JP3235445B2 - Pitch detection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately find a pitch cycle fast with inexpensive constitution even when a speech waveform is a complicated waveform containing an overtone component. SOLUTION: A binarization part 8 and a timer 9 find and store successive zero-crossing intervals of a digital speech signal supplied from the precedent stage. A pitch arithmetic part 11 assumes that a pitch cycle is the sum 2n pieces of zero-crossing interval data as to n=1-4, calculates a reproduction rate as the degree of matching in each pitch cycle of each zero-crossing interval data constituting one pitch cycle, and employs the assumption by which the highest reproduction rate is obtained to find the pitch cycle.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pitch detecting device for detecting a pitch period or a pitch frequency of an audio waveform.

[0002]

2. Description of the Related Art A pitch period (or a pitch frequency) is one of parameters that characterize a speech waveform, and a technique for detecting the pitch period of the speech waveform is generally used in speech analysis / synthesis systems, speech encoding systems, and the like. It is used. Recently, some karaoke systems also detect a pitch period of a singer's voice, and are used for scoring singing.

Conventionally, there have been the following methods for detecting a pitch period of a voice. (1) Zero-cross method Assuming that the speech waveform is very close to a sine wave, the speech waveform crosses the zero level line from the negative direction to the positive direction, then from the positive direction to the negative direction, and again from the negative direction to the positive direction. In order to repeat the monotonous change of traversing the zero level line, the pitch period is given by the time interval traversing the zero level line in the same direction. The zero-cross method is a method of simply measuring the interval between two zero-crosses to determine the pitch period in accordance with this idea. Further, as a similar idea, there is a method of measuring an interval between timings at which the instantaneous value of the audio waveform becomes a maximum value or a minimum value, and uses the measured interval as a pitch cycle.

(2) Autocorrelation method In this autocorrelation method, time-series samples x (1), x (2),... Obtained by sampling a speech waveform at fixed sampling periods are used, and The pitch period is obtained by calculating the autocorrelation function R (r). R (r) = 1 / N · Σ {x (n) · x (n + r)} (where, in the above formula, Σ is an operator for obtaining a sum within ｛｝ within a range of n = 1 to N · r). That is, r is variously changed, an autocorrelation function R (r) is obtained for each r, and the pitch period of the speech waveform is calculated from r when R (r) becomes maximum (ie, the autocorrelation becomes maximum). calculate.

[0005]

By the way, the above-mentioned zero-crossing method can detect the pitch period relatively inexpensively and at high speed, but on the other hand, since human speech contains many harmonic components, it is accurate. There is a problem that a proper pitch period cannot be detected. Although the above autocorrelation method can detect the pitch period with some accuracy, it requires a large amount of calculation and a long detection time. In addition, the cost is increased.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a pitch detecting apparatus which overcomes the problems of the above two pitch detecting methods and which can detect a pitch period accurately and at high speed with an inexpensive configuration.

[0007]

The invention according to claim 1 is
A zero-crossing interval measuring means for measuring continuous zero-crossing intervals t ₁ , t ₂ ,... Of the voice waveform, and changing n (n is an integer of 1 or more) variously. Assuming that the total sum T = (t ₁ + t ₂ +... T _2n ) is the pitch period, and based on the zero-cross intervals t ₁ , t ₂ ,..., M adjacent periods (m is an integer of 2 or more) Pitch detection means for calculating the degree of coincidence of the voice waveforms between each pitch period and selecting n having the highest degree of coincidence of the voice waveforms to determine the pitch period. The device is the gist. The invention according to claim 2 is characterized in that the pitch calculation means sets the m
Is set to 4, and when the n is larger than a predetermined value, the m is set to 3
The gist of the present invention is a pitch detection device according to claim 1. The invention according to claim 3, wherein the pitch calculation means has a time length that matches each of the 2n zero cross intervals constituting the pitch cycle within a predetermined error range in the m pitch cycles. 3. The pitch detecting apparatus according to claim 1, wherein the number of times that the sound waveform is reproduced is obtained, and the degree of matching of the audio waveforms is calculated based on the total number of times.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments will be described to make the present invention easier to understand. Such an embodiment shows one aspect of the present invention, and does not limit the present invention, and can be arbitrarily changed within the scope of the present invention.

A. Configuration of Embodiment FIG. 1 is a block diagram showing a configuration of an embodiment in which the present invention is applied to a karaoke system. The present embodiment relates to a part for grading a singer's song among constituent parts of a karaoke system. In FIG. 1, reference numeral 1 denotes a CD (compact disk) on which a digital music signal is recorded. The digital music signals recorded on the CD 1 are sequentially reproduced in synchronization with a clock having a sampling frequency fs = 44.1 kHz. Reference numeral 2 denotes a vocal extraction unit, and CD1
A signal corresponding to the vocal sound (hereinafter, referred to as a digital sample signal) is extracted from the digital music signal reproduced from the digital music signal. As an example, a digital sample signal can be obtained by extracting a signal in a frequency band including the audio band of the digital music signal reproduced from the CD 1 with a band-pass filter. If a medium on which only vocal sound is recorded can be used, a digital music signal reproduced from the medium may be used as a digital sample signal. 3 is a microphone, C
The singing voice of the singer who sings along with the reproduction of D1 is collected and output as an analog audio signal. Reference numeral 4 denotes an A / D converter, which converts an analog audio signal from the microphone 1 into the same sampling frequency fs = 44.1 as in the case of reproducing the CD1.
Sampling is performed in synchronization with a clock of kHz and converted into a digital audio signal.

