JPH07104777A - Pitch detecting method and speech analyzing and synthesizing method - Google Patents

Abstract

PURPOSE:To make it possible to properly detect pitch for a voice which contains the original pitch component and a pitch component twice as large as it almost to the same extent and to prevent a synthesized sound from deteriorating in quality by judging the pitch of a current block being twice as large as the original pitch when specific conditions are met. CONSTITUTION:It is judged that the pitch P of the current block is twice as large as the original pitch when at least one of the condition, that is., the pitch P of the current block P is more than twice as large as the typical pitch and the block mean level is larger than a specific threshold value, and the typical pitch is larger than the typical pitch of the last block and the pitch P is larger than the specific threshold value, etc. Namely, it is judged whether or not a value is a half as large as the pitch extracted by a pitch extraction part 13 at the time of voice pitch extraction should be used. According to the judgement result, a P/2 selection part 26 sends pitch data P from a high-precision pitch search part 16 to an encoding part 21 as they are or after halving them.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention divides an input voice signal into blocks, and a pitch detecting method for detecting a pitch from an input voice signal when performing voice analysis processing in units of the divided blocks, and a method for detecting the pitch. The present invention relates to a speech analysis / synthesis method using a pitch detection method.

[0002]

2. Description of the Related Art Various coding methods are known in which signal compression is performed by utilizing the statistical properties of audio signals (including voice signals and acoustic signals) in the time domain and frequency domain and human auditory characteristics. ing. As this encoding method,
Broadly speaking, time domain coding, frequency domain coding,
Examples include analysis and synthesis coding.

As an example of high-efficiency encoding of a voice signal or the like, M
BE (Multiband Excitation) coding, SBE (Singleband Excitation) coding, Harmonic coding, SBC
(Sub-band Coding), LPC (Linear
Predictive Coding: Linear predictive coding) or DC
T (Discrete Cosine Transform), MDCT (Modified DC)
In T), FFT (Fast Fourier Transform), etc., when quantizing various information data such as spectrum amplitude and its parameters (LSP parameter, α parameter, k parameter, etc.), conventionally, scalar quantization is performed. There are many.

In the voice analysis / synthesis system such as the PARCOR method, since the excitation source is switched at each block (frame) on the time axis, voiced sound and unvoiced sound cannot be mixed in the same frame. , As a result, high quality voice was not obtained.

On the other hand, in the above MBE coding, for each voice in one block (frame), each harmonic (harmonic) of the frequency spectrum or each band (band) in which a few harmonics are grouped together. Since voiced / unvoiced sound discrimination (V / UV discrimination) is performed for each band divided for each band or for each band divided by a fixed bandwidth (for example, 300 to 400 Hz), sound quality Is recognized. V / U for each band
V discrimination mainly looks at how strongly the spectrum in the band has a harmonic structure.

The voiced sound waveform generally has a periodic structure. This basic period is called a pitch period, and its reciprocal is called a pitch frequency. Detecting or extracting this pitch from the input speech signal is important in general speech coding processing, particularly in speech analysis and synthesis for analyzing and synthesizing speech. As pitch data in speech analysis and synthesis, a so-called pitch lag, which represents the pitch period by the number of samples, is often used.

[0007]

By the way, in the voice uttered by the same speaker, the speaker intentionally suddenly pitches (pitch lag) for a short time (for example, about 100 ms).
It is not very possible to double (decrease by one octave), but it has been acknowledged that some language (eg English) unknowingly happens to be approximately equivalent to this. Frequency analysis of this state has confirmed that, in addition to the peak of the spectrum that appears in synchronism with the original pitch, a new small peak exists in the valley between the peaks. When the peak of this valley becomes large, it can be interpreted that the pitch is doubled. However, when the spectrum envelope is vector-quantized after that, and there is no codebook corresponding to this state. As a result of the quantization, the peaks of the valleys become almost the same as the original peaks, which causes the quality of the synthesized sound to deteriorate.

That is, normal voice has a frequency structure having a fundamental frequency corresponding to the pitch and a peak at a point that is an integral multiple thereof. However, some speakers may have other peaks under certain conditions, for example, as shown in FIG.
It is recognized that a voice having a frequency structure as shown in FIG. This FIG. 10 is a diagram showing a voice power spectrum, in which the peak at the position indicated by the solid line arrow is the peak corresponding to the original pitch component, and the peak at the position indicated by the broken line arrow is the peak corresponding to the double pitch component. Shown respectively.

When the peaks other than the original peaks become large to some extent, it is inevitable that the pitch extraction method on the time axis has a completely doubled pitch. Here, in the case of vector quantization of (the peak envelope of) this spectrum, for example, it is usual that no code vector corresponding to FIG. 10 is prepared, and the code vector of the peak envelope of the most approximate spectrum. Is selected and vector-quantized, the voice power spectrum obtained by inverse-quantization becomes, for example, as shown in FIG. 11, and the valley peak (the peak corresponding to the double pitch component) becomes large.

The synthesized sound in such a state is a very hard-to-hear sound (so-called buzzy sound) because unnecessary components all cause energy concentration in synchronization with the pitch, especially when the phase is not transmitted. Has become.

The present invention has been made in view of such a situation, and an appropriate pitch detection can be performed for a voice in which an original pitch component and a pitch component twice as much as the original pitch component are included in the same degree. An object of the present invention is to provide a pitch detection method and a voice analysis / synthesis method that can be performed and can prevent quality deterioration of synthesized speech.

