JPH0344312B2 - - Google Patents

Description

[Detailed Description of the Invention] Industrial Application Field The present invention is used in conjunction with an audio signal recording and reproducing device such as a karaoke machine, and compares the audio signal sung by a user with the audio signal reproduced from a standard magnetic tape or the like. This invention relates to a scoring device that automatically scores a user's singing ability.

Configuration of conventional examples and their problems As a field of audio equipment, a music signal played by a musical instrument recorded on a recording medium such as a magnetic tape is played back and amplified, and when a user sings along with it, the music is reproduced. Amplify the sound by mixing it with the signal. There is something commonly called a "karaoke device",
It is widely used for general household or commercial use.

By singing using the above-mentioned "karaoke device," the user can gain joy and satisfaction, but in recent years, the number of people who want to improve their singing ability has increased, and the number of people who want to improve their singing ability has increased. Some people receive guidance from a singing teacher, but this is not possible for everyone, and one way to study singing on your own is to use an audio multiplexing method such as magnetic tape called ``audio multiplex tape.'' recording media are rapidly becoming popular. For example, in the case of a magnetic tape, this audio multiplexing recording medium is a first track 101 on a magnetic tape 1, as shown in FIG.
A vocal signal from a singer or the like is sent to the second track 1.
02, musical signals played by musical instruments, etc. are recorded. When using this magnetic tape, an audio multiplexing type ``karaoke apparatus'' having the configuration shown in FIG.
01 and an amplifier 202, and a second tape reproducing means 3 consisting of a magnetic head 301 and an amplifier 302,
These two outputs are mixed and power amplified by a mixing amplifier 5 together with the output of a microphone input means consisting of a microphone 401 and an amplifier 402, and outputted from a speaker 6 as an acoustic signal.

It is said that you can improve your singing ability by listening to vocal signals recorded on a recording medium using the above device and practicing singing along with the vocal signals yourself, but no matter how much you practice, However, the drawback is that users themselves cannot tell how close their own singing style is to the modeled vocal signal, in other words, how much their singing ability has improved. Also, even if the user sings in the wrong way, the user may not be aware of it, and when practicing individually, there will naturally be a limit, and the user will lose interest and desire to practice. It had the disadvantage that there were many

Purpose of the invention The present invention solves the above-mentioned conventional problems.
Compares the vocal signal recorded on an audio multiplex recording medium with the user's singing voice signal, calculates and displays the degree of match as a score, and provides an objective evaluation method for the user's singing ability. The purpose is to

Structure of the Invention The scoring device of the present invention includes a first waveform converting means for converting an inputted first audio signal into a pulse signal, and a first waveform converting means that converts the first audio signal into a pulse signal based on the output pulse signal of the first waveform converting means. a first pitch detection means for detecting the fundamental frequency of the first speech signal; and a data group of the fundamental frequency of the first speech signal that is an output of the first pitch detection means for a certain time width. a first pitch storage means for converting an inputted second audio signal into a pulse signal; and a second waveform conversion means for converting an input second audio signal into a pulse signal; a second pitch detection means for detecting a fundamental frequency of an audio signal; a level change detection means for detecting a change in the signal level of the second audio signal; and a level change detection means for detecting a change in the signal level of the second audio signal; a second pitch storage means for storing and holding the fundamental frequency of the first audio signal which is the output of the first pitch detecting means at the time when a change in the pitch is detected; The fundamental frequency of the second audio signal that is being recorded is compared with the content of the data group of the fundamental frequency of the first audio signal stored in the first pitch storage means to determine the fundamental frequency of the first audio signal. After selecting the fundamental frequency of the first audio signal that is most similar to the fundamental frequency of the second audio signal from a data group of fundamental frequencies,
The fundamental frequency of the selected first audio signal and the second pitch stored in the second pitch storage means.
and a score calculation means for calculating a score based on the degree to which the first audio signal matches the second audio signal by comparing the fundamental frequency of the audio signal with the fundamental frequency of the audio signal. By using the user's singing voice signal as the first voice signal and the reproduced voice signal of the vocal signal recorded on the recording medium that serves as a song model as the second voice signal, the user's singing voice can be reproduced. The degree to which the signal matches the reproduced audio signal of the vocal signal on the recording medium is displayed as a score, so the user can recognize the level of his or her singing ability compared to the vocal signal on the recording medium. It is possible.

