JP2015031738A - Chord progression estimation and detection device and chord progression estimation and detection program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a chord progression estimation and detection device that can estimate and detect more correct chord progressions by musical estimation that considers cadence of performance information.SOLUTION: A chord progression estimation and detection device that estimates a chord from performance information comprises: a chord detection part 32 for detecting a plurality of chords and a likelihood of each chord from the performance information at each specific section; a chord function estimation part 34 for estimating a chord function of each chord at each specific section; a chord function determination part 35 for determining a final chord function progression of the chord function progression obtained by the chord function estimation part 34 according to a plurality of basic cadences; and a chord allocation part 37 for allocating a chord to a chord function of each cadence.

Description

  The present invention relates to a chord progression estimation detecting device that analyzes performance information or an acoustic signal and estimates a chord progression being used, and in particular, a basic cadence (progress of chord function) stored in advance for a performance by an acoustic signal. The present invention relates to a chord progression estimation detection apparatus and a chord progression estimation detection program that estimate chord progression based on

  When popular or rock music is played in a small band, the chord is composed using chords composed of chords, or a sheet that contains only a melody and chord progression called a lead sheet. Is called. Since it is difficult to record the chord progression of a song unless you are an expert with special musical knowledge, a computer is used based on the sound signal output from a sound source such as a music CD that contains the performance sound. The code is detected by the code detecting device.

  For example, as described in Patent Document 1, this type of chord detection apparatus receives an acoustic signal as electronic information, so that each tone level, average beat interval, and The beat position, time signature and bar line position, bass sound, and chord of each bar are detected. When detecting a chord, a plurality of chord candidates are detected, and the detected chord is configured so that a chord can be played with a built-in sound source of the computer.

According to the chord detection device, a plurality of chords are detected for one chord detection section (measures, etc.) in the acoustic signal, and the likelihood (probability) is determined from the strength and duration of the performance sound of the chord. Then, the code having the highest likelihood is detected as the code in the section. Therefore, the chord progression of the music that is the subject of chord detection is a sequence of chords with high likelihood.
On the other hand, generally, the chord progression of music often consists of a combination of a plurality of cadences. Cadence is an indicator of the progress of chord function, and there are several typical ones.

Japanese Patent No. 4767691

However, since the chord progression detected by the above-described conventional chord detection apparatus is determined only by the likelihood of the chord, there is a problem that the chord may be different from the original cadence.
Further, in the conventional method described above, the chord progression has not been detected by musical estimation (considering cadence).

  The present invention has been proposed in view of the above circumstances, and the chord progression estimation detection that can detect and detect the more accurate chord progression by musical estimation in consideration of cadence with respect to performance information or sound signal. An object is to provide a device and a chord progression estimation detection program.

In order to achieve the above object, claim 1 of the present invention is a chord progression estimation detecting device for estimating and detecting a chord being played in performance information or an acoustic signal, and includes the following components: Yes.
Key detection means for detecting the key of a song from the entire performance information or sound signal.
Code detecting means for detecting a plurality of chords and likelihood of each chord for each specific section from the performance information or the sound signal.
Code function estimation means for estimating the code function of each code classified according to the key detected by the key detection means for each specific section.
A chord function determining means for determining a chord function progress obtained by the chord function estimating means based on a plurality of basic cadences.
Code assignment means for assigning codes to the code functions of each cadence.

  2. The chord progression estimation and detection apparatus according to claim 1, wherein the chord function determining means includes reference cadence storage means for storing a plurality of reference cadences, and the chord function progression and the reference cadence are compared and matched. In this case, the code function of the portion is determined as the same code function as the reference cadence.

  3. The chord progression estimation and detection apparatus according to claim 1, wherein the chord function determining means compares the chord function progress with the chord function progress, and if the chord function progression does not coincide, The chord function of the chord function progression is changed so as to match the cadence.

  According to a fourth aspect of the present invention, in the chord progression estimation detecting device according to the first aspect, the main chord progression finally determined by the chord function determining means and a plurality of chord progression candidates different from the main chord progression can be displayed. And a chord progression display means for assigning chords displayed in the chord progression candidates to main chord progressions.

According to a fifth aspect of the present invention, in the chord progression estimation and detection device according to the fourth aspect, the apparatus includes performance means for playing performance information or an acoustic signal in a specified range and playing a chord with a built-in sound source,
The performance information or the acoustic signal and each chord of the main chord progression displayed on the chord progression display means can be played simultaneously.