[0012]

A pitch detecting method according to the present invention is a pitch detecting method for detecting a pitch corresponding to a fundamental period of a voice in a voice analysis, in which an input voice signal is divided into blocks on a time axis. The step of dividing, the step of detecting the typical pitch Pt unique to the speaker and the pitch P of the current block for each signal of each divided block, and the pitch P of the current block being twice the original pitch. This double pitch determination step includes (c1) the pitch P of the current block is equal to or larger than a value in the vicinity of twice the typical pitch Pt, and the block average level LV is Greater than a predetermined threshold LV _t1 , (c
2) The typical pitch Pt is larger than the typical pitch Pt 'in the previous block and the pitch P of the current block is larger than a predetermined threshold th _p1 , (c3) The pitch P of the current block is the pitch P of the previous block. At least one of the three conditions (c1) to (c3) that is greater than a predetermined threshold value th _p2 and the block average level LV is greater than a predetermined threshold value LV _t2. When satisfied, the above problem is solved by determining that the pitch P of the current block is twice the original pitch.

The speech analysis and synthesis method according to the present invention is
When it is determined in the double pitch determination step of the pitch detection method that the pitch P of the current block is twice the original pitch, the pitch P / 2 is half the pitch P.
The above-mentioned problem is solved by performing the synthesis based on.

Here, as a specific example of the speech analysis / synthesis method to which the present invention is applied, there is a speech coding method using multi-band excitation. In this multi-band excitation coding, V / UV (voiced sound / unvoiced sound) is discriminated for each band, and V / UV discriminated for each of these bands.
Depending on UV, the portion determined to be V synthesizes voiced sound by synthesizing a sine wave or the like, and the portion determined to be UV synthesizes unvoiced sound by transforming the frequency component of the noise signal. When analyzing a voice signal in such a voice analysis / synthesis method, the input signal may be divided into bands based on the pitch P of the current block, and V / UV determination may be performed for each of the divided bands. preferable.

[0015]

[Function] It is possible to surely determine whether or not the voice has the same pitch as the original pitch and twice the pitch.
It is possible to effectively take measures when such a state occurs. In addition, since P / 2 is used as pitch data for voice synthesis, strong concentration of energy generated by analysis / synthesis with a double pitch is reduced by halving the pitch to improve the quality of synthesized voice. Can be increased.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A high efficiency coding method which is an embodiment of a voice analysis / synthesis method to which the pitch detection method according to the present invention is applied will be described below. The high-efficiency coding method is, for example, MBE (Multiband Excitation) coding described later, in which a signal for each block is converted on a frequency axis, divided into a plurality of bands, and V ( Voiced sound) or UV
It is possible to use an encoding method for determining whether or not (unvoiced sound).

That is, as a general speech analysis / synthesis method to which the present invention is applied, the speech signal is divided into blocks for every fixed number of samples (for example, 256 samples), and the FFT is performed.
While converting to spectrum data on the frequency axis by orthogonal transformation such as, to extract the pitch of the voice in the block,
The spectrum on the frequency axis is band-divided at intervals according to this pitch, and V (voiced sound) / U is obtained for each of the divided bands.
V (unvoiced sound) is discriminated. This V / UV discrimination information is encoded and transmitted together with spectrum amplitude data. However, in the embodiment of the present invention, V / UV is performed in units of frames (for example, 160 samples) described later
Various information such as discrimination information and spectrum amplitude data is transmitted.

Here, when a speech synthesis analysis system such as an MBE vocoder is assumed, the sampling frequency fs for the speech signal on the input time axis is usually 8 kHz.
The total bandwidth is 3.4 kHz (the effective bandwidth is 200-
3400Hz), the pitch lag from the higher female voice to the lower male voice (the number of samples corresponding to the pitch period)
Is about 20 to 147. Therefore, the pitch frequency is 8000/147 ≒ 54 (Hz) to 8000/20 = 400 (Hz)
It will fluctuate within the range. Therefore, about 8 to 63 pitch pulses (harmonics) stand up to 3.4 kHz on the frequency axis.

In this way, considering that the number of divided bands (the number of bands) for each block (frame) changes between about 8 and 63 when the bands are divided at intervals according to the pitch, It is preferable to reduce or reduce the number of divided bands to a fixed number (for example, about 12).

Further, in the embodiment of the present invention, when detecting the pitch corresponding to the basic period of the voice in the voice analysis, a typical voice signal is unique to the speaker for each signal of each block (frame). After detecting the pitch Pt and the pitch P of the current block, respectively, (c1) the pitch P of the current block is equal to or more than twice the typical pitch Pt and the block average level LV is a predetermined threshold value LV _t1. Greater than (c2) the typical pitch Pt is greater than the typical pitch Pt 'in the previous block and the pitch P of the current block is greater than a predetermined threshold th _p1 , (c3) the pitch P of the current block is It is larger than the block pitch P ′ and the current block pitch P is a predetermined threshold value th.
_Greater than _p2 and the block average level LV is a predetermined threshold LV
_When at least one of the three conditions (c1) to (c3) that is larger than _t2 is satisfied, it is determined that the pitch P of the current block is double the original pitch.

Next, a specific example of a kind of MBE (Multiband Excitation) vocoder of a speech signal synthesis analysis coding apparatus (so-called vocoder) to which the above speech analysis synthesis method can be applied will be described. Will be described with reference to.

The MBE vocoder described below is a DW G
riffin and JS Lim, "MultibandExcitation Vocode
r, "IEEE Trans.Acoustics, Speech, and Signal Process
ing, vol.36, No.8, pp.1223-1235, Aug. 1988, the conventional PARCOR (PARtial au
In a to-CORrelation (partial autocorrelation) vocoder or the like, a voiced sound section and an unvoiced sound section were switched for each block or frame when modeling a voice, whereas in the MBE vocoder, at the same time (same block or frame). The model is based on the assumption that there are voiced and unvoiced intervals in the frequency domain of (in).

