Malay Wikipedia Audio Synthesis with Azure TTS

Azure TTS Yasmin Wikipedia contains speech audio generated from Wikipedia text using Microsoft Azure's text-to-speech technology. The dataset was created by the user 'mesolitica' and uploaded to Hugging Face in July 2022. It is categorized as containing over 100,000 rows of data.

Use Cases

• Train a text-to-speech model on paired Malay text and corresponding 'Yasmin' voice audio waveforms.
• Benchmark speech synthesis quality by comparing generated audio from this dataset's text prompts against its provided audio samples.
• Fine-tune a voice conversion model using the consistent 'Yasmin' speaker identity present across all audio clips.
• Analyze prosody and pronunciation patterns for the Malay language using the aligned text transcripts and synthesized speech.

Strengths

• Dataset contains over 100,000 data points as indicated by its '100K<n<1M' size category.
• Provides aligned text-audio pairs, a fundamental structure for speech synthesis tasks.

Limitations

• Audio quality and characteristics are constrained by the Azure TTS service's capabilities and may not represent natural human speech.
• Limited metadata available; specific text sources, audio specifications, and license details are unknown.

Provenance

Source

Text likely sourced from Wikipedia, with audio synthesized via Microsoft Azure TTS.

Collection Method

Automated synthesis using the Azure Text-to-Speech API, presumably with the 'Yasmin' voice profile.

Freshness

Last updated in July 2022; may not reflect the latest Azure TTS service features.

Geography

Region tag suggests a US focus, but content is in Malay, indicating a potential mismatch or broad scope.