JPH11305794A - Pitch detecting device and information medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a pitch from features of a voice signal waveform with a small amount of arithmetic operations without obtaining the correlation of the voice signal. SOLUTION: This device has a buffer memory 4 which segments a digital voice signal by frames of about 30 milliseconds, a low-pass filter circuit 3 which filters the voice signal, a peak search circuit 5 which detects a local maximum and a local minimum from the low-pass filtered voice signal waveform, a data analyzing circuit 7 which obtains the local maximum and local minimum having the largest absolute values of amplitudes among respective local maximums and local minimums, obtains local maximums and local minimums within specified ranges among the obtained local maximums and local minimums, and obtains as common histograms the histograms of time intervals obtained among the local maximums and local minimums, and a pitch determining circuit 8 which determines the highest frequency of the histograms as the pitch of the voice signal.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is suitable for, for example, converting the fundamental period (pitch) of a singer's voice or a chorus's voice, or the nature of a voice. In addition, the present invention relates to a pitch detection device for detecting a pitch frequency corresponding to a voice pitch or a reciprocal of the pitch period with a small amount of calculation, and an information medium for recording or transmitting program data for realizing the pitch detection.

[0002]

2. Description of the Related Art Conventionally, when extracting speech features, in many cases, instead of directly treating speech waveforms, the speech waveforms are converted into spectrum-related features such as a frequency spectrum and an autocorrelation function. In other words, the speech waveform can be considered to be composed of the sum of sine waves whose amplitude and phase gradually change over time, and an important feature in the perception of speech by human hearing is mainly amplitude information. Since the phase information is included and the phase information usually does not play an important role, when extracting the feature of the voice, the voice signal is often converted into a spectrum-related feature. Note that the power spectrum density of each short-time section of voice, that is, the short-time spectrum, is a spectrum envelope that is a component that changes slowly with frequency, a component that changes finely and periodically in the case of voice, and a case that there is no voice. And a non-periodically changing component can be considered.

Here, one of the features of speech is a frequency (pitch frequency) corresponding to the pitch of the voice, or a cycle (pitch cycle) that is the reciprocal thereof, and a typical technique for detecting the pitch. As the method, a so-called autocorrelation method and a modified correlation method are well known.

[0004]

However, these conventional pitch detection methods require the calculation of the correlation of the audio signals, so that the amount of calculation becomes enormous, and it is necessary to detect the pitch frequency and the pitch period in real time. Requires an expensive arithmetic processing device capable of executing high-speed arithmetic processing.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and makes it possible to detect a pitch with a small amount of calculation from the characteristics of an audio waveform without finding a correlation between audio signals.
It is another object of the present invention to provide a pitch detection device that can also determine whether a voice signal is a vowel or another signal, and an information medium that records or transmits program data for realizing the pitch detection.

[0006]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, a pitch detection device according to the present invention cuts out a digitally input audio signal every unit time of a predetermined length, and extracts a desired signal from the audio signal. Passes only the frequency band, detects a plurality of maximum points and minimum points from the audio signal waveform of the desired frequency band, and finds a maximum point at which the absolute value of the amplitude is maximum from among the plurality of maximum points in a unit time. From among a plurality of minimum points in a unit time, a minimum point where the absolute value of the amplitude is maximum is determined, a maximum point where the amplitude falls within a predetermined range with respect to the maximum maximum point is determined, and the maximum minimum point is determined. On the other hand, a minimum point where the amplitude falls within a predetermined range is determined, a time interval is determined between each maximum point falling within the predetermined range, and a time interval is determined between each minimum point falling within the predetermined range. Determine the interval histogram as a common one, The mean value and the Sutoguramu taken out as the pitch of the speech signal.

[0007] In order to solve the above-mentioned problems, a pitch detection device according to the present invention cuts out a digitally input audio signal every unit time of a predetermined length and passes only a desired frequency band from the audio signal. Detecting a plurality of local maximum points and local minimum points from an audio signal waveform in a desired frequency band, finding a local maximum point at which the absolute value of the amplitude is maximum from among the local maximum points within a unit time, and From among the local minimum points, the local minimum point where the absolute value of the amplitude is maximum is determined, the local maximum point whose amplitude is larger than a predetermined range with respect to the maximum local maximum point is obtained, and the amplitude is predetermined with respect to the maximum local maximum point. Obtain a local maximum point smaller than the range, obtain a local minimum point whose amplitude is larger than a predetermined range for the maximum local minimum point, obtain a local minimum point whose amplitude is smaller than the predetermined range for the maximum local minimum point, It between each maxima greater than the range of The time interval is determined between each local maximum point smaller than the predetermined range, the time interval is determined between each local minimum point larger than the predetermined range, and the local time interval between the local minimum points smaller than the predetermined range. , A time interval is obtained, a histogram of the time interval is obtained as a common one, and the most probable value of the histogram is extracted as the pitch of the audio signal.

Here, in the pitch detecting apparatus according to the present invention, the most probable value of the histogram is compared with its surrounding values, and the vowel / non-vowel of the voice is determined based on the comparison result.

Further, in order to solve the above-mentioned problems, the information medium according to the present invention records or transmits program data for realizing the pitch detecting device according to the present invention.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of a pitch detecting device and an information medium according to the present invention will be described below in detail with reference to the drawings.

FIG. 1 shows a schematic configuration of a pitch detecting device according to a first embodiment of the present invention.

In FIG. 1, for example, a waveform signal as shown in FIG. 2 is supplied to an input terminal 1 as an analog audio signal, and supplied to an analog / digital (A / D) converter 2. The analog / digital converter 2 converts the analog audio signal waveform in FIG. 2 into a digital signal by performing a sampling process at a sampling frequency of, for example, 44.1 kHz. The audio data digitized by the analog / digital converter 2 is sent to the filter circuit 3.

The filter circuit 3 has an IIR (circular type).
Alternatively, it is an FIR (non-recursive) digital filter having a cutoff frequency of 300 as shown in FIG.
It is a low-pass filter (low-pass filter) having a frequency characteristic of Hz. By performing low-pass filtering in the filter circuit 3, the waveform signal as shown in FIG.
A high-frequency component as shown in FIG.
Although the output of the filter circuit 3 is actually digital data, it is represented as an analog waveform signal in the example of FIG. The output data from this filter circuit 3 is
The data is sent to the buffer memory 4.

The buffer memory 4 accumulates data from the filter circuit 3 and outputs the data for each processing unit time in the peak search circuit 5 at the subsequent stage, ie, outputs data for each sample number corresponding to the processing unit time. . In the present embodiment, the processing unit time is set to about 30 ms (more precisely, about 29 ms).
(m seconds, every 1280 samples in terms of the number of samples). The data for each frame (1280 samples) output from the buffer memory 4 is sent to the peak search circuit 5.

The peak search circuit 5 compares each sample data supplied from the buffer memory 4 for each sample, and extracts local peaks (maximum points and minimum points) of the audio data based on the comparison output. .

More specifically, the peak search circuit 5 compares, for each sample of the audio data, the current audio data with the audio data one sample before the current data, and determines that the current sample value is one sample before the sample data. After the comparison result indicating that the current sample value is smaller than the sample value of the immediately preceding sample after the comparison result indicating that the current sample value is smaller than the sample value of the current sample value, or Is followed by a comparison result indicating that the current sample value is smaller than the sample value of the previous sample, and then when a comparison result indicating that the current sample value is larger than the sample value of the previous sample is obtained, The time of change of those comparison results is
Local peak of audio data (maximum point or minimum point)
Detected as

That is, after the comparison result indicating that the current sample value is larger than the sample value of the immediately preceding sample followed by the comparison result indicating that the current sample value is smaller than the sample value of the immediately preceding sample. The comparison result indicating that the sample value immediately before the current sample value indicates the local maximum point (hereinafter, referred to as local maximum MA), while the current sample value is smaller than the sample value immediately before the current sample value. After that, the sample value immediately before the current sample value when the current sample value becomes larger than the sample value immediately before the current sample value indicates a minimum point (hereinafter, referred to as a local minimum MI).