Reference numeral 5 denotes a DC removing unit which removes DC from the digital audio signal and the digital sample signal supplied sequentially.
The digital audio signal and the digital sample signal from which components in a low frequency band that can be regarded as DC, for example, a band of 0 Hz to 50 Hz are removed are output. 6
Is an LPF (Low Pass Filter) which removes, for example, a component having a frequency of 500 kHz or more from each of the digital audio signal and the digital sample signal output by the DC removing unit 5 and outputs the signal. These DC removing units 5 and L
The PF 6 selects and outputs only the components within the band of 50 to 500 Hz for each of the digital audio signal and the digital sample signal.

Reference numeral 7 denotes a 4 times oversampling unit,
Digital audio signal and digital sample signal that have passed through LPF 6 (both sampling frequency fs = 44.1)
kHz), the signal is converted into a signal having a quadruple sampling frequency and output.

FIG. 2 shows the quadruple oversampling unit 7.
Of the digital audio signal or the digital sample signal (hereinafter, referred to as an input digital signal). In this figure, a latch 71 receives and holds an input digital signal when a clock corresponding to a sampling frequency fs is applied. Are cascaded after the latch 71 as shown. Each of these delay units 72 is supplied with a clock having a frequency four times the sampling frequency fs, so that the latch 7
The input signals held at 1 are sequentially shifted, and the input signals are sequentially delayed by one clock cycle to output delayed signals. Are multipliers, and 74, 74,... Are adders.
An interpolation operation for convolving a predetermined interpolation coefficient sequence with each of the output signals 72,... Is performed. With the above configuration, the interpolation operation is performed in synchronization with the clock having a frequency four times the sampling frequency fs, and the interpolated digital signal is sequentially output from the adder 74 at the final stage.

The quadruple oversampling unit 7 is a means provided for improving the accuracy in obtaining the pitch period. That is, in the present embodiment, the pitch period of each digital signal is obtained by measuring the time interval between the zero cross points of each of the digital audio signal and the digital sample signal. For this reason, in order to increase the measurement accuracy of the pitch period, it is necessary to increase the detection accuracy of the position of the zero cross point on the time axis. Therefore, by interposing this quadruple oversampling unit 7, the time density of each sample of the digital audio signal and the digital sample signal is quadrupled, and the detection accuracy of the position of each zero cross point is enhanced. In this example, oversampling is performed by curve interpolation. However, in view of the problem of cost, linear interpolation that can obtain a certain degree of accuracy can be used.

Reference numeral 8 denotes a binarizing unit which binarizes the levels of the digital audio signal and the digital sample signal output from the quadruple oversampling unit 7. This binarization basically determines whether the input digital signal is positive or negative based on the zero level, and outputs “1” when the input digital signal is positive and outputs “0” when the input digital signal is negative. is there. That is, the binarizing section 8 is a means for outputting a binary signal in which "0" / "1" is inverted every time the input digital signal crosses the zero level. However, in the present embodiment, when binarization is performed, the range of ± Δ around the zero level is used as a masking band, and even if there is a minute vibration in the ± Δ masking band in the input digital signal, the masking band is applied. The binary signal is prevented from being inverted by the minute vibration.

FIG. 3 exemplifies a circuit configuration necessary for processing one of the digital audio signal and the digital sample signal (hereinafter, referred to as an input digital signal) in the binarization section 8. In this figure, reference numeral 81 denotes an absolute value detection unit that detects the absolute value of an input digital signal. 82
Is a comparing unit, which compares the absolute value of the input digital signal detected by the absolute value detecting unit 81 with a predetermined value Δ, and when the absolute value exceeds Δ, “1”; Outputs “0”. Reference numeral 83 denotes a sample-and-hold unit, which outputs the input digital signal as it is (in a sample state) while "1" is output from the comparison unit 82, and outputs a comparison unit during which "0" is output from the comparison unit 82 The input digital signal immediately before the output signal of 82 changes from “1” to “0” is held and output (hold state). Reference numeral 84 denotes a comparison unit which performs a positive / negative determination of the output signal of the sample-and-hold unit 83 based on the zero level.
, And outputs a binary signal of “0” in the case of a negative value.

According to the above configuration, when the input digital signal is out of the range of ± Δ, it is output as it is via the sample and hold section 83. When the input digital signal falls within the masking band of zero level ± Δ, the value of the immediately preceding input digital signal is stored in the sample and hold section 83.
The binary signal output from the comparing section 84 is not inverted during the period in which the holding operation is performed. Therefore, when the input digital signal changes across the masking band of zero level ± Δ, the binary signal is inverted at the time when the crossing of the masking band is completed. On the other hand, if the input digital signal enters the masking band of zero level ± Δ but moves up and down within the masking band without crossing it, even if the input digital signal crosses the zero level, the sample hold unit Since the output signal value of 83 does not cross the zero level, the inversion of the binary signal does not occur.

In FIG. 3, the circuit preceding the comparing section 84 may be replaced with the circuit shown in FIG. This figure 4
, Reference numerals 85 and 86 denote comparison units, each of which compares the input digital signal with a reference level and outputs “1” when the input digital signal is higher than the reference level and outputs “0” when the input digital signal is lower than the reference level. The comparator 85 is given + Δ as a reference level, and the comparator 86 is given −Δ as a reference level. 87 is a latch for holding the input digital signal, and 88 is a selector for selecting and outputting the input digital signal or the output signal of the latch 87. Reference numeral 89 denotes a control unit.
Latch 87 and selector 88 based on each output signal of
Control. That is, it is as follows.

A. When the output signals of the comparison units 85 and 86 are both "1" or both are "0" This is the case where the input digital signal is outside the masking band of zero level ± Δ. In this case, the control unit 89 controls the latch 8
7 is set to a sample state, and the selector 88 outputs an input digital signal. b. When the output signal of the comparing unit 85 is “0” and the output signal of the comparing unit 86 is “1” This is the case where the input digital signal is inside the masking band of zero level ± Δ. In this case, the control unit 89 puts the latch 87 into the hold state when the input digital signal enters the masking band, and causes the selector 88 to output the output signal of the latch 87.

