JP2638817B2 - Accompaniment line fundamental tone determination device - Google Patents

Description

Description: TECHNICAL FIELD [0001] The present invention relates to an accompaniment line generator that automatically generates a line different from a melody line, that is, an accompaniment line, and in particular, determines a basic sound (fundamental tone) of an accompaniment line. Related to technology.

[Background of the Invention] A technique for automatically generating an accompaniment line such as a bass and an obligato is already known. This type of technology has evolved in the field of electronic musical instruments. It is incorporated as an automatic accompaniment function for electronic musical instruments. In a conventional automatic accompaniment line generation device, a root tone is used as a fundamental sound (fundamental tone) of an accompaniment line. In a real-time device, the source of the root note is given from a performance input device such as a keyboard. For example, a certain area of the keyboard is assigned as a key area for inputting a chord, and the type of chord and its root note are determined based on construction information input from this key area. The determined root tone is used as a fundamental tone of the accompaniment line, and an accompaniment pattern is output based on the fundamental tone. The automatic accompaniment line generation technique is also applied to non-real time types. In this non-real-time device, code progress information is prepared in a memory. At the time of operation, chord information is read for each chord section, and the root tone included in the information becomes the fundamental tone of the accompaniment pattern.

The root approach described above has the following problems. In other words, the basic sound of the accompaniment line is always the root tone, so that the line has very little change.

In view of this point, the present applicant has filed a patent application for an invention for automatically generating an accompaniment line by another approach (Japanese Patent Application No. 6-207174, Japanese Patent Application No. 60-207175, filing date, 1985). On September 19, the name "Electronic Musical Instrument with Automatic Accompaniment Function"). The inventions of these applications use adjacent sound progression instead of root sound progression. That is, the automatic accompaniment line generation function includes an adjacent sound logic in which the sound closest to the immediately preceding accompaniment sound among the current chord constituent sounds is used as the current accompaniment sound. Further, Japanese Patent Application No. 60-207175 discloses an improved logic for forming an accompaniment line with adjacent sounds that run counter to the bass line in order to create an accompaniment line other than the bass line. The adjacent sound logic is a logic that guarantees that a sound other than the root sound can also be a fundamental sound, and improves the conventional monotony. Also, the improved logic is a counterpoint approach and helps generate moving lines.

However, there are still many things to be improved when compared to the accompaniment of a good performer. One of the characteristics of accompaniment by human composers and musicians is "musical situation"
The conventional automatic accompaniment line generation techniques only take this point inadequately.

[Object of the Invention] Accordingly, an object of the present invention is to provide a non-real time type accompaniment line fundamental tone determination device that simulates the characteristics of the accompaniment line determination process by a performer or a composer to a level that could not be achieved conventionally. That is.

[Summary of the Invention] According to the present invention, an accompaniment line base tone determination device that determines a sound (referred to as a board) that is a basis of an accompaniment line on a non-real time basis for each of a plurality of music sections constituting an accompaniment line , Selectively operate depending on a music situation in a first section of interest out of the plurality of music sections, estimating a base tone for the first section based on a past base tone, and And selectively operates depending on the music situation in the second section of interest out of the plurality of music sections to output a first output signal representing the first output signal representing the second output signal. Past independent fundamental estimating means for estimating a fundamental tone for a section independently of a past fundamental tone and outputting a second output signal representing the estimated fundamental tone; the past dependent fundamental estimating means; and the past independent fundamental tone estimating means Works with The base tone for each section of the accompaniment line is determined from the first and second output signals, and the fundamental tone determined for a section of the accompaniment line is influenced by the fundamental tone determined before that section. Means for receiving a fundamental tone determined for another section of the accompaniment line so as not to be affected by any fundamental tone determined before the another section. Is provided (first invention).

In one aspect, the past-dependent fundamental tone estimating means converts a chord constituent sound having a predetermined relationship to a fundamental tone of the immediately preceding section among chord constituent sounds assigned to the first section, into a base note of the first section. Immediately before depending means.

In one aspect, the past independent fundamental estimating means estimates a specific constituent sound of the chord as a fundamental tone of the second section when the chord assigned to the second section is a specific chord type. Includes type-dependent means.

In one aspect, the past independent base tone estimating means collectively estimates a base tone for a plurality of continuous sections when a chord progression covering a plurality of continuous sections including a section of interest has a specific pattern. Including batch estimation means.

Further, according to the present invention, for each of a plurality of music sections constituting an accompaniment line, an accompaniment line basic tone determining apparatus for determining a sound (hereinafter referred to as a fundamental tone) which is a basis of the accompaniment line on a non-real time basis. In, estimating the fundamental tone for a single section according to the music situation of the single section,
A single-segment fundamental sound estimating means for generating a first output signal representing the fundamental sound; and estimating fundamental sounds for a plurality of continuous segments in a lump when the music conditions of the continuous segments are predetermined music conditions. Operatively coupled to the collective fundamental sound estimating means for generating a second output signal representing these fundamental sounds, and to the single section fundamental sound estimating means and the collective fundamental sound estimating means.
And a means for determining a fundamental tone for each of the plurality of sections from the second output signal. An accompaniment line fundamental tone determining apparatus is provided (second invention).

In one embodiment, the collective fundamental tone estimating means includes a conditional collective fundamental tone estimating means that operates conditionally, and the conditional collective fundamental tone estimating means (a) refers to a condition parameter regarding the plurality of continuous sections, b) Only when this condition parameter has a specific value, the fundamental tone for the plurality of continuous sections is collectively estimated.

In one embodiment, the conditional collective fundamental tone estimating means includes:
(A) operating when the condition parameter is a specific value and the chord progression of the plurality of continuous sections includes a specific chord constituent sound pattern, and It is estimated as a series of fundamental tones for the plurality of continuous sections.

[Operation and Development of the Invention] According to the first aspect, in a certain section, a fundamental tone affected by a fundamental tone in the past than that section is estimated as a fundamental tone in the section, and in a certain section, a fundamental tone in the past is affected. Instead, the fundamental tone of that section is estimated. That is, depending on the musical situation given to each section, in some cases, a sound that depends on a past fundamental tone is a fundamental tone, and in another case, a fundamental that is independent of the past fundamental tone is a fundamental tone. This situational approach well expresses one of the characteristics of accompaniment line determination by musicians and the like.

In one configuration, chord progression is provided as a known musical situation. The past-dependent fundamental tone estimating means immediately before estimating a tone having a predetermined relationship (for example, adjacent relationship) with the fundamental tone in the immediately preceding section from among the constituent sounds of the chord given to the section of interest. Includes fundamental tone dependent means. Further, as a component of the past independent fundamental tone estimating means, when the code of the section of interest is a specific code type (for example, a dominant code), one of the constituent sounds (for example, a dominant that is a root note) May be used as a fundamental tone of the section. The immediately preceding fundamental tone dependent means and the chord type dependent type means belong to a single interval type for estimating a fundamental tone in one interval.

