The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

About This OTS Dataset

Unlock the potential of AI development with the Dutch General Utterances Speech Dataset, tailored for multiple general topics. This specialized collection of voice data is meticulously curated to enhance the understanding and analysis of general conversational topics in Dutch.

Metadata Availability: Insights into Participant Details

While transcripts are not included, this dataset comes with comprehensive metadata for each recording, offering:

  • Participant demographics
  • Call characteristics

This metadata improves the dataset’s value for analysis and model training in the general conversational domain.

Audio Recording Specifications

  • Audio Duration: 300 hours
  • Format: WAV (maintains optimal fidelity)
  • Sample Rate: Variable, adaptable to project needs
  • Bit Depth: 16-bit for high audio clarity
  • Recording Environments: Real-world settings
  • Equipment: Standard recording devices for authenticity

These specifications ensure maximum compatibility for various AI/ML development purposes across different sectors.

Speech Data Overview

Our dataset comprises 300 hours of Dutch audio recordings, representing authentic conversational scenarios across various general topics.

  • Format: WAV
  • Regional Focus: Native dialects and accents from Belgium
  • Use Case: Model training, NLP, speech recognition, conversational AI

License

This dataset is exclusively provided by Macgence and is available for commercial use under transparent licensing.

Updates and Customization

  • Frequent Updates: Ensures relevance for ongoing research and development
  • Tailored Versions Available: Customized datasets can be created upon request

👉 Submit your request to discuss your requirements, check pricing, and acquire the dataset.

Keywords

keywords: Dutch speech dataset, general conversation dataset, Dutch audio dataset, Dutch language corpus, conversational AI Dutch, speech recognition dataset Dutch, Dutch voice recordings, spoken Dutch dataset, Dutch utterances dataset, audio dataset for AI, NLP dataset Dutch, language modeling Dutch, regional speech dataset Belgium, Belgian Dutch dataset, AI training data Dutch, voice AI Dutch, Dutch call center data, multilingual dataset, speech dataset Europe, general topic audio corpus

Downloads last month
84