In FIG. 1, reference numeral 9 denotes a time interval at which the inversion of each output signal of the binarization unit 8 corresponding to the digital audio signal and the digital sample signal occurs, that is, a time at which a zero cross point of each digital signal occurs. Reference numeral 10 denotes a timer for measuring an interval, and reference numeral 10 denotes a RAM for storing the result of the timer 9.

FIG. 5 is a block diagram showing the timer 9 and the RAM 10 together with their control systems. In this figure, only a part necessary for processing corresponding to one of the digital audio signal and the digital sample signal is shown. FIG.
In the figure, reference numeral 91 denotes a delay unit, and 92 denotes an exclusive OR circuit. These constitute a differentiating circuit 90 for differentiating the binary signal output from the binarizing unit 8, and output a pulse every time the inversion of the binary signal occurs. The timer 9 is reset every time an output pulse from the differentiating circuit 90 is given, and counts a clock having a constant frequency of 4 fs after this reset until the next reset.

The count value of the timer 9 is given to the latch 93 as input data. The latch 93 receives and holds the count value of the timer 9 immediately before reset by receiving the output pulse from the differentiating circuit 90. The count value held in the latch 93 is the number of clocks of the frequency 4fs output between the time when the previous inversion of the binary signal is detected and the time when the current inversion is detected. It can be said that it represents the time interval in which. Therefore, hereinafter, the data held by the latch 93 will be referred to as zero cross interval data.

Each time the output pulse from the differentiating circuit 90 is applied, the write control unit 94 sequentially reads the zero-cross interval data in the latch 93, and the zero-cross interval data within a certain range is equal to or greater than a predetermined value (timer 9). When the count value is large, a limit is provided and the data is written into the RAM 10. When the count value is smaller than a predetermined value (the count value of the timer 9 is small), a limit is provided and the data is not written into the RAM 10 and discarded. As described above, only the zero-cross interval data within a certain range is stored in the RAM 10.
The reason why the data is written in is to prevent the use of zero-crossing interval data that is not appropriate as the time interval of the zero-crossing point of the audio signal in the calculation, thereby preventing an erroneous pitch period from being calculated.

The pitch calculation unit 11 in FIG.
The pitch cycle of each of the digital audio signal and the digital sample signal is calculated by referring to the zero-crossing interval data accumulated in 10.

Here, assuming that a digital audio signal or the like is a sine wave, it crosses the zero level line at the start point and the end point of one cycle of the sine wave, and only once at the intermediate point between these zero cross points. Cross with line. Therefore, the pitch period can be obtained by adding two consecutive zero-cross interval data.

However, since a digital audio signal or the like representing a human audio waveform contains many harmonic components, a waveform for one pitch period has three or more zeros between the start point and the end point of the pitch period. A cross point may be included. In such a case, a correct pitch period cannot be obtained even if two consecutive zero cross interval data are added.

Therefore, in the present embodiment, it is assumed that one pitch period has a length corresponding to the sum of 2n zero-crossing interval data for each of a plurality of types of integers n. Then, a pitch period is obtained under each assumption, and 1
The degree to which the occurrence timing of each zero crossing point in the pitch cycle matches between the pitch cycles is determined. The details of the detection of the degree of coincidence of the occurrence timing of the zero cross point will be described later. Then, the pitch cycle having the highest degree of coincidence is selected as the true pitch cycle. This is based on the property of the audio signal that a large waveform change does not occur within a short time.

Next, in FIG. 1, reference numeral 12 denotes a level detector, which detects the levels of the digital audio signal output by the A / D converter 4 and the digital sample signal output by the vocal extractor 2. , And outputs a signal representing each level.

Numeral 13 denotes a scoring unit, which shifts the pitch cycle of each of the digital audio signal and the digital sample signal obtained by the pitch calculator 11 and the level detector 12.
Comprehensively evaluate the difference between both signal levels obtained by
Score the singer's song. This scoring result is displayed on the display unit 14.

B. Operation of Embodiment Hereinafter, the operation of this embodiment will be described. When a singer selects a song, digital music signals are sequentially reproduced from the CD 1 corresponding to the song. And the vocal extraction unit 2
Thus, a digital sample signal is extracted from the digital music signal, and is output to the DC removing unit 5 and the level detecting unit 12. On the other hand, the singer starts singing by playing the CD1, and the singing voice is collected by the microphone 3 and output as an analog audio signal. The analog audio signal is converted to a digital audio signal via the A / D converter 4 and output to the DC removing unit 5 and the level detecting unit 12.

The digital audio signal and the digital sample signal sequentially pass through the DC removing unit 5 and the LPF 6 to remove a signal in an unnecessary frequency band, and represent a waveform composed of only components in the frequency band of a human voice. The signals are output to the four-times oversampling unit 7 as digital signals.

The digital audio signal and the digital sample signal are converted by the 4 × oversampling unit 7
Each is interpolated on the time axis, converted into a signal having a quadruple sampling frequency, output, and converted into a binary signal by the binarization unit 8.

FIG. 6 shows the quadruple oversampling unit 7.
Is an example of the operation. In FIG. 6A, a horizontal straight line is a zero level line. A plot of a circle is shown along a sinusoidal signal waveform. The latter plot is a digital audio signal (digital sample signal).
, And the former represents the original signal waveform which is the base of these original samples. In addition, three plots of x are interposed between plots of ○ representing each original sample, and these plots respectively represent the interpolated samples obtained by the quadruple oversampling unit 7. .

