ODEON仿真与实际测量语言传输指数实验对比

近年来音质仿真技术的快速发展为语言传输指数STI的预测提供了一个潜在的解决方案。但这种方法的有效性如何,则是在使用该技术之前应该考虑的问题。本文对3个房间内音质仿真与实际测量STI进行实验对比,研究表明:在仿真模型与实际空间的声学等效较为准确的情况下,使用音质仿真软件ODEON计算得到的STI误差较小;混响时间的变化在背景噪声较高时可能会对仿真STI的准确性带来显著影响,随着混响时间的增加,仿真与实际测量STI的差值可能变大;信噪比的变化并不会给仿真STI的准确性带来显著影响;仿真脉冲响应与实际测量脉冲响应的频谱有一定差别,时域上的反射声序列也不相同,但这些差别对仿真STI的影响并不大;仿真过程中比较容易产生的信噪比误差对仿真STI产生了较大影响。由于影响音质仿真结果的因素较多,仿真模型与实际空间的声学等效也比较复杂,尤其是对于没有实际参照校准的房间来说,想要获得准确的STI预测结果是比较困难的。

Experimental comparison between ODEON simulated and in-situ measured speech transmission index

The rapid progress in acoustic simulation technique in recent years offers a potential solution to the prediction of speech transmission index (STI). However, before it can be used with confidence, the acoustic simulation technique must be validated. Acoustic simulated and in-situ measured STI were systematically compared at three general rooms in this study. The results reveal that: the STI can be predicted accurately based on the ODEON acoustic simulation software on the assumption that the virtual models are equivalent to the real rooms perfectly. The change of reverberation time (RT) may exert obvious influence on the simulated STI in the case of high noise, with the increase of RT, difference between simulated and in-situ measured STI may increase. But the change of signal-to-noise ratio (SNR) may not exert obvious influence on the simulated STI. Furthermore, there are certain differences on spectrums and reflectogram between the simulated and the in-situ measured IR, while these differences exert little influence on the simulated STI. However, the easy to produced bias of simulated SNR are likely to exert great influence on the simulated STI. Given that many factors may influence the simulation result and to identify the simulated acoustic conditions to those in real rooms is really complicated, it is often difficult to obtain accurate STI by acoustic simulation, especially for a room without actual acoustic calibration.