FIG. 1 is a block diagram showing an overall schematic configuration of an embodiment in which the present invention is applied to the MBE vocoder. In FIG. 1, an audio signal is supplied to the input terminal 11, and the input audio signal is HP
Sent to a filter 12 such as F (high-pass filter),
The so-called DC (direct current) offset is removed and at least the low frequency component (200 Hz or less) is removed for band limitation (for example, 200-3400 Hz). The signal obtained through this filter 12 is output to the pitch extraction unit 13
And the windowing processing unit 14, respectively. In the pitch extraction unit (pitch detection unit) 13, the input voice signal data is divided into blocks in units of a predetermined number N (for example, N = 256) (or cut out by a rectangular window), and the voice signals in this block are divided. Pitch extraction is performed. Such a cut block (256 samples)
, For example, as shown in A of FIG.
(= 160) frame intervals, the blocks are moved in the time axis direction, and the overlap between blocks is NL samples (96 samples, for example). In addition, the windowing processing unit 14 applies a predetermined window function, for example, a Hamming window, to one block of N samples, and sets this windowing block to 1
The frames are sequentially moved in the time axis direction at intervals of L samples.

When this windowing process is expressed by a mathematical expression, x _w (k, q) = x (q) w (kL-q) (1) In this equation (1), k is a block number,
q represents the time index (sample number) of the data, and the q-th data x (q) of the input signal before processing is windowed by the window function (w (kL-q)) of the kth block. It is shown that the data x _w (k, q) can be obtained by doing so. The window function w _r (r) in the case of the rectangular window as shown in A of FIG. 2 in the pitch extraction unit 13 is w _r (r) = 1 0 ≦ r <N (2) = 0 r < 0, N ≦ r Further, the window function w _h in the case of Hamming window as shown in B of FIG. 2 in the windowing processing section 14 (r) _{is, w h (r) = 0.54} - 0.46 cos (2πr / (N-1)) 0 ≦ r <N (3) = 0 r <0, N ≦ r. Such a window function w _r (r) or w (1) when using the _h (r) formula of the window function w (r) (= w (KL-
q)), the zero-zero interval is 0 ≦ kL−q <N, and is modified as follows: kL−N <q ≦ kL Therefore, for example, in the case of the above rectangular window, the window function w _r (kL−q) = 1
This is when kL−N <q ≦ kL, as shown in FIG. In addition, the above formulas (1) to (3) have a length N
It shows that the window of (= 256) samples advances by L (= 160) samples. Below, above (2)
N points (0≤r
The non-zero sample sequences of <N) are respectively x _wr (k, r) and x _wh
We will denote it as (k, r).

In the windowing processing section 14, as shown in FIG. 4, 179 is applied to the sample sequence x _wh (k, r) of one block of 256 samples to which the Hamming window of the equation (3) is applied.
Two samples of 0 data are added (so-called zero padding) to form 2048 samples, and the orthogonal transform unit 15 performs orthogonal transform such as FFT (Fast Fourier Transform) on the time-axis data sequence of 2048 samples. Processing is performed. Alternatively, FF with 256 points without zero padding
There is also a method of applying T to reduce the processing amount.

In the pitch extraction section (pitch detection section) 13,
Sample sequence of x _wr (k, r) above (1 block N samples)
Pitch extraction is performed based on The pitch extraction method is known to include periodicity of time waveform, periodic frequency structure of spectrum, and one using autocorrelation function,
In this embodiment, the center correlation waveform autocorrelation method is adopted. Regarding the center clip level in the block at this time, one clip level may be set for each block, but the peak level of the signal of each part (each sub-block) obtained by subdividing the block is detected and When there is a large difference in peak level between the sub-blocks, the clip level is changed stepwise or continuously within the block. The pitch period is determined based on the peak position of the autocorrelation data of this center clip waveform. At this time, a plurality of peaks are obtained from the autocorrelation data belonging to the current frame (the autocorrelation is obtained for the data of N samples of one block), and the maximum peak of the plurality of peaks is equal to or larger than a predetermined threshold value. When, the maximum peak position is the pitch period,
In other cases, peaks within a pitch range that satisfies a predetermined relationship with the pitches obtained in frames other than the current frame, for example, the preceding and following frames, for example, within a range of ± 20% centered on the pitch of the previous frame. The pitch of the current frame is determined based on the obtained peak position. In this pitch extraction unit 13, a relatively rough pitch search is performed by an open loop, and the extracted pitch data is sent to a high precision (fine) pitch search unit 16 and a closed loop high precision pitch search (pitch) is performed. Fine search) is performed.
Note that the input waveform is LPC
A method of obtaining the pitch from the autocorrelation of the analyzed residual waveform may be used.

Further, this pitch extraction section (pitch detection section)
In No. 13, it was also detected that the pitch was momentarily doubled (the peak of the original peak of the voice power spectrum was increased) in the utterance of a constant speaker as described above. There is. This detection method will be described below. In the description of this detection method, the pitch extracted in the frame currently being analyzed (current frame) is P, the pitch in the previous frame is P ′, the typical pitch in the current frame is Pt, and the previous pitch is Pt. The typical pitch in the frame is Pt ', and the input voice level in the current frame is LV.

Here, the typical pitch Pt, which is the basic pitch, is referred to by the present inventor in Japanese Patent Application No. 4-9226.
No. 1 is disclosed in the specification and drawings. That is, in order to extract the pitch, the autocorrelation of the voice signal of one frame is obtained, and when the maximum value of a plurality of peaks appearing in this autocorrelation exceeds a predetermined value k, the pitch corresponding to this maximum peak is typically set. It is registered as the pitch Pt. Also, this typical pitch Pt becomes the pitch P of the current frame. If the maximum peak value of the autocorrelation does not exceed the predetermined value k, the pitch P of the current frame is determined based on the pitch of the previous frame and the size of the maximum peak value. Do not register as Pt. As the typical pitch Pt at this time, a typical pitch already registered is used.

Such pitches P, P ', Pt, P
The following conditions are set for the values of t ′ and the input sound level LV. (C1) The pitch P of the current frame is equal to or larger than twice the typical pitch Pt and the frame average level LV is larger than a predetermined threshold value LV _t1 , (c2) the typical pitch Pt is typical of the previous frame. Pitch Pt '
The pitch P of the current frame is larger than the predetermined threshold value.
larger than th _p1 , (c3) the pitch P of the current frame is larger than the pitch P ′ of the previous frame, the pitch P of the current frame is larger than a predetermined threshold th _p2 , and the frame average level LV is larger than a predetermined threshold LV _t2 . When at least one of the three large conditions (c1) to (c3) is satisfied, it is determined that the pitch P of the current frame is twice the original pitch.