Claim 6 is a chord progression estimation detection program for estimating chord progression from performance information or an acoustic signal,
A chord detection procedure for detecting the likelihood of each chord and a plurality of chords for each specific section from the performance information or the acoustic signal;
A code function estimation procedure for estimating the code function of each code;
About the estimated chord function progress, a chord function determination procedure for determining the final chord function progress based on a plurality of basic cadences,
A code assignment procedure to assign codes to the code functions of each cadence;
It is characterized by having a computer execute.

  According to the chord progression estimation detection apparatus and chord progression estimation detection program of the present invention, the chord progression estimation detection program estimates the function of each chord detected from performance information or an acoustic signal, and selects a chord according to a pre-stored reference cadence. , You can get more musical chord progression.

  Also, by making it possible to select different chords for the same cadence, a plurality of chord progressions can be proposed.

It is a block diagram which shows the hardware constitutions of the chord progression estimation detection apparatus of this invention. It is a block diagram which shows the main structures of the chord progression estimation detection apparatus of this invention. It is a block diagram which shows the main structures in the other Example of the chord progression estimation detection apparatus. It is the track | truck screen displayed corresponding to the performance information on the display part of the automatic performance apparatus. It is a flowchart which shows the procedure of the chord progression estimation by this invention. It is a table | surface which shows an example of the detection code data arrangement | sequence detected by the chord progression estimation detection apparatus. It is a table | surface which shows the correspondence of the chord function with respect to the chord candidate of each measure detected with the chord progression estimation detection apparatus, and the chord function progress with respect to each measure. It is a flowchart which shows the detailed procedure of the chord function confirmation process in chord progression estimation of a chord progression estimation detection apparatus. It is a flowchart which shows the detailed procedure of the chord allocation process in the chord progression estimation of a chord progression estimation detection apparatus. It is a chord progression screen displayed on the chord progression display means of the chord progression estimation detecting device. It is the track | truck screen displayed corresponding to the performance information on the display part of the automatic performance apparatus.

  Hereinafter, a chord progression estimation detection apparatus according to an example of an embodiment of the present invention will be described with reference to the drawings.

  The hardware of the chord progression estimation detection device can be realized by a general-purpose information processing device such as a personal computer. FIG. 1 is a block diagram showing the hardware configuration of a chord progression estimation / detection device constructed on a computer. A display 11, a mouse 12, a keyboard 13, a ROM 14, a RAM 15, a CPU 16, an HDD 17, a disk The drive 18, MIDI interface 19, audio interface 20, and network interface 21 are connected.

  The computer's HDD 17 is obtained by capturing a sound signal via the MIDI interface 19, the audio interface 20 or the network interface 21 to detect a chord, and estimating the chord function based on a plurality of pre-stored cadences. A chord progression estimation detection program for estimating chord progression is installed from a recording medium attached to the disk drive 18 or downloaded from a predetermined URL via the Internet.

The CPU 16 executes various processes (each step) according to a predetermined control program (chord estimation detection program) installed or downloaded by the above-described procedure, and controls the entire chord estimation apparatus. The performance information storage unit 31, The chord detection unit 32, chord function estimation unit 34, chord allocation unit 37, and chord performance information file creation unit 38 are provided as main functions, so that the captured performance information is stored, and the chord progression is estimated from the stored information. A chord sound corresponding to a chord estimated using the software sound source 42 installed or downloaded in the RAM 15 is provided by being able to display on the display (display unit) 11 and including a reproduction device (sound reproduction means) 46. Performance (chord performance) is possible.
The code estimation detection program and the software sound source are read into the RAM 15 at the time of execution.

FIG. 2 is a block diagram showing a main configuration of a chord progression estimation detection apparatus constructed on a computer by incorporating a chord progression estimation detection program. The chord progression estimation detecting device includes chord progression estimating means 30 for detecting chords from performance information and performance means (automatic performance apparatus) 40 for performing chord performances using the detected chords.
The performance information is information including note data (notes), and represents one note data by a sounding event (sounding start time and strength) and a sounding stop event (sounding stop time). The performance information is obtained by creating from an audio signal, inputting from a MIDI interface, or reading from a performance information file such as a standard MIDI file on a computer or network.