DESCRIPTION OF THE EMBODIMENT FIG. 3 is a block diagram showing an embodiment of the present invention. 4 is a microphone input means for converting the user's singing voice into an electrical signal and amplifying it; 401 is a microphone; and 402 is an amplifier. 2 is a first magnetic tape reproducing means for reproducing a vocal signal recorded on an audio multiplexing recording medium; 201 is a magnetic head; and 202 is an amplifier. Reference numeral 7 denotes a first waveform converting means, which converts a voice signal sung by a user into a pulse signal. Reference numeral 8 denotes a second waveform converting means, which converts the vocal signal on the recording medium into a pulse signal. Reference numeral 9 denotes a first pitch detection means, which detects the fundamental frequency of the voice sung by the user. Reference numeral 10 denotes a second pitch detection means for detecting the fundamental frequency of the vocal signal. 1
Reference numeral 1 denotes a first pitch storage means, which stores a data group of fundamental frequencies of voices sung by a user over a certain time width. Reference numeral 16 denotes level change detection means for detecting level changes in the vocal signal. Reference numeral 12 denotes a second pitch storage means for storing the fundamental frequency of the vocal signal at the time when the level of the vocal signal changes. 13
is a score calculation means that calculates a data group of the fundamental frequency of the voice sung by the user for a certain time width at the time when the signal level of the vocal signal changes and the fundamental frequency of the vocal signal at the time when the signal level of the vocal signal changes. A score is calculated based on the degree to which the voice signal sung by the user matches the vocal signal through comparison calculations. Reference numeral 14 denotes a score display means, which displays the score in order to inform the user of the score calculated by the score calculation means 13.

FIG. 4 is a block diagram showing the specific configuration of this embodiment, which includes detecting the pitch of the voice sung by the user, detecting the pitch of the vocal signal, storing the data group of the fundamental frequency of the voice signal sung by the user, and detecting the pitch of the vocal signal sung by the user. The microcomputer 15 realizes the functions of storing the fundamental frequency and calculating the score.

FIG. 5 shows an actual circuit example of the first waveform converting means 7. Normally, the same circuit is often used for the first waveform converting means 7 and the second waveform converting means 8. , the circuit of the first waveform converting means 7 will be representatively explained with reference to the operational diagram of FIG.

701 is an input terminal, 702, 704, 705,
708, 710, 711 are resistors, 703, 70
6 and 709 are capacitors, 707 is an operational amplifier (hereinafter abbreviated as OP amplifier), 712 is a transistor, and 713 is an output terminal.

OP amplifier 707 and resistors 702, 704, 7
05 and capacitors 703 and 706 constitute a reduced-pass type active filter, which removes high-frequency components of the audio electrical signal as shown in FIG. Necessary signal amplification is performed by the amplification action of , and furthermore, a time constant circuit composed of a resistor 708 and a capacitor 709 supplementally removes high-frequency components that are not sufficiently removed by the active filter. In this way, the audio electrical signal shown in FIG.
12, the pulse waveform is converted into a pulse waveform as shown in FIG. 6c. In this way, the first waveform converting means 7 converts the user's singing voice signal, which is the output of the microphone input means 4, into a pulse waveform, and the second waveform converting means 8 similarly converts the voice signal sung by the user into a pulse waveform. The vocal signal that is the output of is also converted into a pulse waveform.

Further, the level change detection means 16 can be realized by comparing the current level and the previous level using an analog-to-digital converter or a sample-and-hold circuit using conventional analog technology or digital technology.

The operation of this embodiment will be explained below based on the float chart shown in FIG. 7 which shows the main part of the processing operation of the microcomputer.

First, it is assumed that the power of the device is turned on and that the memory elements and the like inside the microcomputer 15 have also been initialized. The user's singing voice signal becomes an electric voice signal by the microphone input means 4,
The signal is amplified and converted into a pulse signal by the first waveform conversion means 7, which is input to the microcomputer 15. In step 17, the time width of the input pulse is converted into a digital quantity and stored. That is, by counting the "H" period of the pulse signal shown in FIG. 6c using the microcomputer's own clock signal, the time width of the input pulse can be converted into a digital quantity. In this way, conversion is performed in the order of the time width from _t1 to _t2 , the time width from _t3 to _t4 , the time width from _t5 to _t6 , etc. in FIG. 6c. Note that this time width is information that is inversely proportional to the fundamental frequency of the voice, and when it increases, it indicates that the pitch has become lower, and when it decreases, it indicates that the pitch has become higher.

