dbschaeffer/schaeffer_thesis_corrected

---
dataset_info:
features:
- name: audio
dtype: audio
- name: username
dtype: string
- name: Processes
dtype: string
- name: PulseTypology
dtype: string
- name: Complexity
dtype: string
- name: Onset
dtype: string
- name: Offset
dtype: string
- name: Type
dtype: string
- name: MassType
dtype: string
- name: Direction
dtype: string
- name: description
dtype: string
splits:
- name: train
num_bytes: 1918141228
num_examples: 788
download_size: 1608587794
dataset_size: 1918141228
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: cc-by-4.0
task_categories:
- text-to-audio
- audio-classification
language:
- en
size_categories:
- n<1K
---

The SCHAEFFER dataset (Spectro-morphogical Corpus of Human-annotated Audio with Electroacoustic Features for Experimental Research), is a compilation of 788 raw audio data accompanied by human annotations and morphological acoustic features.
The audio files adhere to the concept of Sound Objects introduced by Pierre Scaheffer, a framework for the analysis and creation of sound that focuses on its typological and morphological characteristics.
Inside the dataset, the annotation are provided in the form of free text, while the labels are pre-chosen from a list of classes, making the sound description fit into a suitable framework for digital analysis.
All the sounds within the dataset are under a "CC-By-4.0-attribution" license.

原始信息汇总

数据集概述

数据集信息

• 特征列表:

  • audio: 音频数据
  • username: 用户名
  • Processes: 处理过程
  • PulseTypology: 脉冲类型
  • Complexity: 复杂度
  • Onset: 起始
  • Offset: 结束
  • Type: 类型
  • MassType: 质量类型
  • Direction: 方向
  • description: 描述

• 数据分割:

  • train: 训练集，包含788个样本，占用1918141228字节

• 数据大小:

  • 下载大小: 1608587794字节
  • 数据集大小: 1918141228字节

• 配置:

  • default: 默认配置，包含训练集数据文件路径为data/train-*

• 许可证:

  • CC-BY-4.0

• 任务类别:

  • 文本到音频
  • 音频分类

• 语言:

  • 英语

• 规模类别:

  • 样本数小于1K

数据集描述

• 名称: SCHAEFFER 数据集
• 内容: 包含788个原始音频数据，附有人工注释和形态声学特征。
• 音频文件: 遵循Pierre Scaheffer提出的声音对象概念，专注于声音的类型和形态特征。
• 注释形式: 自由文本形式的注释和预选类别的标签，适用于数字分析。
• 版权: 所有声音数据均在CC-By-4.0-attribution许可证下。