Phonetic Transcription using Speech Recognition Technique Considering Variations in Pronunciation

We propose a new approach for performing phonetic transcription of speech and text that combines automatic speech recognition (ASR) and grapheme-to-phoneme (G2P) techniques. By augmenting the text with speech and using automatic speech recognition with a sausage searching net constructed from multiple text pronunciations corresponding to human speech utterance, we are able to reduce the effort for phonetic transcription. By using a multiple pronunciation lexicon, a transcription error rate of 12.74% was achieved. Further improvement can be achieved by adapting the pronunciation lexicon with pronunciation variation (PV) rules and an error rate reduction of 17.11 % could be achieved.