Therefore, in the peak search circuit 5, based on the change in the sample comparison result, the sample data of the local maximum MA of the audio data,
And sample data of the local minimum MI. The sample data of the local maximum MA and the local minimum MI detected by the peak search circuit 5 are sent to the code data conversion circuit 6.

The sign data converting circuit 6 adds a positive (+) sign to the sample data of each local maximum MA supplied from the peak search circuit 5, as shown in FIGS. Then, a negative (-) sign is added to the sample data of each local minimum MI. FIG. 5 shows only the sample data of the local maximum MA and the sample data of each local minimum MI as an example. The local maximum MA converted into code data by the code data conversion circuit 6
And the respective sample data of the local minimum MI are sent to the data analysis circuit 7.

In the data analysis circuit 7, first, each local maximum (maximum value) MA to which a plus (+) sign is added.
And the absolute value of the amplitude of each local minimum (minimum value) MI to which a negative (-) sign is added by the sign data conversion circuit 6, and obtains the absolute value of these amplitudes in the frame. The local maximum MA and the local minimum MI having the largest absolute values are obtained. In the example of FIG. 5, the local maximum MA having the largest absolute value of the amplitude is represented as the first local maximum MAG, in particular.
The local minimum MI having the largest absolute value of the amplitude is particularly represented as a first local minimum MIG.

Next, the data analysis circuit 7 performs all local maximum MAs whose amplitude values fall within a predetermined range RT in FIG. 5 set corresponding to the first local maximum MAG and the first local maximum MA. All local minimums MI whose amplitude values fall within a predetermined range RU in FIG. 5 set corresponding to the local minimum MIG are obtained.

In the example of FIG. 5, the first local maximum M
The local maximum MA whose amplitude value falls within the predetermined range RT corresponding to the AG is represented in particular as a second local maximum MAR (MAR1, MAR2,...), And a first local minimum MIM. A local minimum MI whose amplitude value falls within the corresponding predetermined range RU, particularly a second local minimum MIR (M
IR1, MIR2,...).

Here, the predetermined range RT corresponding to the first local maximum MAG is the second local maximum MA.
R is a range in which a plurality of R can be obtained, and similarly, a predetermined range R corresponding to the first local minimum MIG.
U is a range in which a plurality of second local minimum MIRs can be obtained. That is, in the data analysis circuit 7, as will be described later, the statistical analysis of the time interval is performed. To improve the accuracy of the statistical analysis, the second local maximum MAR and the second local minimum MIR are used. This is because each requires a plurality.

The predetermined range RT can be set to any range between the amplitude value 0 in FIG. 5 and the amplitude value of the first local maximum MAG.
U can be set to an arbitrary range from the amplitude value 0 in FIG. 5 to the amplitude value of the first local minimum MIG.

As an example of the predetermined range RT and the predetermined range RU, among all the local maximums MA and the local minimums MI in an arbitrary frame, for example, the local maximum MA, which most specifically represents the pitch of the audio signal,
It is conceivable to set the range so that only the local minimum MI can be extracted. For example, when it is not preferable to use a local maximum MA or a local minimum MI having an extremely large absolute value of the amplitude, or a local maximum MA or a local minimum MI having an extremely small absolute value of the amplitude for the pitch detection, for example, It is desirable to set a predetermined range RT and a predetermined range RU from which the extremely large local maximum MA and local minimum MI and the extremely small local maximum MA and local minimum MI can be excluded. By using the predetermined range RT and the predetermined range RU, it is possible to extract only the local maximum MA and the local minimum MI suitable for pitch detection from all the local maximums MA and the local minimums MI in the frame. , Accurate pitch detection and all local maximums M in the frame
A, It is possible to detect the pitch with a much smaller amount of calculation than when pitch detection is performed using the local minimum MI.

Further, in the data analysis circuit 7, the first local maximum MAG and all the second local maximums MAR (MAR1, MAR) obtained as described above are obtained.
2, MAR3,...) To measure the time interval (for example, the number of samples) between the respective local maximums,
Similarly, the first local minimum MIG and all the second local minimums MIR (MIR1, MIR2, MIR)
3,...), The time interval (for example, the number of samples) between the local minimums is measured.

That is, this will be described with reference to the example shown in FIG. 5. In the data analysis circuit 7, the time interval between the first local maximum MAG and the second local maximum MAR is calculated from the second local maximum MAR1 to the next local maximum MAR. Time interval (number of samples) TD1 to 1st local maximum MAG, 2nd local maximum MAR
Time interval (number of samples) TD2 from 1 to the second local maximum MAR2, the second local maximum MAR
Like the time interval (number of samples) TD3 from 1 to the second local maximum MAR3, each time interval from the second local maximum MAR1 to each of the other second local maximum MARs or the first local maximum MAG is set. In the same manner, the time interval from the first local maximum MAG to the next second local maximum MAR2, the first local maximum MA
Time intervals from G to the second local maximum MAR3, time intervals from the first local maximum MAG to the second local maximum MAR4,..., From the first local maximum MAG to each other second local maximum MAR Find each time interval.
Each of the second local maximums MAR3, MAR4, MAR5,... (M
Similarly, the respective time intervals are obtained for AR6 and thereafter (not shown).

Similarly, the data analysis circuit 7 sets the second local minimum MIR as a time interval for the first local minimum MIG and the second local minimum MIR.
Time interval (number of samples) AD1 from 1 to the next first local maximum MIG, second local minimum MIR
Time interval (number of samples) AD2 from 1 to the second local minimum MIR2, second local minimum MIR1
From the second local minimum MIR3 to the second local minimum MIR3,...
From IR1 each other second local minimum MIR or first
The respective time intervals for the local minimum MIG are obtained, and similarly, the first local minimum MI
A time interval from G to the next second local minimum MIR2, a time interval from the first local minimum MIG to the second local minimum MIR3,..., Etc. from the first local minimum MIG to each other second local minimum. For each time interval. Each second local minimum MIR after the second local minimum MIR2
, MIR4 (not shown),.

Thereafter, in the data analysis circuit 7, the first local maximum MAG and the second local maximum MA
R, each time interval (the number of samples), the first local minimum MIG and the second local minimum MIR
Statistical analysis is performed using both time intervals (number of samples) determined using

More specifically, in the data analysis circuit 7, each time interval (the number of samples) obtained by using the first local maximum MAG and the second local maximum MAR, and the first local minimum MIG and the second local maximum MIG are used. Time interval (sample number) obtained using the minimum MIR
, A histogram is created, and the closest value of the histogram is determined.

That is, for example, as shown in FIG.
Each time interval (the number of samples) obtained using the local maximum MAG and the second local maximum MAR, the first local minimum MIG and the second local minimum M
Among the time intervals (the number of samples) obtained using IR, the time interval (the number of samples) having the highest occurrence frequency is obtained. In the example of FIG. 6, the time interval between them is two times with a time interval of 81 samples, twice with a time interval of 82 samples, and 3 with a time interval of 83 samples. Times,..., 5 times with a time interval of 192 samples, 11 times with a time interval of 193 samples, 6 times with a time interval of 194 samples,. The most probable values of the histogram are 11 times having a time interval of 193 samples, and therefore, the data analysis circuit 7
The time interval of the 193 samples is output as a statistical analysis result.

The time interval output obtained by the statistical analysis in the data analysis circuit 7 is sent to the pitch determination circuit 8.
The pitch determination circuit 8 determines the pitch of the input audio signal based on the time interval output supplied from the data analysis circuit 7. In other words, using the statistical analysis result shown in FIG. 6 as an example, the pitch determination circuit 8 determines the time interval of 193 samples as the pitch period of the input audio signal supplied to the input terminal 1.

The pitch cycle determined by the pitch determination circuit 8 is output from an output terminal 9 as a pitch detection result.

FIG. 7 is a flowchart showing the flow of the pitch detecting operation in the pitch detecting device according to the first embodiment shown in FIG.

In FIG. 7, in step S1, the audio signal supplied to the input terminal 1 is converted into digital data by the analog / digital converter 2, and only the desired frequency band is extracted by the filter circuit 3. After low-pass filtering, unit time (frame) in buffer memory 4
Cut out each time.

