Title:
AUDIO SYNTHESIZING METHOD, AUDIO SYNTHESIZING DEVICE, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2019/107379
Kind Code:
A1
Abstract:
This audio synthesizing device is provided with: a first learned model for generating a frequency spectrum time-sequence of a harmonic component which is in accordance with control data containing phoneme designation; a second learned model for generating a waveform signal of an anharmonic component which is in accordance with the control data; and a synthesis processing unit for generating an audio signal containing the harmonic component and the anharmonic component from the frequency spectrum time-sequence and the waveform signal.
More Like This:
Inventors:
DAIDO RYUNOSUKE (JP)
SHIMIZU MASAHIRO (JP)
SHIMIZU MASAHIRO (JP)
Application Number:
PCT/JP2018/043656
Publication Date:
June 06, 2019
Filing Date:
November 28, 2018
Export Citation:
Assignee:
YAMAHA CORP (JP)
International Classes:
G10L13/06; G10L13/04
Foreign References:
| US20130262098A1 | 2013-10-03 | |||
| JP2006084967A | 2006-03-30 |
Other References:
BLAAUW MERLIJN ET AL.: "A NEURAL PARAMETRIC SINGING SYNTHESIZER", 17 August 2017 (2017-08-17), XP055627665, Retrieved from the Internet [retrieved on 20190205]
TACHIBANA KENTARO: "nonofficial translation: "Statistical parametric speech synthesis system free to determine V/UV based on separation into periodicity and aperiodicity"", LECTURE PROCEEDINGS OF 2017 SPRING RESEARCH CONFERENCE OF THE ACOUSTICAL SOCIETY OF JAPAN, pages 209 - 212
NISHIMURA, MASANARI: "Review on singing-voice synthesis based on DEEP NEURAL NETWORK", LECTURE PROCEEDINGS OF 2016 SPRING RESEARCH CONFERENCE OF THE ACOUSTICAL SOCIETY OF JAPAN, pages 213 - 214
MERLIJN BLAAUWJORDI BONADA: "A NEURAL PARAMETRIC SINGING SYNTHESIZER", ARXIV, 12 April 2017 (2017-04-12)
OORD ET AL.: "WaveNet: A Generative Model for Raw Audio", ARXIV, 19 September 2016 (2016-09-19)
See also references of EP 3719795A4
TACHIBANA KENTARO: "nonofficial translation: "Statistical parametric speech synthesis system free to determine V/UV based on separation into periodicity and aperiodicity"", LECTURE PROCEEDINGS OF 2017 SPRING RESEARCH CONFERENCE OF THE ACOUSTICAL SOCIETY OF JAPAN, pages 209 - 212
NISHIMURA, MASANARI: "Review on singing-voice synthesis based on DEEP NEURAL NETWORK", LECTURE PROCEEDINGS OF 2016 SPRING RESEARCH CONFERENCE OF THE ACOUSTICAL SOCIETY OF JAPAN, pages 213 - 214
MERLIJN BLAAUWJORDI BONADA: "A NEURAL PARAMETRIC SINGING SYNTHESIZER", ARXIV, 12 April 2017 (2017-04-12)
OORD ET AL.: "WaveNet: A Generative Model for Raw Audio", ARXIV, 19 September 2016 (2016-09-19)
See also references of EP 3719795A4
Attorney, Agent or Firm:
OHBAYASHI, Akira et al. (JP)
Download PDF: