Corpus design for a unit selection TtS system with application to Bulgarian

In this paper we present the process of designing an efficient speech corpus for the first unit selection speech synthesis system for Bulgarian, along with some significant preliminary results regarding the quality of the resulted system. As the initial corpus is a crucial factor for the quality delivered by the Text-to-Speech system, special effort has been given in designing a complete and efficient corpus for use in a unit selection TTS system. The targeted domain of the TTS system and hence that of the corpus is the news reports, and although it is a restricted one, it is characterized by an unlimited vocabulary. The paper focuses on issues regarding the design of an optimal corpus for such a framework and the ideas on which our approach was based on. A novel multi-stage approach is presented, with special attention given to language and speaker dependent issues, as they affect the entire process. The paper concludes with the presentation of our results and the evaluation experiments, which provide clear evidence of the quality level achieved. © 2011 Springer-Verlag.

Statistical parametric speech synthesis based on speaker and language factorization

An increasingly common scenario in building speech synthesis and recognition systems is training on inhomogeneous data. This paper proposes a new framework for estimating hidden Markov models on data containing both multiple speakers and multiple languages. The proposed framework, speaker and language factorization, attempts to factorize speaker-/language-specific characteristics in the data and then model them using separate transforms. Language-specific factors in the data are represented by transforms based on cluster mean interpolation with cluster-dependent decision trees. Acoustic variations caused by speaker characteristics are handled by transforms based on constrained maximum-likelihood linear regression. Experimental results on statistical parametric speech synthesis show that the proposed framework enables data from multiple speakers in different languages to be used to: train a synthesis system; synthesize speech in a language using speaker characteristics estimated in a different language; and adapt to a new language. © 2012 IEEE.

Improving LVCSR system combination using neural network language model cross adaptation

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple subsystems developed at different sites. Cross system adaptation can be used as an alternative to direct hypothesis level combination schemes such as ROVER. The standard approach involves only cross adapting acoustic models. To fully exploit the complimentary features among sub-systems, language model (LM) cross adaptation techniques can be used. Previous research on multi-level n-gram LM cross adaptation is extended to further include the cross adaptation of neural network LMs in this paper. Using this improved LM cross adaptation framework, significant error rate gains of 4.0%-7.1% relative were obtained over acoustic model only cross adaptation when combining a range of Chinese LVCSR sub-systems used in the 2010 and 2011 DARPA GALE evaluations. Copyright © 2011 ISCA.

Transcription of multi-genre media archives using out-of-domain data

We describe our work on developing a speech recognition system for multi-genre media archives. The high diversity of the data makes this a challenging recognition task, which may benefit from systems trained on a combination of in-domain and out-of-domain data. Working with tandem HMMs, we present Multi-level Adaptive Networks (MLAN), a novel technique for incorporating information from out-of-domain posterior features using deep neural networks. We show that it provides a substantial reduction in WER over other systems, with relative WER reductions of 15% over a PLP baseline, 9% over in-domain tandem features and 8% over the best out-of-domain tandem features. © 2012 IEEE.

Language model cross adaptation for LVCSR system combination

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple sub-systems that may even be developed at different sites. Cross system adaptation, in which model adaptation is performed using the outputs from another sub-system, can be used as an alternative to hypothesis level combination schemes such as ROVER. Normally cross adaptation is only performed on the acoustic models. However, there are many other levels in LVCSR systems' modelling hierarchy where complimentary features may be exploited, for example, the sub-word and the word level, to further improve cross adaptation based system combination. It is thus interesting to also cross adapt language models (LMs) to capture these additional useful features. In this paper cross adaptation is applied to three forms of language models, a multi-level LM that models both syllable and word sequences, a word level neural network LM, and the linear combination of the two. Significant error rate reductions of 4.0-7.1% relative were obtained over ROVER and acoustic model only cross adaptation when combining a range of Chinese LVCSR sub-systems used in the 2010 and 2011 DARPA GALE evaluations. © 2012 Elsevier Ltd. All rights reserved.