In another configuration example, the past independent fundamental tone estimating means focuses not only on the current section but also at least one section following the current section as a section of interest, and the chord progression in the plurality of continuous sections is of a specific type. Means for estimating fundamental sounds in the plurality of continuous sections collectively. In the embodiment described later, the code signal of two consecutive sections is V
Means for estimating the fundamental sound progression of the lead sound and the tonic at the time of -I progression (tenth calculating means described later) is shown. As another example, when the chord progression of three or more consecutive sections includes a sequence of chord tones of a progression progression including a progression sound as an intermediate sound, a transitional sound means for estimating the progression of the progression sound as a fundamental progression is provided. May be included. However, in the illustrated embodiment described later, the elapsed sound means is of a type that refers to past conditions,
It is checked whether or not the past condition of the section of interest depends on the section in the past. When the condition does not depend on the section of interest, a sequence of fundamental sounds including a passing sound is estimated.

In general, a feature that is often found with respect to the sequence of fundamentals in a line is that the fundamental in a section is highly correlated with the fundamental in an earlier section (eg, the immediately preceding section or the corresponding section of a similar passage). They have less and sometimes do not. According to the first invention, the degree and presence or absence of the correlation between the sections can be controlled by the cooperation between the past-dependent fundamental tone means and the past independent fundamental tone means.

In addition, by means of the immediately preceding fundamental-dependent means,
When the fundamental tone of the focused section is determined, that section is a section that is strongly affected by the immediately preceding section with respect to the accompaniment line. However, in a configuration in which all sections subsequent to the first section always have a fundamental tone according to the immediately preceding fundamental tone, insufficient results are often obtained. The opposite approach, that is, the logic for giving all sections a fundamental tone that is independent of the past fundamental tone, is not enough, and it is easy to create a monotonous and bland tone progression. The contextual approach of the first invention, including past-independent logic and past-dependent logic, has the potential to fundamentally solve these problems.

Furthermore, the second invention includes a single section type fundamental tone estimating unit and a collective type fundamental tone estimating unit. As the single section type fundamental tone estimating means, a past dependent type or past independent type means can be used. The collective fundamental tone estimating means collectively estimates the fundamental tone of not only the current section but also a continuous section including one or more sections subsequent to the current section. One or more sections following the current section can be considered as future sections. Therefore, it can be said that the collective fundamental tone estimating means considers the future. This function focuses on special musical conditions (for example, chord progression) for several continuous sections when a player or the like considers how to attach an accompaniment, and a base tone suitable for the special musical conditions. It simulates the process of recalling patterns. When a player or the like views several consecutive sections one after another, the musical situation typified by the consecutive sections (for example,
When there is a familiar musical situation), it is not rare to determine the arrangement of fundamental tones in a plurality of sections at once instead of determining the fundamental tones in each section. The collective fundamental tone estimating means realizes this typological approach. As a configuration example,
The collective fundamental tone estimating means may be a past independent type that does not consider past fundamental sounds at all, but desirably considers past conditions (past dependent parameters) regarding a continuous section of interest and focuses on the result of the consideration. It is configured with a past dependency parameter reference type for selectively estimating a fundamental tone pattern for a continuous section. For example, even when it is better to make a musical break in the middle of the continuous section of interest, if a fundamental tone pattern is issued, the musical break becomes inconspicuous, and a desired result cannot be obtained. Therefore, in such a case, the fundamental tone estimating means, which is both the past dependence parameter reference type and the collective estimation type, cancels the issuance of the fundamental tone pattern. The past dependency parameter may be information indicating the degree to which a break is better. In this case, the fundamental tone estimating means can selectively estimate the fundamental tone pattern at a rate of 80% according to the degree of the degree, for example, when 80% of the break orientation is indicated. This can be realized by using an electronic random number generator controlled according to the degree of past dependence as an element for controlling the operation of the fundamental tone pattern generation unit of the fundamental tone estimating means. In a simpler example, the past dependency parameter is represented by binary information indicating whether or not the section of interest depends on the past. In this case, the collective fundamental tone estimating means does not issue the fundamental tone pattern unless the information indicates that the information does not depend on the past.

In still another configuration example, several fundamental tone patterns that can be used for a certain chord progression are prepared. In operation,
One of these fundamental tone pattern candidates is selected.
Which fundamental tone pattern is selected is controlled by musical conditions such as the style of music.

The operation and effect of the second invention can be explained as follows. Now, it is assumed that there is only a single section type means as a fundamental tone estimating means. The generation of the above-mentioned fundamental tone pattern cannot be guaranteed only with such a single section type fundamental tone estimating means. Even if a fundamental pattern is created, it is merely a coincidence. In other words, the collective fundamental tone estimating means can generate a fundamental tone pattern that cannot be created by the single section fundamental tone estimating means. However, irrespective of the musical situation, if we always generate a fixed base tone pattern, we can only do something unsatisfactory.
In the present invention, these problems are solved by combining a single section type fundamental tone estimating means and a collective type fundamental tone estimating means, and estimating a single fundamental tone or a sequence of fundamental sounds for each musical situation.

[Example] Hereinafter, an example of the present invention will be described. The present embodiment is an apparatus for determining (generating) a base tone of a bass line as an accompaniment line on a non-real time basis. In the case of the present embodiment, a melody line, chord progression information, and information on the structure (musical expression) of a music piece are given in advance as musical information.

1. Overall Configuration FIG. 1 shows the overall configuration of the non-real-time baseline fundamental tone determination device according to the present embodiment. In the figure, 1 is a CPU,
2 is a chord progression storage device, 3 is a chord composition sound storage device,
4 is a phrase section storage device, 5 is a melody data storage device,
Numeral 6 denotes an input device, 7 to 17 denote arithmetic means constituting various fundamental tone estimating means, 18 denotes a CRT, 19 denotes a tone generator, and 20 denotes a printer.

The melody data storage device 5 stores melody data constituting a melody line of a song input from the input device 6.

The phrase section storage device 4 stores information on the musical style of the music. An example of a data representation of this information is shown in FIG. In this example, the bar number at which the phrase starts is written in the upper digit A1, and the type of the phrase is written in the lower digit A2.

The chord progression storage device 2 stores information related to the chord progression of a song, and the chord constituent sound storage device 3 stores information related to the sounds constituting each chord. In order to express one sound (note), pitch data and pitch data are required. The pitch data expresses the pitch. FIG. 2 shows an example of pitch data assigned to each pitch on the scale. In FIG. 3B, pitch data is set to an integer value that is incremented by one for each semitone. In FIG. 9A, an integer value is assigned to a white key position (pitch on a diatonic scale composed of a dormimi facsimile), and a random number is assigned to a black key position. In addition to this, any pitch can be assigned, and a unique correspondence is sufficient in principle. However, in the flow example described later, the pitch is assigned as shown in FIG. 2 (b). In this case, one octave corresponds to "12". FIG. 3 exemplifies data expressions relating to chord progression and chord constituent sounds. These pieces of information are stored in the chord progression storage device 2 and the chord component sound storage device 3.

