Speech synthesis using stochastic Markov graphs

Eichner, Matthias; Wolff, Matthias; Ohnewald, S.; Hoffmann, Rüdiger

doi:10.1109/icassp.2001.941043

Cited by 8 publications

(5 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This enables an inverse function of the recognizer system, offering the potential of applying analysis-by-synthesis approach to improve performance of speech recognition systems. Synthesis experiments, as well as recognition experiments, confirmed the suitability of this model for both tasks [16]. The system is used to create artificial adaptation data, and later itself is adapted on the transformed data (coding, noise etc.)…”

Section: Introductionmentioning

confidence: 58%

“…Baseline performance evaluation of the UASR system [16] was performed initially, in order to assess the effects of speech and audio compression on recognition performance. For training and evaluation, the natural speech data subset VM2 of Verbmobil German Database was used [19].…”

Section: Evaluation Resultsmentioning

confidence: 99%

“…The structure is built iteratively during the training process by splitting the Gaussian densities from the initial HMM model [16]. Applying these topologies as acoustic models in a speech synthesis process requires finding of optimal state sequence within a phoneme model in a given context.…”

Section: A Unified Asr and Synthesismentioning

confidence: 99%

“…x Selecting a path through the HMMs according to a demanded length and modeling the duration of each state in the path [16]; x Assembling the feature sequence for the chosen path;…”

Section: A Unified Asr and Synthesismentioning

confidence: 99%

See 3 more Smart Citations

Analysis-by-synthesis approach for acoustic model adaptation

et al. 2013

View full text Add to dashboard Cite

This paper presents an analysis-by-synthesis approach for acoustic model adaptation. Using artificial speech data for speech recognition systems adaptation, has the potential to address the problem of data sparseness, to avoid speech recordings in real conditions and to provide the capability of performing large number of development cycles for Automatic Speech Recognition (ASR) systems in shorter time. The proposed adaptation framework uses unified ASR and synthesis system to produce artificial adaptation speech signals. In order to confirm the usability of the proposed approach, several experiments were performed where the artificial speech data was coded-decoded by different speech and waveform coders and the acoustic model used for synthesis was adapted for each coder.The recognition results show that the proposed method could be used successfully in the process of speech recognition systems performance assessment and improvement, not only for coded speech effects evaluation and adaptation, but also for other environment conditions.

show abstract

Section: Introductionmentioning

confidence: 58%

Section: Evaluation Resultsmentioning

confidence: 99%

Section: A Unified Asr and Synthesismentioning

confidence: 99%

“…x Selecting a path through the HMMs according to a demanded length and modeling the duration of each state in the path [16]; x Assembling the feature sequence for the chosen path;…”

Section: A Unified Asr and Synthesismentioning

confidence: 99%

See 2 more Smart Citations

Analysis-by-synthesis approach for acoustic model adaptation

et al. 2013

View full text Add to dashboard Cite

show abstract

“…Related to this topic, Eichner et al applied stochastic Markov graphs, which have enhanced capabilities for modeling trajectories, to statistical parametric synthesis (Eichner et al, 2000). Although this offers a flexible topology, it requires a search process for the state sequence at the synthesis stage (Eichner et al, 2001) because we need to determine a single-state sequence to generate speech parameters efficiently using the speech parameter generation algorithm (the Case 1 algorithm in (Tokuda et al, 2000)). Although we can skip this process by marginalizing all possible state sequences using an EM-type parameter generation algorithm (the Case 3 algorithm in (Tokuda et al, 2000)), this further increases computational complexity.…”

Section: Model Topologymentioning

confidence: 99%

Statistical Parametric Speech Synthesis

Black

Zen

Tokuda

2007

2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07

211

View full text Add to dashboard Cite

This review gives a general overview of techniques used in statistical parametric speech synthesis. One instance of these techniques, called hidden Markov model (HMM)-based speech synthesis, has recently been demonstrated to be very effective in synthesizing acceptable speech. This review also contrasts these techniques with the more conventional technique of unit-selection synthesis that has dominated speech synthesis over the last decade. The advantages and drawbacks of statistical parametric synthesis are highlighted and we identify where we expect key developments to appear in the immediate future.

show abstract

Analysis of Verbal and Nonverbal Acoustic Signals with the Dresden UASR System

Hoffmann

Eichner

Wolff

Verbal and Nonverbal Communication Behaviours

View full text Add to dashboard Cite

Speech synthesis using stochastic Markov graphs

Cited by 8 publications

References 8 publications

Analysis-by-synthesis approach for acoustic model adaptation

Analysis-by-synthesis approach for acoustic model adaptation

Statistical Parametric Speech Synthesis

Analysis of Verbal and Nonverbal Acoustic Signals with the Dresden UASR System

Contact Info

Product

Resources

About