptual biasing rules can also be used to determine the perceptual weighting filter in speech coders.