README.md

metadata

dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: text
      dtype: string
    - name: pinyin
      dtype: string
    - name: raw_text
      dtype: string
    - name: raw_pinyin
      dtype: string
    - name: domain
      dtype: string
  splits:
    - name: train
      num_bytes: 87216120
      num_examples: 179
  download_size: 79503496
  dataset_size: 87216120
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-nc-nd-4.0
task_categories:
  - text-to-speech
language:
  - zh

TTS-SCCusSerFSC: A Scripted Chinese Customer Service Female Speech Corpus

TTS-SCCUSSERFSC：中文客服场景女性语音语料库

This open-source dataset consists of 22 minutes of annotated female voices in Mandarin Chinese that is applicable for Text-to-Speech Synthesis especially in customer service scenes, where 250 utterances collected from a 22-year-old woman were contained.

此数据集由22分钟的普通话带注释的女声组成，包含了从一名22岁女性那里收集的适用于语音合成的250条中文普通话客服场景语音数据，内容为客服对话用语。

Dataset Overview

数据集概述

Dataset Type	speech corpus for TTS
Language	zh-CN, Mandarin Chinese (China)
Speech Style	scripted monologue
Content	customer service conversations
Audio Parameters	48 kHz, 24 bits, mono
File Format	WAV (PCM) TXT (UTF-8)
Recording Equipment	Nuemann U87-Neve 1073-RME Fireface
Recording Environment	recording studio

数据集类型	语音合成（TTS）语音数据集
语种	zh-CN，中文普通话（中国）
语音类型	朗读式独白
内容	客服对话
音频参数	48 kHz, 24 bits, 单通道
文件格式	WAV (PCM) TXT (UTF-8)
录音设备	Nuemann U87-Neve 1073-RME Fireface
录音环境	recording studio

Cite

引用

@online{TTS-SCCusSerFSC,
title = {TTS-SCCusSerFSC: A Scripted Chinese Customer Service Female Speech Corpus},
author = {Magic Data Technology},
url = {https://magichub.com/datasets/mandarin-chinese-speech-corpus-for-tts-customer-service-scene/},
urldate = {2025-04-26}