As a specific example, th _p1 is 60 and th _p2 is 100.
And the above condition is (c1) P> Pt and LV> LV
_t1 , (c2) Pt> Pt 'and P> 60 (= th _p1 ),
(C3) P> 1.5P 'and P> 100 (= th _p2 ) and
If LV> LV _t2 and either of these is satisfied, it is determined that the half pitch P / 2 of the currently detected pitch P is the original pitch.

When any of the above three conditions (c1) to (c3) is satisfied and P> 1.5P 'is satisfied, the condition P>1.5P' in the next frame is as follows: It becomes true by inheriting the same conditions of the current frame. On the contrary, when all of the above three conditions (c1) to (c3) are not satisfied, the above condition P> 1.5P ′ becomes false, and P is newly added in the next frame.
And P ′, the condition P> 1.5P ′ is determined.

The determination result of P or P / 2 is P, P /
2 is sent to the 2 selection unit 26, and the P, P / 2 selection unit 2
6, the high-precision (fine) pitch search unit 1 described later
Transmit pitch data from 6 as it is,
It is selected whether to transmit by half (P / 2).

The high-precision (fine) pitch search section 16 includes coarse (rough) pitch data of integer (integer) values extracted by the pitch extraction section 13 and, for example, on the frequency axis FFTed by the orthogonal transformation section 15. And the data of. In this high-precision pitch search unit 16, with the coarse pitch data value as the center, ± several samples are shaken in increments of 0.2 to 0.5 to give an optimum decimal point (floating).
To the value of the fine pitch data of. As a fine search method at this time, so-called synthesis analysis
Using the (Analysis by Synthesis) method, the pitch is selected so that the synthesized power spectrum is closest to the power spectrum of the original sound.

The fine search of the pitch will be described. First, in the MBE vocoder, the FF
Let S (j) as spectrum data on the frequency axis orthogonally transformed by T etc. be expressed as S (j) = H (j) | E (j) | 0 <j <J (4) It is assumed that the model. Here, J corresponds to ω _s / 4π = f _s / 2, and the sampling frequency f
_{When s} = ω _s / 2π is, for example, 8 kHz, it corresponds to 4 kHz. In the equation (4), when the spectrum data S (j) on the frequency axis has a waveform as shown in FIG.
(j) indicates the spectrum envelope (envelope) of the original spectrum data S (j) as shown in B of FIG.
(j) shows the spectrum of the excitation signal (excitation) which is periodic at the same level as shown in C of FIG. That is, the FFT spectrum S (j) is the spectrum envelope H (j) and the power spectrum | E of the excitation signal.
(j) | is modeled as |.

Power spectrum of the excitation signal | E (j)
| Is a spectral waveform corresponding to the waveform of one band (band) for each band on the frequency axis in consideration of the periodicity (pitch structure) of the waveform on the frequency axis determined according to the pitch. It is formed by arranging it repeatedly. The waveform for one band is obtained by FFT by regarding a waveform obtained by adding (filling with 0) 1792 samples of 0 data to a Hamming window function of 256 samples as shown in FIG. 4 as a time axis signal. It can be formed by cutting out an impulse waveform having a certain bandwidth on the frequency axis according to the pitch.

Next, for each band divided according to the above pitch, a value (a kind of amplitude) | A that represents the above H (j) (minimizes the error for each band) | A _m
Ask for |. Here, for example, when the lower limit point and the upper limit point of the m-th band (band of the m-th harmonic) are a _m and b _m , respectively, the error ε _m of the m-th band is

[0037]

[Equation 1]

Can be expressed as | A _m | that minimizes this error ε _m is

[0039]

[Equation 2]

Thus, when | A _m | in this equation (6),
Minimize the error ε _m .

Such an amplitude | A _m | is obtained for each band, and an error ε _m for each band defined by the above equation (5) is obtained using the obtained amplitude | A _m |. Next, the sum Σε of all the bands of the error ε _m for each band as described above.
_{Find m} . Further, such an error sum value Σε _m of all bands is obtained for some slightly different pitches, and a pitch that minimizes the error sum value Σε _m is obtained.

That is, with the rough pitch determined by the pitch extraction unit 13 as the center, several types are prepared in the upper and lower portions, for example, in steps of 0.25. The error sum value Σε _m is obtained for each of these plural kinds of slightly different pitches. In this case, if the pitch is determined, the bandwidth is determined, and from the above formula (6), the power spectrum | S (j) |
(j) | can be used to find the error ε _m in the above equation (5), and the total sum Σε _m of all the bands can be found. This error sum total value Σε _m is obtained for each pitch, and the pitch corresponding to the minimum error sum value is determined as the optimum pitch. As described above, the high-precision pitch search unit obtains the optimum fine (for example, 0.25 step) pitch, and the amplitude | A _m | corresponding to this optimum pitch is determined. The calculation of the amplitude value at this time is performed in the voiced sound amplitude evaluation unit 18V.

In the above description of the pitch fine search, in order to simplify the description, all bands are voiced (Vo
Assuming the case of iced), MBE as described above
In the vocoder, unvoiced sound (Un
Since a model in which a voiced) region exists is used, it is necessary to distinguish voiced sound / unvoiced sound for each band.

The optimum pitch from the high precision pitch search section 16 and the amplitude from the amplitude evaluation section (voiced sound) 18V | A
The data of _m | is sent to the voiced sound / unvoiced sound discrimination unit 17,
The voiced sound / unvoiced sound is discriminated for each band. NSR (noise to signal ratio) is used for this determination. That is, the NSR that is the NSR of the m-th band
_m is

[0045]

[Equation 3]

This NSR _m is a predetermined threshold value Th ₁
When larger than (for example, Th ₁ = 0.2) (that is, when the error is large), | A _m || E (j) in that band
The approximation of | S (j) | by | is not good (the above excitation signal |
E (j) | is unsuitable as a base) and the band is determined to be UV (Unvoiced). In other cases, it can be determined that the approximation is performed to a certain extent, and the band is determined to be V (Voiced, voiced sound).

