Rate determination based on perceptual loudness

We describe a perceptually-motivated rate determination algorithm (RDA) for variable bit rate speech coding. Unlike existing rate selection strategies that are based on a voice activity detector and energy thresholds, the proposed method employs a perceptual loudness (PL) measure. The TIA IS-127 enhanced variable rate codec (EVRC) has been chosen as the test-bed for evaluating the performance of the PL-based rate selection strategy relative to three existing methods. In particular, the comparative study includes the following rate determination algorithms: voice activity detection; speech frame energy-thresholding; phonetic segmentation. Experimental results demonstrate that the proposed PL-based RDA compares well against other rate selection techniques in terms of average bitrate and speech quality.