A new source‐filter model audio bandwidth extension using high frequency perception feature for IoT communications
A new source‐filter model audio bandwidth extension using high frequency perception feature for IoT communications
复制标题
使用高频感知功能进行物联网通信的新源滤波器模型音频带宽扩展
DOI：
10.1002/cpe.4638
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Tonghan Wang
中科院分区：
文献类型：
--
作者：
Lin Jiang;Shaoqian Yu;Xiaochen Wang;Chao Wang;Tonghan Wang
音频编码是物联网应用系统的共性技术。音频带宽扩展是当代音频编解码器中的一种标准技术，用于以低比特率高效地编码音频信号。在现有方法中，高频信号是通过复制相应的低频和一些高频参数来产生的。然而，如果高频和低频之间的相关性变弱，编码的感知质量将显著下降。在本文中，我们提出了一种新的源-滤波模型音频带宽扩展方法。在我们的方法中，提取高频信号的感知特征来恢复编码的感知质量。在解码端，在高频感知参数的约束下，得到峰值因子和噪声水平。实验结果表明，该方法的性能优于经典方法。与现有方法相比，由于编码码率显著降低，并且保持了接近的编码感知质量，因此该方法也具有可比性。本文还为车联网等低码率、高质量的物联网通信提供了一种新的解决方案。
Audio coding is a generic technology to IoT applications system. Audio bandwidth extension is a standard technique within contemporary audio codecs to efficiently code audio signals at low bitrates. In existing methods, high frequency signal is generated by a duplication of the corresponding low frequency and some parameters of high frequency. However, the perception quality of coding will significantly degrade if the correlation between high and low frequency becomes weak. In this paper, we proposed a new source‐filter model audio bandwidth extension method. In our method, a perception feature of the high frequency signal is extracted to restore the perception quality of coding. In the decoder side, the crest factor and noise level are obtained under the constraint of the high frequency perception parameter. The performance shown in our experiment results are superior to the classic methods. Compared with the state of art method, the proposed method is also comparative because of the coding bitrate is significantly reduced and keeps a close perception quality of coding. This paper also provided a new solution for IoT communications that requires low bitrates and high quality of coding, especially like Internet of Vehicles.