  The chord progression estimation means 30 includes a performance information storage unit 31 in which performance information is stored, a chord detection unit 32 that detects chords with respect to the performance information, and a key detection unit 33 that detects the key of a song from the entire performance information. A chord function estimating section 34 for estimating the chord function of each chord for each specific section, a chord function determining section 35 for deciding chord function progress, a chord assigning section 37 for assigning chords to chord functions, and a detected chord A chord performance information file creation section 38 for storing the above information, and a display section 39 for displaying a detection result of chords detected from the performance information on a liquid crystal display or the like.

  The chord detector 32 detects a plurality of chords and the likelihood of each chord for each specific section from the performance information. The likelihood of a chord is an index indicating the likelihood of the chord, and is a known value calculated from “sound intensity”, “length of sound duration”, and the like. As a general method for detecting a chord from performance information, a chord is estimated by comparing a sound that is generated in a detection section with a chord constituent sound. In performance information, the likelihood of a chord containing a note that is strong and has a long duration is high.

The chord function estimated by the chord function estimation unit 34 is a cadence that expresses the character of the chord, and includes “tonic”, “dominant”, “subdominant”, and the like.
The code function classifies the detected code name into each code function according to the key detected by the key detection unit 33. The classification rule is determined by the key and the degree No.

“Tonic” in the chord function means a tonic chord that is a chord that is a main chord with respect to a key (key). It is the most stable chord and is usually the chord used at the beginning and end of a song. For example, if the key is “C”, the tonic is “C” or “CM7”. This is the most important chord that determines the key of the song.
“Dominant” of the chord function is a sound that is five times higher than the main sound, and is also called a genus. A chord with this sound as the root is called a dominant chord and is abbreviated as a dominant chord. For example, if the key is “C”, the dominant is “G”. The power to proceed to tonic works, and usually progresses from dominant to tonic at the end of a phrase or song.
The “subdominant” chord function is a sound that is four times higher than the main sound, and is also called a subgenus sound. A chord with this sound as a root is called a subdominant chord, and is abbreviated as subdominant. For example, if the key is “C”, the sub-dominant is “F”. In response to the progress from tonic, it has the effect of changing the atmosphere of the song and creating a feeling of inflection.

A plurality of basic cadences (reference cadences) are stored in the reference cadence storage unit 36 of the chord function determination unit 35, and the progress of the chord function obtained by the chord function estimation unit 34 based on the reference cadence is finally determined. Determine the progression of chord function. As the reference cadence, for example, the following three types of chord function progress are stored.
(A) "Tonic""Subdominant""Tonic" (TST)
(B) “Tonic” “Dominant” “Tonic” (TDT)
(C) "Tonic""Subdominant""Dominant""Tonic" (TSDT)

  The chord function determination unit 35 compares the chord function progress with the reference cadence, and if they match, the chord function of that portion is confirmed as the same chord function as the reference cadence. Further, the reference cadence and the chord function progress are compared, and if they do not match, the chord function of the chord function progress is changed so that the chord function of the chord function progress matches the reference cadence.

The display unit 39 includes a chord progression display unit 39a that can display a main chord progression finally determined by the chord function determination unit 35 and a plurality of chord progression candidates different from the main chord progression. .
The chord assignment unit 37 assigns a chord corresponding to the chord function determined by the chord function determination unit 35 (chord progression chord) to the main chord progression.

  In correspondence with the hardware qualities of FIG. 1, the chord progression estimation means 30 includes a display 11, a mouse 12, a keyboard 13, a ROM 14 for reading data relating to chords, a RAM 15 serving as a working memory, a CPU 16 for executing a program, a chord progression estimation detection It is composed of an HDD 17 or the like in which a program is stored.

As shown in FIG. 2, the chord progression estimation detection apparatus records performance information in the performance information storage unit 31 and detects chords from the data in the performance information storage unit 31 by the chord detection unit 32 to create a chord performance information file. The chord performance information file is recorded in the section 38.
In the performance means 40, the music data corresponding to the chord created by the software sound source 42 is reproduced by the reproduction device (sound reproduction means) 46 via the reproduction buffer 45 based on the data created by the chord performance information file creation section 38. Is done.
The code detection result is displayed on the display unit 39.

  The performance means 40 includes a display 11, a mouse 12, a keyboard 13, a CPU 16 for executing a program, a RAM 15 as a working memory, and an audio interface 20 in correspondence with the hardware qualities shown in FIG.