In the next step S 2, the local maximum MA (maximum point) and the local minimum MI (minimum point) are extracted by the peak search circuit 5, and the local maximum MA is positive (+) by the code data conversion circuit 6. ) Is added and the local minimum MI is added with a negative (−) sign to generate coded data.

In step S3, the absolute value of the amplitude of each local maximum MA (maximum value) to which the positive (+) sign is added and the negative (-) The absolute value of the amplitude of each local minimum MI (minimum value) to which the sign is added is obtained, and the first local maximum MAG and the first local minimum MIG having the highest absolute value of the amplitude in the frame are obtained.

In step S4, in the data analysis circuit 7, the first local maximum M having the largest absolute value of the amplitude is set.
All second local maximum MARs whose amplitude values fall within a predetermined range RT corresponding to AG are obtained, and each time interval (the number of samples) is determined using the first local maximum MAG and the second local maximum MAR. Is measured, and a histogram is created from the time interval.

In step S5, in the data analysis circuit 7, the first local minimum MI having the maximum absolute value of the amplitude is obtained.
All second local minimum MIRs whose amplitude values fall within a predetermined range RU corresponding to G are obtained, and each time interval (the number of samples) is determined using the first local minimum MIG and the second local minimum MIR. Is measured, and the time interval is added to the histogram in step S4.

In step S6, the data analysis circuit 7 finds the histogram's most probable value, and then the pitch determination circuit 8 determines the pitch period from the histogram's most probable value.

According to the first embodiment, the pitch can be detected with high accuracy with a small amount of calculation from the characteristics of the waveform without finding the correlation of the audio signal. Further, according to the first embodiment, pitch detection can be performed with a smaller amount of calculation than when measuring a time interval using all local maximums MA and local minimums MI.

Next, a description will be given of a pitch detecting apparatus according to a second embodiment to which the second pitch detecting apparatus of the present invention is applied.

In the pitch detecting device according to the second embodiment, as shown in FIG. 8, the amplitude value is larger than the predetermined range RTA corresponding to the first local maximum MAG having the largest absolute value in the frame. Seeks a local maximum MA (MAB) with a large
A (MAB) is used to measure a time interval to create a histogram, and a local area having a larger amplitude value than a predetermined range RUI corresponding to a first local minimum MIG having the largest absolute value in the frame. Minimum MI (MI
B) and their local minimum MI (MIB)
Is used to measure the time interval and add it to the histogram.

Further, in the pitch detection device of the second embodiment, a local maximum MA (MAS) whose amplitude value is smaller than a predetermined range RTA corresponding to the first local maximum MAG having the largest absolute value is obtained. A time interval is measured using the local maximum MA (MAS) to generate a histogram, and the amplitude value is smaller than a predetermined range RUI corresponding to the first local minimum MIG having the largest absolute value of the amplitude. Local minimum M
I (MIS) and their local minimum MI
The time interval is measured using (MIS) and added to the previous histogram.

The schematic configuration of the pitch detector of the second embodiment is the same as that of FIG. 1 and is not shown, but the data analysis circuit 7 of the pitch detector of the second embodiment is not shown. Do the following:

That is, in the data analysis circuit 7 of the pitch detection device according to the second embodiment, as shown in FIG. 8, first, each local maximum (maximum value) MA to which a positive (+) sign is added Absolute value of amplitude and code data conversion circuit 6
, The absolute value of the amplitude of each local minimum (minimum value) MI to which a negative (-) sign is added, and the first local maximum MAG and the first local maximum MAG having the largest absolute value of the amplitude in the frame. 1. Find a local minimum MIG.

Next, the data analysis circuit 7 sets a predetermined range RT corresponding to the first local maximum MAG.
All local maximums MA whose amplitude value is larger than A and all local minimums MI whose amplitude value is larger than a predetermined range RUI set corresponding to the first local minimum MIG are obtained. In the example of FIG.
A predetermined range RT corresponding to the first local maximum MAG
Local maximum MA whose amplitude value is larger than A
In particular, the third local maximum MAB (MAB1, M
AB2,...), And the first local minimum M
The local minimum MI whose amplitude value is larger than the predetermined range RUI corresponding to the IM is particularly represented as a third local minimum MIB (MIB1, MIB2,...).

Similarly, data analysis circuit 7 has a predetermined range RT set corresponding to the first local maximum MAG.
All local maximums MI whose amplitudes are smaller than A and all local minimums MI whose amplitudes are smaller than a predetermined range RUI set corresponding to the first local minimum MIG are obtained. In the example of FIG.
A predetermined range RT corresponding to the first local maximum MAG
Local maximum MI whose amplitude value is smaller than A
In particular, the fourth local maximum MAS (MAS1, M
AS2,...), And the first local minimum M
A local minimum whose amplitude value is smaller than a predetermined range RUI corresponding to the IM, in particular, a fourth local minimum M
IS (MIS1, MIS2,...).

Here, the predetermined range RTA corresponding to the first local maximum MAG is the third local maximum M
AB and a plurality of fourth local maximum MAS can be obtained. Similarly, the predetermined range RUI corresponding to the first local minimum MIM is a plurality of third local minimum MIB and fourth local minimum MIS. This is the range that can be obtained. That is, since the data analysis circuit 7 performs the statistical analysis of the time interval, in order to improve the accuracy of the statistical analysis, the third local maximum MAB, the third local minimum MIB, and the fourth local maximum MAS And fourth
This is because a plurality of local minimum MISs are required.

The predetermined range RTA can be set to any range from the amplitude value 0 in FIG. 8 to the amplitude value of the first local maximum MAG, and the predetermined range RUI is as shown in FIG. An arbitrary range can be set between the middle amplitude value 0 and the amplitude value of the first local minimum MIG. In the first embodiment described above, FIG.
As a specific example of the predetermined range RT and the predetermined range RU, only the local maximum MA and the local minimum MI that most characteristically represent the pitch of the audio signal among all the local maximum MA and the local minimum MI in the frame are extracted. Although the example in which the range is set so as to be able to be performed has been described, the predetermined range RT shown in FIG.
As a specific example of A and RUI, for example, a range that can exclude the local maximum MA and the local minimum MI that occur most frequently in a frame can be considered. That is, if the local maximum MA and the local minimum MI that occur most frequently in a frame are used, the accuracy of pitch detection can be increased, but the amount of calculation increases. Therefore, if the local maximum MA and the local minimum MI that occur most frequently in those frames are excluded from the pitch detection calculation, the pitch detection accuracy is slightly reduced, but the calculation amount can be reduced. By using such a predetermined range RTA and a predetermined range RUI, all the local maximums MA,
The pitch can be detected with a much smaller calculation amount than when the local minimum MI is used, or when, for example, the local maximum MA and the local minimum MI occur most frequently in a frame.

Further, in the data analysis circuit 7, the first local maximum MAG and all the third local maximum MABs (MAB1, MAB) obtained as described above are obtained.
2, MAB3,...) To measure the time interval between the respective local maximums, and similarly, the first local minimum MIG and all the third local minimum MIBs
(MIB1, MIB2, MIB3,...)
The time interval (for example, the number of samples) between these local minimums is measured.

Similarly, the data analysis circuit 7 uses the first local maximum MAG and all the fourth local maximum MASs (MAS1, MAS2, MAS3,...) To measure the time interval between the respective local maximums. Similarly, the first local minimum MIG and all fourth local minimums MIS (MIS1, MIS2, MI
S3,...), The time interval (for example, the number of samples) between these local minimums is measured.

