A classifier-based target cost for unit selection speech synthesis trained on perceptual data
نویسندگان
چکیده
Our goal is to automatically learn a perceptually-optimal target cost function for a unit selection speech synthesiser. The approach we take here is to train a classifier on human perceptual judgements of synthetic speech. The output of the classifier is used to make a simple three-way distinction rather than to estimate a continuously-valued cost. In order to collect the necessary perceptual data, we synthesised 145,137 short sentences with the usual target cost switched off, so that the search was driven by the join cost only. We then selected the 7200 sentences with the best joins and asked 60 listeners to judge them, providing their ratings for each syllable. From this, we derived a rating for each demiphone. Using as input the same context features employed in our conventional target cost function, we trained a classifier on these human perceptual ratings. We synthesised two sets of test sentences with both our standard target cost and the new target cost based on the classifier. A/B preference tests showed that the classifier-based target cost, which was learned completely automatically from modest amounts of perceptual data, is almost as good as our carefullyand expertly-tuned standard target cost.
منابع مشابه
Study on Unit-Selection and Statistical Parametric Speech Synthesis Techniques
One of the interesting topics on multimedia domain is concerned with empowering computer in order to speech production. Speech synthesis is granting human abilities to the computer for speech production. Data-based approach and process-based approach are the two main approaches on speech synthesis. Each approach has its varied challenges. Unit-selection speech synthesis and statistical parametr...
متن کاملDesigning Target Cost Function Based on Prosody of Speech Database
This research aims to construct a high-quality Japanese TTS (Text-to-Speech) system that has high flexibility in treating prosody. Many TTS systems have implemented a prosody control system but such systems have been fundamentally designed to output speech with a standard pitch and speech rate. In this study, we employ a unit selectionconcatenation method and also introduce an analysis-synthesi...
متن کاملAutomatic feature selection for acoustic-visual concatenative speech synthesis: towards a perceptual objective measure
We present an iterative algorithm for automatic feature selection and weight tuning of target cost in the context of unit selection based audio-visual speech synthesis. We perform feature selection and weight tuning for a given unit-selection corpus to make the ranking given by the target cost function consistent with the ordering given by an objective dissimilarity measure. We explicitly perfo...
متن کاملMaximum likelihood unit selection for corpus-based speech synthesis
Corpus-based speech synthesis systems deliver a considerable synthesis quality since the unit selection approaches have been optimized in the last decade. Unit selection attempts to find the best combination of speech unit sequences in an inventory so that the perceptual differences between expected (natural) and synthesized signals are as low as possible. However, mismatches and distortions ar...
متن کاملThe VUB Blizzard Challenge 2009 Entry
In this paper we describe the voices we submitted to the 2009 Blizzard Challenge, a yearly challenge to evaluate auditory speech synthesis on common data. Since it is the second time we participate in this challenge, in this paper we focus on the changes we made to our unit selection-based system. The weighted sum of symbolic target costs has been replaced by a single statistical target cost; t...
متن کامل