FIG. 3 is a block diagram showing the main configuration of another embodiment of the chord progression estimation detecting apparatus constructed on the computer by incorporating the chord progression estimation detecting program. In this example, instead of the performance information in FIG. 2, a chord is detected from an acoustic signal in which music is expressed by electronic information.
The chord progression estimation detecting device is constituted by chord progression estimating means 30 for detecting a chord from an acoustic signal, and performance means (automatic performance apparatus) 40 for performing a data performance by the acoustic signal and performing a chord performance by the detected chord. Yes.
Parts having the same configuration as the chord progression estimation and detection apparatus of FIG. 2 are assigned the same reference numerals and description thereof is omitted.

  The chord progression estimating means 30 detects a tone of a tune from the entire performance based on the sound signal, a waveform storage unit 81 in which an acoustic signal corresponding to the performance music is stored, a chord detection unit 32 that detects a chord with respect to the acoustic signal A key detection unit 33, a code function estimation unit 34 that estimates the chord function of each chord for each specific section, a chord function determination unit 35 that determines chord function progress, a code assignment unit 37 that assigns codes to chord functions, The chord performance information file creation section 38 stores information on the detected chords, and the display section 39 displays the detection results of the chords detected from the acoustic signals on a liquid crystal display or the like.

In the chord progression estimation detecting device, as shown in FIG. 3, the sound signal is recorded in the waveform storage unit 81, the chord detection unit 32 detects the chord from the data in the waveform storage unit 81, and the chord performance information file creation unit 38 Is recorded as a chord performance information file.
In the performance means 40, the waveform corresponding to the chord created by the software sound source 42 for the data input from the waveform storage unit 81 in which the acoustic signal is recorded in the adder 44 through the data waveform buffer 41. As the data is added via the chord waveform buffer 43, the superimposed music data is reproduced by the reproduction device (sound reproduction means) 46 via the reproduction buffer 45. Playback by the playback device (sound playback means) 46 is performed in stereo or monaural, but in the case of stereo, the right channel side of the sound signal and chord performance with the right channel speaker and the sound signal and chord performance with the left channel speaker are performed. The left channel side is played back. The code detection result is displayed on the display unit 39.

The operations of the chord progression estimation means 30 and the performance means 40 in FIGS. 2 and 3 will be described below.
The chord progression estimation means 30 analyzes the audio information captured from an audio track or audio interface such as a music file, CD, etc., and determines the tempo, time signature, first beat position (start position of each measure), and chord of the song. Detect using technology. The code detection is performed by the technique described in Japanese Patent No. 4767691, for example, and will be briefly described below.

In the chord detection unit 32, a chord detection period is set for the performance information stored in the performance information storage unit 31 (or the acoustic signal stored in the waveform storage unit 81), and each of the chord detection sound ranges, for example, C3 to A6, is set. Calculate the average level of the frequency data during the chord detection period of the scale, detect several pitch names in order from the scale that has the largest value, and based on this and the pitch of the bass sound being detected Extract code candidates.
At this time, since a sound with a high level is not necessarily a chord constituent sound, for example, five sounds having a plurality of pitch names are detected, and two or more of them are extracted in all combinations, and this is used as a base. Chord candidates are extracted from the note names of the notes. After the code candidates are extracted, one code is selected and extracted based on a preset criterion.

When the code is extracted, the display unit 39 displays a detection result screen (track screen) as shown in FIG.
In the master track of the track screen, tempo = 120, time signature = 4/4, and key = G (G major) detected from performance information or sound signals are designated. In the material track 1, performance information 1 from which chords have been extracted is designated. The performance information is displayed in a format compliant with the MIDI standard, which is a well-known standard.
Nothing is specified for the chord track, rhythm track, and material track 2. A dotted line indicates a bar line.

A procedure for performing chord progression estimation from this screen state will be described with reference to FIG.
Right-click on the performance information displayed on the track screen to display a pop-up menu, and select the “Estimate chord progression from performance information” command (not shown) from the menu. Be started.
In the chord progression estimation by the chord function estimation unit 34, the key is detected from all the note information included in the performance information (step 51). The key is used in subsequent processing. When the key is not detected, the key of the master track may be used or specified at the start of processing.
The key detection by the key detection unit 33 is performed by a general method such as estimation from a note having the highest appearance frequency.

  Subsequently, a chord is detected for each measure (step 52). Bar breaks follow the master track. When no time signature is specified for the master track, the time signature may be specified at the start of processing or estimated from performance information. A musical note included in a specific section (within a bar) is compared with the table showing the constituent sounds of each chord from the performance information, and a plurality of possible chords are detected for each bar.

