Title:
IMAGING DEVICE AND VOICE SIGNAL PROCESSING DEVICE
Document Type and Number:
Japanese Patent JP2015050530
Kind Code:
A
Abstract:
PROBLEM TO BE SOLVED: To separate voice from a main subject and voice from a photographer which are captured with a single stereo microphone.SOLUTION: A FFT (207) performs fast Fourier transform on a voice signal from a stereo microphone (205A, 205B). A voice band extraction filter (208) extracts a band component of human voice. A subtracter (213) calculates difference between L and R channels; an adder (214) calculates a summation of them. A level detection unit (215) detects output levels of the subtracter (213) and adder (214); a determination unit (216) determines the propriety of a correction unit (209)'s correction processing on the basis of output of the level detection unit (215). The correction unit (209) subtracts an amount corresponding to the difference between L and R channels from output of each channel of the FFT (207). An IFFT (210) performs inverse Fourier transform on output of the correction unit (209). An IFFT (219) performs inverse Fourier transform on output of the subtracter (213).
More Like This:
Inventors:
SONODA KEITA
Application Number:
JP2013179437A
Publication Date:
March 16, 2015
Filing Date:
August 30, 2013
Export Citation:
Assignee:
CANON KK
International Classes:
H04R3/00; H04N5/225; H04N5/91; H04R1/40
Attorney, Agent or Firm:
Tsuneo Tanaka