首页> 外文会议>Conference on Speech Technology and Human-Computer Dialogue >GENETIC OPTIMIZATION OF WAVELET PACKET COMPRESSION FOR SPEECH SIGNAL
【24h】

GENETIC OPTIMIZATION OF WAVELET PACKET COMPRESSION FOR SPEECH SIGNAL

机译:语音信号小波包压缩的遗传优化

获取原文

摘要

The compression of speech signals using Wavelet Packet (WP) consider fixed values of threshold for elimination of small coefficients from resulting subspaces, but none of them take into account the speech specific characteristics. This leads to rude distortions of high frequencies of speech signal and small compression degrees. In this paper, we propose the use of genetic algorithms (GA) for finding multiple threshold values (one for each subspace of WP decomposition) to achieve the maximum compression and the maximum quality of the synthesized signal. We transformed the multiobjective optimization algorithm with controversial fitness functions (quality F1 and compression F2) into an simple one by building a self-adaptive function witch take into account the both targets. The final solution will be determined from the Pareto space generated by the F1 and F2.
机译:使用小波包(WP)的语音信号的压缩考虑从得到的子空间中消除小系数的固定值,但是它们都不考虑语音特定特征。 这导致粗鲁的语音信号和小压缩度的高频扭曲。 在本文中,我们提出了使用遗传算法(GA)来查找多个阈值(对于WP分解的每个子空间一个)以实现最大压缩和合成信号的最大质量。 通过构建自适应函数女巫考虑两个目标,通过构建自适应函数女巫将多目标优化算法(质量F1和压缩F2)转换为简单的函数。 最终解决方案将从F1和F2产生的帕累托空间确定。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号