Voice converting system and method using blind voice separation

A voice conversion and voice separation technology, which is applied in voice analysis, voice recognition, instruments, etc., can solve the problems of large amount of calculation in the separation process, cannot meet the real-time requirements of the voice conversion system, and high algorithm complexity, so as to optimize the learning process , optimization of separation results, and the effect of overcoming dependencies

Inactive Publication Date: 2012-07-18
BEIJING JIAOTONG UNIV
View PDF4 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the current popular ICA-based blind signal source separation method has high algorithm complexity and a large amo

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Voice converting system and method using blind voice separation
  • Voice converting system and method using blind voice separation
  • Voice converting system and method using blind voice separation

Examples

Experimental program
Comparison scheme
Effect test

Example Embodiment

[0037] Combine below figure 1 The specific steps of the speech conversion synthesis method of the present invention are described in detail.

[0038] Step 1: Use subband decomposition to obtain two-channel data

[0039] The voice of the source speaker or the target speaker required by the voice conversion system is usually obtained by a microphone. The voice of the noisy speaker collected by the microphone can only constitute a single-channel signal, while independent component analysis (ICA) requires an observation signal x i (t) (ie the mixed signal of background noise and desired speech) is greater than or equal to the number of source signals (ie each independent source signal before mixing), so the collected voice signal is divided into two parts using the existing subband decomposition technology : Low frequency part and high frequency part.

[0040] The speech obtained by the low frequency part conversion is used as the first observation signal of ICA x 1 (t), in the same way...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a voice converting system and method using blind voice separation. The voice converting method comprises the steps of: applying a blind voice separation method which combines sub-band decomposition with ICA (Independent Component Analysis) in a voice input module at the front end of the voice converting system to separate a mixed signal including background noise and expected voice, and converting the expected voice obtained through separation and subjected to an unequal-length framing process to realize conversion of specific voice of people in a noise environment.

Description

technical field [0001] The invention relates to speech conversion, analysis and signal processing, in particular to a speech conversion system and method using blind speech separation. Background technique [0002] Speech conversion is to change a speaker's voice to make it sound like another person's voice, that is, to realize the conversion of timbre characteristics from a specific source speaker to a target speaker. The voice conversion system needs to perform two voice inputs. For the first time, it needs to collect a certain amount of source speaker's voice and target speaker's voice to establish two parallel voice databases, and then perform feature parameter extraction and Training, establish a source-to-target conversion function; the second time you need to input the source speaker's voice with any content into the conversion function, the conversion system can output the target speaker's voice with the same content. Speech conversion is a relatively new branch in ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G10L21/02G10L15/14G10L11/06G10L19/04G10L21/0224
Inventor 申艳汶跃龙张嘉驰范礼乾杨柳蒋诗慧
Owner BEIJING JIAOTONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap