Skip to Main content Skip to Navigation
Conference papers

Error detection of grapheme-to-phoneme conversion in text-to-speech synthesis using speech signal and lexical context

Abstract : In unit selection text-to-speech synthesis, voice creation involved a phonemic transcription of read speech. This is produced by an automatic grapheme-to-phoneme conversion of the text read, followed by a manual correction. Although grapheme-to-phoneme conversion makes few errors, the manual correction is time consuming as every generated phoneme should be checked. We propose a method to automatically detect grapheme-to-phoneme conversion errors by comparing contrastives phonemisation hypothesis. A lattice-based forced alignment system is implemented, allowing for signal-dependent phonemisation. We implement also a sequence-to-sequence neural network model to obtain a context-dependent grapheme-to-phoneme conversion. On a French dataset, we show that we can detect to 86.3% of the errors made by a commercial grapheme-to-phoneme system. Moreover, the amount of data annotated as erroneous is kept under 10% of the total evaluation data. The time spent for phoneme manual checking can thus been drastically reduced without decreasing significantly the phonemic transcription quality.
Complete list of metadatas

Cited literature [20 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01585770
Contributor : Yannick Estève <>
Submitted on : Tuesday, September 12, 2017 - 1:19:48 AM
Last modification on : Monday, April 9, 2018 - 4:55:19 PM
Document(s) archivé(s) le : Wednesday, December 13, 2017 - 4:20:41 PM

File

kv_asru17.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Kévin Vythelingum, Yannick Estève, Olivier Rosec. Error detection of grapheme-to-phoneme conversion in text-to-speech synthesis using speech signal and lexical context. 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), Dec 2017, Okinawa, Japan. ⟨10.1109/ASRU.2017.8269004⟩. ⟨hal-01585770⟩

Share

Metrics

Record views

637

Files downloads

624