A high-quality, open-source dataset for Arabic Text-to-Speech (TTS) research, containing paired audio and text samples from both male and female speakers. All audio is provided in 24kHz WAV format, with rich metadata and phonetic transcriptions.
This dataset is designed for training and evaluating neural TTS systems in Modern Standard Arabic. It includes:
Audio: Clean, studio-quality WAV files at 24,000 Hz.
Text: Original Arabic… See the full description on the dataset page:
https://huggingface.co/datasets/NeoBoy/arabic-tts-wav-24k.
