音频编码器中导谱频率系数的差错隐藏算法

发布时间：2018-03-02 17:16

本文选题：差错隐藏　切入点：导谱频率　出处：《计算机工程与应用》2017年02期 　论文类型：期刊论文

【摘要】：为了降低移动网络丢包、抖动失真等现象对音频质量的影响,提出一种基于帧间相关性的导谱频率系数的差错隐藏算法。该方法根据近邻成功接收若干好帧的导谱频率(Immittance Spectral Frequencies,ISF)距离预测与当前坏帧的相关性,再基于相关性进行导谱频率系数的自适应坏帧恢复。解决了现有算法采用单一的坏帧恢复技术,在面临多样化的音频信号输入时,坏帧恢复效果不佳的问题。在1%~10%的误码条件下的客观实验结果表明,该算法谱失真明显优于AVS-P10(AVS移动音频编码标准)的差错隐藏算法,性能平均得到42%的提升。主观实验也同时表明了该算法的优势。该方法可扩展应用到音频编码器的自适应码本及固定码本等参数的差错隐藏。
[Abstract]:In order to reduce the impact of packet loss and jitter distortion on audio quality in mobile networks, This paper presents an error concealment algorithm based on inter-frame correlation of conductance frequency coefficients, which is based on the correlation between distance prediction and current bad frames of several good frames successfully received by the nearest neighbor. Then adaptive bad frame recovery based on correlation is carried out, which solves the problem that the existing algorithms adopt a single bad frame recovery technique, and face a variety of audio signal input. The objective experimental results show that the spectral distortion of the algorithm is obviously superior to that of the AVS-P10(AVS mobile audio coding standard. The performance is improved by an average of 42%. Subjective experiments also show the advantages of the algorithm. This method can be extended to the error concealment of adaptive codebook and fixed codebook of audio encoder.
【作者单位】：武汉大学计算机学院国家多媒体软件工程技术研究中心;武汉大学深圳研究院;
【基金】：国家自然科学基金(No.61102127,No.61231015,No.61201340,No.61201169,No.61272278) 国家科技重大专项项目(No.2010ZX03004-003-03) 湖北省自然科学基金(No.2012FFB04205) 武汉市晨光科技计划(No.201150431104)
【分类号】：TN919.81

【相似文献】