Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02301995 2000-03-29
HIGH QUALITY SPEECH CODER AT LOW BIT RATES
BACKGROUND OF THE INVENTION
This application is a division of Canadian
Patent Application Serial Number 2,213,909 filed on
August 25, 1997. The claims of this application are
directed to a high quality speech coder at low bit
rates in order to satisfy unity of invention
requirements of the Canadian Patent Office. However,
in order to assist the reader to readily understand
the overall invention including all features which are
inextricably bound up in one and the same inventive
concept, the teachings of those features and the broad
objects relating thereto, are all retained in the
present disclosure as they were in the disclosure of
the parent application aforesaid but the claims are
limited to the features identified hereinbefore.
The present invention relates to a speech
coder for high quality coding speech signals at low
bit rates.
Systems for high quality coding speech signals
are well known in the art, as described in, for
instance, W. Schroeder and B. Atal., "Code-Excited
Linear Prediction: High Quality Speech at Very Low Bit
Rates", Proc. ICASSP, pp. 937-940, 1985 (Literature
1), and Kleijn et al., "Improved Speech Quality and
Effective Vector Quantization in SELP:, Proc. ICASSP,
pp. 155-158, 1988 (Literature 2). In these prior art
systems, on the transmitting side spectral parameters
1
CA 02301995 2000-03-29
representing a spectral characteristic of a speech
signal is extracted from the speech signal for each
frame (of 20 ms, for instance) by using linear
prediction (LPC). The frame is split into a plurality
of sub-frames (of 5 ms, for instance), and adaptive
codebook parameters (i.e., a delay parameter
corresponding to the pitch period and a gain
parameter) are extracted for each sub-frame on the
basis of a past excitation signal. The sub-frame
speech signal is then pitch predicted using the
adaptive codebook. The pitch predicted excitation
signal is quantized by selecting an optimum excitation
vector from an excitation codebook (or vector
quantization codebook),which
la
CA 02301995 2000-03-29
consists of predetermined different types of noise
signals, and computing an optimum gain. The optimum
excitation code vector is selected such that error
power between a synthesized signal from selected
noise signals and an error signal is minimized. A
multiplexer combines an index representing the type
of the selected codevector and a gain, the spectral
parameters, and the adaptive codebook parameters,
and transmits the multiplexed data to the receiving
side for de-multiplexing.
The above prior art process has a problem that
the selection of the optimum excitation codevector
from the excitation codebook requires a great deal
of computation. This is so because in the methods
shown in Literatures 1 and 2 the optimum excitation
codevector is selected by making filtering or
convolution with respect to each of a plurality of
codevectors stored in the codebook, that is,
executing the filtering or convolution iteratedly a
number of times corresponding to the number of the
stored codevectors. With bit number of B and degree
of N of a codebook,,for instance, the filtering or
convolution should be executed N x K x 2B x 8000/N
times per second, where K is the filtering or
impulse response length in the filtering or
convolution. With B = 10, N = 40 and K = 40, for
instance, the necessary computational effort is
81,920,000 times per second, which is very enormous
2
CA 02301995 2004-05-25
indeed.
To reduce the computational effort that is
necessary for the excitation codebook retrieval,
various systems have been proposed. Among the proposed
systems is an ACELP (Algebraic Code Excited Linear
Prediction system, which is described in, for
instance, C. Laflamme et al., "16 kbps Wide-Band
Speech Coding Technique Based on Algebraic Celp",
Proc. ICASSP, pp. 13-16, 1991 (Literature 3)). In
this system, an excitation signal is represented by a
plurality of pulses, and the position of each pulse is
represented by a predetermined number of bits that are
transmitted. Since the amplitude of each pulse is
either "+1.0" or "-1.0", the computational effort for
the pulse retrieval can be greatly reduced.
This prior art system described in Literature
3, however, has a problem that the sound quality is
not sufficient, although it is possible to obtain
great reduction of the computational effort. This is
attributable to the fact that each pulse always has
the absolute amplitude of "1.0" irrespective of its
position and has only either positive or negative
polarity. This means that very coarse amplitude
quantization is made, and therefore the sound quality
is deteriorated.
Moreover, in the systems described in
Literatures 1 to 3, the retrieval of the excitation
3
CA 02301995 2004-05-25
codebook or pulses is executed under the assumption
that the speech signal is multiplied by a fixed gain.
Therefore, the performance is deteriorated in the case
where the excitation codebook size is reduced by
reducing the bit rate or where the number of pulses is
small.
SUMMARY OF THE INVENTION
An object of the present invention is therefore to
provide a speech coding system, which can solve the
above problems and is less subject to sound quality
deterioration with relatively less computational
effort even at a low bit rate.
According to an aspect of the present
invention, there is provided a speech coder comprising
a spectral parameter computer for obtaining a
plurality of spectral parameters from an input speech
signal, and quantizing the spectral parameters thus
obtained, and an excitation quantizer for retrieving
the positions of M non-zero amplitude pulses together
constituting an excitation with different gains for
multiplification for each group of pulses less in
number than M. The excitation quantizer includes a
codebook for jointly quantizing the amplitudes or
polarities of a plurality of pulses.
According to another aspect of the present
invention, there is provided a speech coder comprising
a spectral parameter computer for
4
CA 02301995 2004-05-25
obtaining a plurality of spectral parameters from an
input speech signal, and quantizing the spectral
parameters thus obtained, an excitation quantizer for
retrieving positions of M non-zero amplitude pulses
which constitutes an excitation signal of the input
speech signal with a different gain for each group of
the pulses less in number than M, and a second
excitation quantizer for retrieving the positions of a
predetermined number of pulses by using the spectral
parameters, the outputs of the first and second
l0 excitation quantizers being used to compute
distortions of the speech so as to select the less
distortion one of the first and second excitation
quantizers. The excitation quantizer includes a
codebook for jointly quantizing the amplitudes or
polarities of a plurality of pulses. The speech coder
further comprises a mode judging circuit for obtaining
a feature quantity from the input speech signal,
judging one of a plurality of different modes from the
obtained feature quantity and outputting mode data,
the first and second excitation quantizers being
switched according to the mode data.
According to another aspect of the present
invention, there is provided a speech coder comprising
a spectral parameter computer for obtaining spectral
parameters from an input speech signal and quantizing
the spectral parameters thus obtained, an impulse
5
CA 02301995 2004-05-25
response computer for computing impulse responses
corresponding to the spectral parameters, a first
correlation computer for computing correlations of the
input speech signal and the impulse response, a second
correlation computer for computing correlations among
the impulse responses, a first pulse data computer for
computing positions of first pulses from the outputs
of the first and second correlation computers, a third
correlation computer for correcting the output of the
first correlation computer by using the output of the
1o first pulse data computer, and a second pulse data
computer for computing positions of second pulses from
the outputs of the third and second correlation
computers, the pulse data computation being made by
executing the correlation correction and the pulse
data computation in an iterated manner a predetermined
number of times.
According to a further aspect of the present
invention, there is provided a speech coder comprising
a spectral parameter computer for obtaining a
plurality of spectral parameters from an input speech
signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
quantizer for forming an excitation signal of the
6
CA 02301995 2004-05-25
input speech signal with M non-zero amplitude pulses,
obtaining a sample position corresponding to a pulse
position meeting a predetermined condition with
respect to the computed pitch prediction signal,
setting a pulse position retrieval range on the basis
of a position obtained by shifting the obtained sample
position by a predetermined number of samples,
extracting the polarity of a cross-correlation
calculated between input speech signal and impulse
response obtained from said spectral parameter
computer, for each position candidate in the retrieval
range, retrieving a best position in the pulse
position retrieval range thus set, and outputting data
of the retrieved best position.
