首页> 外国专利> FEATURE SOUND EXTRACTION METHOD, FEATURE SOUND EXTRACTION DEVICE, COMPUTER PROGRAM, AND DISTRIBUTION SYSTEM

FEATURE SOUND EXTRACTION METHOD, FEATURE SOUND EXTRACTION DEVICE, COMPUTER PROGRAM, AND DISTRIBUTION SYSTEM

机译:特征声音提取方法,特征声音提取装置,计算机程序和分配系统

摘要

PROBLEM TO BE SOLVED: To provide a feature sound extraction method, etc., for specifying a most characteristic part even when a plurality of feature parts exist in a daily life sound, and extracting a sound relative to a specific part.SOLUTION: A feature sound extraction method executed by a computer, performs processes of: sectioning sound data for each prescribed time; calculating a feature quantity including a frequency component of sound data for each sectioned period; causing a function f(x) in which a function subjected to differential calculus or subdifferential calculus by x is a monotonous decrease in a section a≤x≤a(0≤aa≤∞) and there exists the lower bound T of a function value satisfying an expression below to act on the component value x of one or more feature quantities including the frequency component of sound data calculated for each period; calculating a feature quantity for each period on the basis of the result of the action; and extracting feature sound data on the basis of the calculated feature quantity for each period.SELECTED DRAWING: Figure 2
机译:解决的问题:提供一种特征声音提取方法等,即使在日常生活的声音中存在多个特征部分时,也能指定最有特征的部分,并提取相对于特定部分的声音。由计算机执行的声音提取方法,执行以下处理:在每个规定的时间分割声音数据;计算每个分段时段包括声音数据的频率分量的特征量;导致函数f(x),其中经过x进行微积分或亚微积分的函数在a≤x≤a(0≤a

著录项

  • 公开/公告号JP2016126479A

    专利类型

  • 公开/公告日2016-07-11

    原文格式PDF

  • 申请/专利权人 FUJITSU LTD;

    申请/专利号JP20140266102

  • 发明设计人 OKABAYASHI MIWA;ODAJIMA SHIGEYUKI;

    申请日2014-12-26

  • 分类号G06F17/30;G10L15/10;G10L25/51;

  • 国家 JP

  • 入库时间 2022-08-21 14:47:29

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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