Indonesian Text-to-Speech Audio Dataset with 16.38 Hours of Speech

16.38 hours of high-quality audio for Indonesian Text-to-Speech (TTS) applications. The dataset contains 4,531 audio segments with an average duration of 13.01 seconds each. It was created by Muhammad Arief from Universitas Muhammadiyah Sorong.

Use Cases

• Train Text-to-Speech models based on the 16.38 hours of Indonesian speech audio.
• Benchmark speech synthesis quality based on the high-quality, lossless WAV format.
• Fine-tune pre-trained TTS models based on the dataset's 4,531 distinct audio segments.
• Develop voice cloning applications based on the dataset's technical specifications like 22 kHz sample rate.

Strengths

• Contains 16.38 hours of total audio duration, providing substantial material for model training.
• Comprises 4,531 individual audio files, offering a high number of distinct speech segments.
• Audio is stored in a lossless WAV format with a 22 kHz sample rate and PCM_16 bit depth, indicating high technical quality.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-01-13 19:48:00; freshness should be verified.

Provenance

Source

Muhammad Arief, S.Kom., Universitas Muhammadiyah Sorong

Geography

Indonesia