Deep Neural Network-Based Statistical Parametric Speech Synthesis System Using Improved Time-Frequency Trajectory Excitation Mo

International Conference
2015-09-01 00:47
Authors : Eunwoo Song, Hong-Goo Kang

Year : 2015

Publisher / Conference : INTERSPEECH

This paper proposes a deep neural network (DNN)-based statistical parametric speech synthesis system using an improved time-frequency trajectory excitation (ITFTE) model. The ITFTE model, which efficiently reduces the parametric redundancy of a TFTE model, improved the perceptual quality of the vocoding process and the estimation accuracy of the training process. However, there remain problems related to training ITFTE parameters in a hidden Markov model (HMM) framework, such as inefficiency of representing cross-dimensional correlations between ITFTE parameters, over-smoothed outputs caused by statistical averaging, and an over-fitted model due to a decision tree-based state clustering paradigm. To alleviate these limitations, a centralized DNN replaces the decision trees of the HMM training process. Analysis of trainability confirms that the DNN training process improves the model accuracy, which results in improved perceptual quality of synthesized speech. Objective and subjective test results also verify that the proposed system performs better than the conventional HMM-based system.
전체 319
239 International Conference Kyungguen Byun, Eunwoo Song, Hong-goo Kang "A constrained two-layer compression technique for ECG waves" in Enegineering in Medicine and Biology Society (EMBC), 2015
238 International Conference Eunwoo Song, Hong-Goo Kang "Deep Neural Network-Based Statistical Parametric Speech Synthesis System Using Improved Time-Frequency Trajectory Excitation Mo" in INTERSPEECH, 2015
237 International Journal Taegyu Lee, Hyun Oh Oh, Jeongil Seo, Young-Cheol Park, Dae Hee Youn "Scalable Multiband Binaural Renderer for MPEG-H 3D Audio" in IEEE Journal of Selected Topics in Signal Processing, vol.9, issue 5, pp.907-920, 2015
236 International Conference Heejin Ahn, Eunwoo Song, Won-Suk Jun, Hong-goo Kang "A Compression Algorithms for Hidden Markov Model-Based Speech Synthesis Systems" in ITC-CSCC, pp.942-945, 2015
235 International Conference JeeSok Lee, Sejin Oh, Hong-Goo Kang "Coherent channel based subband multichannel dereverberation" in ICASSP, pp.2704-2708, 2015
234 International Conference Eunwoo Song, Young-Sun Joo, Hong-Goo Kang "Improved time-frequency trajectory excitation modeling for a statistical parametric speech synthesis system" in ICASSP, 2015
233 Domestic Journal 박영철, 이태규, 윤대희 "MPEG-H 3D 오디오 바이노럴 렌더링 기술 표준화" in 대한전기학회, 전기의 세계, vol.64, 제 2호, pp.27-31, 2015
232 International Journal Taegyu Lee, Yonghyun Baek, Young-Cheol Park, Dae Hee Youn "Stereo upmix-based binaural auralization for mobile devices" in IEEE Transactions on Consumer Electronics, vol.60, issue 3, pp.411-419, 2014
231 International Conference Eunwoo Song, Hong-Goo Kang, Joonil Lee "Fixed-point implementation of MPEG-D unified speech and audio coding decoder" in 19th International Conference on Digital Signal Processing (DSP), pp.110-113, 2014
230 International Journal Soonho Baek, Hong-Goo Kang "Selection of spectral compressive operator for vector Taylor series-based model adaptation in noisy environments" in The Journal of the Acoustical Society of America, vol.135, 2014