JPS58132798A - Recognition method and apparatus for performance or singing - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is a performance or singing recognition method that recognizes input information from a performance or singing based on a music teacher's sense, thereby making it possible to remotely control various objects, for example. and regarding equipment.

Conventionally, in the field of speech recognition, the target of recognition is generally limited to speech 8H, and it is customary to pattern-recognize input speech words using various complex algorithms.

However, in general, the uses of voice recognition do not only require such high-level language recognition, but also include, for example, instructing a recording device that stores multiple songs by singing or playing the first phrase. As in the case of reading out one of these words, there are cases in which the object is identified by pronouncing a melody without necessarily using eight words.

5- The purpose of this invention is to
Rather than playing or singing, for example, the purpose is to have the object recognized by the arrangement order of musical notes and pitches.

In particular, it is an object of this invention that, when performing normal singing or playing, if a performer or the like notices a mistake during singing or playing, he or she should correct the mistake or perform it again after re-singing. The goal is to make it possible for a music teacher to recognize a performance or singing by taking advantage of the characteristic of moving forward, and by removing such replayed parts from the recognition target.

In other words, the object of the first invention of this application is not only when a singer or performer sings or plays a predetermined phrase accurately (but also when a singer or performer sings or plays a predetermined phrase accurately). It is an object of the present invention to provide a method for recognizing musical performance or singing information that can be recognized with a relatively wide range of tolerance even in the case of singing.

In order to achieve the above-mentioned object, the present invention provides performance 6- or singing information that has at least the dimension of pitch.
Convert the sound sequence data into sound sequence data with a dimension higher than that of the standard sound sequence; then, compare the sound sequence data with the sound sequence data corresponding to the reference sound sequence with respect to predetermined characteristics between the sound sequence patterns; Based on the comparison result, A system characterized by extracting similar phrases from an input sound string; comparing sound string data corresponding to the extracted phrase with sound string data corresponding to a reference sound string to recognize input performance or singing. It is.

Next, the object of the second invention related to this application is to be able to achieve the above-mentioned 1121 processing with a simpler circuit configuration, and to obtain a recognition result that is closer to the sense of a music teacher. An object of the present invention is to provide a recognition device for performance or singing information.

In order to achieve the above object, in the second invention, as a component thereof, each constituent sound sequentially generated by playing or singing is sequentially converted into single sound data of one or two dimensions or more having at least the dimension of pitch. 1 to 2 a tone sequence data registration memory in which reference tone sequence data of each dimension corresponding to the above performance or singing melody is registered; a tone sequence of each dimension corresponding to one performance or singing read from the tone sequence data registration memory; The data and the input sound sequence data of the corresponding dimension are compared between the parts of the sound sequence that exist in the same time period while shifting the reference points of the time axes of both, and the ones with the highest degree of similarity are selected in order. a similar sound string part extracting means for extracting one or more similar sound string parts extracted from the input sound string data; a matching part with the registered sound string data of the corresponding dimension from each of the extracted similar sound string parts; and superimposing them on each other to form the most similar sound sequence data corresponding to the most similar phrase. Find the ratio between the total number of sound data included in the similar sound sequence data and the total number of each component sound data of the registered sound sequence data for the necessary dimensions,
similarity score calculation means for calculating a similarity score for an input performance or singing using the values of these ratios as scoring elements; and input information discrimination for discriminating the input performance or singing by comparing the similarity score with a predetermined reference value. It is characterized by comprising means.

Embodiments including the first and second inventions of this application will be described in detail below with reference to the accompanying drawings.

First, before proceeding to a detailed explanation of the operation of the embodiment, the basic operation of this device will be schematically explained with reference to FIG.

Note that although melody performance will be described below as a representative example, the same applies to other performance modes such as rhythm performance and chord performance.

In the figure, a reference pitch pitch sequence generation circuit 1 and a reference note length tone sequence generation circuit 12 each contain reference pitch pitch sequence data PDlin corresponding to a plurality of types of melodies.
e (ref) and reference note length string data LDlin
e(ref) are stored respectively, and these tone sequence data are used as a melody designation signal 5SeleCt-1, which will be described later.
~5select-4 is configured to be read out alternatively.

Further, the pitch detection circuit 5 is supplied with an audio signal detected via the microphone 4 or an output signal of an electronic musical instrument detected via the input terminal 2, etc. Pitch data PD included in the input signal (
in) and a timing signal 5kOn' indicating the output timing of each pitch data.

The pitch string formation circuit 6 discriminates the length of each note based on the signal S kon- for each of the original pitch data (see FIG. 27 (A)) that is sequentially supplied in a time-division manner, and calculates the musical note length. Pitch data that is less than a minute note length and cannot be considered as a jitter is removed.

