JPH10149160A - Sound signal analyzing device and performance information generating device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily analyze an effective section even when a level of an input sound is attired delicately by dividing a mean sound pressure level of successively inputted signals to valid, invalid section based on a first prescribed value and further specifying the valid section based on first, second prescribed lengths. SOLUTION: The section where the average sound pressure level (A) of the successively inputted signals is the first prescribed value (e.g. 20%) or above is defined as the valid section where a musical sound exists, and the section of the first prescribed value or below is defined as the invalid section (B). Then, when the time length of the invalid section held between both side valid sections is the first prescribed length or below, these both sides are synthesized to obtain the valid section (C), and thereafter, when the time length of the valid section held between both side invalid sections is the second prescribed length or below, these both sides are synthesized to obtain the invalid section (D). At this point of time, the mean value of the average sound pressure level is calculated, and when it is the second prescribed value or below, its valid section is changed to the invalid section (E).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sound signal analyzing apparatus for analyzing a section (effective section) in which a musical sound exists based on an input voice or a musical sound, and a stationary part of the musical sound, and a sound signal analyzing apparatus. The present invention relates to a performance information generation device that generates performance information such as MIDI information based on voices and musical sounds.

[0002]

2. Description of the Related Art Recently, MIDs have been
A computer performance system that generates performance information such as I information and reproduces performance sounds based on the performance information has attracted attention as a new musical performance apparatus. In this type of computer performance system, a method for inputting data for generating performance information includes a real-time input method, a step input method, a numerical value input method, a score input method, and the like.

[0003] The real-time input method is a method of converting operation information of a performance operator, such as a keyboard, actually played by a player like a tape recorder into performance information in real time. The numerical value input method is a method in which performance information such as a pitch (pitch), a sound length, and the strength of a sound is directly input as numerical data from a computer keyboard. The score input method is a method in which simplified music symbols and the like are arranged in a score (5-line notation) on a display using a function key of a computer, a mouse, or the like. In the step input method, the musical score is input on a MIDI keyboard or software keyboard,
In this method, the length of a sound is input using a function key of a computer, a mouse, or the like.

[0004] Among the above-mentioned input methods, the real-time input method can store an actual performance operation state as performance information as it is, so that human nuances can be easily expressed and a short-time input is possible. It has the advantage that. However, this method requires the player to have a high level of musical instrument playing ability, and is not suitable for beginners. Therefore, as a performance information generator that allows beginners to easily input performance information in a short time by taking advantage of the real-time input method, a human voice or a musical instrument of a natural instrument is directly input via a microphone, and the input sound is input. In some cases, performance information is generated in response to a request. In other words, this is just to input the sound (single sound) of human voice and guitar etc. from the microphone,
A MIDI signal can be easily generated, and a MIDI device can be controlled without using a MIDI keyboard or the like.

[0005]

In a conventional performance information generating apparatus, MIDI information is generated by performing the following processing for a pitch change of a sound input from a microphone. That is, the first method detects a pitch change in semitone units,
Only note information of the pitch is generated. The second method detects a pitch change in semitone units, and generates note information of the pitch and pitch bend information (pitch change information) relating to a pitch change therebetween. In the third method, the pitch change is generated as pitch bend information that changes in the range of one octave up and down. To generate note information (note on or note off), the level of the input sound is compared with a predetermined reference value, and when the level of the input sound becomes higher than the reference value, the note-on is reduced. A note-off has occurred at that point.

However, in the case of detecting a pitch change in semitone units as in the first and second methods,
If the pitch of the input sound is slightly changed, there is a problem that many unintended note information (note on or note off) is generated. Further, when the pitch change is generated by the pitch bend information as in the third method, the pitch change can be faithfully followed by the pitch bend information, but is not suitable for purposes such as transcription. Further, when note information is generated according to the input level, there is a problem that a large number of unintended note information is generated according to the fluctuation of the level of the input sound.
By the way, in the real-time input method, since a plurality of sounds are input to the microphone in a time series at arbitrary time intervals, it is required to efficiently analyze a portion where the sound exists. In other words, if the analysis of the pitch etc. was constantly performed on the signal input to the microphone,
Actually, useless analysis processing is performed even during a time when no sound is input, which is not preferable. Therefore, it is efficient to extract a section (effective section) where a sound actually exists from a signal input to the microphone, and to perform a complicated analysis process such as pitch analysis only on the extracted effective section. is there. The conventional method of extracting valid sections for that purpose is as follows:
Since the effective section is extracted simply by comparing the input signal level with the predetermined reference level, the effective section is extracted when the input sound level fluctuates slightly, especially when the level fluctuates near the reference level. Was inaccurate and had problems.

The present invention provides a sound signal analyzing apparatus capable of easily analyzing a section (effective section) in which a musical sound exists, even when the pitch or level of an input sound from a microphone or the like slightly fluctuates. The purpose is to provide. According to the present invention, even when the pitch or level of an input sound from a microphone or the like is slightly fluctuated, a sound capable of analyzing a stationary part of a musical sound other than the fluctuated part, that is, a part corresponding to one note. It is an object to provide a signal analyzer. More specifically, a steady part is effectively analyzed from a series of inputted sounds, and the pitch of the sound can be accurately analyzed based on the steady part. The present invention has been made in view of the above points, and provides a performance information generating device capable of reliably generating note information for a pitch even when the pitch or level of an input sound from a microphone or the like slightly fluctuates. The purpose is to provide.

[0008]

According to a first aspect of the present invention, there is provided a sound signal analyzing apparatus, comprising: input means for inputting an arbitrary sound signal from outside; and a sample amplitude value of a signal sequentially input from the input means. Calculating means for obtaining an average value over a predetermined number of samples and outputting the result as time-series average sound pressure level information; and a section in which the average sound pressure level obtained by the calculating means is equal to or more than a first predetermined value. Is a valid section in which a musical sound is present, and a section less than the first predetermined value is an invalid section in which no musical sound is present, and the invalid section sandwiched between the valid sections on both sides. If the time length of the section is less than the first predetermined length in
An effective section which changes the invalid section into a valid section, combines the changed valid section and the valid sections on both sides thereof into a new valid section, and a point in time when the processing by the valid section forming means is completed. If the time length of the valid section sandwiched between the invalid sections on both sides is less than the second predetermined length, the valid section is changed to the invalid section, and the changed invalid section and the A first invalid section generating unit that combines the invalid sections on both sides into a new invalid section; and the average sound pressure for each of the valid sections at the time when the processing by the first invalid section ends. A second invalid section conversion means for calculating an average value of the levels and, when the average value is less than a second predetermined value, changing the valid section into an invalid section. In the sound signal analyzer according to the first aspect, since the average value of each sample amplitude value of the signal sequentially input from the input means over a predetermined number of samples is obtained, the average value is sensitive to the level fluctuation of the input sound signal. You can get the responding smoothly changing average sound pressure level information. The average sound pressure level obtained in this manner is divided into an effective section and an invalid section based on the first predetermined value, and the divided effective section and invalid section are further divided based on the first and second predetermined lengths. Finally, the valid section is specified. As a result, even when the level of the input sound from the microphone or the like slightly fluctuates, a section in which a musical sound exists (an effective section)
Can be easily analyzed.

According to a second aspect of the present invention, there is provided a sound signal analyzer, comprising: input means for inputting an arbitrary sound signal from the outside; and an average value over a predetermined number of samples of a sample amplitude value of a signal sequentially input from the input means. Calculating means for outputting the result as time-series average sound pressure level information, and a section in which the average sound pressure level obtained by the calculating means is equal to or more than a first predetermined value is defined as an effective section, First
A section that is less than the predetermined value of and a section sandwiched between the effective sections on both sides as an invalid section, and other sections on both ends of the average sound pressure level as undetermined sections; When the time length of the invalid section between the valid sections is shorter than the first predetermined length, the invalid section is changed to the valid section, and the valid section after the change and the valid sections on both sides are changed. An effective section forming means for combining a section and a new effective section; and, when the processing by the effective section forming means is completed, the time of the section in the effective section sandwiched between the invalid sections on both sides. If the length is less than the second predetermined length, the valid section is changed to an invalid section, the changed invalid section and the invalid sections on both sides are combined to form a new invalid section, and When the time length of the adjacent effective section is the If the length is less than the predetermined length of the first invalid section, the first invalid section is composed of a valid section, an invalid section adjacent thereto and an undetermined section, and a new undetermined section. Calculating an average value of the average sound pressure level for each of the effective section and the undetermined section at the time when the processing by the converting means is completed, and if the average value is less than a second predetermined value, the effective section or the A second invalid section conversion means for changing the undetermined section to an invalid section and changing the undetermined section to a valid section when the value is equal to or more than the second predetermined value. A sound signal analyzer according to a second aspect is basically the same as that of the first aspect, and converts the determined average sound pressure level into an effective section and an invalid section based on a first predetermined value. At this time, both ends of the average sound pressure level, that is, the rising part and the falling part of the average sound pressure level are set as undetermined sections, and the subsequent processing for specifying the effective section is performed. As a result, it is possible to analyze whether the rising portion and the falling portion of the input sound signal are correctly valid sections.

The sound signal analyzer according to claim 3 is the sound signal analyzer according to claim 1 or 2, wherein
At the time when the processing by the invalid section conversion means is completed, the effective section is extended using the average sound pressure level obtained by the calculation means and a second predetermined value smaller than the first predetermined value. Means are further provided. In the sound signal analyzer according to claim 1 or 2, since the boundary between the effective section and the invalid section of the average sound pressure level is determined based on the first predetermined value, the first predetermined value is It depends greatly on whether you choose such a value,
By extending the effective section length specified by the series of processing by using a second predetermined value smaller than the first predetermined value, it is possible to accurately specify an effective section in which a musical sound exists. I have.

According to a fourth aspect of the present invention, there is provided a sound signal analyzer, comprising: input means for inputting an arbitrary sound signal from the outside; and a plurality of candidate positions serving as a cycle reference for the sound signal input from the input means. Cycle reference detecting means for detecting the same signal, and calculating the degree of coincidence between waveforms of adjacent sections represented by the candidate positions with respect to the sound signal, and connecting those having higher degrees of coincidence to detect the same waveform section And a steady section determining means for detecting a steady section based on the same waveform section detected by the section detecting means. The sound signal analyzing apparatus according to claim 4, wherein a candidate position serving as a period reference is detected on both the positive and negative sides of the sound signal, the same waveform section is detected based on the candidate position, and the same waveform section on both the positive and negative sides is overlapped. Since the pitch is adjusted, even if the pitch or level of the sound signal slightly fluctuates on the positive or negative side, the error can be reduced. Then, the steady-state section corresponding to one final note is analyzed based on the rapid change of the pitch and the rapid change of the sound pressure in the same tone section determined in this way. As a result, even when the pitch or level of the input sound from the microphone or the like slightly fluctuates, it is possible to analyze a stationary part corresponding to one note of a musical sound other than the fluctuated part.