The first calculating means 7 to the thirteenth calculating means 17 constitute fundamental tone estimating means having different estimation logics. The CPU 1 controls the estimation between the plurality of arithmetic means 7 to 17 by giving a priority. Basically, each fundamental tone estimating means is composed of a condition inspection unit for examining whether a musical condition is satisfied and an estimated fundamental tone generating unit (action unit) for generating an estimated fundamental tone.

Briefly, the functions of the respective calculation means 7 to 17 are as follows.

First calculation means 7: presumes that the base fundamental at the beginning of the music is the root.

Fourth calculation means 8: Estimates the chord constituent sound closest to the base fundamental in the immediately preceding section, that is, the adjacent sound, as the base fundamental in the current section.

Fifth arithmetic means 9: When a presumed current fundamental tone (here, the output of the fourth arithmetic means 8) is used and a predetermined relationship is established between the assumed current fundamental tone, the present root tone, and the immediately preceding fundamental tone (in short, the When the root is closer to the immediately preceding fundamental to some extent), the current root is estimated as the current fundamental.

Sixth calculation means 10: Estimate the fundamental tone as the root note at the start measure of a phrase repeating the motif.

Seventh arithmetic means 11: When the chord progression of three consecutive sections includes a bass progression (progression of chord constituent sounds) which is not close to the start measure of the passage, and the chord progression of the three sections, the base fundamental note of these three sections is Issue all at once. A transitional sound is created between the root sounds.

Eighth operation means 12: There is no special condition checker in this means.
Extract the dominant sound of the melody line.

Ninth calculating means 13: When the dominant sound of the melody line is not dominant when the current section is the dominant code (V), the dominant is estimated as the current fundamental sound. In this case, the dominant is the root note.

Tenth calculating means 14: In the case of VI progression, it estimates the lead sound (S) and the main sound (D) as the fundamental tones in the V section and the I section, respectively.

Eleventh arithmetic means 15: In cases other than VI progression, when the assumed current fundamental is a lead, a sound closest to the immediately preceding fundamental among the chord constituents other than the lead is estimated as the fundamental.

The twelfth calculating means 16: when the assumed current fundamental tone is the same as the ruler of the melody line (has the same note name), among the chord constituent sounds other than that tone, estimates the tone closest to the immediately preceding substrate as the fundamental tone.

Thirteenth arithmetic means 17: The last fundamental tone of the music is a root note.

Overall Flow FIG. 7 exemplifies the overall flow of the base fundamental generation of the non-real time bass fundamental determination device according to the present embodiment. In this example of the flow, each arithmetic unit activates the estimated fundamental tone generation unit when the internal condition inspection unit indicates that the condition is satisfied. Therefore, when the condition checking unit indicates that the condition is not satisfied, the estimated fundamental sound generation unit does not change the assumed current fundamental sound. In this configuration, it is necessary for the fourth calculation means to always estimate the current fundamental tone. The inference here estimates the chord constituent sound closest to the sound as the current base note.

Before starting the detailed description of the operation of the individual calculation means from FIG. 8, the main variables used in those flows are shown in FIG. 6 and should be referred to.

For other convenience reasons, in the flow, the range of the melody line is from the lowest note indicated by S = 1 to the highest note indicated by S = 25. Further, it is assumed that the base fundamental sound moves in a range of three octaves, that is, one and a half octaves before and after the pitch of the note described in the chord constituent sound storage device 3 as a chord constituent sound (however, explicit upper part). No limits are used.) It is also assumed that the chord section and the bar have the same length. One base note is generated for each chord section.

First Computing Means A detailed flow of the first computing means 7-1 is illustrated in FIG. I = 1 (8-1) indicates the beginning of the music, that is, the first measure. In the first measure, the root of the first chord is estimated as the base fundamental (8-2).

Fourth arithmetic means The fundamental tone determination section is the second
It is moved to the bar and the x-th (x
2) Substrates for bar or xth bar and one or two bars following it are determined. The beginning of this loop is the fourth operation means 7-3, and a detailed flow of the operation is illustrated in FIG.

As described above, the outline of this operation is as follows.
This is to estimate the chord constituent tone closest to the fundamental tone of the first measure) as the fundamental tone of the current measure (the Ith measure). J is the number of the chord component, and MIN is a variable or register that contains the closest chord component found. flag
The reason why FL is used is that in this example, absolute pitches are compared with each other instead of pitch names. Alternatively, the comparison may be performed using note names.

9-4 to 9-6 indicate whether the J-th chord component sound is the same as the immediately preceding fundamental tone (note the pitch or pitch within an octave). In other words, the pitch CHORDK (CHORDK) of the Jth chord component of the Ith chord
The pitch name KN (J) of (I), J) is KN (J) = (CHORDK (CHORD (I), J) MOD12). Similarly, the pitch name KN (I-1) of the pitch KION (I-1) of the previous fundamental tone is KN (I-1) = (KION (I-1) MOD12). 9-4 to 9-6 correspond to whether or not KN (I-1) ≠ KN (J), that is, whether or not the immediately preceding fundamental tone and the J-th chord constituent tone are not the same tone name.

In the case of the same note name (may be simply called the same note), the processing shifts to comparison with the next chord constituent note as shown in the figure.

If it is not the same sound, the pitch difference, that is, the difference in pitch between the previous fundamental tone and the Jth chord constituent tone is substituted for D (J) (9-).
7). Then, it is checked whether or not the pitch difference D (J) is larger than half of one octave (9-8). If it is larger, it is checked whether or not the pitch difference D (J) is larger than one octave (9-). 9). These two checks are for evaluating the pitch difference as a difference of not more than half an octave, as can be seen from the processes of 9-10 and 9-11. The other is to set the flag FL to confirm the change of one octave when the route to 9-9 is taken. In other words, when considering the pitch difference D (J) instead of the pitch difference D (J) (semitone U (J) major seventh), D (J) = MIN (U (J), U ( J), which is the one-octave complement of J). For example, in the case of "do" and "so", it is closer to seeing "so" four degrees below ("5" in numerical value) than to see "so" five degrees above "do" (in this case, "7").

Thus, D (J) evaluated less than half an octave is compared 9-12 with the contents of MIN. Where M
The case where IN> D (J) is satisfied is as follows. That is, this is the time when the J-th chord component sound is the closest to the previous fundamental tone.

The purpose of the fourth calculating means 7-3 is to find the sound closest to the previous fundamental tone from the current chord constituent tones and estimate it as the current fundamental tone. Therefore, when MIN> D (J) holds, as shown in 9-13, the D (J)
In the MIN, and substitute the J-th chord constituent tone having the J as an argument into the fundamental tone KION (I) of the I-th section (current section).

