首页> 外文会议>Conference on Speech Technology and Human-Computer Dialogue >AN IMPROVED HYBRID TIME-FREQUENCY ALGORITHM FOR TIME-SCALE MODIFICATION OF SPEECH/AUDIO SIGNALS
【24h】

AN IMPROVED HYBRID TIME-FREQUENCY ALGORITHM FOR TIME-SCALE MODIFICATION OF SPEECH/AUDIO SIGNALS

机译:一种改进的语音/音频信号时尺度修改的混合时频算法

获取原文

摘要

The paper presents a new hybrid algorithm for time-scale modification, based on the Knox, Bailey and Stewart solution (KBS-TSM). The proposed algorithm improves the segmentation and concatenation methods and uses a more elegant spectral analysis involving instantaneous frequency (IF) spectrogram. The efficiency of the new algorithm is increased by including the IF attractors as a better estimation for the frequencies of the partials. Finally, a better balance between temporal and frequency resolution is obtained by decoupling the lengths of analysis and synthesis frames for the sinusoidal modelling.
机译:本文提出了一种新的混合算法,用于时间尺度修改,基于KNOX,Bailey和Stewart解决方案(KBS-TSM)。该算法提高了分割和串联方法,并使用涉及瞬时频率(IF)频谱图的更优异的光谱分析。通过包括IF吸引器的效率增加了新算法,作为偏离部分的频率的更好估计。最后,通过对正弦建模的分析和合成帧的长度分离来获得时间和频率分辨率之间的更好平衡。

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