According to a fifth aspect of the present invention, there is provided a sound signal analyzing apparatus, comprising: input means for inputting an arbitrary sound signal from the outside; and a temporary candidate position serving as a cycle reference for the sound signal input from the input means. First period reference detection means for detecting a plurality, frequency band detection means for detecting a maximum frequency and a minimum frequency of the sound signal based on the tentative candidate position detected by the first period reference detection means, A filter processing unit that performs a band-pass filter process using the detected maximum frequency and minimum frequency as a cutoff frequency on the sound signal input from the input unit, and an output from the filter processing unit. Second period reference detecting means for detecting a plurality of candidate positions serving as a period reference for the sound signal; and the sound signal is represented by the candidate position. Calculating a degree of coincidence between waveforms of adjacent sections and connecting sections having a high degree of coincidence to detect the same waveform section; And a stationary section detecting means for detecting the In the sound signal analyzer according to claim 5, a temporary candidate position serving as a period reference is detected from the input sound signal, and the maximum frequency and the minimum frequency of the sound signal are detected based on the temporary candidate position. By performing band-pass filter processing using the maximum and minimum frequencies as cutoff frequencies on the input sound signal again, unnecessary low-frequency components and harmonics that may cause errors when detecting the same waveform section are generated. Wave components can be effectively removed. By performing the same processing as in claim 4 on the sound signal from which unnecessary low-frequency components and high-frequency components have been removed, more accurate analysis of a steady section can be performed.

According to the sound signal analyzing device of the present invention, it is considered that an input means for inputting an arbitrary sound signal from the outside and a musical sound exists from the sound signals input from the input means. Effective section analyzing means for analyzing a valid section to be used, cycle reference detecting means for detecting a plurality of candidate positions serving as a cycle reference for both the positive and negative sides of the sound signal constituting the effective section, and the sound signal For each of the positive and negative sides, a section detecting means for calculating a degree of coincidence between waveforms of adjacent sections represented by the candidate positions, and connecting those having a high degree of coincidence to detect the same waveform section. A timbre section determining means that sets a section formed by overlapping the same waveform sections on both the positive and negative sides detected by the section detecting means as a same timbre section; In which equipped with a constant interval determining means for detecting the constant section based on the tone interval that. In the sound signal analyzer according to the sixth aspect, a candidate position serving as a period reference is detected on both the positive and negative sides of the sound signal, the same waveform section is detected based on the candidate position, and the same waveform section on both the positive and negative sides is overlapped. Since the pitch is adjusted, even if the pitch or level of the sound signal slightly fluctuates on the positive or negative side, the error can be reduced. Then, the steady-state section corresponding to one final note is analyzed based on the rapid change of the pitch and the rapid change of the sound pressure in the same tone section determined in this way. Thus, even when the pitch or level of the input sound from the microphone or the like slightly fluctuates, it is possible to analyze a stationary section corresponding to one note of a musical sound other than the fluctuated portion.

In the sound signal analyzing device according to the present invention, it is considered that an input means for inputting an arbitrary sound signal from the outside and a musical sound exists among the sound signals input from the input means. Effective section analyzing means for analyzing an effective section to be used, first cycle reference detecting means for detecting a plurality of temporary candidate positions serving as a cycle reference for the sound signal constituting the effective section, and the first cycle Frequency band detecting means for detecting a maximum frequency and a minimum frequency for the entire section or the effective section of the sound signal based on the tentative candidate position detected by the reference detecting means; and a maximum frequency band detected by the frequency band detecting means. A filter processing unit that performs band-pass filtering using a frequency and a minimum frequency as a cutoff frequency for the entire section or the effective section of the sound signal input from the input unit. And second period reference detecting means for detecting a plurality of candidate positions serving as a period reference for the sound signal output from the filter processing means, and for each of the positive and negative sides of the sound signal, A section detecting means for calculating the degree of coincidence of the waveforms of adjacent sections represented by the position, connecting those having a high degree of coincidence to detect the same waveform section, and the same waveform detected by the section detecting means A stationary section determining means for detecting a stationary section based on the section. In the sound signal analyzer according to claim 7, a temporary candidate position serving as a period reference is detected from the input sound signal, and the maximum frequency and the minimum frequency of the sound signal are detected based on the temporary candidate position. By performing band-pass filter processing using the maximum and minimum frequencies as cutoff frequencies on the input sound signal again, unnecessary low-frequency components and harmonics that may cause errors when detecting the same waveform section are generated. Wave components can be effectively removed. By performing the analysis processing in the stationary section on the sound signal from which such unnecessary low-frequency components and harmonic components have been removed, more accurate analysis is possible. In a sound signal analyzer according to an eighth aspect, the effective section detecting means according to the sixth or seventh aspect is configured by the sound signal analyzer according to the first, second or third aspect.

According to a ninth aspect of the present invention, there is provided a performance information generating apparatus, comprising: input means for externally inputting an arbitrary sound signal; and a stationary section corresponding to one note among the sound signals input from the input means. Based on the representative frequency of the stationary section determined by the frequency determining means, the frequency determining means for determining a representative frequency for each stationary section analyzed by the stationary section analyzing means, Cent value conversion means for converting the difference between the representative frequencies of the two preceding and following stationary sections into a value based on cents, and the two stationary values based on the cent value converted by the cent value conversion means. Pitch difference calculating means for calculating relative pitch difference data between sections, and in each steady section based on the pitch difference data calculated by the pitch difference calculating means. In which it equipped the pitch assignment means for assigning pitches on a constant of the scale. In the performance information generating device according to the ninth aspect, a representative frequency is determined for each stationary section corresponding to one note, and the cent is calculated from a difference between the representative frequencies of the preceding and following stationary sections. Pitch difference data is calculated based on the value, and a pitch on a predetermined scale is assigned to each steady section based on the pitch difference data. That is, the representative frequency of the stationary section is an average value of a plurality of waveforms constituting the stationary section, and the pitch difference data is calculated based on a value based on the cent of the consecutive sections adjacent to each other. Even if the pitch of the input sound from the microphone or the like slightly fluctuates, it is possible to absorb the error component in the pitch on the scale finally assigned.

According to a tenth aspect of the present invention, there is provided a performance information generating apparatus.
An input unit for inputting an arbitrary sound signal from outside, a stationary interval analyzing unit for analyzing a stationary interval corresponding to one note from the sound signal input from the input unit, and an analyzing unit for analyzing the stationary interval analyzing unit Frequency determining means for determining a representative frequency for each of the stationary sections, a plurality of the stationary sections analyzed by the stationary section analyzing means, a phrase detecting means for detecting one phrase, and the phrase detecting means. Cent value conversion means for converting the difference between the representative frequencies into values based on cents for all stationary sections existing before the stationary section in one detected phrase, and the phrase detecting means Based on the relative time distances to all the stationary sections existing before the stationary section in one phrase detected by A weight calculation unit that calculates a weight was,
Pitch difference calculating means for calculating relative pitch difference data between the two stationary sections based on the value based on the cent converted by the cent value converting means and the weight calculated by the weight calculating means; And pitch assigning means for assigning a pitch on a predetermined scale to each stationary section based on the pitch difference data calculated by the pitch difference calculating means. A performance information generating device according to a tenth aspect determines a value based on a representative frequency and a cent for a phrase composed of a plurality of stationary sections in the same manner as the performance information generating device according to the ninth aspect. The pitch difference data is calculated by performing weighting based on the time distances for all the pre-positioned sounds, and the pitch on a predetermined scale is assigned to each steady section based on the weighted data. As a result, even when the pitch of the input sound from the microphone or the like slightly fluctuates, it is possible to perform the pitch assignment depending on the sound in the steady section constituting the phrase.

A performance information generating device according to claim 11 is
An input unit for inputting an arbitrary sound signal from outside, a stationary interval analyzing unit for analyzing a stationary interval corresponding to one note from the sound signal input from the input unit, and an analyzing unit for analyzing the stationary interval analyzing unit Frequency determining means for determining a representative frequency for each of the determined stationary sections, phrase detecting means for detecting a single phrase by collecting a plurality of the stationary sections analyzed by the stationary section analyzing means, and the phrase detecting means Cent value conversion means for converting the difference between the detected representative frequency of the first stationary section in one phrase and the representative frequency of each of the other stationary sections in the phrase into a value based on cents; Pitch difference calculating means for calculating relative pitch difference data between the two steady intervals based on a value based on cents converted by In which equipped the pitch assignment means for assigning the pitch on a given scale to each constant interval on the basis of the pitch difference data calculated by said pitch difference calculating means. The performance information generating device according to claim 11 determines a representative frequency for a phrase constituted by a plurality of stationary sections in the same manner as the performance information generating device according to claim 9, and determines a representative frequency for a leading sound in the phrase. A value based on a cent and pitch difference data are calculated, and a pitch on a predetermined scale is assigned to each steady section based on the calculated value. As a result, even when the pitch of the input sound from the microphone or the like slightly fluctuates, the pitch can be assigned depending on the leading sound of the phrase.

[0018] According to a twelfth aspect of the present invention, there is provided a performance information generating apparatus.
The pitch assigning means according to claim 9, 10 or 11,
When assigning a pitch on a predetermined scale to each stationary section, a predetermined pitch is assigned to the first stationary section, and then a pitch on the predetermined scale is sequentially assigned to the remaining stationary sections. Things. According to a thirteenth aspect of the present invention, in the performance information generating device, the pitch allocating means according to the ninth, tenth, or eleventh aspect assigns a pitch on a predetermined scale to each steady interval. Analyzing the sound signal, detecting the average frequency of the stationary section, assigning the pitch based on the detected average frequency as the pitch of the first stationary section,
It is configured such that pitches on a predetermined scale are sequentially assigned to the remaining stationary sections. According to a fourteenth aspect of the present invention, in the performance information generating apparatus, the pitch allocating means according to the ninth, tenth, or eleventh aspect assigns pitches on a plurality of scales to each steady section while shifting note positions. Calculates the total value of the note assignment error at each note position of each scale, determines the optimal scale according to the total value, and sequentially assigns the pitch on the determined scale as the pitch of the stationary section. It is like that. A performance information generating apparatus according to claim 15, wherein the pitch assigning means according to claim 9, 10 or 11 sequentially assigns the pitches on the determined scale as pitches in a steady section. Are assigned to pitches outside the scale according to the value of the note allowable error range.

[0019]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings. FIG. 2 is a hardware block diagram showing the configuration of an electronic musical instrument incorporating a musical sound information analyzer and a performance information generator according to the present invention. The electronic musical instrument is controlled by a microcomputer including a microprocessor unit (CPU) 1, a program memory 2, and a working memory 3. The CPU 1 controls the operation of the entire electronic musical instrument. This CPU1
, A program memory 2, a working memory 3, a performance data memory 4, a key press detection circuit 5, a microphone interface 6, a switch detection circuit 7, a display circuit 8, and a sound source circuit 9 are connected to each other via a data and address bus 1 E. Have been.