According to a still further aspect of the
present invention, there is provided a speech coder
comprising a spectral parameter computer for obtaining
a plurality of spectral parameters from an input
speech signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
quantizer for forming an excitation signal of the
input speech signal with M non-zero amplitude pulses,
obtaining a sample position meeting a predetermined
condition with respect to the pitch prediction signal
7
CA 02301995 2004-05-25
in a time interval equal to the pitch period from the
forefront of a frame, setting a pulse position
retrieval range for retrieving a pulse position on the
basis of a position obtained by shifting the obtained
sample position by a predetermined number of samples,
extracting the polarity of a cross-correlation
calculated between input speech signal and impulse
response obtained from said spectral parameter
computer, for each position candidate in the retrieval
range, retrieving a best position in the pulse
position retrieval range thus set, and outputting data
of the retrieved best position.
According to a still further aspect of the
present invention, there is provided a speech coder
comprising a spectral parameter computer for obtaining
a plurality of spectral parameters from an input
speech signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
quantizer for forming an excitation signal of the
input speech signal with M non-zero amplitude pulses,
obtaining a sample position corresponding to a pulse
position meeting a predetermined condition with
respect to the computed pitch prediction signal in a
time interval equal to the pitch period from the
8
CA 02301995 2004-05-25
forefront of a frame, setting pulse position
candidates through shifting the obtained sample
position by the pitch period on the basis of the
position shifted by predetermined numbers of samples
from the sample position, extracting the polarity of a
cross-correlation calculated between input speech
signal and impulse response obtained from said
spectral parameter computer, for each position
candidate in the retrieval range, retrieving the
position candidates for a best position, and
to outputting data of the retrieved best position.
The excitation quantizer includes a codebook
for jointly quantizing the amplitudes or polarities of
a plurality of pulses.
According to yet another aspect of the present
invention, there is provided a speech coder comprising
a spectral parameter computer for obtaining a
plurality of spectral parameters from an input speech
signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
2o delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
quantizer for forming an excitation signal of the
input speech signal with M non-zero amplitude pulses,
obtaining a sample position meeting a predetermined
condition with respect to the computed pitch
9
CA 02301995 2004-05-25
prediction signal, setting a plurality of pulse
position retrieval ranges on the basis of positions
obtained by shifting the obtained sample position by
corresponding shift extents, extracting the polarity
of a cross-correlation calculated between input speech
signal and impulse response obtained from said
spectral parameter computer, for each position
candidate in the retrieval range, making retrieval of
the pulse position retrieval ranges to select a best
combination of a shift extent and a pulse position,
and outputting data of the selected best combination.
According to a further aspect of the present
invention, there is provided a speech coder comprising
a spectral parameter computer for obtaining a
plurality of spectral parameters from an input speech
signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
2o quantizer for forming an excitation signal of the
input speech signal with M non-zero amplitude pulses,
obtaining a sample pulse position meeting a
predetermined condition with respect to the computed
pitch prediction signal in a time interval equal to
the pitch period from the forefront of a frame,
setting a plurality of pulse position retrieval ranges
l0
CA 02301995 2004-05-25
on the basis of positions obtained by shifting the
obtained sample position by corresponding shift
extents, extracting the polarity of a cross-
correlation calculated between input speech signal
and impulse response obtained from said spectral
parameter computer, for each position candidate in the
retrieval range, making retrieval of the pulse
position retrieval ranges to select a best combination
of a shift extent and a pulse position, and outputting
data of the selected best combination.
to According to a still further aspect of the
present invention, there is provided a speech coder
comprising a spectral parameter computer for obtaining
a plurality of spectral parameters from an input
speech signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
quantizer for forming an excitation signal of the
input speech signal with M non-zero amplitude pulses,
obtaining a sample pulse position meeting a
predetermined condition with respect to the computed
pitch prediction signal in a time interval equal to
the pitch period from the forefront of a frame,
setting pulse position candidates through shifting the
obtained sample position by the pitch period on the
11
CA 02301995 2004-05-25
basis of the position shifted by predetermined numbers
of samples from the sample position, extracting the
polarity of a cross-correlation calculated between
input speech signal and impulse response obtained from
said spectral parameter computer, for each position
candidate in the retrieval range, retrieving the
position candidates for a best position, and
outputting data of the retrieved best position.
The excitation quantizer includes a codebook
for jointly quantizing the amplitudes or polarities of
a plurality of pulses.
According to a still further aspect of the
present invention, there is provided a speech coder
comprising a spectral parameter computer for obtaining
a plurality of spectral parameters from an input
speech signal and quantizing the obtained spectral
parameters, a mode judging means for extracting a
characteristic amount from the input speech signal,
judging a plurality of modes from the extracted
feature quantity, and outputting mode data, an
adaptive codebook means for obtaining a delay
corresponding to a pitch period from the input speech
signal, computing a pitch prediction signal, and
making pitch prediction, and an excitation quantizer
for forming an excitation signal of the input speech
signal with M non-zero amplitude signals, obtaining a
sample position meeting a predetermined condition with
12
CA 02301995 2004-05-25
respect to the pitch prediction signal when the mode
data represents a predetermined mode, setting a pulse
position retrieval range on the basis of the obtained
sample position, extracting the polarity of a cross-
correlation calculated between input speech signal
and impulse response obtained from said spectral
parameter computer, for each position candidate in the
retrieval range, retrieving a best position in the
pulse position retrieval range, and outputting data of
the retrieved best position.
l0 The feature quantity is an average pitch
prediction gain. The mode judging means judges the
modes on the basis of comparison of the average pitch
prediction gain with a plurality of threshold values.
According to a still further aspect of the
present invention, there is provided a speech coder
comprising a spectral parameter computer for obtaining
a plurality of spectral parameters from an input
speech signal and quantizing the obtained spectral
parameters, an adaptive codebook means for obtaining a
delay corresponding to a pitch period from the input
speech signal, computing a pitch prediction signal,
and executing pitch prediction, and an excitation
quantizer for obtaining a position meeting a
predetermined condition with respect to the pitch
prediction signal computed in the adaptive codebook
means, extracting the polarity of a cross-correlation
13
CA 02301995 2004-05-25
calculated between input speech signal and impulse
response obtained from said spectral parameter
computer, for each position candidate in the retrieval
range, setting a plurality of pulse position retrieval
ranges for respective pulses constituting an
excitation signal, and retrieving the pulse position
retrieval ranges for the best positions of the pulses.
Other objects and features will be clarified
from the following description with reference to
attached drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a block diagram showing a first
embodiment of the speech coder according to the
present invention;
Fig. 2 shows a flow chart for explaining the
operation in the excitation quantizer 350;
Fig. 3 is a block diagram showing a second
embodiment of the present invention;
Fig. 4 is a block diagram showing a third
embodiment of the present invention;
Fig. 4 is a block diagram showing a third
embodiment of the present invention;
Fig. 5 is a block diagram showing a fourth
embodiment of the present invention;
Fig. 6 is a block diagram showing a fifth
embodiment of the present invention;
13a
CA 02301995 2004-05-25
Fig. 7 is a block diagram showing a sixth
embodiment of the speech coder according to the
present invention;
Fig. 8 is a block diagram showing construction
of the excitation quantizer 350;
13b
CA 02301995 2000-03-29
Fig. 9 is a block diagram showing a seventh
embodiment of the present invention;
Fig. 10 shows the construction of the
excitation quantizer 450;
Fig. 11 is a block diagram showing an eighth
embodiment of the present invention;
Fig. 12 shows the construction of the
excitation quantizer 550;
Fig. 13 is a block diagram showing a ninth
embodiment of the present invention;
Fig. 14 shows the construction of the
excitation quantizer 390;
Fig. 15 is a block diagram showing a tenth
embodiment of the present invention;
Fig. 16 is a block diagram showing the
construction of the excitation quantizer 600;
Fig. 17 is a block diagram showing an eleventh
embodiment of the present invention;
Fig. 18 is a block diagram showing the
construction of the excitation quantizer 650;
Fig. 19 is a block diagram showing a twelfth
embodiment of the present invention;
Fig. 20 is a block diagram showing the
construction of the excitation quantizer;
Fig. 21 is a block diagram showing a thirteenth
embodiment of the present invention;
Fig. 22 is a block diagram showing the
construction of the excitation quantizer 850; and
14
CA 02301995 2000-03-29
Fig. 23 is a block diagram showing a fourteenth
embodiment of the present invention.