The detection result is stored as an array of detection code data as shown in FIG.
One detection code data includes a code name array, a likelihood array corresponding to the code name, an index of the selected code, a bar number, and a code start time in the bar.
In the case of code detection of measure number 1, “G”, “GM7”, “G7”, “Em”, “Em7”, “Bm”, “Bm7”, “CM7” are detected as codes, and the corresponding likelihoods are “6726”, “3758”. “3432” “2904” “2323” “1848” “1478” “2745”.

Next, the codes of the detected code data array are classified into “tonic”, “dominant”, and “subdominant” chord functions to generate chord function progressions (step 53). The chord function progression is a chord function assigned to each chord detection range.
According to the key detected in step 51, the detected code is classified into each code function. The classification rule is determined by the key and the degree No. In G major, it is classified as follows.

Tonic: G, GM7, Em, Em7, Bm, Bm7
Subdominant: Am, Am7, C, CM7
Dominant: Bm, Bm7, D, D7, F # m (b5), F # m7 (b5)

The detected codes are classified by code function, and a code function in which many code candidates are classified is set as a code function of the code detection section.
For example, in a plurality of chords detected as chord progression candidates for each measure, when the maximum number of chords belonging to “Tonic” is selected, “Tonic” is selected as the chord function for that measure.

Next, processing for determining cadence is performed for each measure of the detection code data array shown in FIG. 6 (step 54).
FIG. 7 shows the chord function classification result and chord function progress corresponding to the detected code data array of FIG. That is, the code candidates in measure number 1 are G, GM7, G7, Em, Em7, Bm, and Bm7 as chords belonging to “Tonic”, and Bm and Bm7 as chords belonging to “Dominant”, CM7 is extracted as a code belonging to “subdominant”. Therefore, for the chord function corresponding to measure number 1, “tonic” having the largest number of chord candidates is selected, and “sub-dominant” is selected in measure number 2 and “sub-dominant” in measure number 3 by the same processing. Is selected, bar number 4 selects "Tonic", bar number 5 selects "Dominant", bar number 6 selects "Tonic", and chord functions corresponding to bar numbers 1-6 Progress is TSSTDT.

Subsequently, for each measure of the chord function progress, processing for determining the cadence from the above-described reference cadences (A) to (C) is performed. This procedure is shown in FIG.
The code function determination process is performed in order from the reference cadence with the highest degree of coincidence.
First, a section corresponding to the length of the reference cadence from the comparison start measure (three measures for the reference cadence (A) and (B) and four measures for the reference cadence (C)) is compared with the reference cadence (step 61). ), It is confirmed whether or not they completely match (step 62). If they match completely, the code function is fixed as it is.

If they do not match the reference cadence, all (three in this case) reference cadences are compared with the chord function progression from the comparison start measure, and the degree of cadence match is calculated. The degree of cadence matching is calculated by the number of matched code functions divided by the total number of codes.
The first comparison start bar is the first bar. Comparing the TST of the reference cadence (A) with the chord function progression, the first tonic matches, the second subdominant matches, and the third tonic does not match. As a result, the degree of coincidence is 0.67 = 2 ÷ 3.
In the comparison with the TDT of the reference cadence (B), only the first tonic matches, so the matching degree is 0.33 = 1 ÷ 3.
In the case of comparison with the TSDT of the reference cadence (C), except for the third dominant, they are in agreement, so the degree of coincidence is 0.75 = 3 ÷ 4.

  In the case of this example, the reference cadence having a high degree of coincidence is T-S-D-T of the reference cadence (C), and subsequently, it is confirmed whether or not the chord function of the part where the chord function progression does not coincide can be changed (step 63). In the case of this example, since the third (dominant) of the reference cadence (C) and the third (subdominant) chord function progression of FIG. 7 are different, the process of whether or not the change is possible is performed.

Specifically, it is checked whether there is a code candidate classified into the same code function as the reference cadence by referring to the classification results of the code functions of different parts. If there is a code candidate, it is considered that the code function can be changed, and a code function change process is performed (step 64).
In this example, since there is a code candidate classified as “dominant” in the third code detection section of FIG. 7, it is determined that the code can be changed, and the code function is changed (step 64). Regarding the chord function progression of measure number 3, the third of the standard cadence (C) is changed from “dominant” to “subdominant”, and the chord function progression of the section up to four measures is determined as the chord function of TSST.

