Supervised Grapheme-to-Phoneme Conversion of Orthographic Schwas in Hindi and Punjabi
This addresses a specific bottleneck in Hindi and Punjabi text-to-speech systems, offering an incremental improvement over rule-based methods.
The paper tackles the problem of predicting schwa deletion in Hindi grapheme-to-phoneme conversion by developing the first statistical classifier based solely on orthography, achieving state-of-the-art performance in Hindi and good results in Punjabi without modification.
Hindi grapheme-to-phoneme (G2P) conversion is mostly trivial, with one exception: whether a schwa represented in the orthography is pronounced or unpronounced (deleted). Previous work has attempted to predict schwa deletion in a rule-based fashion using prosodic or phonetic analysis. We present the first statistical schwa deletion classifier for Hindi, which relies solely on the orthography as the input and outperforms previous approaches. We trained our model on a newly-compiled pronunciation lexicon extracted from various online dictionaries. Our best Hindi model achieves state of the art performance, and also achieves good performance on a closely related language, Punjabi, without modification.