PREFERRED EMBODIMENTS OF THE INVENTION
Embodiments of the present invention will now
be described with reference to the drawings.
Fig. 1 is a block diagram showing a first
embodiment of the speech coder according to the
present invention.
Referring to the figure, a frame circuit 110
splits a speech signal inputted from an input
terminal 100 into frames (of 10 ms, for instance),
and a sub-frame circuit 120 further splits each
frame of speech signal into a plurality of shorter
sub-frames (of 5 ms, for instance).
A spectral parameter computer 200 computes a
spectral parameters of a predetermined order P (for
instance, P = 10) by cutting the speech signal with
a window longer than the sub-frame length (for
instance 24 ms) for each with respect to at least
one sub-frame of speech signal. The spectral
parameters may be calculated in a well-known process
of LPC analysis, Burg analysis, etc. In the instant
case, it is assumed that the Burg analysis is used.
The Burg analysis is detailed in Nakamizo, "Signal
Analysis and System Identification", published by
Corona Co., Ltd., 1988, pp. 82-87 (Literature 4),
and not described in the specification.
The spectral parameter computer 200 also
CA 02301995 2000-03-29
converts linear prediction parameters ai (i = 1, ...,
10) which have been obtained by the Burg process
into LSP parameters suited for quantization or
interpolation. The conversion of the linear
prediction parameters into the LSP parameters is
described in Sugamura et al., "Speech Compression by
Linear Spectrum Pair (LSP) Speech Analysis Synthesis
System", J64-A, 1981, pp. 599-606 (Literature 5).
For example, the spectral parameter computer 200
converts the linear prediction parameters obtained
in the 2-nd sub-frame by the Brug process into LSP
parameters, obtains the 1-st sub-frame LSP
parameters by linear interpolation, inversely
converts the 1-st sub-frame LSP parameters thus
obtained into linear prediction parameters, and
outputs the linear prediction parameters ail (i = 1,
..., 10, 1 = 1, ..., 2) of the 1-st and 2-nd
sub-frames to a perceptual weighter 230, while
outputting the 2-nd sub-frame LSP parameters to a
spectral parameter quantizer 210.
The spectral parameter quantizer 210
efficiently quantizes LSP parameters of
predetermined sub-frames by using a codebook 220,
and outputs quantized LSP parameters which minimizes
a distortion given as:
P
D~ _ ~ W(i)(LSP(i) - QLSP(i)J~2 (1)
t
where LSP(i) is i-th sub-frame LSP parameters before
16
CA 02301995 2000-03-29
the quantization, QLSP(i)~ is a ~-th sub-frame
codevector stored in the codebook 220, and W(i) is a
weighting coefficient.
In the following description, it is assumed
that the vector quantization is used as the
quantization and the 2-nd sub-frame LSP parameters
is quantized. The LSP parameters may be vector
quantized by any well-known process. Specific
examples of the process are disclosed in Japanese
Laid-Open Patent Publication No. 4-171500 (Japanese
Patent Publication No. 2-297600) (Literature 6),
Japanese Laid-Open Patent Publication No. 4-363000
(Japanese Patent Application No. 3-261925)
(Literature 7), Japanese Laid-Open Patent
Publication No. 5-6199 (Japanese Patent Application
No. 3-155049 (Literature 8), and T. Nomura et al.,
"LSP Coding Using VQ-SVQ with Interpolation in 4.075
kbps M-LCELP Speech Coder", Proc. Mobile Multimedia
Communications", B.2.5, 1993 (Literature 9), these
processes being not described in the specification.
The spectral parameter quantizer 210 also
restores the 1-st sub-frame LSP parameters from the
2-nd sub-frame quantized LSP parameters. In the
instant case, the 1-st sub-frame LSP parameters are
restored by linear interpolation between the 2-nd
sub-frame quantized LSP parameters of the present
frame and the 2-nd sub-frame quantized LSP
parameters of the immediately preceding frame.
17
CA 02301995 2000-03-29
Here, the 1-st sub-frame LSP parameters are restored
by the linear interpolation after selecting a
codevector which minimizes the error power between
the non-quantized and quantized LSP parameters.
The spectral parameter quantizer 210 converts
the restored 1-st sub-frame LSP parameters and the-
2-nd sub-frame quantized LSP parameters into the
linear prediction parameters ail (i = 1, .,.., 10, 1
- 1, ..., 2) for each sub-frame, and outputs the
result of the conversion to an impulse response
computer 310, while outputting an index representing
the 2-nd sub-frame quantized LSP parameters
codevector to a mutiplexer 400.
The perceptual weighter 230 receives each
sub-frame non-quantized linear prediction parameters
ai (i = 1, ..., P) from the spectral parameter
computer 200, perceptual-weights the sub-frame
speech signal according to Literature 1, and outputs
a perceptually weighted signal thus obtained.
A response signal computer 240 receives each
sub-frame linear prediction parameters ai and also
each sub-frame linear prediction coefficient a;',
having been restored by quantization and
interpolation, from the spectral parameter computer
200 and the spectral parameter quantizer 210,
computes a response signal corresponding to an input
signal of d(n) - 0 for one sub-frame by using stored
filter memory data, and outputs the computed
18
CA 02301995 2000-03-29
response signal to a subtractor 235. The response
signal x=(n) is expressed as:
P P P
xs(n~ = d(n~ -' ~ ~d(n - i) -~ ~ ai7'~J(n '_ _~ -+' ~ ay'xz(n - t~
i-1 i_1 i_1
When n - 1 <_ 0,
y(n-_~=P(N- (n-t))
xr(n _ i~ - s,~(N -I- (n _ i)) (4~
where N is the sub-frame length, y is a weighting
coefficient for controlling the order of the
perceptually weighting and the same in value as
shown in equation (6) given below, sN(n) is the
output signal of the weighting signal computer 230,
and p(n) is a filter output signal in the divisor of
the first term of the right side of equation (6).
The subtractor 235 subtracts the response
signal from the heating sense weighted signal for
one sub-frame, and outputs the difference xN'(n) to
an adaptive codebook circuit 300.
xw(n) = xw(n) - xZ~n~ (5~
The impulse response calculator 310 calculates
the impulse response hW(n) of the perceptually
weighting filter executes the following z transform:
P
1 _ ~ aiz_i
H,o(z~ - i_-1 1
P P (6J
1 - ~ ai7'z ' 1 - ~ at7'z-'
i-__1 i=1
19
CA 02301995 2000-03-29
for a predetermined number L of points, and outputs
the result to the adaptive codebook circuit 300 and
also to an excitation quantizer 350.
The adaptive codebook circuit 300 receives the
past excitation signal v(n) from the weighting
signal calculator 360, the output signal x'~(n) from
the subtractor 235 and the perceptually weighted
impulse response hW(n) from the impulse response
calculator 310, determines a delay T corresponding
to the pitch such as to minimize the distortion:
N-1 N-1 N-1
DT = ~ x,~2~n) - ~ ~ xwU)yw(n W')~2~~ ~ ~J~~n - Z')~ ~7)
n=o n=o n=o
tJw(tt T) - 'U(n - '~') * hw(tt) C8)
represents a pitch prediction signal, and the symbol
* represents convolution. It also obtains the gain
(3 a s
N-1 N-1
xvW~yw~n-T)~ ~ ywU'-~)
n=0 n=0
In order to improve the delay extraction
accuracy for women's speeches and children's
speeches, the delay may be obtained as decimal
sample values rather than integer samples. For a
specific process, P. Kroon et. al, "Pitch predictors
with high temporal resolution", Proc. ICASSP, 1990,
pp. 661-664 (Literature 10), for instance, may be
referred to.