FIG. 6 (b) shows a binary signal obtained when only the original sample (marked by ○) is given to the binarizing section 8 without performing 4-fold oversampling. ): 4 times oversampling, original sample (○)
2 shows a binary signal obtained when an interpolation sample (marked by x) is given to the binarization unit 8. Note that, for convenience of description, these figures show examples in which the digital audio signal (digital sample signal) does not include vibration at a level smaller than the masking band of the binarization unit 8.

Here, the digital audio signal or the like is sampled at a constant sampling cycle regardless of the signal waveform. Therefore, when the digital audio signal or the like repeats the same waveform, as shown in FIG. 6A, the instantaneous value at which timing is sampled varies depending on each waveform. For this reason, if the sampling period is coarse, as shown in FIG. 6B, when the pitch period is switched, a binary signal having a different waveform may be obtained despite the same waveform. However, when binarization is performed after performing 4 times oversampling of a digital audio signal or the like as in the present embodiment,
As shown in FIG. 6C, a binary signal which is inverted at a timing close to the original zero cross point is obtained, and the problem shown in FIG. 6B is prevented.

FIGS. 7A to 7D exemplify the operation of the binarizing section 8. FIG. First, in FIG. 7A, a sinusoidal signal waveform represents a digital audio signal (digital sample signal) output from the 4 × oversampling unit 7, and a horizontal line represents a zero level line. FIG. 7 (b)
Shows the operation of the sample hold unit 83 in FIG. As shown in FIG.
Reference numeral 3 denotes a sample state when the digital audio signal (digital sample signal) as an input signal is outside the masking band of zero level ± Δ (denoted by “S” in the figure), and masking of zero level ± Δ. If it is inside the band, it will be in the hold state (indicated as "H" in the figure)
Is done. As a result of such control of the sample and hold unit 83, the signal waveform input to the comparison unit 84 is as shown in FIG.
7C, and the binary signal obtained from the comparing unit 84 is as shown in FIG. 7D. As described above, when the digital audio signal (digital sample signal) changes across the masking band of zero level ± Δ, the binary signal is inverted when the crossing of the masking band is completed. Even if the digital audio signal (digital sample signal) includes a minute vibration portion having an amplitude of ± Δ or less, the digital audio signal (digital sample signal) is within the masking band of zero level ± Δ. Since the sample hold unit 83 performs the previous value holding operation, the binary signal is not inverted in the vibration part.

In the present embodiment, the pitch period is calculated using the zero-cross interval. Therefore, if too many zero-cross intervals are detected for the input digital signal waveform corresponding to one pitch period, the pitch period is calculated. The computational burden increases. However, in the present embodiment, since the binary signal is generated by the binarizing unit 8 having the masking band as described above, the fine movement near zero level which is not important for the calculation of the pitch period in the input digital signal is generated. A binary signal that is ignored and does not include the “0” / “1” inversion point more than necessary is obtained, and it becomes possible to detect an appropriate number of zero-cross intervals for the calculation of the pitch period.

As described above, a binary signal is generated based on each of the digital audio signal and the digital sample signal. Then, for each binary signal, the time interval at which "1" / "0" inversion occurs is sequentially timed by the timer 9, and the zero-crossing interval data as the timed result is stored in the latch 93 shown in FIG.
Are sequentially held. The zero-cross interval data sequentially held in the latch 93 in this manner is sequentially written into the RAM 10 under the control of the write control unit 94. That is, the write control unit 94 operates the differentiating circuit 90 by inverting the binary signal.
In response to the output of the pulse from, a write control routine whose flow is shown in FIG. 8 is executed. First, the write control unit 94
Fetches the zero-cross interval data t from the latch 93 (step S1), and determines whether or not the zero-cross interval data t is equal to or larger than a lower limit value “8”. If the result of this determination is "NO"
In this case, the routine ends without writing the zero cross interval data t. If the determination result of step S2 is "YE
In the case of "S", the process proceeds to step S3, and it is determined whether or not the zero cross interval data t is larger than the upper limit value "8192".
If the result of this determination is "NO", the zero cross interval data t
Is written into the RAM 10 (step S4), and the routine ends. On the other hand, if the determination result of step S3 is “YES”, “8” is used instead of the acquired zero cross interval data t.
192 "is written in the RAM 10 (step S5), and the routine ends. By the above control, “8” to “819”
Since only the zero-cross interval data within the range of "2" is written into the RAM 10, the zero-cross interval data that is not appropriate as the time interval of the zero-cross point of the audio signal is used for the calculation, and an erroneous pitch period is calculated. Can be prevented.

The zero-crossing interval data stored in the RAM 10 is referred to by the pitch calculator 11 to determine the pitch periods of the digital audio signal and the digital sample signal. Here, with reference to FIG. 9, an outline of the process of calculating the pitch cycle of the digital audio signal will be described as an example. Assuming that a digital audio signal as illustrated in FIG. 9A is given to the binarizing unit 8, the zero-cross interval data t ₁ , t ₂ ,.
It is stored in 0. The pitch calculation unit 11 makes the following four assumptions regarding the relationship between the zero-cross interval data t ₁ , t ₂ ,... And the pitch period of the digital audio signal, and examines the validity of each. The pitch period is obtained according to

Assumption 1 The pitch period of the digital audio signal has a length T ₁ corresponding to the sum of _two zero-cross interval data t ₁ and t ₂ . That is, the times T ₁₁ , T ₁₂ ,... Shown in FIG. 9 (b1) are the pitch periods of the digital audio signal. Pitch period hypotheses 2 digital audio signal has a length T ₂ corresponding to the sum of the four zero-cross interval data t ₁ ~t _4. That is, the time T _21, T ₂₂ shown in FIG. 9 (b2), ... is the pitch period of the digital audio signal. The pitch period of the hypothetical 3 digital audio signal has a length T ₃ corresponding to the six zero sum of cross interval data t ₁ ~t _6. That is, the times T ₃₁ , T ₃₂ ,... Shown in FIG. 9 (b3) are the pitch periods of the digital audio signal. Pitch period hypotheses 4 digital audio signal has a length T ₄ corresponding to the sum of the eight zero cross interval data t ₁ ~t _8. That is, the times T ₄₁ , T ₄₂ ,... Shown in FIG. 9B4 are the pitch periods of the digital audio signal.

