首页> 外国专利> ESTIMATION MODEL CONSTRUCTION METHOD, PERFORMANCE ANALYSIS METHOD, ESTIMATION MODEL CONSTRUCTION DEVICE, PERFORMANCE ANALYSIS DEVICE, AND PROGRAM

ESTIMATION MODEL CONSTRUCTION METHOD, PERFORMANCE ANALYSIS METHOD, ESTIMATION MODEL CONSTRUCTION DEVICE, PERFORMANCE ANALYSIS DEVICE, AND PROGRAM

机译:估计模型施工方法,性能分析方法,估计模型施工装置,性能分析装置和程序

摘要

PROBLEM TO BE SOLVED: To analyze a performance of musical instruments with high accuracy.;SOLUTION: Provided is a device for constructing an estimation model M which estimates onset data D representing a pitch at which an onset exists from feature amount data F representing a feature amount of a performance sound of an instrument, which includes: a training data preparation unit 21 which prepares a plurality of pieces of training data T; and an estimation model construction unit 22 which constructs the estimation model M by machine learning using the plurality of pieces of training data T. The training data preparation unit 21 prepares a plurality of pieces of training data which include: the feature amount data representing the feature amount of the performance sound of the instrument; first training data which includes the onset data representing the pitch at which the onset exists; feature amount data representing the feature amount of a generated sound by a sound source different from the instrument; and second training data which includes the onset data representing that the onset does not exist.;SELECTED DRAWING: Figure 3;COPYRIGHT: (C)2021,JPO&INPIT
机译:要解决的问题:分析高精度的乐器性能。;解决方案:提供了一种用于构造估计模型M的装置,其估计表示从代表特征的特征量数据F存在发起的音高的发作数据D.仪器的性能声音的量,包括:训练数据准备单元21,其准备多条训练数据T;和估计模型构造单元22通过使用多个训练数据T构造估计模型M的估计模型M.训练数据准备单元21准备多条训练数据,该数据包括:表示该功能的特征量数据仪器性能声音的数量;第一训练数据包括表示存在发起的音高的生病数据;特征量数据表示由仪器不同的声源的生成声音的特征量;和第二训练数据包括表示不存在的生病数据。;所选绘图:图3;版权:(c)2021,JPO和INPIT

著录项

  • 公开/公告号JP2021128297A

    专利类型

  • 公开/公告日2021-09-02

    原文格式PDF

  • 申请/专利权人 YAMAHA CORP;

    申请/专利号JP20200023948

  • 申请日2020-02-17

  • 分类号G10G1;G10H1;

  • 国家 JP

  • 入库时间 2022-08-24 22:24:32

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