CA 02301995 2000-03-29
The adaptive codebook circuit 300 makes the
pitch prediction as:
zw(n) = x~(n) _ w(n - ~~ * hw(n) (10)
and outputs the prediction error signal zw(n) to the
excitation quantizer 350.
An excitation c~uantizer 350 provides data of M
pulses. The operation in the excitation quantizer
350 is shown in the flow chart of Fig. 2.
The operation comprises two stages, one dealing
with some of a plurality of pulses, the other
dealing with the remaining pulses. In two stages
different gains for multiplification are set for
pulse position retrieval.
The excitation signal c(n) is expressed as:
c(rz) - G1 ~ sign(k)b(n - mx) -I- G2 ~ sign(i)b(n - mt) (11)
i=1
where M1 is the number of first stage pulses, Mz is
the number of second stage pulses, sign(k) is the
polarity of a k-th pulse, G1 is the gain of the first
stage pulses, GZ is the gain of the second stage
pulses, and Ml + Mz = M.
Referring to Fig. 2, in a first step zW(n) and
hN(n) are inputted, and a first and a second
correlation function d(n) and ~ are calculated as
N-1
d(n) - ~ z(i)h;~(i - n), n = 0, . . . , N - 1 (12)
c=n
21
CA 02301995 2000-03-29
N-1
~(P~ Q) - ~ ~~(n - P)~~(n - q), P~ q = 0, . . . , N - 1 (13)
a-max(P,?~
In a subsequent step, the positions of the M1
(M1 <_ M)..non-zero amplitude pulses (or first pulses)
are computed by using the above two correlation
functions. To this end, predetermined positions as
candidates are retrieved for an optimal position of
each pulse as according to Literature 3.
In Fig. 2, examples of candidates for each
pulse position where sub-frame length N = 40 and
number of pulses M1 = 5 are as'shown in the following
table 1:
FIRST PULSE 0,5,10,15,20, 30,35
25,
SECOND PULSE 1,6,11,16,21, 31,36
26,
THIRD PULSE 2,7,12,17,22, 32,37
27,
FOURTH PULSE 3,8,13,18,23, 33,38
28,
FIFTH PULSE 4,9,14,19,24, 34,39
29,
For each pulse, each position candidate is
checked to select an optimal position, which
maximizes an equation:
D _ C~ (14)
E,
where
Aii
C~ _ ~ Sign(~;)d(,nt~.) (15)
~;=1
hfi M-1 Ml
E - ~, St9n(~'')2~(m'~e ml:) -I- 2 ~ ~ s9n(~')S9'n(=)~('~lC~ mi) , (16)
=1 k=li=k+1
22
CA 02301995 2000-03-29
M1 pulse positions are outputted.
Then, using the computed positions of M1 pulses
the correlation function d(n) is corrected with the
amplitude as the polarity as:
M1
d~(n) _ d(n) - ~ sigrz(~;)~(mn,mk)~ (17)
~:=i
n=0,...,N-1 (18)
Next, using d'(n) and ~ the positions of the M2
pulses are computed. In this step, d'(n) may be
substituted for d(n) in equation (15), and the
number of pulses may be set to Mz.
The polarities and positions of a total of M
pulses are thus obtained and outputted to a gain
quantizer 365. The pulse positions are each
quantized with a predetermined number of bits, and
indexes representing the pulse positions are
outputted to the multiplexer 400. The pulse
polarities are also outputted to the multiplexer
400.
The gain quantizer 365 reads out the gain
codevectors from a gain codebook 355, selects a gain
codevector which minimizes the following equation,
and finally selects a combination of an amplitude
codevector and a gain codevector which minimizes the
distortion.
It is now assumed that three different
excitation gains G1 to G3 represented by adaptive
23
CA 02301995 2000-03-29
codebook gains and pulses are vector quantized at a
time.
N-1
Dt = ~Ixw(7L)-~tv(7~'-T)*ft,u(n)_
n=o
Mi Mi
Git ~ sign(k)h~,(n-mi;)-GZt ~ sign(i)hw(n-tni)~Z (19)
ki-1 i-1
Denoted ~3t' , Glt' and GZt' are t-th elements of
three-dimensional gain codevectors stored in the
gain codebook 355. The gain quantizer 365 selects a
gain codevector which minimizes the distortion Dt by
executing the above computation with each gain
codevector, and outputs the index of the selected
gain codevector to the multiplexer 400.
The weighting signal computer 360 receives each
index, reads out the corresponding codevector, and
obtains a drive excitation signal V(n) given as:
2 0 M1 Mx
v(n) _ Riv(n - T ) -E- Git ~ sign(k)b(n - m,~) -f- Gzt ~ sign(i)~(n - mi) (24)
k=1 i=1
V(n) being outputted to the adaptive codebook
circuit 300.
The weighting signal computer 360 then computes
the response signal sW(n) for each sub-frame from the
output parameters of the spectral parameter computer
200 and the spectral parameter quantizer 210 by
using the following equation, and outputs the
24
CA 02301995 2000-03-29
computed response signal to the response signal
computer 240.
P P P
Sw(n) _ '~(n') - ~, Qiv('n - i) -f- ~ di'y'P(n - t) ~' ~, ai~Ytsw(~t~ _ t)
(21)
i=1 i=1 i=s
Fig. 3 is a block diagram showing a second
embodiment of the present invention. This
embodiment comprises an excitation quantizer 450,
which is different in operation form that in the
embodiment shown in Fig. 1. Specifically, the sound
source quantizer 450 quantizes pulse amplitudes by
using an amplitude codebook 451.
In the excitation quantizer 450, after the
positions of the M1 pulses have been obtained, Q (Q
1) amplitude codevector candidates are outputted for
maximizing an equation:
(22)
M' (23)
Ci _ ~ 9~:id(rn.~:)
x=i
bil M1-1 Mi
- ~9Lj~('ITbk,77Lk) -'~ 2 ~ ~ fj;j9ij~(TlRk>TTLi~ ~(24)
k=1 i=k+1
were gki' is an j-th amplitude codevector of a k-th
pulse.
Then, the correlation function is corrected
with respect to each of the selected Q amplitude
codevectors using an equation:
CA 02301995 2004-05-25
Rh
d~(n) - d(n) - ~ 9~:j ~(mm,'i'nx) (25)
b=1
Then, for each corrected correlation function
d'(n) the amplitude codevectors in the amplitude
codebook 451 are retrieved with respect to the
remaining MZ pulses, and a pulse which maximizes the
following equation is selected.
(26)
r C; - ~~,91~d~(T~~:) (2T)
k=1
bti M=-1 M=
Ei - ~ 9kWmk,'mk) -" 2 ~ ~ 9IC9Ivm'k, m~) (28)
1=1 t=k+1
The above process is executed in an iterated manner for
the Q corrected functions d'(n), and a combination
which maximizes the accumulated value given as:
29
D - C~ / Ej + C; / E; ( )
is selected.
The excitation quantizer 450 outputs the index
representing the selected amplitude codevector to
the mutiplexer 400. It also outputs position data
and amplitude codevector data to a gain quantizer
460.
The gain quantizer 460 selects a gain
codevector which minimizes the following equation
from the gain codebook 355.
26
CA 02301995 2000-03-29
N-1
Dt- ~(xw(n)_~iv(~n_T)*hw(n)-
a=0
M1 M=
Glt ~ 9Lfw(n-rn'~;)-G2t ~gihw(n-mi)~2
bi=1 i=1
While in this embodiment the amplitude codebook
451 is used, it is possible to use, instead, a
polarity codebook showing the pulse polarities.
Fig. 4 is a block diagram showing a third
embodiment of the present invention.
