中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
基于相位声码器的虚拟低音增强算法研究

文献类型:学位论文

作者郭杰
学位类别博士
答辩日期2008-05-29
授予单位中国科学院声学研究所
授予地点声学研究所
关键词虚拟低音 虚拟音调 音频带宽扩展 相位声码器 EVB Phase Vocoder
学位专业信号与信息处理
中文摘要目前多媒体产品市场绝大多数小型扬声器因受到单元尺寸等物理条件限制存在低频还原能力差的问题。虚拟低音技术(Virtual Bass)指的是针对扬声器低频截止频率高于信号低频截止频率,信号的低频段无法被扬声器重放的问题,利用心理声学中的虚拟音调模型,将低频段能量搬移至倍频处,利用谐波重建低音听感的信号处理技术。本论文探讨了虚拟低音技术的算法实现与改进。 传统的虚拟低音实现是通过非线性器件实现的。此类算法实现简单、运行效率高,但存在互调失真的问题。MaxxBass虚拟低音增强系统是此类算法的典型代表。另一方面,相对于利用非线性器件构造谐波信号的虚拟低音增强算法,基于短时谱分析的二维时频处理方法在算法灵活性和畸变控制等诸多地方有着明显的优势。基于相位声码器的虚拟低音增强算法VB Phase Vocoder有效消除了互调失真,但同时引入了Phase Vocoder方法典型的“Phasyness”信号畸变。 论文深入研究了MaxxBass和VB Phase Vocoder两种虚拟低音增强算法,对比分析了两种算法的理论基础、算法原理、结构设计和实现细节;在VB Phase Vocoder的基础上,提出了改进的谐波相位构造方法以消除合成相位失真带来的信号畸变,同时在改进算法中引入峰值匹配机制并结合MaxxBass算法优化了谐波能量控制策略,进一步提出了改进的基于相位声码器的虚拟低音增强算法EVB Phase Vocoder,并给出了改进算法的参数讨论和有效性分析。 本文最后针对MaxxBass、VB Phase Vocoder、EVB Phase Vocoder三种算法设计了主客观对比实验。实验结果证明,EVB Phase Vocoder算法在低音效果还原尤其是畸变控制方面相对于VB Phase Vocoder算法和MaxxBass算法均有提高。 如何进一步结合心理声学模型优化算法输出,并将算法有效应用于低频截止频率更高的扬声器系统,值得进一步地研究。
英文摘要Virtual Bass refers to the technology that the lowest frequency components of the signal which cannot be reproduced by the loudspeaker are shifted to above the loudspeaker’s low cut-off frequency to recover the bass impression. How to produce bass using a loudspeaker which has no low-frequency capability is a longstanding challenge. To solve the problem, this thesis has thoroughly studied the implementation and improvement issues of Virtual Bass algorithms. The Virtual Bass system is traditionally realized by nonlinear devices which could be easily realized and relatively efficient but could introduce the inherent defect of intermodulation distortion. The MaxxBass bass enhancing system is the typical representative of the kind. On the other hand, compared to nonlinear algorithms, the joint time-frequency processing approaches gain superiority in both flexibility and distortion control. The Phase Vocoder based system, VB Phase Vocoder, could effectively eliminate the nonlinear distortion but in the meantime may bring another type of distortion known as “Phasyness” which is usually related to Phase Vocoder approaches. The thesis extensively analyzes and evaluates the MaxxBass and VB Phase Vocoder systems, including the theoretic basis, algorithm principles, framework and realization details. On the basis of VB Phase Vocoder, an improved Phase Vocoder based Virtual Bass system called EVB Phase Vocoder has been developed. The improved system adopts a new harmonic phase constructing approach, introduces the peak-matching strategy, and chooses the simplified harmonic gain control method to reduce the computational complexity. Finally, the parameter optimization and algorithm effectiveness are discussed. Objective and subjective tests have been carried out to compare the effectiveness of Maxxbass, VB Phase Vocoder and EVB Phase Vocoder. The tests indicate that the EVB Phase Vocoder system achieves better results in the aspects of bass impression, distortion control and computational complexity control. Further research efforts will be devoted to optimizing the system output by further integrating the psychoacoustic models and applying the system to loudspeakers with higher low cut-off frequency.
语种中文
公开日期2011-05-07
页码77
源URL[http://159.226.59.140/handle/311008/412]  
专题声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文
推荐引用方式
GB/T 7714
郭杰. 基于相位声码器的虚拟低音增强算法研究[D]. 声学研究所. 中国科学院声学研究所. 2008.

入库方式: OAI收割

来源:声学研究所

浏览0
下载0
收藏0
其他版本

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。