9-14 to 9-17 are corrections to 9-8 to 9-11 described above. That is, when D (J) has a difference of more than half of one octave in 9-8, D (J)
(9-10, 9-1)
1) By executing the inverse conversion in 9-15 to 9-17, KION (I) is corrected in pitch, that is, the previous fundamental tone
KION (I-1) is converted to a pitch having a pitch narrower than half an octave (octave shift from data described as pitches of chord constituent sounds). No change is necessary when the pitch is originally narrower than half an octave (FL = 0). However, regarding the processing before entering 9-14, note that pitch name comparison may be performed instead of pitch comparison. In this case, note names of chord constituent tones that are higher than half an octave of the previous fundamental note are closer to the lower side (upper / lower flags are set). At the time of octave adjustment, looking at the upper / lower flag, if the upper side is indicated, the pitch difference or pitch difference D (J) between the previous fundamental and the current fundamental is added to the pitch of the previous fundamental. The pitch of the current fundamental tone is determined (KION (I) = KION (I-1) + D (J)), and when it indicates the lower side, subtraction, that is, KION (I) = KION (I-1) -D By executing (J), the process is substantially equivalent to the octave adjustment from 9-14 to 9-17. Alternatively, D (J) of the nearest sound above the last fundamental sound in the search is set to MIN
(Upper), and D (J) of the nearest sound below is MIN
(Bottom), when converting to pitch,
When the height obtained by adding MIN (upper) exceeds the upper limit, a sound having a pitch obtained by subtracting MIN (lower) from the previous base note may be estimated as the current base note. If not, MIN (above)
The sound shifted from the previous fundamental by the smaller value of MIN (lower) and MIN (lower) is set as the current fundamental.

9-18 is checking whether or not all the chord constituent sounds have been read out. If not all the chord constituent sounds have been read out, the process proceeds to the evaluation of the next chord constituent sound. At 9-18, J = CHORDN (CHOR
When D (I)) is satisfied, KION (I) contains the sound closest to the previous fundamental tone among the chord constituent sounds of the current section. Therefore, the inference of the fourth operation means is completed.

Fifth arithmetic means As shown in the overall flow, the fifth arithmetic means 7-4 is next to the fourth arithmetic means 7-3. This is illustrated in detail in FIG. The purpose of this means 7-4 is to use the fundamental tone of the current section estimated by the fourth calculating means 7-3 as a temporary fundamental tone, and to use the temporary fundamental tone, the previous fundamental tone and the root tone of the chord of the current interval. Is to evaluate the root note of the current section as a fundamental tone when there is a specific relation.

The operations from 10-3 to 10-6 are the same as the operations from 9-8 to 9-11 described above, and similarly, from 10-12 to 10-
Step 15 is the same as steps 9-14 to 9-17, and a description thereof will be omitted. As shown in 10-2, the register A stores the difference in pitch between the previous fundamental and the current fundamental, and the register B stores the difference between the previous fundamental and the current chord. Here
KION (I) is an output of the fourth calculating means 7-3, and the priority control means (CPU1) is connected to the fifth calculating means 7-4 by the fourth calculating means 7-3.
It can be seen that the estimated fundamental tone of -3 is being passed.
Also, the operation after 10-10 is the operation of the action unit (estimated fundamental tone generation unit) of the fifth calculating means 7-4, and the preceding part can be regarded as the operation of the condition inspection unit.

That is, 10-7 to 10-10 are the condition check units of the fifth arithmetic unit. The condition elements here are as follows.

(A) Inspection of whether or not the previous fundamental and the present fundamental are allophone names. (B) The difference between the pitch of the previous fundamental and the present assumed fundamental, and the difference between the previous fundamental and the present root. Is less than the whole tone (simply speaking, whether the current root tone is also sufficiently close to the previous fundamental tone) When both of the above condition elements (A) and (B) are satisfied, Inspection passed. Others are not established.

If the inspection is successful, the current root tone is issued as the current base tone as indicated by 10-11. If the test fails, the fundamental tone estimated by the fourth calculating means 7-3 is passed to the next means as it is.

Thus, the fifth calculating means 7-4 is
-3 is rechecked, and when a certain condition is satisfied, the root tone is changed to the root tone. That is, the function as the correcting means of the fourth calculating means 7-3 (adjacent sound means) is realized in cooperation with the priority control means.

Sixth arithmetic means The details of the sixth arithmetic means 7-5 shown in the overall flow are exemplified in FIG. Sixth calculating means 7 in the entire fundamental tone determination device
The role of -5 is to estimate the root note of the first measure of the repetition of the motif (first phrase) as the root note. 11−
Reference numerals 1 to 11-8 denote condition inspection units of the sixth calculation means 7-5, and reference numerals 11 to 9 denote estimated base tone generation units.

In 11-3 to 11-7, in the array SB (x), a repetitive phrase (this means that the R-th phrase segment data SABI
Take the MOD10 of (R) and check the type W of the passage to see if it is a repetition type (W = 0). The number of the bar where this starts (this is indicated in the upper digit of SABI (R)). Is set).

In step 11-8, it is checked whether any of the first bar numbers of the repetitive passages matches the current bar number.
If not, the condition of the condition checking unit of the sixth calculating means 7-5 is not satisfied, and the output of the fourth calculating means 7-3 or the output of the estimated fundamental tone changed by the fifth calculating means 7-4 is still valid. Become. If there is a match, the current measure is the first measure of the repetition, so the estimated fundamental tone generator of the sixth calculating means 7-5 is started, and the current fundamental tone is processed by the processing of 11-9 to 11-13. The root is estimated. Note that 11-10
Steps 11 to 13 are octave adjustment processes, and the pitch of the current fundamental tone is set to a pitch narrower than half of one octave from the previous fundamental tone.

As described above, the sixth computing means 7-5 examines the musical conditions related to the structure of the music, and when the examination passes, estimates the root note as the current fundamental. In other words, it has logic to make the fundamental tone of the current section depend on the fundamental tone of the past section farther than the immediately preceding fundamental tone. It is common to the first calculation means 7-1 in that it depends on the music structure or the progress position of the music.

Seventh arithmetic means The details of the seventh arithmetic means 7-6 are exemplified in FIGS. 12 and 13. The seventh arithmetic means 7-6 is provided with a first condition checker (FIG. 12) for checking whether the current section has a specific relationship with the start position of the passage, and no specific relationship is recognized in this check. At the time of the start, and the root sandwiched by the elapsed sounds in three chord sections (three measures in this example) including the section (I-1), the next section I, and the next section (I + 1). A second condition inspection unit (first half of FIG. 13) that searches for a sequence of chord notes having a sound, that is, a root note → an elapsed sound → a root note, and a case where it is found as a result of the search The root sound, elapsed sound, and root sound corresponding to
Fundamental tone of section (I-1), fundamental tone of section I, section (I + 1)
And a presumed fundamental tone generator (the latter half of FIG. 13).