This embodiment uses a first and a second
excitation quantizer 500 and 510. In the first
excitation quantizer 500, like the above excitation
quantizer 350 shown in Fig. l, the operation
comprises two stages, one dealing with some of the
pulses and the other dealing with the remaining
pulses, and different gains for multiplification are
set for the pulse position retrieval. The two
stages, in which the operation is executed, is by no
means limitative, and it is possible to provide any
number of stages. The pulse position retrieval
method is the same as in the excitation quantizer
350 shown in Fig. 1. The excitation signal cl(n) in
this case is given as:
arl M,
W (n) = G1 ~ 8i9'n(~)b(n -'mi:) + G2 ~ si9n(a)6(n -- m;) (31)
k=1 i=1
After the pulse position retrieval, a
distortion D1 due to a first excitation is computed
27
CA 02301995 2000-03-29
as:
N-1
131 - ~ (xw(n) - cl(n) * h~,(ax)~z (g2)
n=o
, It is possible to replace the above equation
with an equation:
N-1
l - ~ xw(n) - (C; ~E~ + ~: I E~l
n-o
As Cj, Ci, Ej and Ei, values after the pulse position
retrieval are used.
In the second excitation quantizer 510, the
operation comprises a single stage, and a single
gain for multiplification is set for all the M (M >
(MI + MZ)) pulses. A second excitation signal c2(n)
is given as:
M
c2(n) - G ~ sign(~;)6(n - zn~) (34)
z;=1
where G is the gain for all the M pulses.
A distortion DZ due to the second excitation is
computed as:
N-i
Dz - ~ (xw(n) - cz(n) * hw(n)~2
n=0
or as:
N-1
j)2 _ ~ xw(,rt) _ Cl ~E~ (36)
n=0
28
CA 02301995 2000-03-29
As C1 and E1 are used values after the pulse position
retrieval in the second excitation quantizer 510.
A judging circuit 520 compares the first and
second excitation signals cl(n) and cz(n) and the
distortions D1 and DZ due thereto, and outputs the
less distortion excitation signal to a gain
quantizer 530. The judging circuit 520 also outputs
a judgment code to the gain quantizer.530 and also
to the multiplexer 400, and outputs codes
representing the positions and polarities of the
less distortion excitation signal pulses to the
multiplexer 400.
The gain quantizer 530, receiving the judgment
code, executes the same operation as in the above
gain quantizer 365 shown in Fig. 1 when the first
excitation signal is used. When the second
excitation is used, it reads out two-dimensional
gain codevectors from the gain codevector 540, and
retrieves for a codevector which minimizes an
equation:
N-i
DZt = ~ ~x~~n~ - ~i~'~n - ~') * hw(n~ - Gt ~ asgn,(k)hw(n - rn~ ~~2 ~37)
n=o
kt=
It outputs the index of the selected gain codevector
to the multiplexer 400.
Fig. 5 is a block diagram showing a fourth
embodiment of the present invention. This
embodiment uses a first and a second excitation
29
CA 02301995 2000-03-29
quantizer 600 and 610, which different operations
from those in the case of the embodiment shown in
Fig. 4.
The first excitation quantizer 600, like the
excitation quantizer 450 shown in Fig. 3, quantizes
the pulse amplitudes by using the amplitude codebook
451.
After the positions of the M1 pulses have been
determined, it selects Q (Q >_ 1) amplitude
codevector candidates for maximizing an equation:
C~ /E~
(38)
M1
Ci - ~ 9~~d(rnk) (39)
t=i
20
Mi Ml-1 Ml
E''j - ~9ij~('nZIC~mx~-f-2 ~ ~ 9~~9i~~(mlb,~i~ (40~
k=1 k=1 i=k+1
where gkj' is a j-th amplitude codevector of a k-th
pulse according to the following equation.
M1
d~(n) - d(n) - ~ 9~:~ ~('nzn, m.i;) (41)
k=1
Then, with respect to each of the Q corrected
correlation functions d'(n) it retrieves the
amplitude codevectors in the amplitude codevector
451 for the remaining Mz pulses, and selects an
amplitude codevector which maximizes an equation:
Ci /Ei (42)
CA 02301995 2000-03-29
where
M=
( )
C~ - ~ 9ttd~('n~) 43
k=1
MZ M=-1 Mi
Ec - ~ 9ki~(mk~'~k~ "~' 2 ~ ~ 9Ii9Ii~('~'~k~'~i) (44)
1;=1 ~;-1 l=L f 1
It executes above process iteratedly for the Q
corrected correlation functions d'(n) to select a
combination which maximizes an accumulated value
given as:
D - C~IE~ +C'~lEc (45)
It also obtains the first excitation signal
given as:
M1 Ms
~i('rz) _ Gl ~ g~:~(T~ - Tlt~) + GZ ~ 9ia('~ -- mt) (46)
~:-1 . ~_1
It further computes the distortion D1 due to the
first excitation using an equation:
N 1 ( )
Di - ~ (xm(n) - cl(n) * ~w(n)~2 4'l
n=o
and outputs the distortion D1 to the judging circuit
520.
The second excitation quantizer 610 retrieves
for an amplitude codevector which maximizes an
equation:
(4B)
CJ /EI
where
31
CA 02301995 2000-03-29
M
C1 - ~ 9~td('J'n~) (49)
x=1
M M-1 M .
( - ~ 9ki~('mi:~'n'tk) -f- 2 ~ ~ 9kt9it~(mk~ mi) (50)
i;=1 b=1 i=ki-1
It also obtains the second excitation signal
given as:
M
c2(n) = G1;~ 9x~(~ - ~k) (51)
~m
It further computes the distortion DZ due to the
second excitation signal using an equation: '
DZ - ~ (xw('n') -' c2(~'') * hw(n)1 (52)
tt=0
and outputs the distortion D2 to the judging circuit
520.
Alternatively, the distortion DZ may be obtained
as:
N-1
_ ~ x~,(n) - C'i /E1 (53)
n=o
C1 and E1 are correlation values after the second
excitation signal pulse positions have been
determined.
The judging circuit 520 compares the first and
second excitation signals cl'(n) and c2'(n) and also
compares the distortions D1' and Dz' due thereto, and
outputs the less distortion excitation signal to the
32
CA 02301995 2000-03-29
gain quantizer 530, while outputting a judgment code
to the gain quantizer 530 and the multiplexer 400.
Fig. 6 is a block diagram showing a fifth
embodiment of the present invention.
This embodiment is based on the third
embodiment, but it is possible to provide a similar
system which is based on the fourth embodiment.
The embodiment comprises a mode judging circuit
900, which receives the perceptually weighting
signal of each frame from the perceptually weighting
circuit 230 and outputs mode data to an excitation
quantizer 600. The mode judging circuit 900 judges
the mode by using a feature quantity of the present
frame. The feature quantity may be a frame average
pitch prediction gain. The pitch prediction gain
may be computed as:
L
G = 101og1o(1/L ~(P~/E;)] (54)
c=i
where L is the number of sub-frames in the frame, Pi
is the speech power in an i-th sub-frame, and Ei is
the pitch predicted error power.
N-1
1'~ _ ~ x~i('n) (55)
n=a
N-1 N-1
E~ - P' - ( ~ x~i(n)xwi(n - Z')]2/( ~ xwt(n - ~')~ (56)
n=a n=o
33
CA 02301995 2000-03-29
Here, T is an optimum delay which maximizes the
prediction gain.
The mode judging circuit 900 sets up a
plurality of different modes by comparing the frame
average pitch prediction gain G with respective
predetermined thresholds. The number of different
modes may, for instance, be four. The mode judging
circuity 900 outputs the mode data to the
multiplexer 400 as well as to the excitation
quantizer 700.
When a predetermined mode is represented by the
received mode data, the excitation quantizer 700
executes the same operation as in the first
excitation quantizer 500 shown in Fig. 4, and
outputs the first excitation signal to a gain
quantizer 750, while outputting codes representing
the pulse positions and polarities to the mutiplexer
400.. When the predetermined mode is not
represented, it executes the same operation as in
the second excitation quantizer 510 as shown in Fig.
