• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Computer speech and language >Introduction to the Special Issue on Broadening the View on Speaker Analysis
【24h】

Introduction to the Special Issue on Broadening the View on Speaker Analysis

机译：拓宽演讲者分析视野专刊介绍

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In the last five decades, the focus of automatic speech analysis and more recently of automatic singing analysis was on the linguistic and structural content side: words (and note-events) and their semantic interpretation. Yet, when it comes to the human behind speaking and singing, so far, research has been mostly interested in the identity of the person. Only in the last one and a half decades, increasing effort has been invested to computationally analyse a variety of speaker and singer states and traits that do not denote but characterise the person. There are short-term states that are indicated by different speaking and singing styles such as emotions (full blown, prototypical) and emotion-related states or affects (e.g., stress, intimacy, interest, confidence, uncertainty, deception, politeness, frustration, sarcasm, and pain). Medium-term phenomena are between states and traits; they include (partly) self-induced more or less temporary states (e.g., sleepiness, medical, medical and alcohol intoxication, health state, or moods such as depression) and structural (behavioural, interactional, social) signals (e.g., role in dyads or groups, friendship and identity, positiveegative attitude). Finally, there are long-term traits, such as biological trait primitives (e.g., height, weight, age, gender), group/ethnicity membership (race/culture/social class with a weak borderline towards other linguistic concepts, i.e., speech registers such as dialect, regional accent, or nativeness), or personality traits (such as likability and personality in general) - just to mention a few. All these tasks are so far mostly handled in isolation when it comes to automatic analysis; yet, it seems intuitive that they are highly inter-dependent.
机译：在过去的五十年中，自动语音分析以及最近的自动歌唱分析都集中在语言和结构内容方面：单词（和音符事件）及其语义解释。但是，到目前为止，谈到说话和唱歌背后的人时，研究大多对人的身份感兴趣。仅在过去的十五年中，人们才投入了更多的精力来通过计算分析各种说话者和歌手的状态和特征，这些状态并不能表示人的特征。有些短期状态以不同的说话和唱歌风格表示，例如情绪（完全爆发，典型）和与情绪相关的状态或影响（例如压力，亲密感，兴趣，信心，不确定性，欺骗，礼貌，沮丧，讽刺和痛苦）。中期现象介于状态与特质之间；它们包括（部分）自我诱导的或多或少的临时状态（例如，嗜睡，医学，医学和酒精中毒，健康状态或情绪（如抑郁））和结构性（行为，互动，社会）信号（例如，双胞胎的作用）或团体，友谊和身份，积极/消极的态度）。最后，还有一些长期特征，例如生物学特征原语（例如身高，体重，年龄，性别），群体/族裔成员（种族/文化/社会阶层，对其他语言概念（例如，语音记录）的边界较弱）例如方言，区域性口音或原住民）或人格特质（例如一般的喜好和人格）-仅举几例。到目前为止，在进行自动分析时，所有这些任务大多数都是孤立地处理的。然而，似乎直觉上它们是高度相互依赖的。

著录项

  • 来源
    《Computer speech and language》 |2014年第2期|343-345|共3页
  • 作者

    Bjoern Schuller; Stefan Steidl; Anton Batliner; Florian Schiel; Jarek Krajewski;

  • 作者单位

    Imperial College London, Department of Computing, United Kingdom,Technische Universitaet Muenchen, Machine Intelligence & Signal Processing Group, MMK, Germany;

    Friedrich-Alexander University Erlangen-Nuremberg, Pattern Recognition Lab, Germany;

    Technische Universitaet Muenchen, Machine Intelligence & Signal Processing Group, MMK, Germany,Friedrich-Alexander University Erlangen-Nuremberg, Pattern Recognition Lab, Germany;

    Bavarian Archive for Speech Signals, Ludwig-Maximilians-Universitaet Muenchen, Germany;

    University of Wuerzburg, Industrial and Organizational Psychology, Germany;

  • 收录信息 美国《科学引文索引》(SCI);美国《工程索引》(EI);
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

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