The program memory 2 stores various programs of the CPU 1 (such as system programs and operation programs),
It stores various data and the like, and is constituted by a read only memory (ROM). Working memory 3
Is a memory for temporarily storing performance information and various data generated when the CPU 1 executes a program. A predetermined address area of a random access memory (RAM) is assigned to each of the registers, flags, buffers,
It is used as a table or the like. Performance data memory 4
Is for storing performance information (MIDI data) generated based on an input sound from a microphone or the like. Although not shown, a hard disk drive or the like may be connected to the CPU 1 to store various data such as automatic performance data and chord progression data therein.
The operation program may be stored. Also, without storing the operation programs in the ROM 2, these operation programs are stored in the hard disk device and read into the RAM 3, whereby the ROM 2 is read.
The same operation as when the operation program is stored in the CPU
1 can be performed. By doing so, it is possible to easily add an operation program or upgrade the version.
A CD-ROM may be used as one of the removable external storage media. This CD-ROM may store various data and an arbitrary operation program. CD-
The operation program and various data stored in the ROM can be read out by a CD-ROM drive (not shown) and transferred to a hard disk device for storage. This makes it possible to easily perform new installation and version upgrade of the operation program.

Although not shown, a communication interface is connected to the data and address bus 1E, and can be connected to various communication networks such as a LAN (local area network) and the Internet via the communication interface. May be exchanged with another server computer. Thus, when the operation program and various data are not stored in the hard disk device, the operation program and various data can be downloaded from the server computer. In this case, an electronic musical instrument, which is a musical sound generation device serving as a client, transmits a command for requesting download of an operation program and various data to a server computer via a communication interface and a communication network. The server computer transmits predetermined operation programs and data to the electronic musical instrument 1 via the communication network in response to the command. The electronic musical instrument receives these operation programs and data via the communication interface and stores them in the hard disk device. Thus, the download of the operation program and various data is completed.

The keyboard 10 is provided with a plurality of keys for selecting the pitch of a musical tone to be produced, has a key switch corresponding to each key, and detects a key pressing speed as required. It has touch detection means such as a device and a pressing force detection device. The key press detection circuit 5 is configured to include a circuit composed of a plurality of key switches provided corresponding to each key of the keyboard 10 for designating a pitch of a musical tone to be generated, and a new key is pressed. When a key-on event is output,
When a key is newly released, a key-off event is output. In addition, a process of generating touch data by determining a key pressing operation speed or a pressing force at the time of key pressing is performed, and the generated touch data is output as velocity data. As described above, data such as a key-on event, a key-off event, and velocity are data conforming to the MIDI standard (hereinafter referred to as “MI
DI data ”) and includes data indicating the key code and the assigned channel. The microphone 1A converts an audio signal or a musical instrument sound into a voltage signal,
Output to the microphone interface 6. The microphone interface 6 converts an analog voltage signal from the microphone 1A into a digital signal and outputs the digital signal to the CPU 1 via the data and address bus 1E.

Numeric keys and various switches 1B include a numeric keypad for inputting numerical data, a keyboard for inputting character data,
It is configured to include various operators such as a start / stop switch for a note conversion process (sound signal analysis process and performance information generation process). In addition, various controls for selecting, setting, and controlling pitches, timbres, effects, and the like are also included, but details thereof are publicly known, and thus description thereof is omitted. The switch detection circuit 7 detects the operation state of each operator of the numeric keypad & various switches 1B, and outputs switch information corresponding to the operation state to the CPU 1 via the data and address bus 1E. The display circuit 8 displays various information such as the control state of the CPU 1 and the contents of the setting data on the display 1C. The display 1C is a liquid crystal display panel (LCD)
, A CRT or the like, and the display operation thereof is controlled by the display circuit 8. GU is operated by this numeric keypad & various switches 1B and display 1C.
I (Graphical User Interface)
e) is configured.

The tone generator 9 is capable of simultaneously generating musical tone signals on a plurality of channels, and has a data and address bus 1E.
, MIDI data on the musical sound track given via the input device, and generates a musical sound signal based on this data, and outputs it to the sound system 1D. The tone generator circuit 9 can simultaneously generate tone signals on a plurality of channels by using a single circuit in a time-division manner to form a plurality of tone channels, or a single tone channel is constituted by a single circuit. It may be of the type as described below. Also, any tone signal generation method in the tone generator 9 may be used. For example, a memory reading method (waveform memory method) for sequentially reading out tone waveform sample value data stored in a waveform memory in accordance with address data that changes according to the pitch of a musical tone to be generated, or a phase angle parameter An FM method for executing a predetermined frequency modulation operation as data to obtain tone waveform sample value data, or an AM for executing a predetermined amplitude modulation operation using the address data as phase angle parameter data to obtain tone waveform sample value data
A known method such as a method may be appropriately adopted. In addition to these methods, a physical model method that synthesizes a musical sound waveform by an algorithm that simulates the sounding principle of a natural musical instrument, a harmonic synthesis method that synthesizes a musical sound waveform by adding a plurality of harmonics to a fundamental wave, A formant synthesis method of synthesizing a musical tone waveform using a formant waveform having a specific spectral distribution, an analog synthesizer method using VCO, VCF, and VCA may be employed. Further, the present invention is not limited to the case where the tone generator circuit is configured using dedicated hardware.
The tone generator circuit may be configured by using the SP and the microprogram, or the tone generator circuit may be configured by a CPU and a software program. Sound source circuit 9
Is generated through a sound system 1D including an amplifier and a speaker.

Next, an example in which the electronic musical instrument according to the present invention operates as a sound signal analyzer and a performance information generator will be described. FIG. 1 is a diagram showing a main flow when the electronic musical instrument of FIG. 2 operates as a performance information generating device. The main flow is executed sequentially in the following steps. Step 11: First, an initial setting process is performed, and an initial value is set in each register, flag, and the like in the working memory 3 in FIG. At this time, when the musical note processing start switch on the numeric keypad & various switches 1B is turned on, a series of processing from step 12 to step 17 is performed.

Step 12: This step is carried out when it is determined that the musical note processing start switch has been turned on. The input audio signal and the voltage waveform of the musical instrument sound are sampled at a predetermined period (for example, 44.1 kHz).
It is stored as a digital sample signal in an area within the working memory 3. Since this sampling process is performed by a conventionally known method, the details are omitted here.
Steps 13 to 15 correspond to the note conversion process corresponding to the ON operation of the note conversion start switch. In this note conversion process, sampled audio signals and digital sample signals of musical instrument sounds are analyzed in various ways and converted into a pitch sequence, that is, MIDI data that can be displayed in a musical score. Step 13: Based on the digital sample signal obtained as a result of the audio sampling processing in step 12, an effective section detection process for detecting a section where a musical sound exists, that is, an effective section, is performed. Details of the valid section detection processing will be described later. Step 14: As a result of the effective section detection processing in step 13, a steady section detection processing for detecting a steady portion of a musical sound present in each detected effective section is performed. The details of the stationary section detection processing will be described later. Step 15: A pitch sequence determination process for allocating the most optimal note for each steady section obtained as a result of the processes of steps 13 and 14 is performed. That is, in this step, MIDI data is generated. Details of the pitch sequence determination processing will be described later. Step 16: A score creation process for creating a score based on the MIDI data generated by the process of step 15 is performed. Since the score creation processing can be easily realized by the conventional technique, the details are omitted. Step 17: Perform an automatic performance process based on the MIDI data generated by the process of step 15. Since the automatic performance processing can be easily realized by the conventional technique, the details are omitted.

FIG. 3 is a diagram showing details of the valid section detection processing in step 13 of FIG. Hereinafter, how the effective section is detected from the digital sample signal obtained in step 12 will be described with reference to FIGS. 9 and 10. Step 31: An average sound pressure level is calculated based on the digital sample signal obtained in step 12. FIG. 9 is a diagram illustrating an example of a waveform value of a voice signal sampled at a sampling frequency of 44.1 kHz, that is, a digital sample signal. FIG. 9 shows waveform values for about 20 points. In step 31,
The average of the sample amplitude values over a predetermined number of samples (for example, the number of samples corresponding to a time equivalent to 10 msec) is obtained, and the average is used as the average sound pressure level. Therefore, in the case of the sampling period of 44.1 kHz, the predetermined number of samples is “441”, and the average value of a certain sample point is the sum of the points 10 msec before the last point, that is, the point. From 44
This is a value obtained by dividing the sum of the waveform values one point before by 441. In addition, 44 points from 0 points to 440 points
Since there is no waveform value for one point, the average of the waveform values from 0 point to the corresponding point is set as the average value of that point. In this way, time-series average sound pressure level information is obtained for each sample timing. In FIG.
For convenience of explanation, the case where the average value of the waveform values for 15 points is set as the average sound pressure level is shown. Therefore, the first one
Up to 5 points, the total value of the waveform values up to that point is divided by the number of points. The sum of the waveform values is obtained by summing the absolute values. FIG.
(A) is a graph showing the average sound pressure level value obtained in this way, with the sampling point being taken on the horizontal axis. Hereinafter, this average sound pressure level is referred to as an average level curve. When the average sound pressure level is obtained at every 15 points as shown in FIG. 9, a low-pass filter with a cutoff frequency of about 10 Hz is applied to smooth the level fluctuation. Therefore, when actually averaging the waveform values for 441 points, a low-pass filter with a cutoff frequency of about 80 to 100 Hz is applied,
It is desirable to smooth the level fluctuation. Also, here, a case has been described in which the average value of a certain sampling point is obtained, and the average sound pressure level is obtained by summing the waveform values of a predetermined number of points before that point. The waveform values at a predetermined number of points before and after may be totaled, or the waveform values at a predetermined number of points after the sampling point may be totaled.

Step 32: The average level curve as shown in FIG. 10 obtained in step 31 is classified into a valid section and an invalid section based on a predetermined threshold value. In this process, the threshold value is set to 20% of the maximum waveform value in the average level curve. It goes without saying that other values may be used as threshold values. For example, the average value of the average level curve may be set as the threshold value, 80% of the average value may be set as the threshold value, or a half value of the maximum value of the average level curve may be set as the threshold value. The threshold value is indicated by a dotted line as shown in FIG. Therefore, the intersection between the dotted line (threshold value) and the average level curve is a boundary between the valid section and the invalid section. . In FIG. 10 (B), the valid section is indicated by a circle,
Invalid sections are indicated by crosses.

Step 33: If the minimum length required for human recognition of the pitch is 0.05 msec, an invalid section smaller than the minimum length is set as an effective section from the invalid sections determined in step 32. change. For example, when the sampling period is 44.1 kHz, invalid sections of 2205 or less in sampling number are changed to valid sections.
In FIG. 10B, the third and fifth invalid sections from the left correspond to this short invalid section. Therefore,
As a result of the processing in step 33, FIG.
As shown in (C), the effective section is extended. In this process, the invalid sections existing at the beginning and end of the entire section correspond to short invalid sections, but are represented by special symbols that are not changed to valid sections just because they are short, using a triangle. doing.

