Title:
MODEL GENERATION METHOD, SPEECH SYNTHESIS METHOD AND APPARATUS
Document Type and Number:
WIPO Patent Application WO/2018/072543
Kind Code:
A1
Abstract:
The present invention relates to the technical field of speech synthesis and discloses a model generation method, a speech synthesis method and an apparatus. The model generation method comprises: acquiring training speech data (202); extracting, from the training speech data, a training speech segment having a first tag type (204); calculating an average difference matrix according to an adjacent candidate speech segment corresponding to the training speech segment having the first tag type before splicing (206); generating, according to the average difference matrix, a splicing cost model having a target splicing weight (208); and performing speech synthesis by means of the splicing cost model having the target splicing weight to obtain synthetic speech information (210). In this way, a splicing cost model having a target splicing weight is generated according to an average difference matrix, avoiding multiple manual adjustments of a weight in a splicing cost model and final generation of a still insufficiently accurate weight, thereby achieving a reduction in the number of manual adjustments and calculation of a more precise target splicing weight directly by an average difference matrix.
Inventors:
YUAN HAOLEI (CN)
WU FUZHANG (CN)
QIAN BINGHUA (CN)
WU FUZHANG (CN)
QIAN BINGHUA (CN)
Application Number:
PCT/CN2017/097314
Publication Date:
April 26, 2018
Filing Date:
August 14, 2017
Export Citation:
Assignee:
TENCENT TECH SHENZHEN CO LTD (CN)
International Classes:
G10L13/07
Foreign References:
| CN106356052A | 2017-01-25 | |||
| CN104575488A | 2015-04-29 | |||
| CN101131818A | 2008-02-27 | |||
| US6366883B1 | 2002-04-02 | |||
| CN104112444A | 2014-10-22 |
Other References:
SYRDAL, ANN K.: "DATA-DRIVEN PERCEPTUALLY BASED JOIN COSTS", 5TH ISCA SPEECH SYNTHESIS WORKSHOP, 14 June 2004 (2004-06-14) - 16 June 2004 (2004-06-16), Pittsburg, PA, USA, pages 49 - 54, XP055603262
SYRDAL, ANN K.: "Perceptually-based Data-driven Join Costs: Comparing Join Types", INTERSPEECH 2005, 31 December 2005 (2005-12-31), Lisbon, Portugal, pages 2813 - 2816, XP055603269
SYRDAL, ANN K.: "Perceptually-based Data-driven Join Costs: Comparing Join Types", INTERSPEECH 2005, 31 December 2005 (2005-12-31), Lisbon, Portugal, pages 2813 - 2816, XP055603269
Attorney, Agent or Firm:
BEIJING SAN GAO YONG XIN INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF:
Previous Patent: FEEDING DEVICE FOR CRUCIBLE
Next Patent: INFORMATION TRANSMISSION METHOD AND STATION
Next Patent: INFORMATION TRANSMISSION METHOD AND STATION