---
title: 2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus
canonical_url: "https://www.modelscope.cn/datasets/DatatangBeijing/2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus"
md_url: "https://www.modelscope.cn/datasets/DatatangBeijing/2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus.md"
repository: DatatangBeijing/2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus
chinese_name: "2人粤语多情感自由对话合成库"
last_updated: 2025-12-22
license: "Apache License 2.0"
storage_size: "659 KB"
domain:
  - image
tasks:
  - image-to-text
size_scale:
  - 100-10k
downloads: 27
stars: 0
---

# 2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus

> 2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus - DatatangBeijing 在 ModelScope 开源的数据集。2人粤语多情感自由对话合成库，由广州母语发音人录制，自由对话风格。给定话题，发音人自由发挥，情感包含正常、开心、生气、恐惧、厌恶、悲伤等。专业语音学家参与标注，带有情感和副语言标注。精准匹配自然度高且富有情感的语音合成的研发需求。

DatatangBeijing/2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus 是 ModelScope 魔搭社区上的image-to-text数据集，涉及 image 领域，存储大小 659 KB，采用 Apache License 2.0 许可。

- **Repository**: DatatangBeijing/2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus
- **License**: Apache License 2.0
- **Tasks**: image-to-text
- **Domain**: image
- **Storage size**: 659 KB
- **Downloads**: 27
- **Stars**: 0
- **Last updated**: 2025-12-22

Source: https://www.modelscope.cn/datasets/DatatangBeijing/2People_Cantonese_Multi_emotional_Natural_Conversation_Speech_Synthesis_Corpus

---

# 2人粤语多情感自由对话合成库

## 数据堂
详见https://www.datatang.com/dataset/1775?source=modelscope

## 数据集描述
2人粤语多情感自由对话合成库，由广州母语发音人录制，自由对话风格。给定话题，发音人自由发挥，情感包含正常、开心、生气、恐惧、厌恶、悲伤等。专业语音学家参与标注，带有情感和副语言标注。精准匹配自然度高且富有情感的语音合成的研发需求。

### 格式：
48kHz，24bit，wav
### 录音环境：
专业录音棚环境
### 录音内容：
给定话题，发音人自由发挥；情感包含正常、开心、生气、恐惧、厌恶、悲伤等
### 人员：
专业声优，自由对话，非朗读；男声1人，女声1人
### 标注特点：
文字标注，副语言标注，情感标注
### 设备：
专业录音设备及软件
### 语言：
粤语
### 应用场景：
语音合成


## 数据集版权信息
版权归数据堂所有，商用数据。