Step 34: A process for changing a short effective section of 0.05 msec or less from the pattern of the effective section and the invalid section obtained as a result of the processing of the step 33 to an invalid section is performed. This process is performed in the same manner as in step 33. In FIG. 10C, the effective section at the right end corresponds to the short effective section. Therefore, as a result of the processing in step 34, FIG. 10C becomes as shown in FIG. As is clear from FIG. 10D, the effective sections are all four sections from the first section to the fourth section.
Note that the mark at the end of the section is regarded as the fourth effective section. Step 35: The average value of the average level curve of the effective section specified in step 34 is obtained, and if the average value is smaller than a predetermined value, the final effective section is checked to make that part an invalid section. This average value is obtained by dividing the sum of the average sound pressure level values of the points existing in the effective section by the effective section length. The average value of the average sound pressure levels obtained in this way is shown below each section in FIG. The first section is 60, the second section is 25,
The third section is 45 and the fourth section is 15. If the average value of the average sound pressure level falls below 30% of the maximum waveform value of the section, the section is regarded as an invalid section. Here, since the second section and the fourth section correspond to each other, each section becomes an invalid section. FIG. 10E is a diagram showing the valid section and the invalid section specified by the valid section check processing in step 35.

Step 36: A process for extending the valid section specified by the processes of steps 31 to 35 is performed. For example, as shown in FIG.
With 5% as the extension permission level, a line is drawn on that portion, and the boundary line for specifying the valid section is extended to the line of the extension permission level. That is, the extension processing is performed while checking whether the average level curve is below the extension permission level while checking the rise and fall of the average level curve from the end of each effective section to the outside. At this time, when the descending is reversed to the ascending or when the descending is below the extended permission level, the area up to that point is regarded as an effective section. Also,
FIG. 10 (G) is a diagram showing another example of the effective section extension processing. The extension permission level is set to 5% of the maximum waveform value, and the position where the lowering of the average level curve ends is set as the end of the effective section. Alternatively, the position where the rise has started may be set as the end. According to this extension processing, the extension width of the first section and the third section is larger than in the case of FIG. In this way, an effective section that can be recognized as a pitch by a human is finally determined. If the extension permission level is low and the effective section is at a short distance, the extension position at the end of a certain effective section may be close to the extension position at the beginning of the next effective section. It may also be at the same position. In addition, the boundary position changes depending on whether a part where the descent ends or a part where the ascent starts is delimited. If the effective sections overlap as a result of this extension processing, both intermediate positions may be set as boundary positions. 10 (F) and 10 (G), the case where the effective section is extended before and after has been described. However, only the forward direction or the backward direction may be performed. Further, when extending forward and backward, the extension permission level may be different between the forward direction and the backward direction.

FIG. 4 is a diagram showing the details of the stationary section detection processing in step 14 of FIG. Details of how the steady section is detected from the effective sections obtained in step 13 will be described with reference to FIGS. 11 to 18. When analyzing musical audio signals such as voices and musical sounds, it is important to know where the stationary part is. For non-rhythm sounds, the pitch is determined by the periodicity of the stationary part.
This is because the sound value is determined using the stationary part as a skeleton. In this embodiment, the stationary part is a section corresponding to one note when expressed as a musical score. Means to detect on the time axis a section recognized as. Hereinafter, the stationary section detection processing will be described according to the steps in FIG.

Step 41: A process of detecting a reference position for one cycle is performed for all the effective sections obtained by the effective section detection processing of FIG. In order to detect the reference position of the cycle, it is general that either one of the 0 cross position detection and the peak position detection is generally used. In order to detect the reference position of the cycle by detecting the 0 cross position, it is difficult to detect the overtone without removing as much as possible of the overtone with a filter or the like, and band division is also required. In the case of peak position detection, it is desirable to remove harmonics as much as possible. However, since it is not as severe as the detection of the 0 cross position, it is only necessary to apply a band-pass filter with the cut-off frequency of the audible frequency band of the voice or musical instrument. There is no particular need to perform the processing of. Therefore, the peak position detection is a simpler procedure and a method that can obtain a reasonable result, which is desirable. Therefore, in this embodiment, a case where the reference position of the cycle is detected by the peak position detection will be described. First, before performing peak position detection, harmonic overtones are removed by a filter. This is to apply a band-pass filter using a soundable band as a cutoff frequency. In the case of voice, the band that humans can pronounce is 80-1
The frequency is about 000 Hz, which is necessary for almighty analysis without limiting the user. However,
When the number of users is limited, it is possible to reduce errors due to overtones and to improve accuracy by narrowing the band in which sound can be generated to some extent. 80 ~
Although the frequency is about 700 Hz, the accuracy is improved if the pitch frame is determined in advance. The accuracy can be improved by setting in advance the differences for each instrument. As shown in FIG. 11, the peak position of the musical sound waveform in the effective section is detected. This peak position detection method is performed by a known method. The peak level of the musical sound waveform is detected, and this is held by a predetermined time constant circuit.
The peak level is set as a threshold voltage, and the case where the voltage becomes higher than the threshold voltage is held as the next peak level, and the peak level as shown in FIG. be able to. FIG. 11A is a diagram showing a state of a threshold hold voltage when detecting this peak position. From the audio waveform of FIG. 11A, FIG.
Will be detected. The peak positions P1, P3, P5, P7, PA, PC, and PE all regularly appear at a predetermined period, but the peak position P9 appears due to a slight disturbance of the audio waveform. You are. Hereinafter, such an incorrect peak position is corrected, and a stationary section at the correct peak position is detected.

Step 42: Based on the periodic reference position detected in step 41, a basic section starting from a certain peak reference position and a section up to the next peak reference position immediately after the basic section (hereinafter referred to as a moving section). ) Are compared with each other. As shown in FIG.
Considering the peak reference position P7, the first basic section is a section 79 from the peak reference position P7 to the next peak reference position P9. However, since the section 79 has a length equal to or less than the minimum band length, the section 79 is extended to a section 7A from the peak reference position P7 to the next peak reference position PA. Since this section 7A is larger than the minimum band length and smaller than the maximum band length, it is determined to be this. Next, the moving section is the peak reference position P
A section AC from A to the peak reference position PC. The error rate between the section 7A and the section AC is calculated, and if the error rate is equal to or more than a predetermined value, it is determined that the two sections do not match, and the length of the moving section is further increased. Then, an error rate is calculated between the enlarged moving section and the basic section. If the error rate is equal to or more than a predetermined value, it is determined that the basic section at that time does not match, and the basic section is expanded. Therefore, the basic section is a section 7 from the peak reference position P7 to the next peak reference position PC.
Extended to C. However, since this section 7C is larger than the maximum band length, the comparison process is stopped, and it is determined that there is no match in this section. If, as a result of the comparison between the sections 7A and AC, the error rate is smaller than a predetermined value (for example, 10), it is determined that the two sections match, and the sections 3 and 4 starting from the next peak reference position PA The same processing is performed for the moving section of. The method of calculating the error rate will be described later. At this time, the working memory (RA
M) has a data area in which the data of the peak reference position information, the error rate at that time, and the match flag are respectively written. In the case described above, when the section 7A and the section AC match, the peak reference position information P7, the error rate at that time, and the match flag indicating the match are written in the data area. On the other hand, when the sections 2 and 3 do not match, only the mismatch flag is written.

FIG. 12 is a diagram for explaining a method of calculating an error rate performed in the waveform comparison processing. First, it is assumed that two waveforms whose error rates are to be calculated are a comparison wave 1X and a comparison wave 2X as shown in FIG. First, with respect to the comparison wave 1X and the comparison wave 2X, the amplitude values are normalized such that the maximum amplitude value becomes 100%. Here, since the comparative wave 2X has a smaller size in the time axis (horizontal axis) direction than the comparative wave 1X, the comparative wave 2X extends so as to have the same time width as the comparative wave 1X. Thus, the comparison wave 1X becomes the comparison wave 1Y, the comparison wave 2X becomes the comparison wave 2Y, and finally becomes the comparison wave 2Z due to the elongation of the time axis. The error rate is calculated between the comparison wave 1Y and the comparison wave 2Z. FIG. 13 is a diagram illustrating a specific example in the case where the error rate between the comparison wave 1Y and the comparison wave 2Z is calculated. In the figure, a case will be described in which the error rates are calculated for the waveforms of the first cycle of the comparison wave 1Y and the comparison wave 2Z, that is, for 24 sampling waves. The difference is calculated for the same sampling position of the comparison wave 1Y and the comparison wave 2Z, and the sum of the absolute values of the difference is calculated. In the case of FIG. 13, the sum of the absolute values is 122. By dividing this by the sampling number 24, the error rate is obtained. In this case, the error rate is 5. In this case, if the predetermined value is 10, the error rate is 10 or less, so that the same waveform is processed.
In FIG. 13, each waveform is normalized with 1000 as the maximum level. By performing the waveform comparison processing in this manner, the peak reference position P9 in FIG.
Is canceled, and regular peak positions are detected as shown in FIG.

Step 43: Using the result of the waveform comparison processing in step 42, the sections having an error rate smaller than a predetermined value (for example, 10) are connected to each other, and this is used as a pseudo matching section. The maximum value and the minimum value are detected, and the cutoff frequency band is determined based on the maximum value and the minimum value. For example, it is assumed that the minimum value among a plurality of matching sections obtained as a result of the waveform comparison processing is 235 points and the maximum value is 365 points.
In order to allow a margin for this matching section, the minimum value is set to 1
Assuming that the discount is reduced and the maximum value is increased by 10%, the matching section is approximately 2
From 12 points to about 402 points. This corresponds to a frequency band of an audio signal of 110 Hz to 208 Hz when the sampling frequency is 44.1 kHz.
Accordingly, the cutoff frequency band is from 110 Hz to 208 Hz. Step 44: The same processing as in steps 41 and 42 is repeatedly executed using the new cutoff frequency determined in step 43. For example, in the case described above, the cutoff frequency is set in the range of 110 Hz to 208 Hz, and the cycle reference position detection processing in step 41 and the waveform comparison processing in step 42 are repeated to detect a continuous section (coincidence section) of the same waveform. . As a result, low frequencies and harmonics that cause errors are cut off, so that more accurate processing can be performed, and a matching section with higher accuracy than the previous time can be obtained. By the same waveform section detection processing in step 44, the audio waveform as shown in FIG.
Divided into three sections X, Y, and Z, such as
In other words, the section X and the section Z have a form in which continuity is interrupted by the section Y. In FIG. 14B, both the plus side and the minus side of the audio waveform are interrupted at similar places. In the case of the speech waveform as shown in FIG. 15A, since the fundamental components in the fourth and fifth cycles are slightly smaller, the peak on the plus side cannot be recognized well, and as shown in FIG. Will be detected. Therefore, FIG.
When the processing from step 41 to step 44 is performed on the audio waveform as shown in (A), the continuity is broken between the peak position P5 and the peak position PB only on the plus side as shown in FIG. It takes shape. However, when this is a stationary section, the fundamental tone component always appears remarkably on either the plus side or the minus side, and therefore, FIG.
In the case of the voice waveform of, the negative side of the voice waveform is regularly recognized as the same waveform section. Therefore, assuming that the error rate in the waveform comparison process in step 42 is within the allowable range, the steady section becomes a range indicated by a minus arrow as shown in FIG. It will be recognized that it was. In the case of an audio waveform, there is a possibility that the fundamental tone component is disturbed on both the positive side and the negative side.
The superimposing process of the stationary part is performed in order to take countermeasures.

