Urdu Voice Enabled Assistive Technologies for Print Disabled Community of Pakistan - Introduction

Urdu Voice Enabled Assistive Technologies for Print Disabled Community of Pakistan



To improve Urdu TTS in the areas of Natural Language Processing (NLP) and speech synthesis.

To improve date formats that contain English slashes (/), commas (,), dashes (-), dots (.) and English names of months (e.g., March, Mar) will be handled.

Support for handling English text will be added In addition to the 70,000 words vocabulary, the phonetic lexicon used by the TTS system will be enhanced by adding 10,000 transliterated English words.

To enhance Urdu TTS for using a single voice with Urdu accent to read aloud mixed content, containing both Urdu and English.

To improve the response time of the Urdu TTS; the speech quality for naturalness and intelligibility.

Deep learning based voice will be developed for Urdu TTS.

4 hours Urdu speech corpus will be developed and annotated at intonation level to improve the naturalness and intelligibility of existing Urdu voice of TTS.

The usability of Urdu TTS will be enhanced by integrating it with SAPI-5.