4, and outputs the second excitation to the gain
quantizer 750, while outputting codes representing
the pulse positions and polarities to the
multiplexer 400.
When the predetermined mode is represented, the
gain quantizer 750 executes the same operation as in
the gain quantizer 365. Otherwise, it executes the
same operation as in the gain quantizer 530 shown in
34
CA 02301995 2000-03-29
Fig. 1.
The embodiments described above may be modified
variously. As an example, a codebook used for
quantizing the amplitudes of a plurality of pulses,
may be stored in advance by studying the speech
signal. A method of storing a codebook through the
speech signal study is described in, for instance,
Linde et al., "An Algorithm for Vector Quantization
Design", IEEE Trans. Commun., pp. 84-95, January
1980.
In lieu of the amplitude codebook, a polarity
codebook may be provided, in which pulse polarity
combinations corresponding in number to the number
of bits equal to the number of pulses are prepared.
It is possible to obtain the positions of any
number of pulses with gain variations and to switch
adaptive codebook circuits or gain codebooks by
using mode data.
For the pulse amplitude quantization, it is
possible to arrange such as to preliminarily select
a plurality of amplitude codevectors from the
amplitude codebook 351 for each of a plurality of
pulse groups each of L pulses and then permit the
pulse amplitude quantization using the selected
codevectors. This arrangement permits reducing the
computational effort necessary for the pulse
amplitude quantization.
As an example of the amplitude codevector
CA 02301995 2000-03-29
selection, a plurality of amplitude codevectors are
preliminarily selected and outputted to the
excitation quantizer in the order of maximizing
equation (57) or (58).
N-1 L
Dk = ~ ~, z(n') ~ gikb(~'=))2 (57)
n=a ;-i
N-1 L L
D~: _ ~~ z(n)~g~ka('~'~'ts)~2ly9va(mi)~2 (58)
n=o t=i c=1
As has been described in the foregoing,
according to the present invention, the positions of
M non-zero amplitude pulses are retrieved with a
different gain for each group of the pulses less in
number than M. It is thus possible to increase the
accuracy of the excitation and improve the
performance compared to the prior art speech coders.
The present invention comprises a first
excitation quantizer for retrieving the positions of
M non-zero amplitude pulses which constitutes an
excitation signal of the input speech signal with a
different gain for each group of the pulses less in
number than M, and a second excitation quantizer for
retrieving the positions of a predetermined number
of pulses by using the spectral parameters, judges
the both distortion for selecting the better one,
36
CA 02301995 2000-03-29
and uses better excitation in accordance with the
feature time change of the speech signal to improve
the characteristic.
In addition, according to the present invention
a mode of the input speech may be judged by
extracting a feature quantity therefrom, and the
first and second excitation quantizers may be
switched to obtain the pulse positions according to
the judged mode. It is thus possible to use always
use a good excitation corresponding to time changes
in the feature quantity of the speech signal with
less computational effort. The performance thus can
be improved compared to the prior art speech coders.
Fig. 7 is a block diagram showing a sixth
embodiment of the speech coder according to the
present invention.
Referring to the figure, a frame circuit 110
splits a speech signal inputted from an input
terminal 100 into frames (of 10 ms, for instance),
and a sub-frame circuit 120 further splits each
frame of speech signal into a plurality of shorter
sub-frames (of 5 ms, for instance).
A spectral parameter computer 200 computes a
spectral parameters of a predetermined order P (for
instance, P = 10) by cutting the speech signal with
a window longer than the sub-frame length (for
instance 24 ms) for each with respect to at least
one sub-frame of speech signal. The spectral
37
CA 02301995 2000-03-29
parameters may be calculated in a well-known process
of LPC analysis, Burg analysis, etc. The spectral
parameter computer 200 also converts linear
prediction parameters ai (i = 1, ..., 10) which have
been obtained by the Burg process into LSP
parameters suited for quantization or interpolation.
For example, the spectral parameter computer 200
converts the linear prediction parameters obtained
in the 2-nd sub-frame by the Brug process into LSP
parameters, obtains the 1-st sub-frame LSP
parameters by linear interpolation, inversely
converts the 1-st sub-frame LSP parameters thus
obtained into linear prediction parameters, and
outputs the linear prediction parameters ail (i = 1,
..., 10, 1 = 1, ..., 2) of the 1-st and 2-nd
sub-frames to a perceptual weighter 230, while
outputting the 2-nd sub-frame LSP parameters to a
spectral parameter quantizer 210.
The spectral parameter quantizer 210
efficiently quantizes LSP parameters of
predetermined sub-frames by using a codebook 220,
and outputs quantized LSP parameters which minimizes
a distortion given as equation (1).
In the following description, it is also
assumed that the vector quantization is used as the
quantization and the 2-nd sub-frame LSP parameters
is quantized as described before.
The spectral parameter quantizer 210 also
38
CA 02301995 2000-03-29
restores the 1-st sub-frame LSP parameters from the
2-nd sub-frame quantized LSP parameters. In the
instant case, the 1-st sub-frame LSP parameters are
restored by linear interpolation between the 2-nd
sub-frame quantized LSP parameters of the present
frame and the 2-nd sub-frame quantized LSP
parameters of the immediately preceding frame.
Here, the 1-st sub-frame LSP parameters are restored
by the linear interpolation after selecting a
codevector which minimizes the error power between
the non-quantized and quantized LSP parameters.
The spectral parameter quantizer 210 converts
the restored 1-st sub-frame LSP parameters and the
2-nd sub-frame quantized LSP parameters into the
linear prediction parameters ail (i = 1, .,.., 10, 1
- 1, ..., 2) for each sub-frame, and outputs the
result of the conversion to an impulse response
computer 310, while outputting an index representing
the 2-nd sub-frame quantized LSP parameters
codevector to a mutiplexer 400.
The perceptual weighter 230 receives each
sub-frame non-quantized linear prediction parameters
a; (i = 1, ..., P) from the spectral parameter
computer 200, perceptual-weights the sub-frame
speech signal according to Literature 1, and outputs
a perceptually weighted signal thus obtained.
A response signal computer 240 receives each
sub-frame linear prediction parameters a; and also
39
CA 02301995 2000-03-29
each sub-frame linear prediction coefficient ai',
having been restored by quantization and
interpolation, from the spectral parameter computer
200 and the spectral parameter quantizer 210,
computes a response signal corresponding to an input
signal of d(n) - 0 for one sub-frame by using stored
filter memory data, and outputs the computed
response signal to a subtractor 235. The response
signal xZ(n) is expressed as equation (2).
When n - 1 <_ 0, equations (3) and (4) are used.
The subtractor 235 subtracts the response
signal from the perceptually weighted signal for one
sub-frame, and outputs the difference xW'(n) to an
adaptive codebook circuit 300.
The impulse response calculator 310 calculates
the impulse response hw(n) of the perceptually
weighting filter executes the z transform equation
(6), for a predetermined number L of points, and
outputs the result to the adaptive codebook circuit
300 and also to an excitation quantizer 350.
The adaptive codebook circuit 300 receives the
past excitation signal v(n) from the weighting
signal calculator 360, the output signal x'W(n) from
the subtractor 235 and the perceptually weighted
impulse response hW(n) from the impulse response
calculator 310, determines a delay T corresponding
to the pitch such as to minimize the distortion
expressed by equation (7). It also obtains the gain
CA 02301995 2000-03-29
by equation (9).
In order to improve the delay extraction
accuracy for women's speeches and children's
speeches, the delay may be obtained as decimal
sample values rather than integer samples.
The adaptive codebook circuit 300 makes the
pitch prediction according to equation (10) and
outputs the prediction error signal zW(n) to the
excitation quantizer 350.
An excitation quantizer 350 provides data of M
pulses. The operation in the excitation quantizer
350 is shown in the flow chart of Fig. 2.