Step 45: If the section corresponding to the double-headed arrow as shown in FIG. 16A is determined to be the same waveform section by the processing up to step 44, the section is extended. In other words, the beginning of the same waveform section (the part marked with ○)
The error rate is obtained by the same method as the waveform comparison process of step 42, and the waveforms at the end and the end (marked with x) are set as the basic waveforms, and the sections on both sides thereof are set as the moving sections. At this time, the threshold value of the error rate is set higher than in the case of step 42 (for example, the error rate 15
), The same waveform section can be extended to a position shown by a dotted arrow in FIG. 16B. However, if the result of the extension overlaps the adjacent section, the extension process is stopped at that point. The same waveform section extended in this way becomes a stationary section on each of the plus side and the minus side of the voice waveform. Step 46: Since the processes from step 41 to step 45 are also performed on the plus side and the minus side of the audio waveform, the stationary sections independently obtained on both sides are overlapped. For example, as a result of the processing up to step 45, it is assumed that the steady section on the plus side and the minus side is in the range of the arrow as shown in FIG. When these positive side and negative side normal sections are overlapped with each other, FIG.
Is the final stationary section. In this case, each of the five steady sections existing on the plus side and the minus side becomes four steady sections as a result of the steady section overlapping processing in step 46.

Step 47: Subdivision processing based on changes in pitch and sound pressure is further performed on the stationary section obtained by the processing up to step 46. In the stationary section detection processing up to step 46, since the waveform is expanded and compared, even if the pitch change of the voice waveform is caused by a continuous vowel such as "Oh", it is regarded as one and the same sound. It is a mechanism. Therefore, in the case of the musical sound waveform of the musical instrument sound, a situation may occur in which the pitch change of the continuous musical instrument sound cannot be found. Therefore, in this embodiment, the state of the pitch change is checked for each steady part section obtained by the processing up to step 46, and it is determined whether or not it is necessary to further divide the tone according to the state. When it is determined that it is necessary, the stationary part section is further divided. For example, the frequency at the periodic reference position is calculated by calculating the length between the periodic reference positions in a certain stationary section and dividing the length by the sampling frequency. FIG. 18A shows the value of the frequency of each waveform constituting the stationary section, and also shows the note distance obtained by quantifying the difference between the frequency of the section and the frequency of the previous section on a linear axis corresponding to the note. . Here, note distance is log (frequency to be compared / source frequency) / log
It is calculated by the equation (12√2). The value of this note distance is ± 0.
When the value is within the range of 5, it is considered that the pitch does not change suddenly, and when the value is larger than this, it is considered that the pitch has changed suddenly, and that part is defined as a section break and the steady section is defined as the section. Separate further. For example, FIG.
In the case of (A), since the tenth to twelfth note distances in the stationary section are both greater than 0.5, it is considered that the pitch has changed abruptly in that part. It is divided into the first to ninth sections and the thirteenth to twenty-fourth sections. This is equivalent to detecting a section of a note when the tone does not change but the pitch changes.

Next, since the same can be said for the sound pressure as in the case of the sound pitch change, the position where the sound pressure level has sharply changed is detected, and the steady section is further finely divided at that position. FIG. 18B shows the value of the average sound pressure level of each waveform constituting the steady section, and also shows the amplification ratio between the average sound pressure level of that section and the average sound pressure level of the previous section. This amplification ratio is obtained by log (average level of previous section / average sound pressure level of that section). If the value of this amplification ratio is within the range of ± 0.01, it is considered that there is no sudden change in sound pressure, and if the value is larger than this, it is considered that the sound pressure has changed suddenly, and that part is regarded as a section. , The stationary section is further finely separated. For example, in the case of FIG. 18B, since the sixteenth and seventeenth portions of the steady section are larger than 0.01, that portion is regarded as a portion where the sound pressure level changes rapidly. Therefore, it is divided into three sections like the initial division section. However, since a person can sense a sound for about 0.01 to 0.1 seconds, the minimum length of the section is determined accordingly, and a section shorter than that is connected to the next section. Therefore, the second section (the portion marked with a circle) of the initial division section in FIG. 18B is connected to the third section, and finally, the second section as shown on the right side of the initial division section.
It will be divided into sections.

FIG. 19 is a diagram for illustrating the concept of the stationary section detection processing of FIG. In FIG. 19, the valid section is a result detected by the valid section detection processing of FIG.
The same waveform section is a section obtained by the processing from step 41 to step 46. The same pitch section is a section obtained by further subdividing the same pitch section at the portion where the pitch change is abrupt in step 47. The same sound pressure section is a section obtained by further dividing this same sound pitch section by a portion where the sound pressure changes rapidly. Although the case has been described in which the steady section (same waveform section) is divided by the note distance and the amplification ratio, either one of the divided results may be employed, or both may be employed. When both are adopted, adjustment may be performed based on the minimum length of the section in the same manner as described above. Further, when dividing by the note distance and the amplification ratio, one of the dividing processes may be preferentially performed, and the other dividing process may be performed only when the division is not performed as a result.

FIG. 5 is a diagram showing details of the pitch sequence determination processing in step 15 of FIG. An optimal pitch sequence is determined for each stationary section detected in step 14. In this embodiment, four types of pitch sequence determination processing will be described with reference to FIGS. 20 to 25.
When voices, musical sounds, and the like are finally converted to note information, the melody greatly changes depending on the pitch to which a certain specific frequency is rounded, and in many cases, the desired detection cannot be performed. Therefore, in this embodiment, the pitch is determined mainly by the relative sound, and the pitch sequence is determined by selecting the most suitable pitch transition using the key. First, a pitch sequence determination process 1 which is a first embodiment of the pitch sequence determination process will be described with reference to the flowchart of FIG. Step 51: For each stationary section obtained by the effective section detection processing of step 13 (FIG. 3) and the stationary section detection processing of step 14 (FIG. 4), the representative frequency of the section is determined. FIG. 20A is a diagram illustrating an example of a finally obtained steady section. Here, it is assumed that a total of 12 sections have been detected, and each section is enclosed in parentheses.

Section numbers [0] to [12] are assigned. What is important when determining the representative frequency of each stationary section is to identify the trend of the frequency from the periodic position of each stationary section and determine one unique frequency for that section. As a method for this, the first method uses the average frequency of the entire stationary section as the representative frequency of the section. In the second method, a cycle (frequency) near the middle of a stationary section is set as a representative frequency of the section. In the third method, the average frequency of the portion where the pitch is stable is set as the representative frequency of the section. In this embodiment, the representative frequency is calculated using the error rate calculated at the time of the stationary section detection processing in step 14 of FIG. That is, using the same waveform section detection processing before the steady section expansion processing of step 45 in FIG. 4, the average of the frequencies in the stable steady sections in which the error rates are equal to or less than a predetermined value (for example, 10) is calculated. And set it as the representative frequency of the stationary section. For example, it is assumed that in the same waveform section detection processing in step 44 of FIG. 4, when the error rate of the adjacent section is 10 or less, it is determined that the waveforms match, and a steady section is detected. In this case, it is assumed that the information of each waveform section constituting the steady section in this case (before the steady section extension processing) is as shown in FIG. That is, as shown in FIG. 20B, the stationary section is composed of 12 waveform sections having an error rate of 10 or less. The cycle length and error rate of each waveform section are as shown. In this case, the average value of the cycle length in this steady section is 255.833. Here, since the cycle length is represented by the number of samples, the sampling frequency is 4
Since the frequency is 4.1 kHz, the representative frequency of this stationary section is obtained by dividing the sampling frequency by the average value of the period length. In the case of FIG.
z. In this case, the value of the representative frequency treats two decimal places as valid. FIG. 20 (C) thus shows FIG.
It is a figure showing the result of having computed the representative frequency of each steady section like (A).

Step 52: When the representative frequency of each stationary section is determined by the processing of step 51, the note distance between consecutive stationary section numbers of successive stationary sections is determined based on the representative frequency. The note distance is determined in the same manner as the arithmetic expression used in step 47 of FIG. FIG. 20C shows an example of the note distance calculated in this way. Step 53: The calculated note distance is rounded off to one digit after the decimal point, and the note distance is rounded to each pitch on the 12th scale. For example, in the case of FIG. 20C, each note distance is rounded off and becomes a real number in the right column. Since this real number indicates the difference in note number from the previous pitch, the pitch sequence data can be completed by determining the first pitch. The pitch sequence data shown in the rightmost column of FIG. 20C is data indicating a state of a pitch transition when the first pitch is set to 0. That is, FIG.
In the case of (C), 0-2--4-5-2-3... Step 54: Determine the pitch of the first sound. First, the simplest method is to assign a 60 note number (note name C4) sound to the first sound as a default value. That is, in the case of the MIDI standard, the limit of the note number is 0 to 1
27, the note number 60
(Note name C4) is assigned. As a result, the pitch can be increased by 67 semitones on the high frequency side (plus side) and by 60 semitones on the low frequency side (minus side).
In this case, the data indicating the pitch sequence in the rightmost column in FIG. 20C is 60 (C4) -62 (D4) -64 (E4).
-65 (F4) -62 (D4) -63 (D♯4) ・ ・ ・
・ It becomes.

Step 55: Correct the pitch sequence data determined in Step 54. That is, the swing range of the pitch sequence data determined in step 54 is detected, and if the swing range is -60 or less on the low frequency side (minus side), the default value 60 is set according to the minimum swing width. Fix it.
This correction is performed by shifting the default value upward so that the note with the minimum swing is 0 or more. For example, if the minimum swing width is −64, the calculation formula −60−
According to the result of (−62) = 4, the default value 60 is set to 4
Shift up by a note and assign 64 as the first note. Even in the case where the swing is +67 or more on the treble side (plus side), the default value 60 may be similarly corrected in accordance with the maximum swing width. It is impossible to judge from the human utterance band that the swing width exceeds both the bass side (minus side) and the treble side (plus side), and such a case is excluded. If such a situation can occur, the range may be set in a range of 0 to 256. In step 54, the case where the pitch of the first sound is determined as a default value (for example, 60) and the pitch sequence data is created has been described. However, the present invention is not limited to this. It is also possible to detect the frequency of a close genuine scale and apply it to the scale. For example, in the case of FIG.

Since the representative frequency of [0] is 172.38 Hz, the pitch of the first sound is determined to be the note number 53 (note name F3) closest thereto. As a result, the data indicating the pitch sequence in FIG. 20C is 53 (F3) -55 (G3) -5.
7 (A3) -58 (A♯3) -55 (G3) -56 (G
♯3) ...