If the change cannot be made, the process proceeds to a process for determining whether or not all cadences are completed (step 65). In step 65, it is checked whether there is a reference cadence that has not yet been processed. If there is a reference cadence that has not been processed, the unprocessed reference cadence is determined from step 62.
If there is no unprocessed reference cadence, the process is completed.

When the processing for all the reference cadences in step 65 is completed, it is confirmed whether or not the processing for all the sections of the chord function progress has been completed (step 66), and if completed, the processing is completed.
If not completed, the comparison start measure is updated, and the process returns to the process of comparison with the reference cadence (step 61). In this example, the comparison start measure is updated to 4, and the process returns to the reference cadence and comparison (step 61).
Similarly, the chord function confirmation process is performed in the sections 4 to 6 to complete the chord function confirmation process (step 54).

  In this example, only the most basic three cadences described above (TST, TDT, and TSDT) are used as reference cadences, but other commonly used cadences are included. Those using “secondary dominant” may be added as a reference cadence.

Subsequently, in the flowchart of FIG. 5, a code is assigned to the finally determined code function (step 55). Here, referring to the classification result of the code function, a code corresponding to the determined code function is assigned.
Basically, a code with a high likelihood is assigned.
When choosing one chord progression, it is sufficient to assign the one with the highest likelihood, but when proposing multiple chord progressions, the chords assigned once are stored in the assigned chord array, etc., and different chords are selected. You may be made to do.
Also, assignment may be performed in consideration of the connection with the preceding and following codes.

The flow chart of FIG. 9 shows the procedure for code allocation at step 55 in the flowchart of FIG.
A code corresponding to the code function is selected as a candidate for the section (step 71).
The allocated frequency is confirmed by referring to the allocated array (step 72).
If allocation is not performed or if the allocation frequency is lower than other codes, a code allocation process for allocating a candidate code as a code of the section is performed (step 76), and the process of storing in an allocated array (step) 77) is performed.
If the allocation frequency is equal to or higher than other codes, a code allocation process (step 73) at a different root is performed. That is, in step 73, a code with a lower allocation frequency and a different root is selected as a candidate (step 74).

If there is a code corresponding to a different candidate in step 74, the process proceeds to a code assignment process (step 76).
If there is no code corresponding to a different candidate in step 74, the code assigned this time is fixed to the code selected in step 71 (step 75).
In step 74, a code with a lower allocation frequency in the same route is set as a candidate. If there is no corresponding code, a process of storing in the allocated array (step 77) is performed.

In the code assignment process (step 76), the candidate code is assigned as the code of the section.
In the process of storing in the allocated array (step 77), the allocated code is stored in the allocated array.

Next, it is determined whether the candidate selection process has been completed for all bars (step 78). If not, the next section candidate selection process is returned to step 71 and performed. If all measures have been processed, it is determined whether all chord progressions to be presented have been completed (step 79). If not, the process returns to step 71 to perform the next chord progression.
If all the chord progression processes to be presented have been completed, the process of FIG. 9 is terminated.

  In this example, candidates for the first section (first bar) are determined in step 71. G having the highest likelihood in “Tonic” is a candidate. In step 72, since G is not stored in the assignment array, the process moves to step 76 and is assigned as the code of the first measure. In step 77, it is stored in the allocated array. Since all the sections are not completed in step 78, the process returns to step 71 and proceeds to the processing of the second measure.

  In the second measure, the code C having the highest likelihood of “subdominant” is a candidate. In step 72, since C is not stored in the assignment array, the process moves to step 76 and is assigned as the code of the second measure. In step 77, it is stored in the allocated array. Furthermore, it returns to step 71 and moves to the process of the 3rd bar.

  The third measure is processed in the same manner.

  In the fourth measure, G having the highest likelihood in “Tonic” is a candidate, but since it has already been registered in step 72, it is changed to a candidate having a highest likelihood in a different Root and a code with a low allocation frequency. . In this case, Em7 is a candidate. In step 75, Em7 is assigned as the code of the fourth measure, and in step 77, Em7 is stored in the assigned array.

  The fifth measure is “dominant”, Bm7 having the highest likelihood and the lowest allocation frequency is a candidate, and Bm7 is registered as the code of the fifth measure and stored in the assigned sequence.

