Datasets:
audio audioduration (s) 1.23 2.45 |
|---|
BedVibe Emotional Speech Dataset
Studio-quality emotional speech dataset
• Preview samples available on Hugging Face • Full commercial dataset available via BedVibe Studio • Languages currently available: English, Greek • Additional languages can be recorded on request • 6 emotions • 48 kHz / 32-bit float audio • Professionally recorded in studio conditions
Studio-recorded emotional speech datasets designed for training text-to-speech systems.
Currently available on our website: more than 108,000 recordings and growing.
Features
- Multiple speakers
- Expandable multilingual recordings (additional languages can be recorded on request)
- Six emotional states (neutral, happy, angry, whisper, shouting, scared)
- Clean studio speech suitable for TTS training
Languages
Core dataset languages currently available:
- English
- Greek
Additional languages can be recorded and expanded on request, including:
- Norwegian
- Swedish
- Danish
- French
- Russian
- and others
Contact us at bedvibe@bedvibe.studio for custom dataset generation. Audio demonstrations https://tts.bedvibe.studio https://tts.bedvibe.studio/datasets.html
Dataset details and commercial licensing https://tts.bedvibe.studio/datasets.html
Commercial use requires licensing through BedVibe Studio.
- Downloads last month
- 60