The examination of the validity of each of the above assumptions and the calculation of the pitch period based on the examination result are executed according to the flow shown in FIG. First, the pitch calculator 11 calculates the recall rate CR1 of the waveform of the digital audio signal on the assumption of the assumption 1 (step S101). This recall is a numerical value indicating how much each digital audio signal waveform corresponding to each pitch period matches under the above assumptions. In the present embodiment, the zero cross interval data t ₁ , Calculated based on t ₂ ,.

Here, the procedure for calculating the recall rate CR1 performed in step S101 will be described with reference to the flowchart in FIG. First, step S
In step 201, "0" and "1" are respectively set as initial values for the counter CNT and the control variable i.

Next, the process proceeds to step S202, where the control variable i
Is increased by “2”, and i = “3”. Next, the process proceeds to step S203, and it is determined whether or not the condition of 0.9t ₁ −t _i <0 is satisfied, that is, whether or not the zero cross interval data t ₃ is larger than 90% of the zero cross interval data t _1. I do. If the result of this determination is "YES", the counter CNT is increased by "1" (step S204),
Proceed to step S205, and if “NO”, step S205
The process proceeds to step S205 without the intervention of 204. Then proceeds to step _{S205, -1.1t 1 + t i <} 0 satisfies the condition or not made, i.e., the zero cross interval data t ₃
Is smaller than 110% of the zero cross interval data t ₁ . If the result of this determination is “YES”, the counter CNT is increased by “1” (step S
206), and the process proceeds to step S207. If “NO”, the process proceeds to step S207 without passing through step S206.

Next, in step S207, it is determined whether or not the control variable i has become "7". If the result of this determination is "NO", the flow returns to step S202. Thereafter, step S202~S207 over twice is performed, zero for each of the cross interval data t ₅ and t ₇ the determination in steps S203 and S205 are performed, the zero-cross interval data the zero cross interval data t _The counter CN when it is larger than 90% of ₁ or smaller than 110%.
T is incremented (steps S204, S204).
206).

When i = “7”, step S
The result of determination in step 207 is “YES”, and step S20
Proceeding to step 8, the control variable i is set to "2".

Next, proceeding to step S209, the control variable i is increased by "2", and i is set to "4". Next, the process proceeds to step S210, where it is determined whether or not the condition 0.9t ₂ −t _i <0 is satisfied, that is, whether or not the zero cross interval data t ₄ is greater than 90% of the zero cross interval data t _2. I do. If the result of this determination is "YES", the counter CNT is increased by "1" (step S21).
1) The process proceeds to step S212. If “NO”, the process proceeds to step S212 without going through step S211.
Next, when the process proceeds to step S212, -1.1t ₂ + t _i <0
Satisfying whether made, i.e., to determine the zero or cross interval data t ₄ is less than 110% of the zero-cross interval data t _2. If the result of this determination is "YES"
In the case of, the counter CNT is increased by “1” (step S213), and the process proceeds to step S214. In the case of “NO”, the process proceeds to step S21 without going through step S213
Proceed to 4.

Next, in step S214, it is determined whether or not the control variable i has become "8". If the result of this determination is "NO", the flow returns to step S209. Thereafter, step S209~S214 over twice is performed, zero for each of the cross interval data t ₆ and t ₈ determines in step S210 and S212 are performed, the zero-cross interval data the zero cross interval data t If it is larger than 90% of ₂ or smaller than 110%, the counter CN
T is incremented (steps S211, S2).
213).

When i = 8, step S
The result of the determination at 214 is “YES”, and step S21
Proceeding to 5, the value of the counter CNT is normalized by the number of determinations on the zero-cross interval data, and the result is set as the recall rate CR1. In the case of this flow, since the determination is performed 12 times, CNT / 12 is set as the recall rate CR1.

Here, the assumption that the length of the pitch period is assumed to be the sum T _{1 of} two zero-cross interval data is correct, and that the waveform of the digital audio signal does not change even if the pitch period is switched four times. , T ₁ = t ₃ = t ₅ = t ₇ and t ₂ = t ₄ = t ₆ = t ₈ . Therefore, in this case, the recall rate CR1 obtained by the above processing is 100%.
Also, even if each zero cross interval data has some errors,
If t ₃ , t ₅ and t ₇ fall within the range of t ₁ ± 10%, and t ₄ , t ₆ and t ₈ fall within the range of t ₂ ± 10%, The recall rate CR1 becomes 100%. On the other hand, if the above assumption is erroneous, a large difference occurs between the mutually corresponding zero-cross interval data due to the switching of the pitch period. For this reason, it is easy to make a negative determination in step S203 and the like, and
As the number of times such a negative judgment is made increases, the recall ratio CR1 decreases.