By the way, as described above, the number of bands divided by the fundamental pitch frequency (the number of harmonics) is
The number of V / UV flags for each band similarly varies because the voice varies in the range of about 8 to 63 depending on the pitch (the size of the pitch).

Therefore, in this embodiment, the V / UV discrimination results are put together (or degenerated) for each of a fixed number of bands divided by a fixed frequency band. Specifically, a predetermined band including a voice band (for example, 0 to
4000 Hz) is divided into N _B (for example, 12) bands, and the weighted average value is discriminated by a predetermined threshold value Th ₂ (for example, Th ₂ = 0.2) according to the NSR value in each band, The V / UV of the band is judged.

Next, in the unvoiced sound amplitude evaluation unit 18U, the on-frequency data from the orthogonal transformation unit 15, the fine pitch data from the pitch search unit 16, and the voiced sound amplitude evaluation unit 1 are included.
Data of amplitude | A _m | from 8 V and V / UV (voiced sound / unvoiced sound) discrimination data from the voiced sound / unvoiced sound discrimination unit 17 are supplied. This amplitude evaluation section (unvoiced sound) 1
In 8U, the amplitude is obtained again (amplitude re-evaluation is performed) with respect to the band that is determined to be unvoiced sound (UV) by the voiced sound / unvoiced sound determination unit 17. The amplitude | A _m | _UV for this UV band is

[0050]

[Equation 4]

It is calculated by

The data from the amplitude evaluation unit (unvoiced sound) 18U is sent to the data number conversion (a kind of sampling rate conversion) unit 19. The data number conversion unit 19 has a different number of divided bands on the frequency axis according to the pitch,
This is for keeping the number constant, considering that the number of data (especially the number of amplitude data) is different. That is, for example, when the effective band is set to 3400 kHz, the effective band is divided into 8 bands to 63 bands according to the pitch, and the amplitude | A _m | (UV The number of data m _MX +1 including the band amplitude | A _m | _UV also changes from 8 to 63.
Therefore, in the data number conversion unit 19, this variable number m _MX +
One amplitude data is converted into a fixed number M (for example, 44) of data.

Here, in the present embodiment, for example, for the amplitude data of one block of the effective band on the frequency axis, values from the last data in the block to the first data in the block are interpolated. After adding the dummy data to increase the number of data to N _F , the band-limited O _S
By multiplying (e.g., 8 times) oversampling, the amplitude data of O _S times is obtained, and the amplitude data of this O _S times ((m _MX +1) × O _S ) are linearly interpolated. Expand to many N _M (eg 2048),
The N _M pieces of data are thinned out and converted into the fixed number M of data (for example, 44 pieces).

The data from the data number conversion unit 19 (the above-mentioned fixed number M of amplitude data) is the vector quantization unit 20.
To a vector, and a predetermined number of pieces of data are combined into a vector, and vector quantization is performed. The quantized output data from the vector quantizer 20 (main part of the quantized output data) is transmitted from the highly accurate pitch search unit 16 to the P, P / 2 selector 2
The high precision (fine) pitch data obtained via 6 and the voiced sound / unvoiced sound (V / UV) discrimination data from the voiced sound / unvoiced sound discrimination unit 17 are sent to the encoding unit 21 for encoding. .

Each of these data is obtained by processing the data in the block of N samples (for example, 256 samples), but the block is a frame of L samples on the time axis. , The data to be transmitted is obtained in the frame unit. That is, the pitch data, the V / UV discrimination data, and the amplitude data are updated at the above frame period. The V / UV discrimination data from the voiced sound / unvoiced sound discrimination unit 17 is reduced (degenerated) to about 12 bands as necessary as described above, and one or less voiced sounds in all bands ( V) and V are unvoiced (UV) regions, and V (voiced sound) on the low frequency side is expanded to the high frequency side when a predetermined condition is satisfied.
/ UV discrimination data pattern.

Here, the P / P / 2 selection unit 26 satisfies one of the conditions (c1) to (c3) described above, and the pitch P / 2 is half the pitch P currently detected.
Is determined to be the original pitch, the pitch data P from the high-precision pitch search unit 16 is halved (P /
2), this is sent to the encoding unit 21.

Explaining the reason why such a half pitch P / 2 is transmitted, first, the above conditions (c1) to (c).
When one of 3) is established, in the time waveform,
As shown in A of FIG. 6, it can be interpreted that there is a small peak between the peaks corresponding to the pitch P. When the spectrum of B of FIG. 6 corresponding to A of FIG. 6 is analyzed at the pitch P, vector quantization is performed and transmission is performed as described above, and the synthesized voice after dequantization has the result of FIG.
As shown in C and D, the small peak disappears, and instead, all the energy concentrates on the peak of the pitch P. Therefore, in order to reproduce this small peak, the above P is halved (P / 2) and used.

It should be noted that the pitch P is P / P
It is the timing to change to 2. That is, the above V / U
If this replacement is performed before the determination of V, the peak including a valley is evaluated when matching the original (original) spectrum and the base (base) spectrum. This is because it is judged as unvoiced sound). For example, FIG. 7 shows a specific example of a voice power spectrum in the case of performing V / UV discrimination by band division at an angular frequency ω ₀ corresponding to the pitch P. All bands (at least up to ω ₅ ) are V (voiced sound). ) Has been determined.
On the other hand, when V / UV discrimination is performed by band division at an angular frequency 2ω ₀ corresponding to P / 2, as shown in FIG.
It is determined to be V (unvoiced sound). These FIG. 7 and FIG.
In, the solid line shows the original spectrum and the broken line shows the base spectrum.