Fig. 8 is a block diagram showing the
construction of the excitation quantizer 350.
An absolute maximum position detector 351
detects a sample position, which meets a
predetermined condition with respect to a pitch
prediction signal yW(n). In this embodiment, the
predetermined condition is that "the absolute
amplitude is maximum", and the absolute maximum
position detector 351 detects a sample position
which meets this condition, and outputs the detected
sample position data to a position retrieval range
setter 352.
The position retrieval range setter 352 sets a
retrieval range of each sample position after
shifting the input pulse position by a predetermined
sample number L toward the future or past.
41
CA 02301995 2000-03-29
As an example, where five pulses are to be
obtained in a 5-ms sub-frame (40 samples), with an
input sample position D, position candidates
contained in the retrieval ranges of these pulses
are:
1-st pulse: D-L, D-L+5, ...
2-nd pulse: D-L+1, D-L+6, ...
3-rd pulse: D-L+2, L+7, ...
4-th pulse: D-L+3, L+8, ...
5-th pulse: D-L+4, L+9, ...
Then, zw(n) and hW(n) are inputted, and a first
and a second correlation computers 353 and 354
compute a first and a second correlation function
d(n) and ~, respectively, using equations (12) and
(13).
A pulse polarity setter 355 extracts the
polarity of the first correlation function d(n) for
each pulse position candidates in the retrieval
range set by the position retrieval range setter
352.
A pulse position retriever 356 executes
operation on the following equation with respect to
the above position candidate combinations, and
selects a position which maximizes the same equation
(14) as an optimum position.
If the number of pulses is M, equation (15) and
(16) are employed. The pulse polarities used have
been preliminarily extracted by the pulse polarity
42
CA 02301995 2000-03-29
setter 355. In the above operation, polarity and
position data of the M pulses are outputted to a
gain quantizer 365.
Each pulse position is quantized with a
predetermined number of bits to produce a
corresponding index, which is outputted to the
multiplexes 400. The pulse polarity data is also
outputted to the multilexer 400.
The gain quantizer 365 reads out the gain
codevectors from a gain codebook 367, selects a gain
codevector which minimizes the following equation,
and finally selects a combination of an amplitude
codevector and a gain codevector which minimizes the
distortion.
It is now assumed that three different
excitation gains G' represented by adaptive codebook
gain [3' and pulses are vector quantized at a time.
N-1 hs
Di = ~ ~x,~(n~ - ~3iv(n - T~ * Ic,~(n~ - Gi ~ sign(~;~Iaw(n - ~n~;~~2
12=O
~:=1
Denoted ~it'and Gt' are t-th elements of
three-dimensional gain codevectors stored in the
gain codebook 367. The gain quantizer 365 selects a
gain codevector which minimizes the distortion Dt by
executing the above computation with each gain
codevector, and outputs the index of the selected
gain codevector to the multiplexes 400.
The weighting signal computer 360 receives each
43
CA 02301995 2000-03-29
index, reads out the corresponding codevector, and
obtains a drive excitation signal V(n) given as:
ht
v(n~ - ~3i"(n _ Z') + Gt ~ Sign(k)~(n _ "z~) (60)
~:=i
V(n) being outputted to the adaptive codebook
circuit 300.
The weighting signal computer 360 then computes
the response signal sw(n) for each sub-frame from the
output parameters of the spectral parameter computer
200 and the spectral parameter quantizer 210 by
using the following equation, and outputs the
computed response signal to the response signal
computer 240.
Fig. 9 is a block diagram showing a seventh
embodiment of the present invention. This
embodiment comprises an excitation quantizer 450,
which is different in operation form that in the
embodiment shown in Fig. 7.
Fig. 10 shows the construction of the
excitation quantizer 450. The excitation quantizer
450 receives an adaptive codebook delay T as well as
the prediction signal yH(n)., the prediction error
signal z"(n), and the perceptually weighted pulse
response h"( n ) .
An absolute maximum position computer 451
receives delay time data T corresponding to the
pitch period, detects a sample position which
44
CA 02301995 2000-03-29
corresponds to the maximum absolute value of the
pitch prediction signal yw(n) in a range form the
sub-frame forefront up to a sample position after
the delay time T, and outputs the detected sample
position data to the position retrieval range setter
352.
Fig. 11 is a block diagram showing an eighth
embodiment of the present invention. This
embodiment uses an excitation quantizer 550, which
is different in operation from the excitation
quantizer 450 shown in Fig. 9. Fig. 12 shows the
construction of the excitation quantizer 550.
A position retrieval range setter 552 sets
position candidates of pulses through the delay by
the delay time T positions, which are obtained by
shifting input sample positions by a predetermined
sample number L to the future or past.
As an example, where five pulses are to be
obtained in a 5-ms sub-frame (40 samples), with an
input sample position D, position candidates of the
pulses are:
1-st pulse: D-L, D-L+T, ...
2-nd pulse: D-L+1, D=L+T, ...
3-rd pulse: D=L+2, D-L+T, ...
4-th pulse: D=L+3, D-L+T, ...
5-th pulse: D=L+4, D-L+T, ...
Fig. 13 is a block diagram showing a ninth
embodiment of the present invention. This
CA 02301995 2000-03-29
embodiment is a modification of the sixth embodiment
obtained by adding an amplitude codebook. The
seventh and eighth embodiments may be modified
likewise by adding an amplitude codebook.
The difference of Fig. 13 from Fig. 7 resides
in an excitation quantizer 390 and an amplitude
codebook 395. Fig. 14 shows the construction of the
excitation quantizer 390. In this embodiment, pulse
amplitude quantization is made by using the
amplitude codebook 395.
In the pulse position retriever 356, after the
positions of M pulses have been determined, an
amplitude quantizer 397 selects an amplitude
codevector which maximizes the equations (22), (23)
and the following equation (61) from the amplitude
codebook 395, and outputs the index of the selected
amplitude codevector.
bf hf -1 bf
~9~j~'~T~k,'mk) ~-2 ~ ~ 9~79aJ~~mkf'mi) ~61~
k=1 k=1 i=k~-1
where gkj' is a j-th amplitude codevector of a k-th
pulse.
The pulse position quantizer 390 outputs an
index representing the selected amplitude codevector
and also outputs the position data and amplitude
codevector data to the gain quantizer 365.
While the amplitude codebook is used in this
embodiment, it is possible to use instead a polarity
46
CA 02301995 2000-03-29
codebook showing the polarities of pulses for the
retrieval.
Fig. 15 is a block diagram showing a tenth
embodiment of the present invention. This
embodiment uses an excitation quantizer 600 which is
different in operation for the excitation quantizer
350 shown in Fig. 7. The construction of the
excitation quantizer 600 will now be described with
reference to Fig. 16.
Fig. 16 is a block diagram showing the
construction of the excitation quantizer 600. A
position retrieval range setter 652 shifts, by a
plurality of (for instance Q) different shifting
extents, a position represented by the output data
of the absolute maximum position detector 351, sets
retrieval ranges and pulse position sets of each
pulse with respect to the respective shifted
positions, and outputs the pulse position sets to a
pulse polarity setter 655 and a pulse retriever 650.
The pulse polarity setter 655 extracts polarity
data of each of a plurality of position candidates
received from the position retriever 652, and
outputs the extracted polarity data to the pulse
position retriever 656.
The pulse position retriever 656 retrieves for
a position, which maximizes equation (14), with
respect to each of the plurality of position
candidates by using the first and second correlation
47
CA 02301995 2000-03-29
functions and the polarity. The pulse position
retriever 656 selects the position which maximizes
equation (14) by executing the above operation Q
times, corresponding to the number of the different
shifting extents, and outputs position and shifting
extent data of the pulses, while also outputting the
shifting extent data to the multiplexer 400.
Fig. 17 is a block diagram showing an eleventh
embodiment of the present invention. This
embodiment uses an excitation quantizer 650 which is
different in operation from the excitation quantizer
650 shown in Fig. 7. The construction of the
excitation quantizer 650 will now be described with
reference to Fig. 18.
