A Phonetic-Level Analysis of Different Input Features for Articulatory Inversion

Abdolreza Sabzi Shahrebabaki, Negar Olfati, Ali Shariq Imran, Sabato Marco Siniscalchi, Torbjørn Svendsen

The challenge of articulatory inversion is to determine the temporal movement of the articulators from the speech waveform, or from acoustic-phonetic knowledge, e.g. derived from information about the linguistic content of the utterance. The actual position of the articulators is typically obtained from measured data, in our case position measurements obtained using EMA (Electromagnetic articulography). In this paper, we investigate the impact on articulatory inversion problem by using features derived from the acoustic waveform relative to using linguistic features related to the time aligned phone sequence of the utterance. Filterbank energies (FBE) are used as acoustic features, while phoneme identities and (binary) phonetic attributes are used as linguistic features. Experiments are performed on a speech corpus with synchronously recorded EMA measurements and employing a bidirectional long short-term memory (BLSTM) that estimates the articulators’ position. Acoustic FBE features performed better for vowel sounds. Phonetic features attained better results for nasal and fricative sounds except for /h/. Further improvements were obtained by combining FBE and linguistic features, which led to an average relative RMSE reduction of 9.8%, and a 3% relative improvement of the Pearson correlation coefficient.

 DOI: 10.21437/Interspeech.2019-2526

Cite as: Shahrebabaki, A.S., Olfati, N., Imran, A.S., Siniscalchi, S.M., Svendsen, T. (2019) A Phonetic-Level Analysis of Different Input Features for Articulatory Inversion. Proc. Interspeech 2019, 3775-3779, DOI: 10.21437/Interspeech.2019-2526.

  author={Abdolreza Sabzi Shahrebabaki and Negar Olfati and Ali Shariq Imran and Sabato Marco Siniscalchi and Torbjørn Svendsen},
  title={{A Phonetic-Level Analysis of Different Input Features for Articulatory Inversion}},
  booktitle={Proc. Interspeech 2019},