  In the sixth measure, G having the highest likelihood in the tonic becomes a candidate, but since it has already been assigned, a different root candidate is searched. Since Em7 of different Root is also assigned Bm7 and the allocation frequency is the same as G, the code of the same Root (G) that is not allocated is used as a candidate. In this case, G7 with the next highest likelihood is a candidate. In step 76, it is assigned as a code of the sixth measure and registered in the assigned array.

This completes the assignment of the chord progression candidate 1 chord.
Further, when another chord progression candidate is obtained, the above procedure is repeated to generate chord progression candidates corresponding to each measure.

A chord progression screen obtained by the above procedure is shown in FIG.
The chord progression screen 90 is displayed on the chord progression display section 39a of the display section 39. In this example, four chord progression candidates are generated as other chord progression candidates.
The main chord progression is initially the same as chord progression candidate 1.
When the audition button 91 on the chord progression screen 90 is clicked, the original performance information and the main chord progression displayed on the main are played, and it can be confirmed whether or not the generated chord progression is intended.

When changing the main chord, clicking each chord of the chord candidates 1 to 4 in the same bar becomes the main chord.
When the candidates 1 to 4 are clicked, the chord progression becomes the main chord progression.
When the completion button 92 is clicked, as shown in FIG. 11, the main chord progression is pasted on the chord track on the track screen displayed on the display unit 39 of the automatic performance device corresponding to the performance information.

  According to the chord progression estimation and detection apparatus described above, the function of each chord detected from the performance information (or sound signal) is estimated, and a chord according to a pre-stored reference cadence is selected. You can get chord progression.

  Also, as shown in the chord progression screen 90 of FIG. 10, by creating a plurality of chord candidates (candidates 1 to 4) for the same cadence (function) of each measure, it is possible to select different chords. Multiple chord progressions can be proposed.

  DESCRIPTION OF SYMBOLS 30 ... Chord progress estimation means, 31 ... Performance information storage part, 32 ... Chord detection part, 33 ... Key detection part, 34 ... Chord function estimation part, 35 ... Chord function determination part, 36 ... Reference cadence storage part, 37 ... Chord Allocation unit, 38 ... Chord performance information file creation unit, 39 ... Display unit, 39a ... Chord progress display unit, 40 ... Performance means (automatic performance device), 41 ... Data waveform buffer, 42 ... Software sound source, 43 ... Chord waveform Buffer 44... Adder 45. Reproduction buffer 46. Reproduction device (sound reproduction means) 81 waveform storage unit

Claims (6)

A chord progression estimation detecting device for estimating chords from performance information or sound signals,
Key detection means for detecting the key of the song from the performance information or the entire acoustic signal;
Chord detection means for detecting a plurality of chords and likelihood of each chord for each specific section from the performance information or the sound signal;
Code function estimating means for estimating the code function of each code classified according to the key detected by the key detecting means for each specific section;
About the chord function progress obtained by the chord function estimating means, chord function determining means for determining the final chord function progress based on a plurality of basic cadences;
Code assignment means for assigning codes to the code functions of each cadence;
A chord progression estimation detection apparatus comprising:   The chord function determining means includes a reference cadence storing means for storing a plurality of reference cadences. When the chord function progress and the reference cadence are compared, if they coincide, the code function of the portion is determined as the same chord function as the reference cadence. The chord progression estimation detection apparatus according to claim 1.   The chord function determining means compares the reference cadence with the chord function progress, and if not, changes the chord function of the chord function progress so that the chord function of the chord function progress matches the reference cadence. Item 2. The chord progression estimation detection device according to Item 1.   A chord progression display unit that can display a main chord progression finally determined by the chord function determination unit and a plurality of chord progression candidates different from the main chord progression, and is displayed in the chord progression candidate The chord progression estimation detection apparatus according to claim 1, wherein chords can be assigned to main chord progressions. Playing performance information or acoustic signals in a specified specific range, with performance means for playing chords with a built-in sound source,
5. The chord progression estimation and detection device according to claim 4, wherein the chord progression estimation and detection device according to claim 4, wherein the chord progression display unit and the chord progression display chord are displayed simultaneously. A program for estimating chord progression from performance information or sound signals,
A chord detection procedure for detecting the likelihood of each chord and a plurality of chords for each specific section from the performance information or the acoustic signal;
A code function estimation procedure for estimating the code function of each code;
About the estimated chord function progress, a chord function determination procedure for determining the final chord function progress based on a plurality of basic cadences,
A code assignment procedure to assign codes to the code functions of each cadence;
A program for detecting and estimating chord progression characterized by causing a computer to execute.