Next, a pitch sequence determining process 2 which is a second embodiment of the pitch sequence determining process will be described with reference to the flowchart of FIG. Steps 61 and 62 in the pitch sequence determination processing 2 are the same as steps 51 and 52 in FIG. 5 described above, and thus the processing from step 63 will be described here. Step 63: Using the calculated note distance, a process of accumulating note assignment errors when rounding to each pitch on a plurality of scales is performed. That is, in this embodiment, three types of scales, a natural scale, a harmony scale, and a melody scale,
Calculate the fitness of each rounding. As shown in FIG. 21, in the natural scale, assignable sounds are arranged in the order of whole tone, half tone, whole tone, whole tone, half tone, whole tone, and half tone. The harmony scale is arranged in the order of a whole tone, a semitone, a whole tone, a whole tone, a half tone, three semitones (a whole tone + half tone), and a semitone. In addition, the melody scale, when ascending, is a whole tone, a semitone, a whole tone, a whole tone, a whole tone, a whole tone,
The order of the semitones is the same as the natural scale when descending. In FIG. 21, a mark “も の” indicates a sound that can be adopted as a scale constituent sound, and a mark “X” indicates a sound that cannot be adopted as a scale constituent sound. For each scale in FIG. 21, assuming that the first sound starts from the pitch of each circle, the sound of each stationary section number is sequentially assigned so as not to select the cross. At this time, a pitch difference, that is, a note allocation error between the sound of the stationary section number and the allocated sound is calculated, and the total is calculated. For example, stationary section number

When the note distance to [0] is pitch sequence data as shown in FIG.
Let's assign up to the regular section number [5] to the natural scale of FIG. First, the stationary section number shown in FIG.

Since the sound [0] is the first sound, it is assigned to the note position (0) of the natural scale in FIG. Next, the sound of the stationary section number [1] in FIG.

Since the note distance is 1.7158 for the sound [0], a semitone or two semitones (whole sound) should be selected as the note distance. At this time,
In the natural scale, a semitone sound at the note distance, that is, the note position (1) is marked with a cross, so that it is not selected, and is assigned to a two semitone (whole sound) sound at the note distance, that is, the note position (2). Accordingly, since the note with the note distance of 1.7158 is assigned to the note position (2), the note assignment error of the sound of the stationary section number [1] is the note distance of 2 to the actually assigned note position (2). And stationary section number

Note distance 1.715 between [0] and steady section number [1]
8 and the value is 0.2842. Next, since the note distance of the sound of the steady section number [2] is 2.1557 with respect to the sound of the steady section number [1], the whole note or 3 semitones (full note + semitone) is selected as the note distance. Since the sound of the steady section number [1] has been assigned to the note position (2) in the previous processing, the steady section number [2]
Is assigned to the note position (4) or (5) at the note distance of a full tone or three semitones (full tone + half tone).
At this time, the sound at the note position (4) is marked with a cross in the natural scale, so the sound of the stationary section number [2] is assigned to the note position (5). Therefore, note distance 2.1
Note position where the sound of 557 corresponds to note distance 3 (5)
Will be assigned to Therefore, the note allocation error of the sound of the steady section number [2] is 0.8443. The cumulative total of the case of the previous steady section number [1] and the case of the current steady section number [2] is 0.2842 + 0.8443.
= 1.1285. In this way, the calculation is performed for the remaining sounds from the steady section number [3] to the steady section number [5], and the total value of the note assignment errors is calculated.
2.233. This value is obtained when the note position (0) of the natural scale is set to sound output. Therefore, the calculation of the total value of the note assignment error is performed for the note positions (2), (3), (5), (7), (8), and (10) of the natural scale, and the harmonic scale is also calculated. And for each note position of the melodic scale. FIG. 22A is a diagram showing the cumulative value of the note assignment error when each note position is set to output sound for each scale.

Step 64: Calculate the total value of note assignment errors of 0.5 or more when rounded to each pitch on a plurality of scales. That is, in the case of step 63, the total value of all values of the note assignment error was calculated,
In this step, when the note assignment error is 0.5 or more, that is, in the case of FIG.
Is ideally rounded to the position of note distance (2) and stationary section number [2] is rounded to the position of note distance (3), ideally with a small note allocation error. There are cases where they cannot be assigned. In such a case, the allocation is corrected to a pitch other than the pitch closest to the note distance. Therefore, in such a case, that is, when the note allocation error is 0.5 or more, the total value is calculated as the note correction error. FIG. 22B is a diagram showing the cumulative value of the note correction errors, which corresponds to FIG. Step 65: The number of notes having a note assignment error of 0.5 or more in step 64, that is, the total number of notes used for calculating the total value in step 64 is calculated. FIG.
FIG. 22C shows the total number of notes, and FIG.
(B).

Step 66: The result of the processing from step 63 to step 65, that is, FIG.
By using the calculation results of (C) to (C), the most suitable scale and its starting sound are determined. This determination method is shown in FIG.
FIG. 22 (A) shows the smallest note assignment error, FIG. 22 (B) shows the smallest note assignment error of 0.5 or more, and FIG. 22 (C) shows the smallest note assignment error of 0.5 or more. The one with the smallest number of notes, the combination of these as appropriate, and the like can be considered. However, it is not always the case that the number is finally determined to be one depending on the relative tone (related tone). In such a case, no matter which method is used, the same melody transition results, so any decision method may be selected. Therefore,
In FIG. 22 (A), the minimum value of the cumulative value of the note assignment error is 1.688, which is present at four places in total: two places in the natural scale, one place in the harmony scale, and one place in the melody scale. I do. Next, a minimum value in FIG. 22B is searched for these four locations. Then, the values are 0.891, which is the same value in all four places. Therefore, any one of these four locations may be selected. Here, the upper scale is prioritized, and the left note position is prioritized. Therefore, the scale is finally determined by the natural scale, and the starting sound is determined by the third. Note that FIG.
In 2nd, the scales are displayed in minor scales to unify the display of scales.
This is the first (tonic) sound of the scale. Step 67: Based on the scale and the starting sound determined in the step 66, the same processing as the note assignment error calculation processing is performed again to determine a pitch sequence. Thus, FIG.
FIG. 2 shows a pitch sequence assigned to a stationary section number of 0 (C).
2 (D). In the case of FIG. 20C, the pitch sequence is 0-
2-4-5-2-3..., But this time the pitch sequence is 0-2--4-5-2-4. As is clear from this, the pitch sequence of the stationary section number [5] is “4”,
It can be clearly understood that this is different from “3” in FIG.

In the pitch sequence determination processing 2 shown in FIG. 6, since the unstable sound is rounded to the scale sound by rounding the sound using the scale, the melody approaches a relatively stable melody, so that the desired sound of the user can be obtained. Very likely to approach sound. However, when a user with a good sense of sound inputs a melody by intentionally shifting a sound other than the scale-constituting sound, such a process of rounding the sound to the scale-constituting sound is inappropriate. Ultimately, there may be cases where it is necessary to judge the quality of the melody, but this is not possible with the current technology. Therefore, as in pitch sequence determination processing 2 in FIG. Only special sounds may be recognized as sounds other than scale constituent sounds. If the note distance to a sound other than the scale-constituting sound is smaller than a certain value, a process for rounding the sound other than the scale-constituting sound to that sound is performed. This corresponds to the pitch sequence determination process 1 in FIG. 5 and the pitch sequence determination process 2 in FIG.
This corresponds to an intermediate sound rounding process located between. After the pitch sequence determination process 2 of FIG. 6 determines that the third natural tone is the first sound (ie, after the process of step 66), a process of determining a note error allowable range is performed. This processing includes a method of setting a constant such as 0.2 as a note error allowable range in advance, and a method of calculating by a calculation. The method of obtaining by calculation is to calculate the average value of the note distance when the note distance is rounded to 0.5 or less when performing the process of rounding to the scale component sound in the process of step 63 in FIG. A method is conceivable in which the pitch deviation tendency of the person) is grasped, and a constant multiple of the average value is set as a note error allowable range. In this embodiment, a description will be given of what pitch sequence is assigned to each stationary section in FIG. 20C when a constant value of 0.2 is adopted as the note allowable error range. First, by the processing of step 66 in FIG. 6, it is determined that the third sound of the natural minor scale is the first sound, so that the note distance of the stationary section number [1] is 1.
Since it is 7158, the closest pitch is the sound at the fourth note, that is, the note position (5). In this case, there is no problem if the pitch is rounded to a close pitch, and the sound at the note position (5) is determined. Next, note positions (7), (8), and (5) are sequentially determined as sounds in the scale for the steady section numbers [2], [3], and [4]. However,
For the steady section number [5], the note distance is 1.10
93, and the steady interval number [4] is determined at the note position (5), so the closest pitch is the note position (6). The pitch at the note position (6) is a sound other than the scale component sound. Therefore, in this case, it is determined whether or not the value is within the note allowable error range. In this case, since the note distance is 1.1093, the error is 0.1093, which is less than or equal to the note allowable error range of 0.2. Will be. If the note distance of the steady section number [5] is, for example, 1.2093, the error is 0.2093, which is larger than the note allowable error range of 0.2. Is determined to be the note at the note position (7), which is the musical note that is one level higher than that. By setting the note tolerance range in this way, it is possible to add sounds other than scale constituent sounds to the pitch sequence, so that it is possible to consider sounds other than scale constituent sounds while using scales, It becomes possible to determine a pitch sequence close to the melody sung and imaged.

Next, a pitch sequence determination process 3 which is a third embodiment of the pitch sequence determination process will be described with reference to the flowchart of FIG. In the pitch sequence determination processes 1 and 2 described above, the case where the pitch sequence is determined based on the note distance from the immediately preceding sound has been described. Since the pitch is not determined and the flow of the phrase, that is, the sound that makes up the pitch sequence, affects the first sound of the phrase,
When a phrase is detected and a pitch sequence is determined, a pitch sequence is determined in consideration of a pitch difference from the first sound of the phrase. Step 71: It is determined how long the length of the stationary section detected by the stationary section detection process in FIG. 4 is when it is represented by the number of grids of a sound value sequence (current value sequence). When the steady section is as shown in FIG. 23A, the section from the beginning of each steady section to the beginning of the next steady section is defined as one section in FIG.
Create a note value standard like If one hundredth of a second is defined as one grid, it is determined how many of these grids are composed of these pitch standards. FIG.
(C) is a grid for determining a tone value sequence. Therefore, the position of each sound value reference is corrected in order to adapt the sound value reference as shown in FIG. 23 (B) to the grid of FIG. 23 (C). For example, when the boundary position of the note value reference is located between the grids, the boundary position of the note value reference is changed to the closest grid. If the grid is located exactly in the middle between the grid and the grid, the boundary position based on the note value is changed to the grid on the front side. FIG. 23D shows a sound value section in which the boundary position of the sound value reference is changed in this manner. The upper part of the note value section in FIG. 23 (D) shows the same note value section number as the steady section number, and the lower part shows the grid number of each note value section. The grid number is
4,4,5,6,3,6,11,4,7,3,5,3
10, etc.