That is, with reference to the example of FIG.
In the data analysis circuit 7 of the pitch detection device according to the second embodiment, the time interval for the first local maximum MAG and the third local maximum MAB is set to the first local maximum MAG and the third local maximum MAB.
A time interval (number of samples) tdb1 from the local maximum MAG to the next third local maximum MAB1,
A time interval (number of samples) tdb2 from the first local maximum MAG to the third local maximum MAB2,
Time interval (number of samples) tdb3 from the first local maximum MAG to the third local maximum MAB3,
.., The respective time intervals (the number of samples) are obtained from the first local maximum MAG for each of the other third local maximum MABs.
A third time interval (number of samples) from the third local maximum MAB1 to the next third local maximum MAB2, a time interval (number of samples) from the third local maximum MAB1 to the third local maximum MAB3, and so on. Local maximum MAB1 to each other third
The respective time intervals (the number of samples) are obtained for the local maximum MAB. 3rd Local Maximum MA
Each third local maximum MAB3, MAB after B2
, MAB5,... (Not shown after MAB4) in the same manner for each time interval (number of samples)
Ask for.

Similarly, in the data analysis circuit 7 of the pitch detection device according to the second embodiment, the fourth local maximum MAS1 is set as the time interval (the number of samples) for the first local maximum MAG and the fourth local maximum MAS. From the first local maximum MAG to the next first local maximum MAG (number of samples) tds1, the time interval from the fourth local maximum MAS1 to the fourth local maximum MAS2 (number of samples) tds2, the fourth local maximum MAS1 to the fourth local maximum MAS3
As shown in a time interval (number of samples) to (not shown), each time interval is obtained from the fourth local maximum MAS1 for each of the other first local maximum MAGs or the fourth local maximum MAS.
Time interval (number of samples) from the first local maximum MAG to the next fourth local maximum MAS2, the first
Time interval (number of samples) from the local maximum MAG to the fourth local maximum MAS (not shown)
.., The respective time intervals (the number of samples) are obtained from the first local maximum MAG for each of the other fourth local maximum MAS. Each fourth local maximum MAS3 after the fourth local maximum MAS2
The same applies to MAS4, MAS5,... (Not shown after MAB3).

In the data analysis circuit 7 of the pitch detection device according to the second embodiment, the first local minimum M
As the time interval for the IG and the third local minimum MIB, the time interval adb1 from the first local minimum MIG to the next third local minimum MIB1, the first
Local minimum MIG to 3rd local minimum MI
Time interval adb2 to B2, first local minimum M
IG to 3rd local minimum MIB3 (not shown)
The time interval for the other third local minimum MIB is obtained from the first local minimum MIG as in the time interval up to...
Time interval from the local minimum MIB1 to the next third local minimum MIB2, the third local minimum MI
B1 to third local minimum MIB3 (not shown)
The time interval from the third local minimum MIB 1 to each of the other third local minimum MIBs is obtained as in the time interval up to. Third local minimum MIB3, MI after the third local minimum MIB2
B4, MIB5, ... (illustration omitted after MIB3)
Similarly, the respective time intervals are obtained.

Similarly, in the data analysis circuit 7 of the pitch detecting device according to the second embodiment, the time interval between the first local minimum MIG and the fourth local minimum MIS is set to the next local minimum MIS1 to the next local minimum MIS1. A time interval ads1 from the first local minimum MIG, a time interval ads2 from the fourth local minimum MIS1 to the fourth local minimum MIS2, a time interval ads3 from the fourth local minimum MIS1 to the fourth local minimum MIS3, and so on. From the fourth local minimum MIS1, a time interval for each of the other first local minimums MIG or the fourth local minimum MIS is obtained, and similarly, the first local minimum M
A time interval from the IG to the next fourth local minimum MIS2, a time interval from the first local minimum MIG to the fourth local minimum MIS3, and so on, from the first local minimum MIG to each other fourth local minimum MIS. For each time interval. 4th local minimum M after 4th local minimum MIS2
Similarly, the time intervals of IS3, MIS4, MIS5,... (Illustration is omitted after MIS4) are obtained.

Thereafter, in the data analysis circuit 7 of the pitch detection device according to the second embodiment, the first local maximum M
AG, each time interval obtained using the third local maximum MAB and the fourth local maximum MAS, and a time interval obtained using the first local minimum MIG, the third local minimum MIB, and the fourth local minimum MIS. Perform statistical analysis.

More specifically, in the data analysis circuit 7 of the second embodiment, the first local maximum MA
G, time intervals determined using the third local maximum MAB and the fourth local maximum MAS, and time intervals determined using the first local minimum MIG, the third local minimum MIB, and the fourth local minimum MIS, respectively. Is used to create a histogram, and the closest value of the histogram is determined.

That is, similarly to the first embodiment, each time interval obtained using the first local maximum MAG, the third local maximum MAB, and the fourth local maximum MAS, and the first local minimum MIG, Local minimum MIB and fourth local minimum MIS
Among the time intervals obtained by using, the time interval with the highest occurrence frequency is obtained.

The time interval output obtained by the statistical analysis in the data analysis circuit 7 of the second embodiment
Is sent to the pitch determination circuit 8 of the embodiment. The pitch determination circuit 8 determines the pitch of the input audio signal based on the time interval output supplied from the data analysis circuit 7.

The pitch period determined by the pitch determination circuit 8 is output from an output terminal 9 as a pitch detection result of the pitch detection device of the second embodiment.

FIG. 9 is a flowchart showing the flow of the pitch detecting operation in the pitch detecting device according to the second embodiment. Note that the processing from step S1 to step S3 in FIG. 9 is the same as the processing from step S1 to step S3 in the flowchart in FIG. 7, and a description thereof will be omitted.

In FIG. 9, in step S14,
In the data analysis circuit 7, the first absolute value of the amplitude
All third local maximum MABs whose amplitude values are larger than a predetermined range RTA corresponding to the local maximum MAG are obtained, and the first local maximum MAGs are obtained.
Then, each time interval is measured using the third local maximum MAB, and a histogram is created from the time interval.

In step S15, in the data analysis circuit 7, all the fourth local maximum MASs whose amplitude values are smaller than the predetermined range RUI corresponding to the first local maximum MAG having the maximum absolute value of the amplitude are obtained. Each time interval is measured using the first local maximum MAG and the fourth local maximum MAS, and the time interval is added to the histogram in step S14.

In step S16, in the data analysis circuit 7, the first local minimum M having the maximum absolute value of the amplitude is obtained.
All third local minimum MIBs whose amplitude values are larger than a predetermined range corresponding to IG are obtained, and the first local minimum MIG and the third local minimum MIB are obtained.
Is used to measure each time interval, and a histogram is created from the time intervals.

In step S17, in the data analysis circuit 7, the first local minimum M having the maximum absolute value of the amplitude is obtained.
All the fourth local minimums MIS whose amplitude values are smaller than the predetermined range corresponding to the IG are obtained, and the first local minimum MIG and the fourth local minimum MIS are obtained.
Is used to measure each time interval, and the time interval is added to the histogram in step S16.

In step S18, the data analysis circuit 7 determines the closest value of the histogram, and then the pitch determination circuit 8 determines the pitch period from the closest value of the histogram.

According to the second embodiment, the pitch can be detected with high accuracy with a small amount of calculation from the characteristics of the waveform without finding the correlation of the audio signal. Further, according to the second embodiment, pitch detection can be performed with a smaller amount of calculation than when measuring time intervals using all local maximums MA and local minimums MI.

The pitch detecting device described in each embodiment of the present invention can be applied to, for example, a so-called karaoke device for detecting the pitch of a singer's voice or a chorus voice.

That is, in the karaoke apparatus, the key of the song and the pitch of each sound by the singer are obtained from the pitch detected by the present invention, and the deviation from the key and the pitch of the original music is determined. If it can be shown to the singer, it is considered to be very effective for practicing the song.

Hereinafter, as a third embodiment of the present invention,
Applicable to a karaoke apparatus, the pitch of a singer or a chorus is detected by the pitch detection device of the present invention (first and second embodiments), and based on the detected pitch, the singer is selected. FIG. 10 shows a main configuration of a determining device for determining the quality of a voice of a chorus or a chorus. Note that components of a general karaoke apparatus are well-known, and FIG. 10 does not show them, and shows only a portion related to the determination apparatus of the present invention. Also, in FIG.
The same components as those in FIG. 1 are denoted by the same reference symbols.

