声品质客观评价指标的有效性随评价对象和应用场景的不同而存在差异性，而传统的声品质客观评价参量多关注时域特征，难以有效地对啸叫、刺耳特征进行表征。针对上述问题，构建了音噪比（tone-to-noise ratio，TNR）和突出比（prominence ratio，PR）MAP频域衍生特征参量，并结合时、频域特征完善了声品质客观评价体系。本研究构建的柴油机客观参量相对传统客观参量具有更高的声品质预测精度，预测结果为MSE=0.015 3，R²=0.981 9。

The effectiveness of objective evaluation indicators for sound quality varies with the evaluation object and application scenario. Traditional objective evaluation parameters mainly focus on time domain features, which is difficult to effectively characterize howling and harsh features. To address these issues, the frequency domain derived feature parameters of tonetonoise ratio(TNR) and prominence ratio(PR) were constructed. The objective evaluation system of sound quality was improved by combining time and frequency domain features. The proposed objective parameters of diesel engine have higher prediction accuracy with MSE of 0.015 3 and R² of 0.981 9 compared to the traditional objective parameters.