zaiduplift/tts_training_text_clean_chars_only_audio

Urdu Speech Synthesis Dataset

This dataset contains synthesized Urdu speech audio files along with their original and romanized transcripts.

Dataset Structure

The dataset contains the following columns:

• audio: The synthesized audio file in WAV format.
• original_transcript: The original Urdu text used for synthesis.
• romanized_transcript: The romanized version of the Urdu text.
• duration_ms: The duration of the audio file in milliseconds.
• synthesis_job_id: The ID of the synthesis job that produced this audio.

Usage

This dataset can be used for tasks such as:

1. Evaluating text-to-speech models for Urdu
2. Training speech recognition models for Urdu
3. Studying the relationship between Urdu text and its romanized form

Creation

This dataset was created using a batch synthesis process with multiple TTS engines.