Step 72: Since the length of each note value section is defined by the number of grids, a plurality of note value sections are combined into one phrase based on the number of grids. The method of constructing a phrase is described in Japanese Patent Application No. 7-123105 previously filed by the applicant of the present application, and will be briefly described here. First, since one note value section corresponds to one note, the average length of the note value sections (average note value section length) is calculated based on the length of each note value section. A multiplied value is obtained by multiplying the calculated average duration section length by a predetermined coefficient K (a value of 1 or more, for example, 2). A note value section length of a value equal to or larger than the multiplied value obtained in this way is detected. Delimiter data indicating a phrase break is inserted after the detected note value section. The note value sections delimited by the delimiter data constitute one phrase. This time, the average pitch section length is calculated for each of the phrases compiled in this way. The calculated average duration section length is multiplied by a predetermined coefficient L (a value of 1 or more, for example, 2). If the length of the last note value section of each phrase, that is, the end note value section length is smaller than this multiplication value, the phrase delimiter data inserted after the last note value section of the phrase is deleted. If the end note value section length is not less than the multiplication value, nothing is performed. Such a phrase deletion process is performed for all phrases. For example, in the case of FIG. 23D, the total number of grids is 4 + 4 + 5 + 6 + 3 + 6 + 11 + 4 + 7 + 3 + 5 + 3 +
10 = 71. When this is divided by the number of sections 13, 71 ÷ 13 = 5.
46 is obtained. If the value is rounded off to one decimal place, the average pitch length becomes 5. The value obtained by multiplying this 5 by the predetermined coefficient 2 is 10
Becomes Therefore, the note value sections whose note value section length is 10 or more are note value section [6] and note value section [12]. Accordingly, since the delimiter data is inserted after these note value sections [6] and [12], the first phrase is not included in the note value section as shown in FIG.

The second phrase is composed of six pieces of note value sections [7] to [12].

Step 73: The representative frequency of each tone value section of the phrase determined in step 72 is determined. Here, the note value section corresponds to the stationary section of the above-described pitch sequence determination processing 1 of FIG. Step 74: A note distance is determined between the head sound of the phrase, that is, the representative frequency of the first note value section of the phrase. In the pitch sequence determination processes 1 and 2, the note distance is calculated only from the immediately preceding stationary section.

With reference to the representative frequency of [0], the note distance of each note value section of the phrase is determined. FIG. 24 shows the note value section of the first phrase.

[0]-[6]
And the value of the first note value of the first phrase

It is a figure which shows the value of the note distance between [0] and each tone value section [1]-[6] which comprises a 1st phrase. Step 75: Based on the note distance calculated in step 75, step 53 of the above-described pitch sequence determination processing 1
To the step 55 or the processing from the step 63 to the step 66 of the pitch sequence determination process 2 to determine a predetermined pitch sequence.

Next, a description will be given of a pitch sequence determination process 4, which is a fourth embodiment of the pitch sequence determination process, with reference to the flowchart of FIG. In the above-described pitch sequence determination processing 3, the case where the pitch sequence is determined by calculating the note distance between the head sound of the phrase, that is, the sound of the stationary section located at the head of the phrase has been described. Since it may affect not only the first sound of the phrase but also the sound that was pronounced before the corresponding sound was pronounced,
Here, the phrase is detected and the pitch sequence is determined.
In each phrase, the pitch sequence is determined in consideration of the relationship with the sound that was generated before the sound was generated. The processing from step 81 to step 83 is the same as the processing from step 71 to step 73, and a description thereof will be omitted. Step 84: Determine the note distance between each pre-sound in the phrase. This note distance is obtained for the first phrase in FIG. First, the note value section of the first phrase

For [0], since there is no pre-sound, there is no note distance. A note value section [1] is a note value section as a prefix sound

Since [0] is present, the note distance is 1.7158. The note value section [2] is a note value section as a prefix sound

Since [0] and [1] are present, the respective note distances are determined to be 3.8715 and 2.15.
57. Hereinafter, when the note distance between each note value section and the preceding sound in the same manner is calculated, the result is as shown in FIG. Step 85: Weighting is performed based on the time distance between each pre-sound in the phrase. First, the time difference between each note value section and the preceding sound is represented by the number of grids. When this is calculated for the first phrase in FIG. 23 (E), the value is as shown in FIG. 25 (B). Note value section

For [0], there is no pre-sound, so there is no time distance. A note value section [1] is a note value section as a prefix sound

Since [0] exists, the time distance is 4 grids. The note value section [2] is a note value section as a prefix sound

Since [0] and [1] are present, the respective time distances are 8 grids and 4 grids. Hereinafter, when the time distance between each sound value section and the preceding sound in the same manner is calculated, the result is as shown in FIG. The weight is calculated based on the time distance obtained in this manner. The normalized distance value of the reciprocal of the divided value obtained by dividing the time distance of each note value section by the sum of the note value sections is set to be 100, which is a weight based on the time distance of each note value section. For example, in the case of the note value section [2] in FIG.
The time distance is 8 grids and 4 grids. The value obtained by dividing the time distance 8 of the note value section [2] by 12 which is the sum of the note value sections is 8/12 = 2/3, and the time distance 4 is divided by 12 which is the sum of the note value sections. The value obtained is 4/12 = 1 /
3. The sum of the reciprocals 3/2 and 3/1 of this division value is 1
To obtain 00, the reciprocals 3/2 and 3/1 have 2
00/9. Therefore, note value section [2]
Section number

The weight for [0] is 33.3, and the weight for section number [1] is 66.7. In this way, weighting is performed by time distance for each note value section. FIG. 25C shows the values weighted by the time distance in FIG. 25B.

Step 86: Based on the weights calculated by the process of step 85, 12 tones of each note value section are
A process for rounding to a musical scale or a sound on a predetermined musical scale is performed. The process of rounding to a tone on the 12th scale is performed by the processes of steps 53 to 55 in FIG. In the process of rounding to a note on the scale, the processes of steps 63 to 66 in FIG. 6 are performed. At this time, the weight based on the time distance is referred to as the note distance. For example, in FIG. 25 (C), the preceding sound of the note value section [1] is a note value section.

Since only the sound of [0], the note distance is 1.71
58 is used as it is. Therefore, note distance 1.71
Note value section as the interval closest to 58

A sound that is higher than the sound of [0] by a whole tone is selected. Next, when considering the note value section [2], the sound of the note value section [2]

[0] is affected by a weight of 33.3%, and note value section [1] is affected by 66.7%. At this time, the note value section [1] is already a note value section.

Since "2" is determined as the note distance between [0] and the note value interval,

2. Note distance between [0] and note value section [2]
8715 minus the note distance “2”.
8715. On the other hand, the note distance between the note value section [2] and the note value section [1] is 2.1557. Therefore,
The note distance of the note value section [2] is calculated as follows in consideration of the weight. (1.8715 × 33.3 + 2.1577 × 66.6)
/100=2.06 Therefore, the note distance in the note value section [2] is 2.06. Using this note distance of 2.06, processing to round to a note on the 12th scale (the processing of steps 53 to 55 in FIG. 5) or processing to round to a sound on the scale (step 63 in FIG. 6)
(Step 66).

[0053]

According to the sound signal analyzing apparatus of the present invention, even when the pitch or level of the input sound from the microphone or the like slightly fluctuates, a section (effective section) in which a musical sound exists can be easily formed. Can be analyzed. According to the sound signal analyzing device of another invention, even when the pitch or level of the input sound from the microphone or the like slightly fluctuates, it corresponds to a stationary part of a musical sound other than the fluctuated part, that is, one note. It is possible to provide a sound signal analyzer capable of analyzing a portion to be reproduced. According to the performance information generating device of another invention, even when the pitch or level of the input sound from the microphone or the like is slightly fluctuated, note information corresponding to the pitch can be reliably generated.

[Brief description of the drawings]

FIG. 1 is a diagram showing a main flow when the electronic musical instrument of FIG. 2 operates as a performance information generating device.

FIG. 2 is a hardware block diagram showing a configuration of an electronic musical instrument incorporating a musical sound information analyzer and a performance information generator according to the present invention.

FIG. 3 is a diagram showing details of an effective section detection process in step 13 of FIG. 1;

FIG. 4 is a diagram showing details of a steady section detection process in step 14 of FIG. 1;

FIG. 5 is a diagram showing details of a pitch sequence determination process in step 15 of FIG. 1;

FIG. 6 shows a second example of the pitch sequence determination process in step 15 of FIG.
It is a figure which shows the detail of the pitch row | line determination process 2 which is a Example.

FIG. 7 shows a third example of the pitch sequence determination process in step 15 of FIG.
FIG. 14 is a diagram showing details of a pitch sequence determination process 3 which is an embodiment of the present invention.

8 is a fourth example of the pitch sequence determination process in step 15 of FIG.
FIG. 9 is a diagram illustrating details of a pitch sequence determination process 4 according to the embodiment of FIG.

FIG. 9 is a diagram illustrating an example of a waveform value of an audio signal sampled at a sampling frequency of 44.1 kHz, that is, a digital sample signal.

FIG. 10 is a diagram showing a concept of an operation example of an effective section detection process in step 13 of FIG. 1;

11 is a diagram showing a concept of an operation example of a peak position detection process of a musical sound waveform in an effective section, which is an example of a period reference position detection process in step 41 of FIG.

FIG. 12 shows how the calculation method of the error rate performed in the waveform comparison processing in step 42 of FIG. 4 is performed.
FIG. 9 is a diagram illustrating a specific example using two comparison waves.

13 is a diagram showing a specific example of how an error rate is calculated from the two comparison waves in FIG. 12 by the waveform comparison processing in step 42 in FIG. 4;

FIG. 14 is a diagram showing a state in which the peak reference position in FIG. 11B is corrected by the waveform comparison processing in step 42 in FIG. 4, and a regular peak position is detected.

FIG. 15 is a diagram showing a concept for another musical sound waveform in an operation example of a peak position detecting process of a musical sound waveform in an effective section, which is an example of the periodic reference position detecting process in step 41 of FIG.

FIG. 16 is a diagram illustrating an operation example of a steady section extension process in step 45 of FIG. 4;

FIG. 17 is a diagram illustrating an operation example of a steady portion overlapping process in step 46 of FIG. 4;

FIG. 18 is a diagram showing an operation example of a subdivision process based on a change in pitch and sound pressure in step 47 of FIG. 4;

FIG. 19 is a diagram showing a concept of the stationary section detection process in FIG. 4, showing how a stationary section is detected from the effective sections obtained in step 13.

20 is a diagram showing a concept of an operation example of pitch sequence determination processing 1 of FIG.

21 is a diagram illustrating an example of a plurality of scales used in pitch sequence determination processing 2 in FIG.