On the other hand, the vocal signal stored on the magnetic tape 1, which is an audio multiplexing recording medium, is reproduced by the first magnetic table reproduction means 2, and the signal level is increased in step 18 based on the output of the level change detection means 16. If it has increased, in step 19, the pulse time width representing the value of 1/2 of the reciprocal of the fundamental frequency of the voice, which is the output of the second waveform converting means 8 input to the microcomputer 15, is checked. converted into digital quantities and stored.

Following step 19, step 20 and step 21
In step 22, data on the time width of the pulse signal of the first audio signal for a certain period of time T after entering step 19 is collected, and then in step 22, data on the time width of the pulse signal of the first audio signal is collected for the past 2T from the time when step 22 is entered. A data group of the time width of the pulse signal of the first audio signal for a time period of , that is, from the point of entering step 19, the time width data of the pulse signal of the audio signal of the user's song from time −T to +T. The time width of the pulse signal of the audio signal of the vocal signal obtained in step 19 is selected from the group in step 23, and the one that most closely approximates the time width of the pulse signal of the audio signal of the vocal signal obtained in step 19 is selected. and the time width of the pulse signal of the audio signal of the user's song obtained in step 22, the degree of matching is calculated, and a score is calculated according to the degree of matching.

An example of the method for calculating the score in step 23 will be described below. For example, if the pulse signal time width of the first audio signal is 9.5 [mS] and the pulse signal time width of the second audio signal is 10 [mS], the degree of matching is (10-9.5)/10×100=5 The degree of agreement is calculated as [%], and the distribution of points reflected in the score will be changed depending on the degree of agreement. For example, if the degree of match is within 3%, 100 points, if it is 3-5%
Scores are given as 80 points, 60 points for 5-7%, 0 points for 7% or more, and the number of times the degree of matching of duration was calculated in a song to be scored is N [times]. , P _o [points] is the score obtained each time the degree of matching of the time width is calculated.
Then, the score P indicating the degree of matching between the first audio signal and the second audio signal is expressed as follows, as an example.

P=Σ ^N _o=1 (P _o )/N [points] To give an example of score calculation using the above formula, calculate the degree of matching of duration within a song to be scored once. Assume that the match was performed 10 times, and the match was within 3% for 3 times, 3 to 5% for 4 times, 5 to 7% for 1 time, and 7% or more for 2 times, and the points were distributed as described above. The score P is out of 100 points: P = (3 x 100 + 4 x 80 + 1 x 60 + 2 x 0
)/10=680/10=68 (points).

Next, in step 24, it is determined whether it is time to end the scoring. The decision to end the scoring may be based on information from a push button switch (not shown) that specifies the end of the scoring, or whether there is a performance music signal recorded on the magnetic tape 1. It is also possible to detect this and start scoring when the performance music signal disappears. It is also possible to record the end signal of the song in advance and use the point in time when the end signal is detected or the point in time when the end of the magnetic tape is detected.

If it is not yet time to finish scoring, proceed to step 24.
Therefore, the process proceeds to step 17 in preparation for the next collection of time width information, comparison, and score calculation.

When the scoring is finished, the process proceeds from step 24 to step 25, where the scores are displayed.

Now, the meaning of collating and comparing the pulse time widths of the two types of audio signals around the time when the signal level of the vocal signal increases will be explained using the illustration of changes in singing style shown in FIG. 8. FIG. 8A shows an example of a user's song inputted from a microphone, and FIG. 8B shows an example of a vocal signal of an audio multiplexed medium by a professional singer. As shown in Figure 8 a and b, amateur users sing at a later timing than the vocal signal of a professional singer, and as shown in Figure 8 b and e, the user sings with a vocal signal that is similar to that of a professional singer. In many cases, it is not possible to freely change the frequency or signal level within a single vocalization called fist or vibration.

