farabi-lab/kazakh-stt

Kazakh Speech Dataset (KSD)

1. Dataset Summary

• Purpose: High-quality, open-source Kazakh speech dataset for Automatic Speech Recognition (ASR) system development.
• Developed by: Department of Artificial Intelligence and Big Data, Al-Farabi Kazakh National University.
• Total Duration: 554 hours of recorded speech.
• Total Number of Speakers: 873
• Average Sentences per Speaker: 250 sentences (utterances).
• Total Utterances: 204,250
• File Format: .wav
• Audio Characteristics:
  • Sample Rates: 16 kHz or 22 kHz
  • Bit Depth: 16-bit
  • Channels: Mono

2. Corpus Features

• Speaker Diversity: Speakers from diverse regions, age groups, and genders.
• Recording Devices: Recorded using mobile devices (iOS and Android).
• Transcription Quality: Verified by native Kazakh speakers for accuracy.

3. Applications

• Primary Use Case: Automatic Speech Recognition (ASR) system training.
• Additional Use Cases:
  • Speech-to-text systems.
  • Voice-activated assistants.
  • Speaker identification.
  • Linguistic research on Kazakh phonetics and dialects.

4. Technical Characteristics

• File Format: .wav
• Sample Rates: 16 kHz, 22 kHz, or 44 kHz.
• Bit Depth: 16-bit
• Channels: Mono
• Recording Devices: Mobile devices (iOS, Android).

5. Citation

If you use this dataset, please cite it as follows:

@article{kadyrbek2023ksd,
  author = {Kadyrbek, N.; Mansurova, M.; Shomanov, A.; Makharova, G.},
  title = {The Development of a Kazakh Speech Recognition Model Using a Convolutional Neural Network with Fixed Character Level Filters},
  journal = {Big Data and Cognitive Computing},
  year = {2023},
  volume = {7},
  number = {3},
  pages = {132},
  doi = {https://doi.org/10.3390/bdcc7030132}
}