Therefore, once V / UV discrimination is performed with the original P as it is, and after quantization, P is set to P / P at the final stage of analysis.
It is replaced with 2, and this is transmitted to the combining side. In the example of FIG. 1, the voiced sound / unvoiced sound (V / UV) discrimination unit 17 and the vector quantization unit 20 use the pitch data (P) from the high precision pitch search unit 16 as they are to perform processing.
Separately from these systems, the pitch data from the high-precision pitch search unit 16 is sent to the P / P / 2 selection unit 26 and is replaced with P / 2 only when the above condition is satisfied and sent to the encoding unit 21. There is no problem.

In the encoding unit 21, for example, CR
C addition and rate 1/2 convolutional code addition processing is performed. That is, important data among the pitch data, the voiced sound / unvoiced sound (V / UV) discrimination data, and the quantized output data is subjected to CRC error correction coding and then convolutional coding. It The encoded output data from the encoding unit 21 is the frame interleave unit 2
2, is interleaved with part (eg, less important) data from the vector quantizer 20, extracted from the output terminal 23, and transmitted (or recorded / reproduced) to the combining side (decoding side). .

As described above, the judgment that the pitch is momentarily doubled in the utterance of a certain speaker is made in the above (c1).
~ (C3) can be performed by a relatively simple condition determination, and when it is determined that the pitch is doubled, the pitch data (P) from the high precision pitch search unit 16 is halved (P / 2). Therefore, the processing is simple and real-time processing is easy.

Next, FIG. 9 shows a schematic structure of a synthesizing side (decoding side) for synthesizing an audio signal based on the above-mentioned respective data obtained by being transmitted (or recorded / reproduced).
Will be described with reference to.

In FIG. 9, a data signal at the input terminal 31 is substantially equal to the data signal taken out from the output terminal 23 on the encoder side shown in FIG. 1 (ignoring signal deterioration due to transmission and recording / reproduction). Is supplied. The data from the input terminal 31 is sent to the frame deinterleave unit 32, subjected to deinterleave processing which is the reverse processing of the interleave processing of FIG. 1 above, and subjected to CRC and convolutional coding on the main portion (encoder side). In the portion, generally, the data portion having high importance) is subjected to decoding processing by the decoding unit 33, and the remaining portion (data of low importance not subjected to the encoding processing) is directly subjected to bad frame mask processing in the bad frame mask processing unit 34. To the section 34. In the decoding unit 33, for example, so-called Viterbi decoding processing or error detection processing using a CRC check code is performed. Bad frame mask processing unit 34
Performs processing such as obtaining a parameter of a frame with many errors by interpolation, and also performs the above pitch data, voiced sound /
Unvoiced sound (V / UV) data and vector-quantized amplitude data are separated and extracted.

Here, as described above, the pitch data is obtained from the high precision pitch search section 16 of FIG.
Data obtained through the / 2 selection unit 26, and the pitch P from the high-precision pitch search unit 16 under normal conditions, and P / when any one of the above conditions (c1) to (c3) is satisfied. 2 are transmitted and taken out as pitch data from the bad frame mask processing unit 34.

The vector-quantized amplitude data from the bad frame mask processing unit 34 is sent to the inverse vector quantization unit 35 and inversely quantized, and the data number inverse transform unit 3
It is sent to 6 and is inversely transformed. This data number inverse conversion unit 36
Then, the inverse conversion, which is in contrast to the above-described data number conversion unit 19 of FIG.
And the unvoiced sound synthesizer 38. The pitch data from the mask processing unit 34 is sent to the voiced sound synthesis unit 37 and the unvoiced sound synthesis unit 38. The V / UV discrimination data from the mask processing section 34 is also sent to the voiced sound synthesis section 37 and the unvoiced sound synthesis section 38.

In the voiced sound synthesizer 37, for example, cosine
The voiced sound waveform on the time axis is synthesized by wave synthesis, and the unvoiced sound synthesis unit 38 synthesizes the unvoiced sound waveform on the time axis by filtering white noise, for example, with a bandpass filter, and synthesizes each voiced sound waveform and unvoiced sound. The waveform and the waveform are added and synthesized by the adder 41 and output from the output terminal 42. In this case, the amplitude data, the pitch data, and the V / UV discrimination data are updated and given for each frame (L sample, for example, 160 samples) at the time of the analysis, but the continuity between the frames is improved (smoothed). To set each value of the above amplitude data and pitch data to 1
For example, each data value at the center position in the frame,
Each data value up to the center position of the next frame (one frame at the time of composition) is obtained by interpolation. That is, in one frame at the time of composition (for example, from the center of the above analysis frame to the center of the next analysis frame), each data value at the leading end sample point and each data value at the end (leading edge of the next composition frame) sample point Are given, and each data value between these sample points is obtained by interpolation.

In addition, according to the V / UV discrimination data, the voiced sound (V) region and the unvoiced sound (U
V) area, and according to this division,
V / UV discrimination data for each band can be obtained.
Regarding this division position, as described above, V on the low frequency side
May be expanded to the high frequency side. If the analysis side (encoder side) reduces (degenerates) the number of bands to a certain number (for example, about 12), it is solved (restored) and changed at intervals according to the original pitch. Of course, the number of bands is set.

The synthesis processing in the voiced sound synthesis section 37 will be described in detail below. M-th discriminated as V (voiced sound)
1 on the time axis in the band (band of the mth harmonic)
When a voiced sound for a synthetic frame (L samples, for example 160 samples) is V _m (n), V _m (n) = A _m (n) using the time index (sample number) n in this synthetic frame ) cos (θ _m (n)) 0 ≦ n <L (9) The final voiced sound V (n) is synthesized by adding (ΣV _m (n)) the voiced sounds of all bands that are determined to be V (voiced sound) of all bands.

