Moore响度的三种计算方法

简单介绍了Moore响度模型的计算过程,针对该模型以参数化方式输入信号频谱的特点,提出了直接FFT(快速傅里叶变换)频谱算法、FFT校正频谱算法和1/3倍频程谱算法等3条Moore响度计算路线,分析对比了3种不同算法对典型信号响度计算精度的影响。分析结果表明:对纯音或复合音信号,可采用FFT校正频谱算法或点数为4096的直接FFT频谱算法计算响度;对带宽较宽的噪声信号,建议采用1/3倍频程谱计算响度。

Three methods for calculating Moore’s loudness

The procedure of Moore’s loudness model was briefly introduced using the spectrum parameters of the sound signal as the input of the model.Three methods were proposed to calculate Moore’s loudness.They are the FFT spectrum method,the correcting FFT spectrum method and the one-third octave band spectrum method.By calculating loudness of standard signals,the accurate and the efficiency of three methods were analyzed and compared.The implications of the results were discussed.For pure tones and complex tones,the correcting FFT spectrum method or the FFT spectrum method with a FFT length of 4096 sample points are recommended and for broadband noise,the one-third octave band spectrum is recommended.