When the condition of the first condition checking unit of the seventh calculating means 7-6 is not satisfied, the process proceeds to the eighth calculating means 7-7 as shown in the overall flow (FIG. 7). Similarly, the second calculation means 7-6
If the condition is not satisfied by the condition checking unit (if the condition is not found as a result of the search), the process proceeds to the eighth calculating means 7-7. When the condition is also satisfied in the second condition inspection section, the condition is not changed as indicated by 7-15 in FIG.
, F ≠ 0 is confirmed, the section I is incremented by two (see 7-16 and 7-14), and the process proceeds to the fundamental tone determination processing of the section (I + 2). In this case, three fundamental tones of section (I-1), section I, and section (I + 1) are determined.

More specifically, in FIGS. 12-1 to 12-14,
It is checked whether the current section (I-1) or the next section I is the first measure of a passage. In the present example, when the current or next section is the first measure of a phrase, it is considered undesirable to apply the elapsed sound base, and the base tone estimated by the sixth computing means 7-5 or earlier is enabled to enable the eighth tone. It is passed to the calculating means 7-7. In other words, it is a necessary condition for applying the elapsed sound base that the current section or the next section is not the first measure of the passage.

At 12-3, put the chord root of section (I-1) in A,
In B, the first chord constituent sound of section I is put, and in C, the chord root of section (I + 1) is put. 12-4 to 12-
At 16, the root A of the section (I-1) is being adjusted by one octave for connection within half an octave with the pitch of the definitive fundamental note of the section (I-2). This 12-2 to 12-
The processing up to 6 may be performed immediately before the flow from FIG. From 12-7 to 12-14, section I or the next section (I
+1) is to see if it is the start measure of each passage. 12-8
To 12-11, the first measure of the R-th phrase stored in the phrase division storage device 4 (FIG. 1) is decoded, and
Put in. Whether I ’ve checked all the passages
See 12-13. If it is found in the middle of the search at 12-12 that the head measure SB of the focused phrase matches the section I or the section (I + 1), the eighth arithmetic means 7-7
Proceed to. As a result of the search, when it is found that the next section (I + 1) does not match the first measure of any section, the flow proceeds to the flow shown in FIG.

In FIG. 13, three sections (I-1), I, (I
At +1), a search is made as to whether there is a sequence of root sound → elapsed sound → root sound, based on all rounds of the chord constituent sounds. And
When a search is found in the search, the arrangement is estimated to be a sequence of fundamental tones in sections (I-1), I, and (I + 1). As a result of the search, when it is found that there is no arrangement satisfying the elapsed sound condition, the task is completed without changing the fundamental tone. More specifically, 13-1, 13-2, 13-3
Checks whether the J-th chord constituent note of interest in section I is upwardly progressive (half-tone or full-tone) with respect to the root note in section (I-1). In this example, since the pitch includes a difference of one octave above and below, three types of comparison are performed. L = 1 in 13-7 means that if the chord sound of section I is raised by one octave, the sound will be in the section (I-
This is a process for showing that the pitch relationship within a half octave is established for the root note whose octave has been adjusted in 1). 13−
L = 3 of 8 is processing indicating that octave conversion is unnecessary, and L = 2 of 13-9 is processing indicating that the chord sound in section I should be lowered by one octave. On the other hand, 13-4,
In 13-5 and 13-6, it is checked whether or not the J-th chord constituent sound in the section I has a downward progression with respect to the root note in the section (I-1). The meanings of 13-10, 13-11, and 13-12 are the same as those of 13-7, 13-8, and 13-9. When the progress from section (I-1) to section I is progressive, the progress from section I to section (I + 1) is upward at 13-13, 13-14, and 13-15. Check whether or not. In the case of upward sequential progression, the J-th chord constituent sound of the section I is an elapsed sound in the upward sequential progression, and the conclusion is issued in 13-19, 13-20, and 13-21. Similarly, when the progress from the section (I-1) to the section I is a downward progress, the progress from the section I to the section (I + 1) is the same at 13-16, 13-17, and 13-18. The polarity, that is, whether it is a downward sequential progress or not, is checked. If it is established, the J-th chord constituent sound of the interval I is a lapsed sound in the downward sequential progress, and the conclusion is made to 13-22, 13-23, Issued on 13-24. Here also, since the pitch is not a pitch name but a pitch comparison, in order to adjust the pitch, the sounds of the section I and the section (I + 1) are connected within a half octave in the processing of 13-19 to 13-24. . That is, the section (I) for the determined section (I-2)
The octave adjustment processing of -1) has already been performed (12-4 to 12-6), and the octave adjustment processing of the section (I + 1) with respect to the section I is performed from 13-19 to 13-24. The remaining processing is performed on the sections I and (I-1) for the section (I-1).
+1) octave adjustment, which is from 13-7 to 13
This can be performed by referring to the octave change instruction flag L set at -12. The process is from 13-27 to 13-30, and in 13-31, L is set to F in order to proceed to (check of flag F of 7-15) in the overall flow.

If the J-th chord constituent sound in the section I does not satisfy the elapsed sound condition, J is incremented as shown in 13-26, and the process proceeds to a check whether the next chord constituent sound is a passing sound. If none of the chord constituent tones is an elapsed sound, keep F = 0 (see 12-1), and in 13-25, J = CHORDN (CHORDN
(J)) is established, and thereafter, the flow proceeds to the check 7-15 of the flag F of the entire flow.

The example of the flow in FIG. 13 is a chord configuration sound brute force method, but a conversion table may be used instead. That is, a sequence of three specific chords is a specific root →
Elapsed sound → Includes progression of root sound. Therefore, for example, the codes of the three sections of interest are used as keys to search the three code fields of the conversion table. In each of the three chord columns, information on a specific root note → elapsed sound → root note is linked as its conclusion. As a result of the search, if three codes that match the key are found, the conclusion is read out and the fundamental tone variables KION (I-1), KION (I), and KION (I-1) of three sections (I-1), I, and (I + 1) are read. Put in KION (I + 1).

12 and 13 use the pitch comparison as a basic operation. Instead of this, after comparing the pitch names and determining three sections by pitch names, an octave name may be added to the pitch name of each section (octave number from pitch of section (I-2)). Can be determined).

As described above, the seventh calculating means 7-6, when indicating the situation where the condition of the music structure is present,
Root → Elapsed → Search for progress of root, if there is,
Issue them as a sequence of fundamental sounds. When the condition of the music structure and the progress conditions of the three chords are satisfied, the sequence of these fundamental tones is a determined fundamental sequence as can be seen from the overall flow. If the condition is not satisfied, the seventh computing means 7-
6, section I from the means positioned upstream
Is estimated as it is and passed to the downstream means. The role of the seventh calculating means 7-6 in the entire determination device is to give the characteristic depending on the distant past and the characteristic of the elapsing sound baseline to the base tone sequence of the baseline.

