A Simplified Early Auditory Model with Application in Speech/Music Classification

The past decade has seen extensive research on audio classification and segmentation algorithms. However, the effect of background noise on the performance of classification has not been investigated widely. Recently, an early auditory model that calculates a so-called auditory spectrum has been employed in audio classification where excellent performance is reported along with robustness in noisy environment. Unfortunately, this early auditory model is characterized by high computational requirements and the use of nonlinear processing. In this paper, by introducing certain modifications we propose a simplified version of this model which is linear except for the calculation of the square-root value of the energy. A speech/music classification task is carried out to evaluate the classification performance wherein a support vector machine (SVM) is used as the classifier. Compared to a conventional FFT-based spectrum, both the original auditory spectrum and the proposed simplified auditory spectrum show more robust performance in noisy test cases. Test results also indicate that, with a reduced computational complexity, the performance of the proposed simplified auditory spectrum is close to that of the original auditory spectrum