Fig. 18 is a block diagram showing the
construction of the excitation quantizer 650.
A position retrieval range setter 652 sets
positions of each pulse with respect to positions,
which are obtained by shifting by a plurality of
(for instance Q) shift extents a position
represented by the output data of the absolute
maximum position detector 451, and outputs pulse
position sets corresponding in number to the number
of the shifting extents to a pulse polarity setter
655 and a pulse position retriever 656.
The pulse polarity setter 655 extracts polarity
data of each of a plurality of position candidates
outputted from the position retriever 652, and
48
CA 02301995 2000-03-29
extracts the extracted polarity data to the pulse
position retriever 656.
The pulse position retriever 656 retrieves for
a position which maximizes equation (14) by using
the first and second correlation functions and the
polarity. The pulse position retriever 656 finally
selects the position which maximizes equation (14)
with Q different kinds by executing the above
operation Q times corresponding to the number of the
different shifting extents, and outputs pulse
position and shifting extent data, while also
outputting the shifting extent data to the
multiplexer 400.
Fig. 19 is a block diagram showing a twelfth
embodiment of the present invention. This
embodiment uses an excitation quantizer 750 which is
different in operation from the excitation quantizer
350 shown in Fig. 11. The construction of the
excitation quantizer 750 will now be described with
reference to Fig. 20.
Fig. 20 is a block diagram showing the
construction of the excitation quantizer.
A position retrieval range setter 752 sets
positions of each pulse by delaying positions, which
are obtained by shifting by a plurality of (for
instance Q) shifting extents a position represented
by the output data of the absolute maximum position
detector 451, by a delay time T. The position
49
CA 02301995 2000-03-29
retrieval range setter 752 thus outputs position
sets of each pulse corresponding in number to the
number of the different shifting extents to a pulse
polarity'setter 655 and a pulse position retriever
656.
The pulse polarity setter 655 extracts polarity
data of each of a plurality of position candidates
from the position retriever 652, and outputs the
extracted polarity data to the pulse position
retriever 656.
The pulse position retriever 656 retrieves for
a position which maximizes equation (14) by using
the first and second correlation functions and the
polarity. The pulse position retriever 656 selects
the position which maximizes equation (14) by
executing the above operation Q times corresponding
to the number of the different shifting extents, and
outputs pulse position and shifting extent data to
the gain quantizer 365, while outputting the
shifting extent data to the multiplexer 400.
Fig. 21 is a block diagram showing a thirteenth
embodiment of the present invention. This
embodiment is obtained as a modification of the
fifth embodiment by adding an amplitude codebook for
pulse amplitude quantization, but it is possible to
obtain modifications of the eleventh and twelfth
embodiments likewise.
This embodiment uses an excitation quantizer
CA 02301995 2000-03-29
850 which is different in operation from the
excitation quantizer 390 shown in Fig. 13. The
construction of the excitation quantizer 850 will
now be described with reference to Fig. 22.
Fig. 22 is a block diagram showing the
construction of the excitation quantizer 850.
A position retrieval range setter 652 sets
positions of each pulse with respect to positions,
which are obtained by shifting by a plurality of
different (for instance Q) shifting extents a
position represented by the output data of the
absolute maximum position detector 351, and outputs
pulse position sets corresponding in number to the
number of the different shifting extents to a pulse
polarity setter 655 and a pulse position retriever
656.
The pulse polarity setter 655 extracts polarity
data of each of a plurality of position candidates
of the position retriever 652 and outputs the
extracted polarity data to the pulse position
retriever 656.
The pulse position retriever 656 retrieves for
a position for maximizing equation (14) with respect
to each of a plurality of position candidates by
using the first and second correlation functions and
the polarity. The pulse position retriever 656
selects the position which maximizes equation (14)
by executing the above operation Q times
51
CA 02301995 2000-03-29
corresponding in number to the number of the
different shifting extents, and outputs pulse
position and shifting extent data to the gain
quantizer 365, while also outputting the shifting
extent data to the multiplexer 400. An amplitude
quantizer 397 is the same in operation as the one
shown in Fig. 14.
Fig. 23 is a block diagram showing a fourteenth
embodiment of the present invention. This
embodiment is based on the first embodiment, but it
is possible to obtain its modifications which are
based on other embodiments.
A mode judging circuit 900 receives the
perceptually weighted signal in units of frames from
the perceptually weighting circuit 230, and outputs
mode data to an adaptive codebook circuit 950, an
excitation quantizer 960 and a gain quantizer 965 as
well as to the multiplexer 400. As the mode data, a
feature quantity of the present frame is used. As
the feature quantity, the frame average pitch
prediction gain is used. The pitch prediction gain
may be computed by using an equation:
(~2)
G - lOIogio(lIL~,(F'~IL~))
c=i
where L is the number of sub-frames contained in the
frame, and P3 and Ei are the speech power and the
pitch prediction error power in an i-th frame,
52
CA 02301995 2000-03-29
respectively given as:
N-i (63)
2
1, _ ~ xwc(n)
,t=o
and
N-1 N_-1
L~ = t'c - ( ~ x,~i(n)~~i~n - Z')~Z~( ~ xu~a'~ - ~')( (
n=o n=o -
where T is the optimum delay corresponding to the
maximum prediction gain.
The mode judging circuit 900 judges a plurality
of (for instance R) different modes by comparing the
frame average pitch prediction gain G with
corresponding threshold values. The number R of the
different modes may be 4.
When the outputted mode data represents a
predetermined mode, the adaptive codebook circuit
950 receiving this data executes the same operation
as in the adaptive codebook 300 shown in Fig. 7, and
outputs a delay signal, an adaptive codebook
prediction signal and a prediction error signal. In
the other modes, it directly outputs its input
signal from the subtractor 235.
At the same time, that is, in the above
predetermined mode, the excitation quantizer 960
executes the same operation as in the excitation
quantizer 350 shown in Fig. 7.
53
CA 02301995 2000-03-29
The gain quantizer 965 switches a plurality of
gain codebooks 3671 to 3678, which are designed for
each mode, to be used for gain quantization
according to the received mode data.
The embodiments described above are by no means
limitative, and various changes and modifications
are possible. For example, a codebook for amplitude
quantizing a plurality of pulses may be
preliminarily studied and stored by using a speech
signal. A codebook study method is described in,
for instance, Linde et al, "An algorithm for Vector
Quantization Design", IEEE Trans. Commun., pp.
84-95, January 1980.
As an alternative to the amplitude codebook, a
polarity codebook may be used, in which pulse
polarity combinations corresponding in number to the
number of bits equal to the number of pulses are
stored.
As has been described in the foregoing,
according to the present invention the excitation
quantizer obtains a position meeting a predetermined
condition with respect to a pitch prediction signal
obtained in the adaptive codebook, sets a plurality
of pulse position retrieval ranges for respective
pulses constituting an excitation signal, and
retrieves these pulse position retrieval ranges for
the best position. It is thus possible to provide a
satisfactory excitation signal, which represents a
54
CA 02301995 2000-03-29
pitch waveform, by synchronizing the pulse position
retrieval ranges to the pitch waveform.
Satisfactory sound quality compared to the prior art
system is thus obtainable with a reduced
bit rate.
In addition, according to the present
invention, the excitation quantizer may perform the
above process in a predetermined mode among a
plurality of different modes, which are judged from
a feature quantity extracted from the input speech.
It is thus possible to improve the sound quality for
positions of the speech corresponding to modes, in
which the periodicity of the speech is strong.
Changes in construction will occur to those
skilled in the art and various apparently different
modifications and embodiments may be made without
departing from the scope of the present invention.
The matter set forth in the foregoing description
and accompanying drawings is offered by way of
illustration only. It is therefore intended that
the foregoing description be regarded as
illustrative rather than limiting.