When the calculation of the recall rate CR1 is completed in this manner, the flow returns to the flow of FIG. 10 and proceeds to step S102, where the recall rate CR2 of the waveform of the digital audio signal under the assumption 2 is calculated. That is, it is assumed to have a length T ₂ in which the pitch period corresponds to the sum of the four zero-cross interval data. Then, based on the zero-cross interval data t _{1 to} t ₄ corresponding to the first pitch period, respectively, the zero-cross interval data t ₅ to t ₅ corresponding to the second, third, and fourth pitch periods, respectively. t _8, t ₉ each ~t ₁₂ and t ₁₃ ~t ₁₅ determines whether or not the match in the reference and the predetermined error range. Then, the number of times a positive determination result is obtained is counted, and normalized by the total number of determinations, thereby obtaining a recall rate CR2.

In an ideal state where the assumption that the pitch period length is the sum of four zero-cross interval data is correct and the digital audio signal waveform does not change even if the pitch period is switched four times, t ₁ = T ₅ = t ₉ = t ₁₃ t ₂ = t ₆ = t ₁₀ = t ₁₄ t ₃ = t ₇ = t ₁₁ = t ₁₅ t ₄ = t ₈ = t ₁₂ = t ₁₆ CR2 becomes 100%. Even if there is some error in each zero-cross interval data, if it falls within the range of ± 10%, the recall rate CR
2 becomes 100%. When the switching of the pitch cycle produces zero cross interval data greatly deviating from the reference (that is, the zero cross interval data corresponding to the first pitch cycle), the recall rate CR2 depends on the number of the zero cross interval data.
Will decrease.

Next, the process proceeds to step S103, where the above assumption 3
Reproducibility C of the digital audio signal waveform on the assumption of
Calculate R3. That is, it is assumed to have a length T ₃ of the pitch period corresponds to the sum of the six zero cross interval data. Then, the zero cross interval data t ₁ ~t ₆ corresponding to the first pitch period, respectively as a reference, the second, third and fourth zero corresponding to the pitch period of the cross interval data t ₇ ~ t _{1 2,} t ₁₃ ~t ₁₈ and t ₁₉ ~t ₂₄
Are determined to be equal to the reference within a predetermined error range. Then, the number of times a positive determination result is obtained is counted, normalized by the total number of determinations, and the recall C
Find R3.

The recall rate CR3 is calculated as follows: t ₁ = t ₇ = t ₁₃ = t ₁₉ t ₂ = t ₈ = t ₁₄ = t ₂₀ t ₃ = t ₉ = t ₁₅ = t ₂₁ t ₄ = t ₁₀ = t ₁₆ = T ₂₂ t ₅ = t ₁₁ = t ₁₇ = t ₂₃ t ₆ = t ₁₂ = t ₁₈ = t _{24 When} all the conditions are satisfied or each zero cross interval data has some error, the range is ± 10%. If the error is within the range, the recall rate CR3 becomes 100%. Further, when zero-cross interval data having a large error occurs, the recall rate CR3 decreases in accordance with the number of the data.

Next, the process proceeds to S104, in which the reproduction rate CR3 of the waveform of the digital audio signal on the assumption of the above assumption 4 is calculated. That is, it is assumed to have a length T ₄ the pitch period corresponds to the sum of eight zero cross interval data.
Then, based on the zero cross interval data t _{1 to} t ₈ corresponding to the first pitch cycle, respectively, the zero cross interval data t ₉ to t ₉ corresponding to the second and third pitch cycles are respectively used.
Each of t ₁₆ and t ₁₇ ~t ₂₄ determines whether or not the match in the reference and the predetermined error range. Then, the number of times a positive determination result is obtained is counted, and normalized by the total number of determinations, thereby obtaining a recall rate CR4.

[0054] In the process up to the step S101~S103 is has been processed the pitch period of the 4 pieces of, T ₄₁ through T ₄₃ in the pitch period corresponding to three in step S104 (FIG. 9 (b4) ) Is to be processed. This is for the following reason. That is,
In step S104, a long period of time corresponding to eight zero-crossing interval data is assumed as the pitch period. Therefore, if it is assumed that four pitch periods are to be processed in step S104, even if assumption 4 is correct, the digital audio signal waveform becomes stable for an extremely long period of four pitch periods. If not, the recall rate CR4 will decrease. However, although the waveform of the digital audio signal can maintain the same waveform for a certain short time, the waveform changes after a certain time. Therefore, if four pitch periods are to be processed, even if assumption 4 is correct,
There is a high possibility that an unduly low recall rate CR4 will be calculated due to the influence of a temporal change in the waveform of the digital audio signal. Thus, in step S104, three pitch periods are to be processed as described above.

In step S104, the recall rate CR4
Is t ₁ = t ₉ = t ₁₇ t ₂ = t ₁₀ = t ₁₈ t ₃ = t ₁₁ = t ₁₉ t ₄ = t ₁₂ = t ₂₀ t ₅ = t ₁₃ = t ₂₁ t ₆ = t ₁₄ = t ₂₂ Reproduced when all the conditions of t ₇ = t ₁₅ = t ₂₃ t ₈ = t ₁₆ = t ₂₄ are satisfied or when each zero cross interval data has an error within ± 10% even if there is some error. The rate CR4 becomes 100%. Further, when zero-cross interval data having a large error is generated, the recall rate CR4 is reduced according to the number thereof.

Next, the process proceeds to step S105, where assumptions 1 to 4 are obtained based on the recall rates CR1 to CR4 obtained as described above.
4 is determined to be appropriate. FIG. 12 shows a detailed flow of this determination. First, step S301
To determine which of the recall rates CR1 to CR4 is the largest. If the recall rate CR1 is the maximum, it is determined whether or not this CR1 is larger than a predetermined reference value ref (step S302).
To follow the assumption 1 in the case of ES ", i.e., the length T ₁ of the 2 pieces of zero crossing interval data and obtaining the pitch period. The same applies to the case where the other recall rates CR2 to CR4 are maximum, and it is determined whether CR2 or the like is larger than a predetermined reference value ref (steps S303 to S30).
5) If the result of this determination is “YES”, the length T ₂ of four zero-crossing interval data and the six zero-crossing interval data according to the assumption on which the calculation of each recall is based. and determining the pitch period by the length T ₃ or a zero cross interval data of 8 pieces of length T ₄ of. If the recall rates are the same, the priority order is CR1>CR2> CR3
> CR4 (CR1 is the highest priority).

