Dataset Viewer
Auto-converted to Parquet
Search is not available for this dataset
audio
audio

French Call Center Speech Dataset (~164 hours)

Description
A large-scale dataset of real French call center recordings.
Telephone-quality conversations between customers and agents.
Useful for ASR, NLP, speech analytics, and training voicebots.

Technical details

  • Language: French
  • Total duration: ~164 hours
  • Format: MP3
  • Channels: Mono (1 channel, mixed client and agent)
  • Sample rate: 8000 Hz
  • Bitrate: 32 kbps
  • Metadata: Not available

License
Commercial license only. Redistribution not allowed.

Contact
📧 [email protected]
💬 Telegram @Marat_DV


📖 Description en français

Description
Un large jeu de données composé d’enregistrements réels de centres d’appels en français.
Qualité téléphonique, conversations « client–agent ».
Convient à l’entraînement des systèmes ASR (reconnaissance vocale), NLP, assistants vocaux et analyse de dialogues.

Détails techniques

  • Langue : français
  • Durée totale : ~164 heures
  • Format : MP3
  • Canaux : mono (client et agent mélangés)
  • Fréquence d’échantillonnage : 8000 Hz
  • Débit binaire : 32 kbps
  • Métadonnées : non disponibles

Licence
Utilisation commerciale uniquement. Redistribution interdite.


Sample

Downloads last month
35