A _m (n) in the equation (9) is the amplitude of the m-th harmonic wave which is interpolated from the beginning to the end of the composite frame. The simplest way is to linearly interpolate the value of the m-th harmonic of the amplitude data updated in frame units.
That is, the m-th frame at the tip (n = 0) of the composite frame
The amplitude value of the harmonic is A _0m , the end of the composite frame (n =
L: the amplitude value of the m-th harmonic at the next composite frame) is A _Lm, and the following equation is used: A _m (n) = (Ln) A _0m / L + nA _Lm / L (10) It suffices to calculate A _m (n).

Next, the phase θ _m (n) in the above equation (9) is calculated by θ _m (n) = mω _O1 n + n ² m (ω _L1 −ω ₀₁ ) / 2L + φ _{0 m} + Δω n (11) You can ask. In the equation (11), φ _0m represents the phase of the m-th harmonic (frame initial phase) at the tip (n = 0) of the composite frame, and ω ₀₁ is at the tip of the composite frame (n = 0). The fundamental angular frequency, ω _L1, indicates the fundamental angular frequency at the end of the composite frame (n = L: leading end of the next composite frame). Δω in the above equation (11) is
Set a minimum Δω such that the phase φ _{Lm at} n = L is equal to θ _m (L).

Hereinafter, a method of obtaining the amplitude A _m (n) and the phase θ _m (n) according to the V / UV discrimination result when n = 0 and n = L in an arbitrary m-th band will be described. To do.

When the m-th band is V (voiced sound) for both n = 0 and n = L, the amplitude A _m (n) is the amplitude value transmitted by the above equation (10). The amplitude A _m (n) may be calculated by linearly interpolating A _0m and A _Lm . Phase θ _m (n)
From n 0, θ _m (0) = φ _{0 m} to n = L, θ _m (L) is φ
Set Δω to be _Lm .

Next, when n = 0, V (voiced sound) and n =
Amplitude A _m (n) when UV (unvoiced sound) when L
Is linearly interpolated so that 0 A _m (L) from the transmission amplitude value A _{0 m} of A _m (0). The transmission amplitude value A _{Lm when} n = L is the amplitude value of unvoiced sound and is used in unvoiced sound synthesis described later. The phase θ _m (n) is θ _m (0) = φ _{0 m} , and Δω =
Set to 0.

Further, when n = 0, UV (unvoiced sound)
When n = L, when V (voiced sound) is used, the amplitude A _m
(n) is linearly interpolated so that the amplitude A _m (0) at n = 0 is 0 and the transmitted amplitude value A _Lm is obtained at n = L. Phase θ
_{For m} (n), using the phase value φ _Lm at the end of the frame as the phase θ _m (0) at n = 0, θ _m (0) = φ _Lm −m (ω _O1 + ω _L1 ) L / 2 ... (12) and Δω = 0.

When both n = 0 and n = L are V (voiced sound), Δω is set so that θ _m (L) becomes φ _Lm.
A method of setting will be described. In the above formula (11), n
= L, then θ _m (L) = mω _O1 L + L ² m (ω _L1 − ω ₀₁ ) / 2L + φ _0m + ΔωL = m (ω _O1 + ω _L1 ) L / 2 + φ _0m + ΔωL = φ _Lm . Then, Δω becomes Δω = (mod2π ((φ _Lm −φ _0m ) −mL (ω _O1 + ω _L1 ) / 2) / L (13). In this equation (13), mod 2π (x) Is the principal value of x
It is a function that returns a value between π and + π. For example, x = 1.3
mod2 π (x) = -0.7π when π, mod2 when x = 2.3π
When π (x) = 0.3π and x = -1.3π, mod2π (x) = 0.7
π, and so on.

The unvoiced sound synthesizing process in the unvoiced sound synthesizing section 38 will be described below. The white noise signal waveform on the time axis from the white noise generation unit 43 is sent to the windowing processing unit 44, and windowed by a suitable window function (for example, Hamming window) with a predetermined length (for example, 256 samples), STF
A power spectrum on the frequency axis of white noise is obtained by performing STFT (Short Term Fourier Transform) processing by the T processing unit 45. This STFT processing unit 4
The power spectrum from No. 5 is sent to the band amplitude processing unit 46, the above UV (unvoiced sound) band is multiplied by the above amplitude | A _m | _UV, and the amplitude of the other V (voiced sound) band is set to 0. To The band amplitude processing section 46 is supplied with the amplitude data, the pitch data, and the V / UV discrimination data.

The output from the band amplitude processing unit 46 is IS
The signal is sent to the TFT processing unit 47, and the phase is converted into a signal on the time axis by performing inverse STFT processing using the phase of the original white noise. The output from the ISTFT processing unit 47 is sent to the overlap adding unit 48, and the overlap and addition are repeated while appropriate weighting (so that the original continuous noise waveform can be restored) on the time axis,
Synthesize continuous time axis waveforms. The output signal from the overlap adder 48 is sent to the adder 41.

As described above, the signals of the voiced sound portion and the unvoiced sound portion which are synthesized by the synthesis units 37 and 38 and are returned to the time axis are added by the addition unit 41 at an appropriate fixed mixing ratio, The reproduced audio signal is taken out from the output terminal 42.

The present invention is not limited to the above-described embodiment. For example, regarding the configuration of the speech analysis side (encoding side) of FIG. 1 and the configuration of the speech synthesis side (decoding side) of FIG. Although each unit is described as hardware, it can be realized by a software program using a so-called DSP (digital signal processor) or the like. Further, it is sufficient to collectively (degenerate) a band for each harmonic (harmonic) into a certain number of bands, and the number of degenerate bands is not limited to 12 bands. Further, the process of dividing the entire band into the low-frequency side V region and the high-frequency side UV region at one or less division positions may be performed as necessary, and need not be performed. Further, the application of the present invention is not limited to the above-mentioned multi-band excitation speech analysis / synthesis method, and can be easily applied to various speech analysis / synthesis methods using sine wave synthesis. It can be applied not only to transmission and recording / reproduction, but also to various applications such as pitch conversion, speed conversion, noise suppression, and the like.

