Audio Processing Method, Audio Processing Apparatus, and
Recording Reproduction Apparatus Capable of Outputting Voice
Having Regular Pitch Regardless of Reproduction Speed
Patent No. 6,360,198(US), 3073942(JP), 303913(KR), ZL98801333.9(CN), 2271463(CA)
T
his invention relates to an audio processing method, audio processing apparatus, and recording and reproduction
apparatus that enables sound of a normal pitch to be output regardless of the reproduction speed from a
commercial-use VTR , 6-mm tape recorder, or the like, of which the pitch of output sound would otherwise change
in proportion to the reproduction speed.
The pitch of reproduced sound that has been recorded in an analog recording medium such as magnetic tape normally
changes in proportion to the reproduction speed of the apparatus. Thus, it has been impossible for a VTR to reproduce
sound simultaneously with images that are to be reproduced in slow motion without that sound appearing to have a
non-realistic pitch.
The changeable speed reproduction apparatus 1 (Fig. 1) modifies an audio signal that is to be reproduced at a speed
different from that at the time of recording. It divides the audio data into blocks, each having a prescribed time length,
and if necessary, performs interpolation or thinning thereof, according to a changeable speed ratio r as determined by the
VTR changeable speed reproduction part (2) and the sound attributes. The sampling frequency conversion part (4)
matches the sampling frequency fi (Hz) of A/D conversion to the sampling frequency fo (Hz) of D/A conversion.
Consequently, high-quality sound with no change in pitch thereof is output in synchronization with the timing of the
image signal from part 2.
In particular, A/D conversion of the audio signal is performed using the relation, fi =r fo (Hz), when the sampling
frequencies fi and fo (Hz) satisfy fi /fo =r. When fi /fo r because fi and fo cannot be set to the given values, the audio
signal is converted into audio data whereby part 4 does sampling using a sampling frequency conversion coefficient c=r
fo /fi (Hz). The series of sound-attribute analyses on the audio signal is performed to divide the audio data into blocks,
each having a prescribed time length, and if necessary, the data are interpolated or thinned in units of a block to lengthen
or shorten data by 1/r, After that, D/A conversion of the audio signal at fo. The sound output from this procedure will
have no change in the pitch compared with the signal input to part 2, yet will be synchronized to the timing of the image
signal output from part 2.
2
VTR CHANGEABLE
SPEED
REPRODUCTION
PART
INPUT OF
CHANGED
SPEED
REPRODUCED
SOUND
3
A/D
CONVERSION
PART
TIME DATA OF REPRODUCED IMAGE
ANALYSIS
PROCESSING
PART
CHANGEABLE
SPEED RATIO DATA
4
SAMPLING
FREQUENCY
CONVERSION PART
DIVISION
DATA
6
5
10
RESPECTIVE
BLOCK LENGTH
BLOCK DATA
7
BLOCK DATA
DIVISION
PART
BLOCK DATA
ACCUMULATION
PART
CONNECTION
8 DATA
SOUND-EQUIPPED VTR CHANGEABLE
SPEED REPRODUCTION APPARATUS 1
CONNECTING
DATA
PRODUCTION
PART
9
CONNECTING
DATA
ACCUMULATION
PART
CONNECTION
SEQUENCE
PRODUCTION
PART
PREVIOUS
CONNECTION
DATA
CHANGED
SPEED
REPRODUCED
IMAGE
CONNECTION
SEQUENCE
AUDIO DATA
CONNECTION
PART
11
D/A
CONVERSION
PART
12
OUTPUT
SOUND
Figure 1
22
Broadcast Technology no.25, Winter 2006
C
NHK STRL