In FIG. 10, an input terminal 1 is supplied with an analog audio signal obtained by converting the voice of a singer or a chorus into sound / electricity by a microphone. This analog audio signal is converted into a digital signal by the analog / digital converter 2 and low-pass filtered by the filter circuit 3. Output data from the filter circuit 3 is sent to the pitch detection device 32.

The pitch detecting device 32 includes first and second
It has a configuration after the buffer memory 4 of any of the pitch detection devices of the embodiments. The pitch detected by the pitch detection device 32 is supplied to a comparison circuit 33.

On the other hand, the terminal 30 has a MIDI (Musical Instrument Digital Interface) of the karaoke apparatus.
Data is supplied. The MIDI data is sent to the MIDI reference sound extraction circuit 31. Here, the MIDI data is data relating to the accompaniment music being reproduced by the karaoke apparatus, and includes data of the key and pitch of the original music (song). The MIDI reference sound extraction circuit 31 extracts original tone and pitch data of the music as reference sound data (hereinafter referred to as reference pitch) from the supplied MIDI data. This reference pitch is supplied to the comparison circuit 33.

The comparison circuit 33 compares the pitch of the singer's voice supplied from the pitch detection device 32 with the reference pitch supplied from the MIDI reference sound extraction circuit 31, and outputs the comparison result. By this comparison, it is possible to know how much the key and pitch of the singer deviate from the key and pitch of the original music.

Here, the input from the microphone and the MID
An example showing how the key and pitch of the singer deviates from the original key and pitch of the music based on the I data will be described with reference to FIG.

In FIG. 11, the MIDI data is
As shown in FIG. 11A, data DT1, DT2,
Data is supplied for each unit frame as shown in DT3,.
MID for each unit frame of 1, DT2, DT3, ...
The reference pitch is extracted from the I data.

In the pitch detecting device 32, FIG.
As shown in (b), the pitch of the input audio signal from the microphone is detected about every 30 ms.

The comparison circuit 33 compares the pitch of the input audio signal detected by the pitch detection device 32 approximately every 30 msec with the reference pitch corresponding to every 30 msec at which pitch detection is performed by the pitch detection device 32. Then, as the comparison result, one of the comparison results of “high”, “low”, and “correct” is output. In addition, “high” in the comparison result indicates that the pitch of the singer's voice is higher than the reference pitch, and “low” in the comparison result indicates that the pitch of the singer's voice is lower than the reference pitch. Indicates that the pitch of the "correct"singer's voice matches the reference pitch.

The comparison result from the comparison circuit 33 is sent to the image data generation circuit 34. The image data generation circuit 34 is for generating image data for causing a monitor for displaying lyrics and image video, which is usually provided in a karaoke apparatus, to perform display according to the comparison result. .

Examples of display image data according to the comparison result include image data for displaying “high”, “low”, and “correct” by characters, and “high”, “low”, and “correct”. Image data for representing a character corresponding to "", or image data for changing the color or brightness of a character corresponding to "high", "low" or "correct". 34,
One of the image data is generated.

The image data generated by the image data generation circuit 34 is imposed and displayed on a predetermined area 41 on a monitor screen 40 shown in FIG. 12, for example.

In the description so far, the comparison circuit 33
Described the example of outputting three types of comparison results of “high”, “low”, and “correct” and displaying corresponding results on the monitor screen. It is also possible to display by a so-called bar graph how much the key and the pitch of the music deviate.

In this case, the comparison circuit 33 outputs, as a comparison result, the amount (level) in which the singer's key and pitch deviates from the original key and pitch of the music, and outputs the image data generation circuit. At 34, bar graph display image data is generated according to the comparison result.

Further, in the third embodiment, an example is described in which image data is generated in accordance with the comparison result of the comparison circuit 33. For example, as shown in FIG.
For example, LED corresponding to "low" and "correct" respectively
(Light emitting diode) It is also possible to provide a lamp according to the comparison result by providing a lamp below the monitor screen 40 or the like and turning on or off the LED lamp according to the comparison result.

FIG. 14 shows a flow from the voice input and the MIDI data input to the display output of the comparison result in the determination device of the third embodiment.

In FIG. 14, in step S41, the pitch detection device 32 detects the pitch from the input voice signal of the singer's voice supplied from the input terminal 1.

On the other hand, in step S42, the MIDI reference sound extraction circuit 31 extracts the reference pitch from the MIDI data input from the terminal 30, and compares the reference pitch with the pitch detected from the singer's voice. Compare with.

In step S43, one of “high”, “low”, and “correct” is determined from the comparison result by the comparison circuit 33.

In step S44, a display output for notifying the singer of the result of the comparison, "high", "low", and "correct" is performed.

As described above, according to the third embodiment, a high-precision pitch can be obtained with a small amount of calculation without obtaining a correlation, and the detected pitch is compared with the reference pitch. The voice of the singer or the chorus can be guided to the correct pitch, and the voice property can be converted.

In the above-described third embodiment, the pitch detected using the pitch detection devices of the first and second embodiments is compared with a reference pitch, and based on the comparison result, the singing is simply performed. Voice pitch is "high", "low",
Although only the singer or the like is informed of any of "correct", the karaoke apparatus according to the fourth embodiment of the present invention shifts the pitch of the accompaniment sound and sings. It is also possible to match the pitch of the voice of the person or the like.

In order to realize such a pitch shift, it is necessary to determine whether the sound of a certain section (frame) among the sounds input from the microphone is a vowel of the voice of a singer or the like.
Alternatively, it is necessary to determine whether the voice of a singer or the like is a consonant, noise, or silence. That is, when it is determined that the vowel is a vowel of a singer or the like, it is effective to perform the pitch shift of the accompaniment sound based on the pitch of the vowel, but when it is determined that the sound is a consonant, noise, or silence. This is because it is not preferable to perform the pitch shift of the accompaniment sound based on the pitch of those sounds.

In the fourth embodiment, the above-described first and second embodiments are described.
Whether the sound input from the microphone is a vowel of the voice of a singer or the like, or other consonants, noise, or silence using the histogram created by the pitch detection device of the second embodiment Is determined, and whether to perform the pitch shift is determined based on the determination result.

FIG. 15 shows a schematic configuration of a karaoke apparatus according to the fourth embodiment. The configuration shown in FIG. 15 is obtained by adding a vowel or the like determination function and a pitch shift function to the third embodiment to which the first and second embodiments are applied. In FIG. 15, the same components as those in FIG. 10 are denoted by the same reference numerals, and description thereof will be omitted.

In FIG. 15, MIDI including accompaniment data of the karaoke apparatus supplied from terminal 30 is provided.
The data is sent to the pitch shift circuit 37. The pitch shift circuit 37 can shift the pitch of the accompaniment sound, and the presence / absence of the pitch shift and the pitch shift amount are controlled by a determination control circuit 36 described later.

The judgment control circuit 36 includes the pitch detection circuit 3
2, the data of the histogram generated as described above and the data of the comparison result output from the comparison circuit 33 are supplied. The judgment control circuit 36 compares the nearest value of the histogram with its peripheral value, and if the nearest value is larger than the peripheral value by a certain threshold or more, the section (frame) is a vowel. On the other hand, if the most probable value is smaller than a value around it by a certain threshold value, it is determined that the section (frame) is a consonant, noise, or silence.

Further, the determination control circuit 36 controls the presence or absence of the pitch shift and the pitch shift amount in the pitch shift circuit 37 based on the determination result of the vowel or other sounds and the comparison result from the comparison circuit 33. To generate a control signal. That is, the judgment control circuit 36 indicates that the comparison result supplied from the comparison circuit 33 indicates “correct” or that the judgment result of the vowel or other sound is a consonant, noise, or silence. Pitch shift circuit 3 so as not to perform pitch shift when
7 when the comparison result from the comparison circuit 33 indicates “high” or “low” and the judgment result of the vowel or the other sound indicates that it is a vowel. The pitch shift circuit 37 is controlled to perform a shift.

As a result, the accompaniment sound data corresponding to the pitch of the voice of the singer or the like is output from the pitch shift circuit 37, and the accompaniment sound data is transmitted from the terminal 38 to a speaker or the like via an amplifier (not shown). Sent.