Therefore, in the case shown in the example in Figure 8, if the respective frequencies of the user's song and the vocal signal's song are compared and scored continuously, the influence of these fists and vibrations will be evaluated. is greatly expressed. For example, as shown after point G in Figure 8 f, the frequency of a professional singer's vocal signal changes greatly due to vibration, but the user's song is not affected by vibration.
Since the frequency is almost constant, it is possible that only a very low evaluation score can be obtained in this part. These fists and vibrations show great individual differences even among professional singers, and even the same person cannot pronounce them in the same way every time. When comparing and scoring songs using vocal signals and songs using vocal signals, the dispersion in the scores becomes very large.

Furthermore, as can be seen from the deviation in the positional relationship of the lyrics in FIGS. 8a and 8c, in the example of FIG. 8, the singing timing of the user's song and the vocal signal are shifted.
Such a timing shift normally occurs, and if the respective frequencies of the user's song and the vocal signal's song are compared consecutively, the first of the vocal signals in Figure 8e. At the beginning of the song, as shown in Figure 8b, the user has not started singing yet, so it is impossible to compare the two types of fundamental frequencies, and the lyrics of the vocal signal shown in Figure 8d are "mo". In the user's lyrics in Figure 8a, the beginning of the word corresponding to ``k'' is at the extension of the first ``ku'' sound, so even if the fundamental frequencies of the two types of signals are compared, they will not match. Even if a person sings at the correct pitch, he or she may only receive a low evaluation score.

The present invention solves the above-mentioned problems, and in the example of FIG. 8, the point after point A is B as shown in FIG. 8f.
The fundamental frequency of the vocal signal at the point where the signal level of the vocal signal increases greatly, such as point B and then point C, where no fists or vibrations occur, and the A data group of the fundamental frequency of the user's song during a period of time T before and after point A, that is, from point A ^- to point A ⁺ , when the signal level of the vocal signal increases significantly. Verify, A
By selecting data on the fundamental frequency of the user's song that is most similar to the fundamental frequency of the vocal signal at the point in time, and comparing these two types of fundamental frequencies, it is possible to determine whether the timing of the user's song and the vocal signal are slightly different. Even if the data deviates from the fundamental frequency of the user's song during the period from point A ^- to point A ⁺ as shown in figure 8 b, the fundamental frequency data in figure 8 e,
If there is data that almost matches the fundamental frequency of the vocal signal at point A, as shown by f, it will be possible to obtain a good evaluation score, and the difference between the user's song and the professional singer's song of the vocal signal. Since it absorbs timing deviations and does not compare areas where individual differences such as fists and vibrations are large, it is possible to perform more accurate scoring that is closer to the evaluation when humans actually listen and evaluate with their ears. The Rukoto.

As described above, according to this embodiment, information on the fundamental frequency of the vocal signal around the time when the signal level of the vocal signal such as a magnetic tape changes and data for a certain period of the fundamental frequency of the audio signal sung by the user are stored. Since it is possible to compare, calculate and display the degree of agreement as a score, it provides a more accurate objective view of the user's singing ability, excluding individual differences in singing style such as fists, vibrations, and timing differences when singing. It is possible to provide a means of evaluation.

In this example, the user's singing voice signal was used as the subject of scoring, and the vocal signal recorded on magnetic tape, which is an audio multiplexing recording medium, was used as the scoring standard. An audio signal such as a sine wave signal or human speech may also be used.

In addition, in this embodiment, a waveform conversion means using a reduced-pass active filter and a transistor to convert an audio signal into a pulse signal was taken up. A circuit that converts the signal into a pulse signal may also be used.

Further, in this embodiment, the pitch detection means, score calculation means, etc. are implemented by a microcomputer, but it goes without saying that these conventional general-purpose logic circuits may be implemented and used.

In addition, in this embodiment, waveform converting means and pitch detecting means are provided separately for processing the user's audio signal and processing the vocal signal, but these are
It is also possible to use only one system, and process the user's audio signal and the vocal signal in a time-sharing manner.