On the other hand, if the largest one of the recall rates CR1 to CR4 is equal to or less than the reference value ref, step S3
No matter which of the steps from 02 to S305, the determination result is “NO”. In this case, it cannot be concluded which of the assumptions 1 to 4 is appropriate, and the result of the determination is that there is no such case.

When the above judgment is completed, the flow returns to the flow shown in FIG. 10, and proceeds to the step corresponding to the judgment result. That is, if it is determined that the pitch period is to be obtained based on the length T ₁ of the two zero-cross interval data, step S1 is performed.
Proceeds to 06, T ₁₁ ~T ₁₄ each corresponding to two zero crossing interval pitch cycle four cycles determined comprising data (FIG. 9 (b1)
), And these average values are used as the pitch period of the digital audio signal. Further, the process proceeds to step S107 when determining that to determine the pitch period by four minutes of the zero cross interval data length T _2, 4 cycles determine the pitch period according to the determination result (FIG. 9 (b2) corresponds to T ₂₁ ~T _24), these average values and the pitch period of the digital audio signal. In addition, the length T ₃ of the zero cross interval data for six pieces
Proceeds to step S108 when it is determined that obtaining the pitch period, the pitch period four periods determined according to the determination result (equivalent to T ₃₁ through T ₃₄ in FIG. 9 (b3)), the
The average of these values is used as the pitch period of the digital audio signal. Then, the process proceeds to step S109 when determining that to determine the pitch period of eight length of content of the zero cross interval data T _4, the three cycles determine the pitch period according to the determination result (FIG. 9 (b4) corresponds to T ₄₁ ~T _43), these average values and the pitch period of the digital audio signal.

When the above processing is completed, step S10
1 and the same processing is repeated. Thus, the pitch period of the digital audio signal is output continuously. On the other hand, in the determination of FIG. 12, when the conclusion of “not applicable” is obtained, the calculation of the pitch period is not performed, and a signal indicating that the calculation of the pitch period is not performed is output.
It returns to step S101. In the above description, the pitch period calculation process has been described by taking the case of a digital audio signal as an example. However, the pitch period is also calculated for a digital sample signal by exactly the same process.

As described above, this embodiment is based on assumptions 1-4.
For all of the above, the recall is determined, the assumption that gives the highest recall is selected, and the pitch calculation based on the selected assumption is performed only when the recall is within the allowable range. It is a careful procedure that does not take place in some cases. The reasons for this careful procedure are as follows.

A. As a procedure other than the above procedure, for example, the respective recall rates corresponding to assumptions 1 to 4 are sequentially calculated, and when the recall rate within the allowable range is obtained, the calculation is terminated, and the assumption that the recall rate is obtained is obtained. There is an alternative that can be selected to determine the pitch period. However, depending on the audio waveform, for example, a situation may occur in which the recall rates corresponding to assumptions 1 and 3 are within an allowable range, and the recall rate corresponding to assumption 3 is higher than that of assumption 1. If this alternative is followed in such a case, Assumption 1 will be selected and an incorrect pitch period will be determined. It is conceivable to set the allowable range narrow so that the selection of the assumption is correctly performed, but in such a case, there is a possibility that a case determined to be “not applicable” may continue.

B. Further, an alternative is also conceivable in which all the recall rates corresponding to assumptions 1 to 4 are calculated, and the assumption in which the maximum recall rate is obtained is unconditionally adopted to obtain the pitch period. However, the recall rate corresponding to any of the assumptions is uniformly low, and there may be cases where the recall rate corresponding to a specific assumption is slightly superior to the others. Even if the pitch period is forcibly obtained by adopting, there is no guarantee that an accurate pitch period can be obtained. For example, when the waveform of the digital audio signal suddenly changes in pitch cycle, the recall is likely to be low under any of the above assumptions.

C. Therefore, in the present embodiment, the calculation of the pitch period is performed according to the above-described procedure, and the output of the inappropriate pitch period is prevented.

The pitch periods of the digital audio signal and the digital sample signal obtained as described above are sequentially reported to the scoring unit 13, and the difference between the pitch periods of the two signals and the two signal levels obtained by the level detection unit 12. The singers' songs are scored by the comprehensive evaluation with the deviation, and the scoring results are displayed on the display unit 14.

C. Evaluation Results of Apparatus According to the Present Embodiment The operating conditions of each part of the above-described pitch cycle detection apparatus were set variously, and the pitch cycle detection time and the detection error were evaluated. 13 to 16 show the results. First, FIG. 13 shows the result of using a circuit that performs linear interpolation as the 4 × oversampling unit 7 and changing the oversampling frequency of this circuit in various ways, and measuring the pitch period detection error in a practical range. From this result, it can be understood that the detection error in the practical range can be sufficiently reduced by performing interpolation of about four times oversampling. Next, FIG. 14 shows a case in which the pitch period is detected by a correlation between three periods (m = 3) and a case in which the pitch period is determined by a correlation between four periods (m = 4). It shows the result of measuring the delay time for each input frequency. As shown by the experimental results, when m = about 3 or 4, the detection delay can be kept within a range in which there is no problem. FIG. 15 shows the relationship between the number of times of averaging and the pitch period extraction error. FIG.
Shows the result of an experiment on how many cycles (pitch cycles) in the past can be compared with a waveform to accurately extract a pitch cycle. The results of this experiment show that comparing the past two cycles has many errors, and comparing the input waveforms of the past five or more cycles results in a waveform that is too old and incorrectly changes the pitch cycle. It indicates that comparing input waveforms over four cycles is optimal for accurate pitch extraction.

