Sparsha: A Comprehensive Indian Language Toolset for the Blind
Anirban Lahiri, Satya Jyoti Chattopadhyay, Anupam Basu · 2005 · Proceedings of the 7th International ACM SIGACCESS Conference on Computers and Accessibility (Assets '05) · doi:10.1145/1090785.1090807
Summary
This paper presents Sparsha ("feeling of touch" in Hindi), the first comprehensive toolset enabling over 13 million visually impaired people in the Indian subcontinent to read and write in their native languages using a computer. Existing Braille translation systems and screen readers were designed for English and other Western languages and could not be extended to Indian languages due to fundamental differences in script structure. Indian languages use complex scripts with consonant conjuncts (conjugates) formed through a special character called halant, where the visual representation of a conjunct can differ substantially from its component characters. Bharati Braille — the standard for representing Indian languages in six-dot Braille format — assigns the same Braille cell to phonetically equivalent characters across different languages, meaning that a single Braille combination may map to different characters depending on the language. The Sparsha toolset addresses these challenges through four integrated components: a Bharati Braille transliteration system supporting forward and reverse translation between text and Braille for Hindi, Bengali, Assamese, Marathi, Gujarati, Oriya, Telugu, and Kannada; a Nemeth-based mathematical expression editor with a GUI that eliminates the need to learn LaTeX; Sparsha Chitra, a tactile graphics tool that converts images to embossable form; and a file reader that provides audio feedback for Indian language text within Microsoft Word.
Key findings
The Braille translation system uses a finite state machine approach with separate code tables per language, making it highly scalable — Urdu and Sinhala extensions were already underway. It handles both Grade 1 and Grade 2 English Braille with a user-extensible contractions database, and supports mixed-language documents through a letter sign convention. The file reader integrates with Microsoft Word via COM Add-Ins, uses keyboard hooks to intercept keystrokes for audio feedback, and employs the Shruti speech synthesis engine (supporting Hindi and Bengali via diphone concatenation) for real-time text-to-speech. Font rendering required special handling through Uniscribe and OpenType Layout Services for correct glyph substitution and positioning of complex Indian scripts. Testing at the Sparsha Chitra tactile graphics tool showed 40% correct guesses, 35% close guesses, and only 5% wrong guesses for geometric figures. The file reader was deployed and tested at the National Association for the Blind (Delhi), Blind Peoples's Association (Ahmedabad), and National Institute for the Visually Handicapped (Dehradun) through a government-sponsored iterative process. Typing speed comparisons showed visually impaired users achieved approximately 18 words per minute with the Indian language file reader versus about 25 wpm with JAWS for English.
Relevance
Sparsha addresses a massive equity gap in digital accessibility. While assistive technologies for English-speaking blind users were well-established by 2005, the 13 million visually impaired people in India were largely excluded because the underlying technologies assumed Latin-script languages. This work demonstrates that accessible technology development must account for linguistic and script diversity — a challenge that remains highly relevant as digital inclusion efforts expand in the Global South. The technical solutions to Indian script complexity (finite state machines for context-dependent Braille mapping, OpenType glyph substitution, language-specific code tables) provide a framework applicable to other complex scripts. For accessibility practitioners, the paper highlights that internationalization and accessibility are deeply intertwined: a screen reader or Braille system that works only in English is not truly accessible for billions of people worldwide. The iterative deployment process involving blind users at multiple organizations across India also models good practice for user-centered assistive technology development in resource-constrained settings.
Tags: braille · visual impairment · Indian languages · multilingual accessibility · screen reader · text-to-speech · tactile graphics · internationalization · digital inclusion · Global South accessibility
Standards referenced: Bharati Braille · Nemeth Code · ISCII · Unicode