Eighth Calculation Means The eighth calculation means 7-7 does not have logic for estimating a fundamental tone. The purpose of the eighth calculating means 7-7 itself is to extract the dominant sound of the melody line in the current section. Accordingly, this processing may be performed completely independently, but for convenience, the processing is performed next to the seventh calculation means 7-6 in the overall flow. FIG. 14 illustrates a detailed flow of the eighth calculating means 7-7. The definition of the melody dominant sound in this flow is as follows. That is, among the pitches included in the melody line of the section of interest, a pitch that maximizes a function using the position and the pitch in the section (measure) as variables. This function SUMO (S) is given by the following equation.

SUMO (S) = ΣZ (SS) × IN (2, J) where SS is a position in the section, Z (SS) is the weight of the position (see FIG. 5), and IN (2, J) is high. The length of the sound having the length S. J indicates the J-th melody note.

The definition of the melody dominant sound is an example, and for example, the melody dominant sound may be evaluated by a pitch name instead of a pitch. The extraction of the melody dominant sound is to consider the melody dominant sound in the base tone estimation, as described later.

In FIG. 14, HN is the number of the first melody note in the current section, EN is the number of the last melody note in the current section, and is obtained by accessing the melody data storage device 5 (FIG. 1). The range of the melody note is S = 1
2 octaves from S to 25 are assumed. The array {SUMO (S)} contains the evaluation value of each pitch. SS represents the pulse position. The sequence from 14-5 to 14-7 is @SUMO
This is the initialization processing of (S)｝. 14-9 and 14-10 to 14-
In S13, the pitch of the J-th melody note is obtained by scanning S and accumulated in SUMO (S). At step 14-16, J is incremented, and the process proceeds to the SUMO (S) evaluation of the next melody note. When J = EN is satisfied in 14-15, the evaluation function value SUMO (S) is obtained for each pitch S.

From 14-17 to 14-21, which evaluation function value is the largest is determined, and S is set as the dominant tone OH.

Ninth arithmetic means The main purpose of the ninth arithmetic means 7-8 is to estimate the fundamental note of the current section as the root note (dominant) when the code of the current section is the dominant code (V). In this example, a condition in which the melody dominant sound in the current section is not dominant is used as an additional condition. FIG. 15 illustrates a detailed flow of the ninth operation means 7-8.

Reference numeral 15-1 denotes an inspection by the condition inspection unit of the ninth operation means 7-8. 15-2 and below are the operations of the estimated root generation unit of the ninth calculating means 7-8. If the condition is not satisfied in 15-1, the fundamental tone estimated by the upstream means is still valid. 15-1
Is satisfied, that is, the code CHORD of the current section I
If (I) is a dominant (= 8) and the melody dominant tone OH of the current section I is not a dominant, 15-2 to 15
At -6, the fundamental tone of the current section becomes dominant. 15-
3 to 15-6 are octave adjustments.

As described above, the ninth arithmetic means has the logic of the root tone, and has the logic of the prohibition rule of complete eight degrees.

Tenth arithmetic means Next to the ninth arithmetic means 7-8, the tenth arithmetic means 7-9 is activated (FIG. 7). The function of the tenth arithmetic means 7-9 is to estimate the VII (conducted) -I (tonic) sequence as the fundamental sequence when the chord progression is V-1 (dominant-tonic). Additionally, a full eight-degree prohibition is used. FIG. 16 shows a detailed flow of the tenth calculation means 7-9.

16-1 is the inspection of the condition inspection unit of the tenth arithmetic means 7-9, and 16-2 to 16-6 are the operations of the estimated fundamental tone generation unit of the same means 7-8. Since there are many points in common with the description of the flow in FIG. 15, further description of FIG. 16 will be omitted.

The tenth calculating means is common to the seventh calculating means in that the fundamental sounds of a plurality of (two in this case) sections are collectively estimated. Further, it is common to the first, fifth, and sixth calculation means in directing the root tone.

Eleventh arithmetic means The eleventh arithmetic means 7-10 prohibits the use of the sound of VII as the fundamental tone when the chord progression is not VI progression.
Then, among the sounds other than the VII-degree sound, the chord constituent sound closest to the previous base tone is estimated as the current base tone. The eleventh arithmetic means 7-10 uses the assumed current fundamental tone in the condition checking section. As is understood from the overall flow of FIG. 7, this assumed current fundamental tone is a current fundamental tone that is effective up to the means upstream of the eleventh arithmetic means 7-10. The condition checker determines that the chord progression of two consecutive sections I and I + 1 is V-
If the current base note in section I is not VII and the assumed base note in section I is VII (conducted sound), the condition is satisfied. The estimated fundamental tone generation unit of the eleventh arithmetic means has a logic of estimating a sound closest to the previous fundamental tone as the current fundamental tone among the chord constituent sounds excluding the lead sound. Therefore, the generation unit includes the fourth arithmetic unit 7
-3 (adjacent sound estimation means) in many places.

FIG. 17 illustrates a detailed flow of the eleventh calculating means 7-10. 17-2 is a condition check by the eleventh condition inspection unit. If the condition is not satisfied, the fundamental tone estimated by means upstream of this means is still valid. When the condition is satisfied, the eleventh estimated fundamental tone generation unit is activated at 17-3 or less. In this case, some of the chord constituent tones should match the hypothetical present base tone (sound of VII). The place where the chord constituent sound is stored is 17-5 following check 17-4.
17-7, 17-8, and 17-9 are processes for performing an adjacent sound search of 17-10 or less only for chord constituent sounds other than the chord constituent sounds. In D (K) shown in 17-10, the pitch difference between the previous fundamental tone KION (I-1) and the K-th chord constituent note of interest is entered. 17-10 to 17-20 correspond to 9-7 in the flow (FIG. 9) of the fourth calculating means of the adjacent sound rule.
To 9-19. When exiting the flow at 17-20, KION (I) contains the pitch of the chord component other than the lead tone that is closest to the previous fundamental tone.

Twelfth arithmetic means The twelfth arithmetic means 7-11 is a means for realizing a complete eight-degree prohibition rule. As shown in the overall flow (FIG. 7), this means 7-11 is activated next to the eleventh arithmetic means 7-10. Twelfth
The calculating means 7-11 serves to change this assumed current fundamental tone when the current fundamental tone valid in the processing up to the eleventh arithmetic means 7-10 is the same kind of tone as the dominant tone of the current melody line. . That is, when the perfect eighth degree condition is satisfied, the sound closest to the preceding fundamental tone is estimated as the current fundamental tone among the chord constituent sounds excluding the perfect eighth degree. An example of the detailed flow of the twelfth arithmetic means is illustrated in FIG.

In FIG. 18, 18-1 and 18-3 indicate the operation of the twelfth condition inspection unit. 18-4 and below show the operation of the twelfth generated fundamental sound estimation unit. Except that the condition check of 18-3 is different from the condition check of 17-2, it is the same as the means of FIG. 17, and a detailed description thereof will be omitted.

