基于灰色Verhulst模型的音频频带扩展方法

发布时间：2018-06-12 07:41

  本文选题：音频编码 + 频带扩展　； 参考：《电子学报》2014年08期

【摘要】：受到传输带宽的限制,宽带音频的自然度和听觉质量会有所下降,因此,本文提出了一种基于灰色Verhulst模型的宽带向超宽带音频频带扩展方法.根据音频信号频谱包络序列的演变趋势,采用灰色Verhulst模型对高频频谱包络进行估计,并利用最近邻匹配方法对高频频谱细节进行预测,最后经过高频频谱包络的调整,所提方法能够有效地恢复7~14kHz频率范围内的高频成分.主客观测试表明,该方法改善了宽带音频的听觉质量,并且优于传统的基于高斯混合模型的音频频带扩展方法.
[Abstract]:Due to the limitation of transmission bandwidth, the naturalness and auditory quality of wideband audio will be reduced. Therefore, this paper proposes a method based on grey Verhulst model to extend wideband to ultra-wideband (UWB) audio frequency band. According to the evolution trend of spectrum envelope sequence of audio signal, the grey Verhulst model is used to estimate the envelope of high frequency spectrum, and the nearest neighbor matching method is used to predict the details of high frequency spectrum. Finally, the envelope of high frequency spectrum is adjusted. The proposed method can effectively recover the high frequency components in the frequency range of 7 ~ 14 kHz. Subjective and objective tests show that this method improves the audio quality of broadband audio and is superior to the traditional audio band expansion method based on Gao Si hybrid model.
【作者单位】： 北京工业大学电子信息与控制工程学院;
【基金】：国家自然科学基金(No.61072089)
【分类号】：TN912.3
，

本文编号：2008912