D. Modifications (1) In the above embodiment, the pitch period is calculated by adding the sum of 2n pieces of zero cross interval data to each other, based on how much each zero cross interval data forming one pitch period matches between each pitch period. A determination was made as to whether the assumptions made were appropriate. Instead of this method, for each n, a predetermined number of pitch periods is obtained by calculating the sum of 2n zero-cross interval data, and n having the smallest variation in these pitch periods is selected. You may make it select. That is, in FIGS. 9 (b1) to 9 (b4), T
₁₁ when the variations of the through T ₁₄ is the smallest and pitch period average value of T ₁₁ through T _14, if the variation of T ₂₁ through T ₂₄ is the smallest and pitch period average value of T ₂₁ through T _24, ... That is, the pitch period is obtained. Further, the determination method based on the zero-cross interval data disclosed in the above embodiment and the determination method for determining the variation of the pitch cycle are used together, and the zero-cross interval data and the pitch cycle length are comprehensively evaluated for the pitch cycle variation, A pitch cycle may be selected.

(2) In the above embodiment, the binarizing section 8
The width Δ of the masking band was fixed. However, since the amplitude of the minute vertical movement of the voice waveform generated near the zero level depends on the amplitude of the entire voice waveform, it may be difficult to determine an appropriate Δ. Therefore, the amplitude Δ of the digital audio signal or the digital sample signal is detected, the amplitude value is multiplied by a predetermined coefficient, and the result is set as Δ. It is preferred to do so. (3) In the above embodiment, the pitch period is obtained by digital processing. However, the zero crossing interval may be obtained directly from the analog voice waveform, and the pitch period may be obtained based on the result.

[0068]

As described above, according to the present invention,
The continuous zero-cross interval of the voice waveform is obtained, and for each n, the pitch period is assumed to be the sum of 2n zero-cross interval data, and between each pitch period of the zero-cross interval data constituting one pitch period. Or the pitch of each pitch cycle, and the pitch cycle is determined by using the assumption that the best match is obtained, so that when the audio waveform is a complex waveform containing harmonic components, However, there is an effect that the pitch period can be quickly and accurately obtained with an inexpensive configuration.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a configuration example of a 4 × oversampling unit in the embodiment.

FIG. 3 is a block diagram illustrating a configuration of a binarizing unit according to the embodiment;

FIG. 4 is a block diagram illustrating a configuration of a binarizing unit according to the first embodiment;

FIG. 5 is a block diagram showing a timer, a RAM, and a control system thereof in the embodiment.

FIG. 6 is a diagram showing an operation of a 4 × oversampling unit in the embodiment.

FIG. 7 is a diagram showing an operation of a binarization unit in the embodiment.

FIG. 8 is a diagram showing an operation of a write control unit in the embodiment.

FIG. 9 is a diagram for explaining an outline of pitch period calculation processing in the embodiment.

FIG. 10 is a flowchart showing a pitch cycle calculation process in the embodiment.

FIG. 11 is a flowchart showing pitch period calculation processing in the embodiment.

FIG. 12 is a flowchart showing pitch period calculation processing in the embodiment.

FIG. 13 is a diagram showing a performance evaluation result of the embodiment.

FIG. 14 is a diagram showing a performance evaluation result of the embodiment.

FIG. 15 is a diagram showing a performance evaluation result of the embodiment.

FIG. 16 is a diagram showing a performance evaluation result of the embodiment.

[Explanation of symbols]

1 ... CD, 2 ... vocal extraction unit, 3 ... microphone, 4 ... A / D converter, 5 ... DC removal unit, 6 ... L
PF, 7: 4 × oversampling section, 8: Binarization section, 9: Timer (zero-crossing interval measuring means), 10:
... RAM (zero-crossing interval measuring means), 11... Pitch calculating section (pitch calculating means), 12.
... Scoring section, 14 ... Display section.

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Yusuke Yamamoto 10-1 Nakazawa-cho, Hamamatsu City, Shizuoka Prefecture Inside Yamaha Corporation (56) References JP-A-62-115499 (JP, A) JP-A-62- 115500 (JP, A) JP-A-62-54296 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G10L 11/04

Claims (3)

(57) [Claims] 1. A continuous zero cross interval t ₁ ,
a zero-cross interval measuring means for measuring t ₂ ,..., n (n is an integer of 1 or more) being varied, and for each n, a sum T of 2n zero-cross intervals T = (t ₁ + t ₂ +...)
t _2n ) is assumed to be the pitch period, and the zero-cross interval t ₁ ,
Based on t ₂ ,..., the degree of coincidence of the audio waveforms between pitch periods of adjacent m periods (m is an integer of 2 or more) is calculated, and n having the highest degree of coincidence of the audio waveforms is calculated. A pitch calculating means for obtaining a pitch cycle by selecting the pitch detecting means. 2. The apparatus according to claim 1, wherein said pitch calculation means sets said m to 4 when said n is smaller than a predetermined value, and sets said m to 3 when said n is larger than a predetermined value. Pitch detection device. 3. The pitch calculating means reproduces, for each of the 2n zero-cross intervals constituting the pitch cycle, those having a time length that coincides within a predetermined error range in the m pitch cycles. 3. The pitch detecting apparatus according to claim 1, wherein the number of times the voice waveforms are obtained is calculated, and the degree of coincidence of the audio waveforms is calculated based on the total number of times.