[Elsnet-list] A position of Speech Synthesis Researcher - Toshiba Research Europe Ltd, Cambridge Research Lab

masa.akamine at toshiba.co.jp masa.akamine at toshiba.co.jp
Wed Sep 21 16:23:39 CEST 2011

A position of Speech Synthesis Researcher
Toshiba Research Europe Ltd, Cambridge Research Lab (CRL)

Toshiba Research Europe Ltd, Cambridge Research Lab (CRL) is seeking a 
research engineer to join our Text-To-Speech (TTS) research team.

CRL has been working on HMM-TTS for several years to improve its prosody 
generation based on parametric models [1][2], to improve excitation models 
[3][4] for better speech quality and to increase speaker and language 
adaptability [5][6]. The research team is very active in leading the 
research area of HMM-TTS.

The successful candidate will lead research work on speaker and language 
adaptation, expressive speech synthesis and unsupervised training of 
expressive speech from audio books in a context of HMM-TTS. The candidate 
will be expected to create innovative ideas on these research topics, 
to implement them into a TTS system and to do evaluation.

The CRL Speech Technology Group (STG), consisting of TTS, ASR and language 
development teams, is a multinational team of dynamic individuals. The team 
carries out mid-to-long term research for Toshiba’s speech recognition and 
synthesis systems and contributes to the development of systems for deploy-
ment in North America and Europe. STG works in collaboration with speech R&D 
groups in Toshiba RDC, Japan, and Toshiba China. Based in the beautiful city 
of Cambridge, England, the team has close links with the University of 
Cambridge. Dr Mark Gales is a part-time Technical Advisor of the group.

Required skills:
- PhD in related subject or equivalent industry expertise
- Good knowledge of HMMs and machine learning
- Good knowledge of HMM-TTS
- Good software skills in C, Python and Linux
- Good English and communication skills

Any of the following are welcomed:
- Experience of speaker adaptation
- Industrial coding experience
- Experience of working in a team
- Knowledge of other major European languages

Applicants should send a CV and a covering letter, stating where you saw 
this advert, to stg-jobs at crl.toshiba.co.uk. Questions or requests for 
further details can be sent to Dr Masami Akamine who is staying in 
Cambridge at masa.akamine at toshiba.co.jp.


Closing date for applications:  21 October 2011 (or until post filled)

[1] J. Latorre and M. Akamine, "Multilevel parametric-base F0 model 
    for speech synthesis." in Proc. Interspeech , pp. 2274-2277, 
    September 2008.
[2] J. Latorre, M. J. F. Gales and H. Zen, "Training a parametric-based 
    logF0 model with the minimum generation error criterion." in Proc. 
    Interspeech, pp. 2174-2177, September 2010.
[3] R. Maia, H. Zen, M. J. F. Gales. "Statistical Parametric Speech Synthesis 
    with Joint Estimation of Acoustic and Excitation Model Parameters,” 
    in Proc. SSW7, pp. 88-93, 2010.
[4] R. Maia, H. Zen, K. Knill, M. J. F. Gales, S. Buchholz. "Multipulse 
    Sequences for Residual Signal Modeling,” in Proc. Interspeech, pp. 1833
    -1836, 2011.
[5] H. Zen. “Speaker and Language Adaptive Training for HMM-Based Polyglot 
    Speech Synthesis,” in Proc. Interspeech, pp. 410-413, 2010.
[6] N. Pilkington, H. Zen, M J. F. Gales. "Gaussian Process Experts for Voice 
   Conversion,” in Proc. Interspeech, pp. 2761-2764, 2011.

More information about the Elsnet-list mailing list