VOICE SIGNAL SEPARATION METHOD AND APPARATUS, COMPUTER DEVICE AND STORAGE MEDIUM

Title:

VOICE SIGNAL SEPARATION METHOD AND APPARATUS, COMPUTER DEVICE AND STORAGE MEDIUM

Document Type and Number:

WIPO Patent Application WO/2020/015270

Kind Code:

A1

Abstract:

A voice signal separation method and apparatus, a computer device and a storage medium, which relate to the field of voice signal processing. The method comprises: sampling an acoustic waveform of an audio file to be separated to obtain an audio signal (201); converting the audio signal from a time domain to a frequency domain to obtain a frequency spectrum of the audio signal, wherein the frequency spectrum is only used to represent the amplitude of the audio signal, and the amplitude is a real number; decomposing the frequency spectrum of the audio signal to obtain an accompaniment frequency spectrum and a vocal frequency spectrum; converting the accompaniment frequency spectrum and the vocal frequency spectrum from the frequency domain to the time domain to obtain an accompaniment audio and a vocal audio (206). By using a transformation algorithm that uses only real numbers to represent the amplitude of an audio frame during conversion, transformations from the time domain to the frequency domain and from the frequency domain to the time domain are carried out. Since the phase will not be transformed before and after the transformation, phase information is not lost. Therefore, on the basis of the foregoing conversion method, the accompaniment and vocals are separated from within an audio file, thereby avoiding the problem of phase distortion of Fourier transform spectrum decomposition.

More Like This:

JP5508550	Devices for generating extended downmix signals, methods and computer programs for generating extended downmix signals
JP6057368	A recording medium which recorded a signal processing method, a device, a program, and a program and in which computer reading is possible
WO/2023/101343	METHOD AND APPARATUS FOR PERFORMING SPEAKER DIARIZATION ON MIXED-BANDWIDTH SPEECH SIGNALS

Inventors:

ZHANG CHAOGANG (CN)

Application Number:

PCT/CN2018/118293

Publication Date:

January 23, 2020

Filing Date:

November 29, 2018

Export Citation:

Click for automatic bibliography generation Help

Assignee:

GUANGZHOU KUGOU COMPUTER TECH CO LTD (CN)

International Classes:

G10L21/0272

Foreign References:

CN106024005A	2016-10-12
CN104134444A	2014-11-05
CN104078051A	2014-10-01
CN103943113A	2014-07-23
US20130121511A1	2013-05-16

Attorney, Agent or Firm:

BEIJING SAN GAO YONG XIN INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)

Download PDF:

View/Download PDF PDF Help

Previous Patent: ANNOTATION DISPLAY METHOD AND APPARATUS, AND INTELLIGENT INTERACTIVE TABLET

Next Patent: ELECTROMAGNETIC FILM MOUNTING STRUCTURE, BACKLIGHT DISPLAY MODULE AND DISPLAY DEVICE