Classifying soundscapes using a multifaceted taxonomy
Classifying soundscapes using a multifaceted taxonomy
复制标题
使用多方面的分类法对音景进行分类
DOI：
--
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
C. Guastavino
中科院分区：
文献类型：
--
作者：
Christopher Trudeau;C. Guastavino
在信息研究领域，人们提出了多种组织原则来对信息进行分类。等级分类系统是最常见的，但它们要求不同级别之间的关系一致性和类之间的互斥性。由于这两个原因，在分层分类中组织音景的尝试具有挑战性。听力测试结果表明，人们根据声源、产生声音的动作和/或感知这些声音的环境对日常声音进行交叉分类。此外，音景是具有多个并发声源的复杂声音场景。本文提出了一个分面分类法，反映了人们交叉分类的声音。在分类法中使用了五个方面，每个方面都反映了声音的基本特征：代理，源，动作，上下文和声学特征。在某些方面中，可以建立术语的层次结构，因此方面必须允许术语之间的层次关系。这一分类法将使用关系数据库（MySQL）和网络界面实施，供用户浏览分类法的内容和结构。拟议的数据库将使用现有的声音数据集（如freesound.org）进行填充。在这种分类法中，声景可以被描述为一组声音，消除了将声景隔离到分类法中的一个位置的挑战。我们将分析自由格式的口头描述伴随着每一个声音的方面提出的分类法，并探讨新的方面从新兴的类别。因此，建议的分类支持正在进行的努力，标准化和报告声景研究提供了一种方法来系统地描述一个给定的声景。
Multiple organizational principles have been proposed to classify information in the field of information studies. Hierarchical classification systems are the most familiar, but they require relational consistency between the different levels and mutual exclusivity across classes. Attempts to organize soundscapes within hierarchical classification are challenging for these two reasons. Listening test results suggest that people cross-categorize everyday sounds according to the sound source, the action generating the sound, and/or the context in which these sounds are perceived . Moreover, soundscapes are complex sound scenes with multiple, concurrent sound sources. This paper proposes a faceted taxonomy that reflects the way people cross-categorize sounds. Five facets are used in the taxonomy, each reflecting a fundamental characteristic of the sound: agent, source, action, context, and acoustic features. Within some facets, a hierarchy of terms can be established, and thus the facets must allow for a hierarchical relationship between the terms. This taxonomy will be implemented using a relational database (MySQL) and a web interface for users to navigate the content and structure of the taxonomy. The proposed database will be populated using existing sound datasets, such as freesound.org. In this taxonomy, a soundscape can be described as a set of sounds, eliminating the challenge of isolating a soundscape to one placement in a taxonomy. We will analyze the free format verbal descriptions accompanying each sound in terms of the facets of the proposed taxonomy, and explore new facets from emerging categories. Thus, the proposed taxonomy supports the ongoing efforts to standardize and report soundscape research by offering a way to systematically describe a given soundscape.