In the flow of FIG. 18, when the hypothesized current fundamental is completely 8th related to the current melody dominant tone, the sound closest to the previous fundamental in the chord constituting sounds excluding the hypothetical current fundamental is estimated as the current fundamental. The logic to do it. Instead of this, the eleventh arithmetic means and the twelfth arithmetic means are combined, and the condition checker is constructed by ORing the conditions of the eleventh arithmetic means and the conditions of the twelfth arithmetic means shown in the figure, and the estimated fundamental tone generator In the above, the sound closest to the previous fundamental is presumed to be the current fundamental among the chord-constituent sounds in which the conduction sound other than the VI progression is excluded and the sound having a perfect eighth degree relation with the melody dominant sound is eliminated. You may. This modification is easy, and according to this modification, the possibility that occurs in the case of a cascade connection configuration in which the twelfth arithmetic means is located downstream of the eleventh arithmetic means can be eliminated. This possibility is the 12th
As a result of the operation of the estimated fundamental tone generation unit of the calculating means, the possibility that the conducted sound is estimated as the fundamental tone, though it is not a perfect eighth degree. Conversely, the configuration shown allows this possibility.

The twelfth arithmetic means 7-11 can change the hypothesized present fundamental to another fundamental tone on condition that the condition of the condition checking section of the seventh arithmetic means 7-6 is not satisfied, not at the beginning of the music and not at the end of the music. Yes, the changed fundamental tone is determined (assuming that there is no change by the later seventh calculating means 7-6). The twelfth arithmetic means 7-
Reference numeral 11 is common to the eleventh calculation means 7-10 and the fifth calculation means 7-4 in that the assumed current fundamental tone is used. Further, the twelfth arithmetic means 7-10 absolutely needs the melody line information to completely prohibit the eighth melody. On the other hand, regarding the ninth operation means 7-8 and the tenth operation means 7-9, the logic may be changed to a logic that does not refer to the melody line information.

Thirteenth Calculation Means The thirteenth calculation means 7-12 is means having logic to make the end of a song a root note. The operation of the thirteenth calculation means 7-12 is
It is shown in the figure. 19-1 is the condition check of the thirteenth condition check unit, and 19-2 represents the operation of the thirteenth estimated fundamental tone generation unit.

The thirteenth arithmetic means 7-12 directs the root note. In addition, the first operation means 7-1,
It is common to the sixth calculating means 7-5 and the seventh calculating means 7-6 (the first condition checking unit).

Priorities From the above description, the priorities among the arithmetic units according to the present embodiment are clear. The fourth operation means is positioned as the weakest means. Next is the fifth arithmetic means, the next is the sixth arithmetic means, the next is the ninth arithmetic means, the next is the tenth arithmetic means, the next is the eleventh arithmetic means, the next is the twelfth arithmetic means, The strongest are the first arithmetic means, the seventh arithmetic means, and the thirteenth arithmetic means. However, you do not have to be limited to this priority. For example, it is possible to upgrade the tenth arithmetic means relating to VI progress. That is, when the condition checking unit of the tenth arithmetic unit declares that the condition is satisfied, the output of the estimated fundamental tone generating unit of the tenth arithmetic unit (combination of the lead sound and the main tone) is determined as the fundamental sound of two continuous sections. be able to.

Past Dependence and Independence Of the calculation means, the fourth calculation means relating to adjacent sound extraction has past dependency. Similarly, the condition checking unit of the fifth calculating unit also refers to the past information (the base tone of the immediately preceding own line). Further, the eleventh calculation means and the twelfth calculation means also refer to and depend on the fundamental tone of the immediately preceding own line (base line). Means that depend on the structure of the song or the distant past are the first
The operation means, the sixth operation means, the seventh operation means and the thirteenth operation means. In these means, the fundamental tone of the current section is largely dependent on information of a section different from the section of the immediately preceding self-line information. Others are independent from the past. However, for intra-octave connection, the octave number of the bass fundamental in the immediately preceding section is used. The means for connecting within an octave may be executed after one or more fundamental note names are determined in each arithmetic means. In this case, the predetermined arithmetic means (particularly, means having rules of adjacent sounds) stores the sound adjacent to the previous fundamental sound in the upward progression and the sound adjacent to the previous fundamental sound in the downward progression. This is used in the means (pitch name / pitch conversion means) (to create the pitch of the nearest adjacent note within the range permitted by the upper limit or bass range).
In the case of the calculating means for estimating only one fundamental note, the note name / pitch converting means converts the pitch into a pitch connected within 1/2 octave to the pitch of the previous fundamental note.

Use of hypothetical current fundamental tone The fifth arithmetic means, the eleventh arithmetic means, and the twelfth arithmetic means use the hypothetical current fundamental tone from upstream. This configuration works in such a way that when there is a problem with the assumed current fundamental tone, a different tone from the assumed current fundamental tone is used as the fundamental tone. It's an exceptional means. On the other hand, for example, the seventh calculating means does not use the assumed current fundamental tone, and estimates the unique tone as the fundamental tone if the unique internal condition is satisfied. Such a complex character is considered to be a good expression of human musical knowledge.

Collective Estimation of Fundamental Tones in a Plurality of Sections The seventh calculating means and the tenth calculating means include logic for collectively estimating the fundamental sounds in a plurality of continuous sections. This also shows a certain characteristic of the way humans decide the line. This is based on the fact that not only the present time but also the future fundamental tone is estimated from the musical information of not only the current section but also the succeeding section (future section). Further, in the seventh computing means, the first condition inspection unit inspects the tuned structure (the effect of a distant past section), and is of the past reference type.

Consideration of melody line The ninth, tenth, and twelfth calculation means include logic for considering a melody line which is a line different from the baseline. As a whole, the baseline fundamental tone determination device considers not only its own line but also another line.

Melody dominant sound The eighth calculating means extracts a sound dominating the melody line by evaluating from the position and length of the melody sound. In other words, it is the extraction of the most prominent sound. Then, the ninth, tenth, and twelfth calculation means use this result.

Please refer to the table for the properties of each calculation means.

This concludes the description of the non-real-time bass fundamental tone determining means according to this embodiment, but it is clear that various changes, modifications, and improvements are possible. For example, it is easy to give a priority order other than the above. In one configuration, the priority is It can be glam. Alternatively, the priority may be changed by an electronic random number generator. For example, if it is desired to give the same rank to the estimation of a certain arithmetic unit and another arithmetic unit, the random number generator uses the estimation of one of the arithmetic units at a rate of 50% during operation.

Further, the present invention can be applied to the generation of a fundamental tone of an accompaniment line other than the bass line.

For reference, FIG. 20 shows a base fundamental tone determination flow substantially equivalent to the flow of the main part of FIG. A circle with a number indicates the number of the calculation means, a diamond-shaped box indicates the operation of the condition checker of that number, and a rectangular box indicates the operation of the estimated fundamental generator of that number.

