Evaluation of the objective evaluation method for enhanced signal based on the psychoacoustic model- within narrow band noise

摘要

When performance of signal enhancement is evaluated, we often experience a difference between impression of audio demonstration and a physical Index such as SNR. This difference might be due to the fact that SNR does not take account for masking effects. Thus, subjective test had been performed. However, the subjective test requires high cost with respect to high human resource and time. In order to reduce the cost, signal evaluation method based on a psychoacoustic model used in MPEG 1/Audio was proposed. In this study, performance of this evaluation method is examined under simplified sound stimuli using sinusoidal signal masked by narrow band noise.