FIG. 22 is a diagram showing the concept of an operation example of pitch sequence determination processing 2 in FIG. 6;

FIG. 23 is a diagram showing the concept of an operation example of pitch sequence determination processing 3 in FIG. 7;

FIG. 24 is a diagram showing a specific example of a note distance determination process with respect to a phrase leading sound in step 74 of FIG. 7;

FIG. 25 is a diagram showing the concept of an operation example of pitch sequence determination processing 4 in FIG. 8;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... CPU, 2 ... program memory, 3 ... working memory, 4 ... performance data memory, 5 ... key press detection circuit, 6 ...
Microphone interface, 7 ... Switch detection circuit, 8 ...
Display circuit, 9: sound source circuit, 10: keyboard, 1A: microphone, 1B: numeric keypad and various switches, 1C: display, 1D: sound system, 1E: data and address bus

Claims (16)

[Claims] An input means for inputting an arbitrary sound signal from the outside, and an average value of a sample amplitude value of a signal sequentially input from the input means over a predetermined number of samples is obtained. Calculating means for outputting the average sound pressure level information as the average sound pressure level information,
A section that is equal to or greater than a predetermined value as a valid section in which a musical sound is present, and a section that is less than the first predetermined value as an invalid section in which no musical sound is present; If the time length of the invalid section between the invalid sections is shorter than the first predetermined length, the invalid section is changed to a valid section, and the valid section after the change and the valid sections on both sides are changed. A valid section forming means for combining and forming a new valid section; and when the processing by the valid section forming means is completed, the time length of the section in the valid section sandwiched between the invalid sections on both sides is the If the length is less than the predetermined length of 2, the valid section is changed to an invalid section, and the changed invalid section and the invalid sections on both sides are combined to form a new invalid section. At the time when the processing by the first invalid section conversion means is completed. For each of the valid sections, an average value of the average sound pressure level is calculated, and when the average value is less than a second predetermined value, a second invalid section converting means for changing the valid section to an invalid section is provided. A sound signal analyzer characterized by comprising: 2. An input means for inputting an arbitrary sound signal from the outside, and an average value of a sample amplitude value of a signal sequentially input from the input means over a predetermined number of samples is obtained. Calculating means for outputting the average sound pressure level information as the average sound pressure level information,
A section that is equal to or greater than a predetermined value is defined as an effective section, a section that is less than the first predetermined value and a section sandwiched between the valid sections on both sides is defined as an invalid section, and both ends of the average sound pressure level other than this are defined as invalid sections. And a section determining means for setting the section as an undetermined section, and, when the time length of the section is shorter than a first predetermined length in the invalid section sandwiched between the valid sections on both sides, the invalid section is set to valid. A valid section which is changed to a section, and combines the valid section after the change and the valid sections on both sides thereof into a new valid section; and when the processing by the valid section forming means is completed, both sides are invalidated. If the time length of the valid section between the sections is shorter than the second predetermined length, the valid section is changed to the invalid section, and the invalid section after the change and the invalid sections on both sides are changed. Are combined into a new invalid section, which is adjacent to the undetermined section. If the time length of the valid section is shorter than the second predetermined length, the valid section is combined with an invalid section adjacent thereto and an undetermined section to form a new undetermined section. And calculating an average value of the average sound pressure level for each of the valid section and the undetermined section at the time when the processing by the first invalid section forming means is completed. If the value is less than a second predetermined value, the valid section or the undetermined section is changed to an invalid section, and if the value is equal to or more than the second predetermined value, the undetermined section is changed to a valid section.
A sound signal analyzing device, comprising: 3. When the processing by the second invalid section conversion means is completed, the average sound pressure level calculated by the calculation means and a second predetermined value smaller than the first predetermined value are used. The sound signal analyzer according to claim 1, further comprising an extension unit that extends the effective section. 4. An input means for inputting an arbitrary sound signal from the outside, a cycle reference detecting means for detecting a plurality of candidate positions serving as a cycle reference for the sound signal input from the input means, A section detecting means for calculating a degree of coincidence between waveforms of adjacent sections represented by the candidate positions with respect to a sound signal, and connecting the sections having a high degree of coincidence to detect the same waveform section; And a stationary section determining means for detecting a stationary section based on the same waveform section detected by the sound signal analyzing apparatus. 5. An input means for inputting an arbitrary sound signal from outside, and a first cycle reference detection for detecting a plurality of temporary candidate positions serving as a cycle reference for the sound signal input from the input means. Means, frequency band detecting means for detecting a maximum frequency and a minimum frequency of the sound signal based on the tentative candidate position detected by the first cycle reference detecting means, and a frequency band detected by the frequency band detecting means. A filter processing unit that performs band-pass filter processing using a maximum frequency and a minimum frequency as a cutoff frequency on the sound signal input from the input unit; and a period reference for the sound signal output from the filter processing unit. Second period reference detecting means for detecting a plurality of candidate positions; and, for the sound signal, the degree of coincidence of waveforms between adjacent sections represented by the candidate positions. A section detecting means for calculating and connecting those having a high degree of coincidence to detect the same waveform section; and a steady section detecting means for detecting a steady section based on the same waveform section detected by the section detecting means. A sound signal analyzer characterized by comprising: 6. An input section for inputting an arbitrary sound signal from outside, and an effective section analyzing section for analyzing an effective section in which a musical sound is considered to be present from the sound signal input from the input section. A period reference detecting means for detecting a plurality of candidate positions serving as a period reference for each of the positive and negative sides of the sound signal forming the effective section; and for each of the positive and negative sides of the sound signal, A section detecting means for calculating the degree of coincidence of the waveforms of adjacent sections represented by the candidate positions, connecting those having a high degree of coincidence to detect the same waveform section, and positive or negative detected by the section detecting means A timbre section determining means that sets a section formed by superimposing the same waveform sections on both sides as a same timbre section; and a steady state based on the same timbre section determined by the timbre section determining means. The sound signal analysis device, characterized in that it comprises a stationary section determining means for detecting between. 7. An input section for inputting an arbitrary sound signal from outside, and an effective section analyzing section for analyzing an effective section in which a musical sound is considered to be present from the sound signal input from the input section. A first cycle reference detecting means for detecting a plurality of temporary candidate positions serving as a cycle reference for the sound signal forming the effective section; and the temporary candidate detected by the first cycle reference detecting means. Frequency band detecting means for detecting a maximum frequency and a minimum frequency for the entire section or the effective section of the sound signal based on the position; Filter processing means for performing band-pass filter processing for all sections or the effective sections of the sound signal input from the input means; and A second period reference detecting means for detecting a plurality of candidate positions serving as a period reference for the sound signal to be input; Calculating the degree of coincidence of waveforms between sections, connecting sections having a high degree of coincidence to detect the same waveform section, and determining a stationary section based on the same waveform section detected by the section detecting means. A sound signal analyzer comprising: a stationary section determining means for detecting. 8. The method according to claim 1, wherein said effective section detecting means comprises:
The sound signal analyzer according to claim 6, wherein the valid section is detected by the sound signal analyzer according to claim 3. 9. An input means for inputting an arbitrary sound signal from the outside, a stationary section analyzing means for analyzing a stationary section corresponding to one note from the sound signal input from the input means, Frequency determining means for determining a representative frequency for each of the stationary sections analyzed by the stationary section analyzing means; and a representative frequency between two stationary sections preceding and succeeding based on the representative frequency of the stationary section determined by the frequency determining means. Value conversion means for converting the difference between the two constant intervals into a value based on cents, and relative pitch difference data between the two stationary sections based on the value based on cents converted by the cent value conversion means. Pitch difference calculating means for calculating the pitch difference data calculated by the pitch difference calculating means, and assigning a pitch on a predetermined scale to each steady section based on the pitch difference data. Performance information generating device, characterized in that it comprises the assigning means. 10. An input means for inputting an arbitrary sound signal from outside, a steady-state analysis means for analyzing a steady-state section corresponding to one note from the sound signal input from the input means, Frequency determining means for determining a representative frequency for each of the stationary sections analyzed by the section analyzing means; phrase detecting means for detecting a single phrase by collecting a plurality of the stationary sections analyzed by the stationary section analyzing means; Cent value conversion means for converting the difference between the representative frequencies into values based on cents for all the stationary sections existing before the stationary section in one phrase detected by the phrase detecting section; Relative to all the stationary sections existing before the stationary section in one phrase detected by the phrase detecting means. Weight calculating means for calculating a weight based on a proper time distance; and a relative value between the two stationary sections based on a value based on the cent converted by the cent value converting means and the weight calculated by the weight calculating means. Difference calculating means for calculating typical pitch difference data, and pitch assignment for assigning a pitch on a predetermined scale to each steady section based on the pitch difference data calculated by the pitch difference calculating means Means for generating performance information. 11. An input means for inputting an arbitrary sound signal from the outside, a stationary section analyzing means for analyzing a stationary section corresponding to one note from the sound signal input from the input means, Frequency determining means for determining a representative frequency for each of the stationary sections analyzed by the section analyzing means; phrase detecting means for detecting a single phrase by collecting a plurality of the stationary sections analyzed by the stationary section analyzing means; Cent value conversion means for converting the difference between the representative frequency of the first stationary section in one phrase detected by the phrase detecting means and the representative frequency of each other stationary section in the phrase to a value based on cents, The relative pitch difference data between the two stationary sections is calculated based on the value based on the cent converted by the cent value converting means. Pitch difference calculating means, and pitch assigning means for assigning a pitch on a predetermined scale to each steady interval based on the pitch difference data calculated by the pitch difference calculating means, Performance information generator. 12. When assigning a pitch on a predetermined scale to each stationary section, the pitch assigning means assigns a predetermined pitch to an initial stationary section, and then assigns a predetermined pitch to the remaining stationary sections in order. The musical performance information generating apparatus according to claim 9, 10 or 11, wherein a pitch on a musical scale is assigned. 13. The pitch assigning means, when assigning a pitch on a predetermined scale to each stationary section, analyzes a sound signal of a first stationary section and detects an average frequency of the stationary section,
The pitch based on the detected average frequency is assigned as the pitch of the first stationary section, and then the pitch on a predetermined scale is sequentially assigned to the remaining stationary sections. 12. The performance information generator according to item 11. 14. The pitch assigning means assigns pitches on a plurality of scales to each steady interval while shifting note positions, and calculates a total value of note assignment errors at each note position on each scale. 12. The performance information according to claim 9, 10 or 11, wherein an optimum scale is determined according to the accumulated value, and pitches on the determined scale are sequentially assigned as pitches of the stationary section. Generator. 15. The pitch assigning means assigns a pitch outside the scale according to a value of a note allowable error range when assigning the determined pitch on the scale in order as a pitch of the stationary section. The method according to claim 9, 10 or 11, wherein
A performance information generator according to any one of the preceding claims. 16. The method according to claim 5, wherein the stationary section analyzing means includes:
9. The sound signal analyzer according to claim 6, 7 or 8, wherein the stationary section is analyzed.
2. The performance information generator according to 1.