[Effects of the Invention] As described in detail above, the device of the present invention determines the basic sound (fundamental tone) of the accompaniment line for each music section as the first stage for forming the accompaniment line. .
A non-real-time type accompaniment line fundamental tone determining apparatus according to claim 1 is a past-dependent type which estimates a fundamental tone of a section of interest in a form depending on a fundamental tone of a past section according to a musical situation of the section. There are provided a fundamental tone estimating means, a past independent fundamental tone estimating means for estimating in a form independent of the fundamental tone of the past section, and a means for determining a fundamental tone of each section based on the estimation result from both estimating means. As a result, the two fundamental tone estimating means function to complement each other, and can generate a musical fundamental tone progression that cannot be obtained by the fundamental tone determining device using the sole estimation logic.

Further, the apparatus for determining an accompaniment line fundamental tone according to claim 5 includes a single-segment type fundamental tone estimating means for estimating a fundamental tone of the single segment according to a musical situation of the single segment, and a certain segment. Not only one or more subsequent sections, and a collective base note estimating means for collectively estimating a sequence of fundamental notes for the plurality of sections based on a musical situation in the plurality of continuous sections; Means for determining the fundamental tone of each section based on the estimation result (first and second output signals) from the control signal. This configuration also has the ability to form fundamental progressions that cannot be obtained with single fundamental estimation logic.

The present invention expresses the characteristics of the determination of the accompaniment line by the composer and the performer well, and generates a fundamental tone progression that cannot be achieved by the conventional accompaniment line fundamental tone determination technology by an artificial intelligence approach according to the music situation. can do.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of a non-real time type accompaniment line fundamental tone determining apparatus according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of pitch data, and FIG. 3 is a data representation of a chord progression relationship. Figure,
FIG. 4 is a diagram showing the expression of phrase section data, FIG. 5 is a diagram showing an example of a pulse scale, FIG. 6 is a diagram showing main variables used in the present embodiment, and FIG. FIG. 8 is a flowchart showing the operation of the first arithmetic means, FIG. 9 is a flowchart showing the operation of the fourth arithmetic means, FIG. 10 is a flowchart showing the operation of the fifth arithmetic means, FIG. Is a flowchart showing the operation of the sixth computing means, FIG. 12 is a flowchart showing the first half operation of the seventh computing means, FIG. 13 is a flowchart showing the latter half operation of the seventh computing means, and FIG. 15 is a flowchart showing the operation of the ninth calculating means, FIG. 16 is a flowchart showing the operation of the tenth calculating means, FIG. 17 is a flowchart showing the operation of the eleventh calculating means, Figure 18 shows the 12th operator Flowchart illustrating the operation, FIG. 19 is a flowchart showing the operation of the thirteenth calculation means, FIG. 20 is substantially equivalent to the flow chart and flow of Figure 7. 1 CPU, 2 chord progression storage device, 3 chord configuration sound storage device, 4 phrase storage device, 5 melody data storage device, 7 first computation means, 8th 4
Arithmetic means, 9 ... fifth arithmetic means, 10 ... sixth arithmetic means,
11 ... 7th calculation means, 12 ... 8th calculation means, 13 ... ninth
Arithmetic means, 14 ... tenth arithmetic means, 15 ... eleventh arithmetic means,
Sixteenth twelfth arithmetic means, seventeenth thirteenth arithmetic means.

Claims (7)

(57) [Claims] 1. An accompaniment line fundamental tone determining apparatus for determining a sound (referred to as a substrate) which is a basis of an accompaniment line for each of a plurality of music sections constituting an accompaniment line on a non-real time basis. The apparatus selectively operates depending on the music situation in the first section of interest in the section, estimates the base tone for the first section based on the past base, and outputs a first output representing the estimated base tone. A past-dependent fundamental tone estimating means for outputting a signal; and selectively operating depending on a music situation in a second section of interest out of the plurality of music sections, to generate a fundamental tone for the second section in the past. A past independent basic tone estimating means for estimating independently of the fundamental tone and outputting a second output signal representing the estimated fundamental tone; operatively coupled to the past dependent basic tone estimating means and the past independent fundamental tone estimating means; The above The base tone for each section of the accompaniment line is determined from the second output signal and the base tone determined for a section of the accompaniment line is affected by the base tone determined before that section, but is different from that of the accompaniment line. Means for preventing a fundamental tone determined for a section from being affected by any fundamental tone determined before the other section, an accompaniment line fundamental tone determining apparatus. 2. The accompaniment line fundamental tone deciding device according to claim 1, wherein said past-dependent fundamental tone estimating means is configured to determine a chord constituent tone assigned to said first section with respect to a fundamental tone in an immediately preceding section. An accompaniment line base tone determining device, which includes means immediately before and before estimating a chord constituent sound having a predetermined relationship as a base tone of the first section. 3. An accompaniment line fundamental tone determining apparatus according to claim 1, wherein said past independent fundamental tone estimating means is configured to determine whether the chord assigned to said second section is a specific chord type. The specific constituent sound of the second
An accompaniment line fundamental tone determining device, comprising: a chord type dependent means for estimating a basic tone of a section of the section. 4. An accompaniment line fundamental tone determining apparatus according to claim 1, wherein said past independent fundamental tone estimating means comprises:
An accompaniment line fundamental tone comprising batch estimating means for collectively estimating a fundamental tone for a plurality of continuous sections when a chord progression covering a plurality of continuous sections including a section of interest has a specific pattern. Decision device. 5. An accompaniment line fundamental tone determining apparatus for determining a sound (hereinafter referred to as a fundamental tone) which is a basis of an accompaniment line on a non-real time basis for each of a plurality of music sections constituting an accompaniment line. A single section fundamental sound estimating means for estimating a fundamental sound for one section in accordance with the music situation of the single section and generating a first output signal representing the fundamental sound; and a plurality of consecutive fundamental sounds for a plurality of continuous sections. A collective base tone estimating means for collectively estimating when the music situation of the section is a predetermined music situation and generating a second output signal representing these fundamental sounds; Means operatively coupled to the first and second output signals to determine a fundamental tone for each of the plurality of sections from the first and second output signals. 6. An accompaniment line fundamental tone determining apparatus according to claim 5, wherein said collective fundamental tone estimating means includes a conditional collective fundamental tone estimating means which operates conditionally. (A) referring to condition parameters relating to the plurality of continuous sections, and (b) estimating fundamental sounds for the plurality of continuous sections collectively only when this condition parameter has a specific value. Accompaniment line fundamental tone determination device. 7. The accompaniment line fundamental tone deciding device according to claim 6, wherein the conditional collective fundamental tone estimating means includes: (a) the conditional parameter is a specific value, and Operating when the chord progression of the section includes a specific chord component sound pattern, and estimating the specific chord component sound pattern as a series of fundamental tones for the continuous plurality of sections. Line fundamental tone determination device.