In addition, in this embodiment, the time width in the case of "H" of the pulse signal which is the output of the waveform conversion means is shown in FIG.
In this case, the fundamental frequency of the audio signal, that is, the pitch, is detected by detecting the time width from t ₁ to t ₂ , followed by the time width from t ₃ to t ₄ , etc. For example, as shown in FIG. In c, the time span from t ₁ to t ₂ is followed by the time span from t ₅ to t ₆ , and so on, one by one.
The time width may be detected intermittently, or the pulse signal that is the output of the waveform conversion means may be detected for a certain period of time that is sufficiently long compared to one time width in which the pulse signal that is the output of the waveform conversion means becomes “H”. Check the time width of "H" for all pulses or some pulses, find the average time width and maximum time width, etc. per pulse, and detect the pitch of the audio signal from this average time width, etc. Alternatively, instead of finding the "H" time width of the pulse signal, the time width for one pulse period, such as the time width from the rising edge of "H" to the rising edge of the next "H", can be calculated. May be processed.

Effects of the Invention As described above, the present invention provides two waveform converting means for converting two audio signals into pulse signals, and two pitch converters for detecting information representing the fundamental frequencies of the two audio signals based on the outputs of the waveform converting means. a detection means, two pitch storage means for storing and holding the fundamental frequency, a level change detection means for detecting a change in the signal level of the audio signal that serves as the basis for scoring, and a change in the signal level of the audio signal that serves as the basis for scoring. It consists of a score calculation means that collates and performs a comparison calculation between the fundamental frequency of the audio signal that serves as the standard for scoring at the point in time and the fundamental frequency data for a certain period of the audio signal to be scored. The degree of agreement can be obtained as a score by excluding individual differences in singing style, such as fists, vibrations, and timing differences when singing.

This means that people who practice singing using audio multiplexed recording media can use the vocal signals recorded on the audio multiplexed recording media as a singing teacher to evaluate the singing ability of the singing teacher. It is possible to provide an objective means of determining how many points one has in singing ability, which has a great effect on practice, etc.

[Brief explanation of drawings]

Figure 1 is an explanatory diagram of an audio multiplex track on a magnetic tape, which is one type of audio multiplex recording medium, and Figure 2 is an illustration of a so-called audio multiplex track using magnetic tape, which is one type of audio multiplex recording medium. 3 is a block diagram of a main part of an embodiment of the present invention, FIG. 4 is a block diagram showing a specific configuration of this embodiment, and FIG. 5 is a block diagram of a karaoke device of this embodiment. FIG. 6 is an operation explanatory diagram for explaining the operation of the first waveform conversion means;
FIG. 7 is a flowchart showing the main part of the processing operation of the microcomputer of this embodiment, and FIG. 8 is an explanatory diagram of changes in the singing style for explaining temporal changes in the singing style. 7...First waveform converting means, 8...Second waveform converting means, 9...First pitch detecting means, 10...
Second pitch detection means, 11...First pitch storage means, 12...Second pitch detection means, 13...Score calculation means, 14...Score display means, 16...Level change detection means.

Claims (1)

[Claims] 1. Detecting the fundamental frequency of the first audio signal based on the first waveform converting means that converts the input first audio signal into a pulse signal and the output pulse signal of the first waveform converting means. a first pitch detection means for storing and holding a data group of the fundamental frequency of the first audio signal which is the output of the first pitch detection means for a certain time width; a second waveform conversion means for converting an input second audio signal into a pulse signal; and detecting the fundamental frequency of the second audio signal based on the output pulse signal of the second waveform conversion means. a second pitch detecting means; a level change detecting means for detecting a change in the signal level of the second audio signal; and a point in time when a change in the level of the second audio signal is detected by the level change detecting means. a second interval storage means for storing and holding the fundamental frequency of the second audio signal which is the output of the second interval detection means; and the second audio signal stored in the second interval storage means. and the content of the data group of the fundamental frequency of the first audio signal stored in the first pitch storage means to select the fundamental frequency from among the data group of the fundamental frequency of the first audio signal. After selecting the fundamental frequency of the first audio signal that is most similar to the fundamental frequency of the second audio signal, the selected first
The fundamental frequency of the audio signal and the fundamental frequency of the second audio signal stored in the second pitch storage means are compared and computed to determine how the first audio signal is different from the second audio signal. A scoring device comprising a score calculation means for calculating a score based on degree of agreement.