FIG. 16 is a flowchart showing the processing in the judgment control circuit 36 of the karaoke apparatus shown in FIG.

Referring to FIG. 16, in step S21, the closest value of the histogram and its peripheral values are fetched.

In the next step S22, the closest value of the histogram is compared with its peripheral values. For example, taking the histogram of FIG. 6 described above as an example, the most likely value is 193
The time interval for samples is 11 times, and the time interval for 192 samples around it is 5 times, and the time interval for 194 samples is 7 times. Therefore, for example, if the threshold value is 5, the values (5 times and 7 times) around the closest value (11 times) are larger than the threshold value, and the frame can be determined to be a vowel. Conversely, if the threshold is 6
If the value is times or seven or more times, it can be determined that the section (frame) is a consonant, noise, or silence.

In step S22, when it is determined that the value around the closest value is equal to or larger than the threshold value, step S23
When it is determined that the value is smaller than the threshold value, the process proceeds to step S24.

In step S23, it is determined that the section (frame) is a vowel.

On the other hand, in step S24, it is determined that the section is a consonant, noise, or silence.

After the processing in step S23, step S25 is performed.
After the processing in step S24, the process proceeds to step S2.
Proceed to step 7.

In the step S25, it is determined whether or not the comparison result of the comparison circuit 33 indicates "high" or "low". If the comparison result indicates “high” or “low” in step S25, step S26
To step S2 when it indicates "correct".
Go to 7.

In step S26, the pitch shift circuit 3
7 so as to perform the pitch shift processing, and in step S27, control is performed so as not to perform the pitch shift processing in the pitch shift circuit 37.

Thereafter, in step S28, the flow shifts to the processing for the next frame.

In the fourth embodiment, the first,
Although the example in which the histogram according to the second embodiment is used in determining a vowel or the like has been described, for example, the local minimum MI and the local minimum MI in a frame are used.
It is also possible to determine a vowel or the like by using a histogram generated from a time interval between them or a histogram generated from a time interval between zero crossings of audio signals (so-called zero crossings).

In the third and fourth embodiments, a dedicated karaoke apparatus has been described as an example. However, the pitch detection apparatus of the present invention can be realized by a so-called personal computer.

FIG. 17 shows a schematic configuration example of a fifth embodiment in which the present invention is applied to a personal computer. The personal computer of the present embodiment installs application data for realizing the operations of the first to fourth embodiments, thereby enabling the detection device described in the first to fourth embodiments, Although any operation of the device is possible, in the following description, as an example, the sound pitch determination function described in the third embodiment and the pitch shift function described in the fourth embodiment are described. The following describes an example in which a karaoke apparatus provided with a personal computer is realized by a personal computer.

In FIG. 17, I / O port 59
Is, for example, an external terminal connected to an external communication line,
Via the I / O port 59 and the communication line, it is possible to connect to an external server described later or a so-called communication karaoke broadcasting center. The I / O port 59 is connected to the I / F circuit 60. Communication karaoke means that a plurality of music data are stored in a communication karaoke broadcasting center, and the music data is transmitted to remote terminal devices connected to the broadcasting center as needed. , A system that allows a terminal device to play (play) music. Therefore, in the fifth embodiment, a personal computer has been described as an example, but a terminal device for the communication karaoke may be used.

The I / F circuit 60 has an I / O port 59
Is an interface between an external communication line via the CPU and an internal CPU (central processing unit) 54.

Here, when the operation of the karaoke apparatus of the third or fourth embodiment is realized by the personal computer of the present embodiment, a data request from the personal computer is transmitted to the I / O port 59. Accordingly, MIDI data for karaoke and application data for implementing the operations of the third and fourth embodiments (hereinafter, referred to as karaoke application data) are supplied via a communication line. Note that the karaoke application data includes the digital filter coefficient setting data of the filter 3 in each of the above-described embodiments and the program data for controlling the peak detection operation in the peak search circuit 5 (the zero-crossing detection operation when the zero-crossing detection is performed). Control program data), code data conversion control program data in code data conversion circuit 6, data analysis control program data in data analysis circuit 7, pitch determination circuit 8
Data for controlling pitch determination operation in the MI
The reference pitch detection operation program data in the DI reference sound extraction circuit 31, the comparison control program data in the comparison circuit 33, the judgment control program data in the judgment control circuit 36, the pitch shift program data in the pitch shift circuit 37, etc. Have

The MIDI data for karaoke and the application data for karaoke are sent to the CPU 54 via the I / F circuit 60 and are temporarily recorded on a hard disk in a hard disk drive (HDD) 56.

When the karaoke MIDI data and the karaoke application data are supplied to the personal computer of the present embodiment in a state of being recorded on an optical disk such as a so-called CD-ROM or a floppy disk, not on a communication line. There is also. In this case, the optical disk or floppy disk is
1 and read by the disk drive 61 and sent to the CPU 54. Of course, MIDI for karaoke read from optical disks and floppy disks
It is also possible to send data and karaoke application data to the hard disk drive 56 for recording. In the present embodiment, karaoke MIDI data and karaoke application data are recorded on the hard disk drive 56 in consideration of the data transfer speed.

The CPU 54 controls the overall operation of the personal computer in response to an operation from the operation unit 55 composed of, for example, a mouse and a keyboard, and operates the personal computer as a karaoke apparatus as in the present embodiment. In this case, first, the karaoke application data recorded (installed) in the hard disk drive 56 is read out and sent to the signal processing circuit 53.

On the other hand, a terminal 50 is an external input terminal for an analog audio signal. Through this terminal 50, an analog audio signal obtained by acoustically / electrically converting a singer's voice by a microphone is supplied. This analog audio signal is converted into a digital signal by an analog / digital converter 51 and sent to a signal processing circuit 53.

The signal processing circuit 53 is a high-speed arithmetic processing circuit capable of realizing various processes described in the first to fourth embodiments on software based on karaoke application data. When the personal computer is operated as the karaoke apparatus according to the third or fourth embodiment as in the embodiment, the signal processing circuit 53 includes the filter circuit 3 and the MIDI reference sound extraction circuit 31 shown in FIG. Element, judgment control circuit 3 in FIG.
6 and a component of the pitch shift circuit 37.
Of course, the signal processing circuit 53 can also generate image data for displaying lyrics and image video usually provided as a karaoke apparatus.

The memory 52 is a memory for temporarily storing data necessary for signal processing in the signal processing circuit 53, data in the middle of calculation, generated image data, and the like, and functions as the buffer memory 4 in FIG. Is also provided.

The image data generated by the signal processing circuit 53 as in the image data generation circuits 34 of the third and fourth embodiments is sent from a terminal 62 to a monitor.
The audio signal of the singer's voice or accompaniment music which is input through the analog signal and converted by the analog / digital converter 51 into a digital signal is returned to an analog audio signal by the digital / analog (D / A) converter 57. The signal is output from the audio output terminal 58 and sent to a speaker or the like.

Next, FIG. 18 shows, as a sixth embodiment, a schematic configuration of a data transmission apparatus for transmitting karaoke MIDI data and karaoke application data in response to an external request.

That is, the transmission apparatus according to the sixth embodiment is provided with a karaoke MID for the personal computer and the communication karaoke terminal apparatus according to the fifth embodiment.
The present invention is applicable to a server or a communication karaoke broadcasting center that transmits I data and karaoke application data.

In FIG. 18, the MIDI storage unit 70
Stores MIDI data of a plurality of songs for karaoke, and a transmission program storage 71 stores application data for karaoke prepared in advance, each of which is connected to a bus. Here, MIDI data and karaoke application data are MIDI
The example in which the MIDI data and the karaoke application data are stored in the storage unit 70 and the transmission program storage unit 71 may be recorded on an optical disk such as a CD-ROM or a floppy disk. The floppy disk is loaded in a disk drive 74 connected to the bus.

The ROM 72, the RAM 73, and the hard disk drive 80 are connected to the CPU 79 via a bus. The ROM 72, the RAM 73, and the hard disk drive 80 store various data when the CPU 79 controls the transmission device of the present embodiment. Or it is for storing.