[0080]

As is apparent from the above description, according to the pitch detecting method of the present invention, in the pitch detecting method for detecting the pitch corresponding to the fundamental period of the voice in the voice analysis, the input voice signal is detected. (C1) The step of classifying the blocks on the time axis, and the step of detecting the typical pitch Pt unique to the speaker and the pitch P of the current block for each of the signals of the partitioned blocks. The pitch P of the current block is equal to or larger than a value near twice the typical pitch Pt and the block average level LV is a predetermined threshold value.
Larger than LV _t1 , (c2) the typical pitch Pt is larger than the typical pitch Pt ′ in the previous block, and the pitch P of the current block is larger than a predetermined threshold th _p1 (c
3) Three conditions: the pitch P of the current block is larger than the pitch P ′ of the previous block, the pitch P of the current block is larger than a predetermined threshold th _p2 , and the block average level LV is larger than a predetermined threshold LV _t2. (C1) to (c
When at least one of 3) is satisfied, there is a double pitch determination step of determining that the pitch P of the current block is double the original pitch. It is possible to surely determine whether or not the voice has the same double pitch, and it is possible to effectively take measures when such a state occurs.

Further, according to the voice analysis / synthesis method of the present invention, when the pitch P of the current block is judged to be double the original pitch in the double pitch judgment step of the above pitch detection method, P / Since the voice synthesis is performed using 2 as the pitch data, the strong concentration of the energy generated by the analysis / synthesis with the double pitch can be alleviated by halving the pitch to improve the quality of the synthesized voice. . Further, as compared with the normal processing, it is only necessary to add a simple condition determination processing and a simple processing of replacing P with P / 2, so that real-time processing can be easily performed.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing a schematic configuration of an analysis side (encoding side) of a speech signal analysis / synthesis coding apparatus as a specific example of a speech analysis / synthesis method to which a pitch detection method according to the present invention is applied. .

FIG. 2 is a diagram for explaining a windowing process.

FIG. 3 is a diagram for explaining a relationship between windowing processing and a window function.

FIG. 4 is a diagram showing time axis data as an object of orthogonal transform (FFT) processing.

FIG. 5 is a diagram showing spectrum data on a frequency axis, a spectrum envelope (envelope), and a power spectrum of an excitation signal.

FIG. 6 is a diagram for explaining energy concentration due to a double pitch.

FIG. 7 is a diagram for explaining V / UV discrimination when band division is performed at an angular frequency ω ₀ corresponding to a pitch P.

FIG. 8 is a diagram for explaining V / UV discrimination when band division is performed at an angular frequency ω ₀ corresponding to a pitch P.

FIG. 9 is a functional block diagram showing a schematic configuration of a synthesis side (decoding side) of a speech signal analysis / synthesis coding apparatus as a specific example of a speech analysis / synthesis method to which a pitch detection method according to the present invention is applied. .

FIG. 10 is a diagram showing an audio power spectrum having a double pitch component.

11 is a diagram showing a power spectrum of a signal obtained by vector-quantizing and dequantizing a signal having the spectrum structure of FIG.

[Explanation of symbols]

13 Pitch extraction unit 14 Window processing unit 15 Orthogonal transform (FFT) unit 16 High precision (fine) pitch search unit 17 Voiced / unvoiced (V / UV) discriminating unit 18V: Voiced sound amplitude evaluation unit 18U: Unvoiced amplitude evaluation unit 19: Data number conversion (data rate conversion) unit 20: Vector quantizer 26: P, P / 2 selector 37: Voiced sound synthesizer 38: Unvoiced sound synthesizer

Claims (3)

[Claims] 1. A pitch detecting method for detecting a pitch corresponding to a fundamental period of a voice in a voice analysis, a step of dividing an input voice signal into blocks on a time axis, and a signal of each divided block. Detecting a typical pitch Pt unique to the utterer and the pitch P of the current block for each of them, and (c1) the pitch P of the current block is equal to or larger than twice the typical pitch Pt. And the block average level LV is larger than a predetermined threshold LV _t1 , (c2) the typical pitch Pt is larger than the typical pitch Pt ′ in the previous block, and the current block pitch P is larger than a predetermined threshold th _p1 . (C3) The pitch P of the current block is larger than the pitch P ′ of the previous block, the pitch P of the current block is larger than a predetermined threshold th _p2 , and the block average is Three conditions (c1) that level LV is greater than a predetermined threshold LV _t2 ~
And a step of determining that the pitch P of the current block is twice the original pitch when at least one of (c3) is satisfied. 2. The input voice signal is divided into blocks on the time axis, the pitch corresponding to the fundamental period of the voice is detected for each block, and the voice signal is analyzed for each block based on this pitch, In the voice signal analysis / synthesis method for synthesizing a voice signal based on this analysis result, in the pitch detection, a typical pitch Pt unique to the speaker for each signal of each block and a pitch P of the current block are respectively detected. (C1) The pitch P of the current block is equal to or larger than a value in the vicinity of twice the typical pitch Pt, and the block average level LV is larger than a predetermined threshold value LV _t1 , (c2) the typical pitch Pt is larger than the typical pitch Pt ′ in the previous block and the pitch P of the current block is larger than a predetermined threshold th _p1 , (c3) the pitch P of the current block is Three conditions (c1) that are larger than the pitch P ′ of the previous block, the pitch P of the current block is larger than a predetermined threshold th _p2 , and the block average level LV is larger than a predetermined threshold LV _t2.
And a step of determining that the pitch P of the current block is twice the original pitch when at least one of (c3) is satisfied, and a double pitch determining step in the pitch detection. When it is determined that the pitch P of the current block is double the original pitch, the half pitch P / 2 of the pitch P / 2 is determined.
A speech analysis and synthesis method characterized in that synthesis is performed based on. 3. When analyzing the voice signal, the input signal is divided into bands based on the pitch P of the current block, and a voiced sound / unvoiced sound is discriminated for each of the divided bands. The speech analysis and synthesis method according to claim 2.