The I / O port 78 is an external terminal connected to an external communication line. The personal computer and the communication karaoke terminal of the fifth embodiment are connected via the I / O port 78 and the communication line. It can be connected to. The I / O port 78 is connected to the I / F circuit 77.

This I / F circuit 77 has an I / O port 78
External communication line through the internal communication data processing circuit 7
6 is an interface with the C.6.

Hereinafter, the flow of transmitting the MIDI data and the karaoke application data to the communication line by the transmission device of FIG. 18 will be described with reference to the flowchart of FIG. In the flowchart of FIG. 19, connection management, traffic management, information collection, fee collection, and the like, which are performed in normal data communication, are omitted.

In the transmission apparatus of FIG. 19 and the transmission apparatus of FIG. 18, first, in step S51, when a transmission request for MIDI data or karaoke application data is received via an external communication line, the transmission request is transmitted and received. The data is sent to the CPU 79 via the processing circuit 76.

Upon receiving the transmission request, the CPU 79 reads out the karaoke application data from the transmission program storage section 71 in step S52, and then reads the MI of the music requested by the transmission request in step S53.
The DI data is read from the MIDI storage unit 70.

The read MIDI data and karaoke application data are transferred to the transmission / reception data processing circuit 76. This transmission / reception data processing circuit 76
Then, the MIDI data and the karaoke application data are packetized, for example,
5 is modulated and sent to the I / F circuit 77.

Thus, the MIDI data and the karaoke application data packetized in step S55 are transmitted from the I / F circuit 77.

Thereafter, confirmation of transmission completion is made in step S56.

In the sixth embodiment, the MID
The MIDI data read from the I storage unit 70 and the karaoke application data read from the transmission program storage unit 71 are packetized and transmitted from the communication line. The MIDI data and the karaoke application data can be recorded on an optical disk or a floppy disk. It is also possible to provide the user with an optical digital or floppy disk that can be recorded on a computer. In this case, the MIDI data read from the MIDI storage unit 70 and the karaoke application data read from the transmission program storage unit 71 are sent to, for example, a disk drive 74 and recorded on an optical disk or a floppy disk recordable by the disk drive 74. I do.

In the above description, an example in which the present invention is applied to, for example, a karaoke apparatus or the like has been described. However, in a recording studio, music or voice slightly longer than the editing time may be played within a predetermined editing time. The present invention can also be applied when inserting. That is, at the time of the editing operation, a process of changing the pitch frequency of the audio signal is performed, but it is necessary to detect the pitch (pitch frequency) of the audio signal prior to the change of the pitch frequency. The present invention can be applied to detection.

[0137]

According to a first aspect of the present invention, there is provided a pitch detecting apparatus for detecting a plurality of maximum points and minimum points from an audio signal waveform, and detecting an absolute value of an amplitude from the plurality of maximum points in a unit time. The maximum point where the absolute value of the amplitude is maximum is found from among the multiple minimum points within a unit time, and the maximum point where the amplitude falls within a predetermined range with respect to the maximum maximum point Is determined, a minimum point where the amplitude falls within a predetermined range with respect to the maximum minimum point, a time interval is determined between each maximum point falling within the predetermined range, and a time interval is determined between the minimum points falling within the predetermined range. Since the time interval is obtained in each case, the histogram of the time interval is obtained as a common one, and the most prominent value of the histogram is extracted as the pitch of the audio signal, so that the characteristic of the audio waveform is obtained without obtaining the correlation of the audio signal. Pitch detection with a small amount of computation Possible it is.

A pitch detecting apparatus according to the present invention detects a plurality of local maximum points and local minimum points from a voice signal waveform, and obtains a maximum absolute value of the amplitude from the plural local maximum points in a unit time. Find the maximum point where the absolute value of the amplitude is the largest from among a plurality of minimum points in a unit time, and find the maximum point whose amplitude is larger than the predetermined range for the maximum maximum point For the maximum local maximum point, find a local maximum point whose amplitude is smaller than a predetermined range, obtain the local minimum point whose amplitude is larger than a predetermined range for the maximum local minimum point, and determine the amplitude for the maximum local minimum point. Determine the minimum point smaller than the predetermined range, determine the time interval between each maximum point larger than the predetermined range,
A time interval is obtained between each local maximum point smaller than the predetermined range, a time interval is obtained between each local minimum point larger than the predetermined range, and a time interval is obtained between each local minimum point smaller than the predetermined range. Since the histogram of the time interval is obtained as a common one and the most prominent value of the histogram is taken out as the pitch of the audio signal, the correlation of the audio signal is not obtained, and a small amount of calculation is performed based on the characteristics of the audio waveform. The pitch can be detected.

The pitch detecting device according to the third aspect of the present invention compares the nearest value of the histogram with its peripheral values, and determines the vowel / non-vowel of the voice based on the comparison result. Therefore, the vowel / non-vowel determination of the voice can be realized by a small device and simple processing.

An information medium according to a fourth aspect of the present invention detects a plurality of maximum points and minimum points from an audio signal waveform,
Find the maximum point where the absolute value of the amplitude is the maximum from among the multiple maximum points in the unit time, find the minimum point where the absolute value of the amplitude is the maximum from the multiple minimum points in the unit time, and The maximum point where the amplitude falls within a predetermined range with respect to the maximum point is found, the minimum point where the amplitude falls within a predetermined range with respect to the maximum minimum point is determined between each maximum point that falls within the predetermined range. Program data for obtaining a time interval, obtaining a time interval between each minimum point falling within a predetermined range, obtaining a histogram of the time interval as a common one, and detecting the most prominent value of the histogram as a pitch of the audio signal. Is recorded or transmitted, and the program data can be supplied to the arithmetic unit. In the arithmetic unit, the pitch can be detected with a small amount of calculation from the characteristics of the audio waveform without finding the correlation of the audio signal. , It is possible to further lead the height of the voice of the singer's voice and chorus at the correct height, to allow nature conversion of voice.

The information medium according to the fifth aspect of the present invention detects a plurality of maximum points and minimum points from an audio signal waveform,
Find the maximum point where the absolute value of the amplitude is the maximum from among the multiple maximum points in the unit time, find the minimum point where the absolute value of the amplitude is the maximum from the multiple minimum points in the unit time, and A maximum point whose amplitude is larger than a predetermined range with respect to a maximum point is obtained, a maximum point whose amplitude is smaller than a predetermined range with respect to the maximum maximum point is obtained, and a range where the amplitude is specified with respect to the maximum minimum point is obtained. Find a minimum point larger than the maximum minimum point, find a minimum point whose amplitude is smaller than a predetermined range, obtain a time interval between each maximum point larger than the predetermined range, and determine a time interval between the maximum points. A time interval is obtained between each small maximum point, a time interval is obtained between each minimum point larger than a predetermined range, a time interval is obtained between each minimum point smaller than a predetermined range, and a histogram of the time interval As a common thing, By recording or transmitting the program data for detecting the most probable value of the stogram as the pitch of the audio signal, and by being able to supply this program data to the arithmetic device, the arithmetic device does not require the correlation of the audio signal. The pitch of the singer's voice or the chorus can be guided to the correct pitch with a small amount of calculation from the characteristics of the voice waveform, and the voice characteristics can be converted.

An information medium according to the present invention is characterized in that the most probable value of a histogram is compared with the values around the histogram, and program data for judging a vowel / non-vowel of a voice is determined based on the comparison result. By recording or transmitting the program data and supplying the program data to the arithmetic device, the arithmetic device can determine the vowel / non-vowel of the voice with a small device and simple processing.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a pitch detection device according to a first embodiment of the present invention.

FIG. 2 is a waveform diagram showing a waveform of an original audio signal input from a microphone.

FIG. 3 is a characteristic diagram illustrating frequency characteristics of a filter circuit.

FIG. 4 is a waveform diagram showing an audio signal waveform after low-pass filtering by a filter circuit, and a local maximum and a local minimum detected from the audio signal in the first embodiment.

FIG. 5 is a diagram used to explain measurement of a time interval using a maximum point and a minimum point obtained by the pitch detection device according to the first embodiment of the present invention.

FIG. 6 is a diagram used for explaining a histogram obtained in the embodiment of the present invention.

FIG. 7 is a flowchart showing a flow of an operation of the pitch detection device according to the first embodiment of the present invention.

FIG. 8 is a diagram used to explain measurement of a time interval using a maximum point and a minimum point obtained by the pitch detection device according to the second embodiment of the present invention.

FIG. 9 is a flowchart illustrating an operation flow of the pitch detection device according to the second embodiment of the present invention.

FIG. 10 is a block diagram illustrating a schematic configuration of a determination device according to a third embodiment to which the pitch detection device according to the present invention is applied.

FIG. 11 is a diagram used to explain the operation of the determination device according to the third embodiment of the present invention.

FIG. 12 is a diagram illustrating an example of a monitor screen of a karaoke device to which the determination device according to the third embodiment of the present invention is applied.

FIG. 13 is a diagram illustrating another example of a monitor screen of a karaoke device to which the determination device according to the third embodiment of the present invention is applied.

FIG. 14 is a flowchart illustrating a flow of an operation of the determination device according to the third exemplary embodiment of the present invention.

FIG. 15 is a block diagram illustrating a schematic configuration of a determination device according to a fourth embodiment to which the pitch detection device according to the present invention is applied.

FIG. 16 is a flowchart illustrating an operation flow of a determination device according to a fourth embodiment of the present invention.

FIG. 17 is a block diagram illustrating a schematic configuration of a personal computer according to a fifth embodiment to which the pitch detection device according to the present invention is applied.

FIG. 18 is a block diagram illustrating a schematic configuration of a transmission device according to a sixth embodiment to which the pitch detection device according to the present invention is applied.

FIG. 19 is a flowchart illustrating an operation flow of the transmission device according to the sixth embodiment of the present invention.

[Explanation of symbols]

2 ... analog / digital converter, 3 ... filter circuit, 4
... Buffer memory, 5 ... Peak search circuit, 6 ... Code data conversion circuit, 7 ... Data analysis circuit, 8 ... Pitch detection circuit, 32 ... Pitch detection device, 33 ... Comparison circuit, 31 ... M
IDI reference sound extraction circuit, 34... Image data generation circuit, 3
Reference numeral 6: judgment control circuit, 37: pitch shift circuit, 51: analog / digital converter, 52: memory, 53: signal processing circuit, 54: CPU, 55: operation unit, 56: hard disk drive, 57: digital / analog conversion Bowl, 6
0 I / F circuit, 72 ROM, 61 disk drive, 70 MIDI storage unit, 73 RAM, 71 transmission program storage unit, 74 disk drive, 77 I
/ F circuit, 75 clock oscillator, 76 transmission / reception data processing circuit, 80 hard disk drive

Claims (6)

[Claims] 1. A unit time generating means for cutting out a digitally input audio signal for each unit time of a predetermined length, a band-pass means for passing only a desired frequency band from the audio signal, and the desired frequency band A maximum and minimum point detecting means for detecting a maximum and a minimum point from the audio signal waveform, and a first maximum point detection for obtaining a maximum point having the maximum absolute value of the amplitude from the plurality of maximum points in the unit time. Means, first minimum point detecting means for obtaining a minimum point having the maximum absolute value of the amplitude from a plurality of minimum points in the unit time, and an amplitude within a predetermined range with respect to the maximum maximum point. A second maximum point detecting means for finding a maximum point that falls within; a second minimum point detecting means for finding a minimum point whose amplitude is within a predetermined range with respect to the maximum minimum point; Time interval between each existing maximum point A maximum point-to-maximum point detection unit to be obtained; a minimum point-to-minimum point detection unit to obtain a time interval between each of the minimum points existing within the predetermined range; a histogram calculation unit to obtain a histogram of the time interval as a common one; Pitch detecting means for extracting the most probable value of the histogram as the pitch of the audio signal. 2. A unit time converting means for cutting out a digitally input audio signal for each unit time of a predetermined length; a band pass means for passing only a desired frequency band from the audio signal; A maximum and minimum point detecting means for detecting a maximum and a minimum point from the audio signal waveform, and a first maximum point detection for obtaining a maximum point having the maximum absolute value of the amplitude from the plurality of maximum points in the unit time. Means, first minimum point detecting means for obtaining a minimum point having the maximum absolute value of the amplitude from a plurality of minimum points in the unit time, and an amplitude for the maximum maximum point falling within a predetermined range. A second maximum point detecting means for obtaining a maximum maximum point, a third maximum point detecting means for obtaining a maximum point having an amplitude smaller than a predetermined range with respect to the maximum maximum point, and On the other hand, the amplitude is Second minimum point detecting means for obtaining a critical minimum point; third minimum point detecting means for obtaining a minimum point having an amplitude smaller than a predetermined range with respect to the maximum minimum point; and larger than the predetermined range. A first inter-maximal point detecting means for obtaining a time interval between each of the maximal points; a second inter-maximal point detecting means for obtaining a time interval between each of the maximal points smaller than the predetermined range; First minimum point detection means for obtaining a time interval between each minimum point larger than the range; second minimum point detection means for obtaining a time interval between each minimum point smaller than the predetermined range; A histogram calculating unit for obtaining a histogram of the time interval as a common one; and a pitch extracting unit for extracting a most probable value of the histogram as a pitch of the audio signal. Pitch detection device. 3. The apparatus according to claim 1, further comprising: comparing means for comparing the most probable value of the histogram with a value in the vicinity thereof; and determining means for determining a vowel / non-vowel of the voice based on the comparison result. The pitch detection device according to claim 1 or 2. 4. A step of cutting out a digitally input audio signal for each unit time of a predetermined length; a step of passing only a desired frequency band from the audio signal; and a step of extracting an audio signal waveform of the desired frequency band. Detecting a plurality of local maximum points and local minimum points; and obtaining a local maximum point whose absolute value of the amplitude is maximum from the local maximum points in the unit time; and Obtaining a minimum point at which the absolute value of the amplitude becomes maximum from among; obtaining a maximum point whose amplitude falls within a predetermined range with respect to the maximum maximum point; and Obtaining a minimum point falling within a predetermined range; obtaining a time interval between each maximum point existing within the predetermined range; and obtaining a time interval between each minimum point existing within the predetermined range. Calculating the time interval, obtaining the histogram of the time interval as a common one, and extracting the most probable value of the histogram as the pitch of the audio signal, to the arithmetic device. An information medium for recording or transmitting program data to be transmitted. 5. A step of cutting out a digitally input audio signal for each unit time of a predetermined length; a step of passing only a desired frequency band from the audio signal; and a step of extracting an audio signal waveform of the desired frequency band. Detecting a plurality of local maximum points and local minimum points; and obtaining a local maximum point whose absolute value of the amplitude is maximum from the local maximum points in the unit time; and Obtaining a minimum point at which the absolute value of the amplitude is maximum from the middle; obtaining the maximum point whose amplitude is larger than a predetermined range with respect to the maximum maximum point; and Obtaining a local maximum point smaller than a predetermined range; obtaining the local minimum point whose amplitude is larger than a predetermined range with respect to the maximum local point; Obtaining a minimum point smaller than a predetermined range; obtaining a time interval between each of the maximum points larger than the predetermined range; and obtaining a time interval between each of the maximum points smaller than the predetermined range. A step of obtaining a time interval between each minimum point larger than the predetermined range; a step of obtaining a time interval between each minimum point smaller than the predetermined range; and a histogram of the time interval. Recording, or transmitting program data for causing an arithmetic device to execute an arithmetic process including a step of obtaining the highest value of the histogram as a pitch of the audio signal, and a step of extracting the closest value of the histogram as the pitch of the audio signal. Medium. 6. At least pitch data of a reference sound is also recorded. The arithmetic processing includes a step of comparing the most probable value of the histogram with values around the histogram, and based on the comparison result, 6. The information medium according to claim 4, further comprising: determining a non-vowel.