Then, each of the remaining pitch data is stored in the order of occurrence, and 10- thereby input pitch string data P D 1ine (f
(see FIG. 27(B)).

Then, this data P D 1ine (full)
C. It is supplied in parallel to the similar tone high pitch string extraction circuit 7 and the -several tone high section valve type matching circuit 8.

The similar tone pitch string extraction circuit 7 receives the φ reference tone pitch string data p[) 1i which is output from the reference tone pitch g string generation circuit 1.
ne (ref) and the input pitch string data P D 1ine (f
The standard after converting ``ull'' into a sound string pattern that exists in the same time period and connects each constituent sound in chronological order while shifting the reference points of each time axis, just as if two trains were passing each other. The sound string parts having the same length as the sound string are compared, and among these sound string parts, those having the highest degree of similarity to the reference tone roar sequence are sequentially extracted coarsely to the maximum extent.

(See FIGS. 27 to 30) Then, these extracted maximum sets of similar pitch string partial data are converted into the input pitch string data P[) 1ine
Shift number data expressed by the number of shifts relative to (full) PD 1ine sml -1 to -k (
(See FIGS. 27(C) to (J)).

At the same time, the similar sound sequence extraction circuit 7 extracts the number of single sounds between the extracted similar sound high pitch sequence partial data and the reference sound sequence data p[) 1ine (ref). Output PDeq-1 to -k.

The matching note partial superimposition circuit 8 uses the shift number data p Q 1in output from the similar note high pitch sequence extraction circuit 7.
e ssl -1 to -board and the input pitch string data PDline (
Each similar sound ^ tone string partial data is played back based on the ``full'', and each reproduced similar tone pitch string partial data is compared with the reference tone ^ tone string data PDline (ref), and if the two match. The pitch data are sequentially superimposed to form the most similar pitch string data p[)smlp1e. (See FIG. 35.) Next, the input note length string forming circuit 9 receives the signal 5kon'' (see FIG. 26) output from the input pitch string forming circuit 6.
C)), input pitch string data P D 1
Input note length note string data LDline (full) is formed by forming a note length data corresponding to each constituent note of ine (full) and sequentially arranging this data in the order of occurrence. (See Figure 31 (8)) Similar note length string extraction circuit 1
0, the reference note length tone string data l Dline (ref ) (25th
(see FIG. 31(C)) and the input note length note string data (see FIG. 31(B)) output from the input note length string forming circuit 9 are arranged on each other's time axes in the same manner as in the case of the pitch. By shifting the reference point of the sound sequence and comparing the parts of the sound sequence that exist in the same time period, the sound sequences are determined in order of similarity. (Figures 32 to 34
) Then, each similar note length note string partial data is converted into input note length note string data LDline (
shift count data l[)line for full)
Convert to ssl -i to -k and output.

At the same time, the similar note long note string extraction circuit 10 extracts each similar note long note string partial data and reference note long note string data [) 1ine
Matching note length 13-note number data LDeq-1 to -k representing the number of matching note length notes with (ref) are output.

The matching note length part valve type matching circuit 11 uses the shift number data LDli output from the similar note length string extraction circuit 1-0.
ne 5sl-1 to - and the input note length string forming circuit L
Based on the input note length note string data L D 1ine (full) output from D 1ine (full), each similar note length note string partial data is played back, and this reproduced similar note length note string partial data is sequentially converted into a reference note. Long string data L
D 1ine (ref) is compared to find the matching note portion between them, and these are superimposed on each other to form the most similar note length note string data L D 381ple. (See Figure 36) Next , the individual note matching discrimination circuit 14 determines each pitch of each note constituting the reference note pitch sequence data based on the reference note note sequence data pQ1ine (ref) and the most similar pitch note sequence data p[)3alple. It is determined whether or not it matches the reference melody, and the determination result is expressed as a 1-bit signal for each line pitch as individual pitch match discrimination data PDe.
Output q-bit.

- The number tone high order detection circuit 15 calculates the matching number N1 for the reference tone ^ sound da] data based on the data PDeQ-b output from the individual pitch matching valve 14 - separate circuit 14, Outputs numerical data 0 (Nl).

In the individual note length coincidence discrimination circuit 16, the reference note length tone string data L
Based on D 1ine (ref) and the most similar note length note string data L D sample, it is determined whether or not both notes match for each note, and the determination result is expressed as a 1-bit signal for each note. Long match discrimination data LDeq-b
Output it.

The matching note length number detection circuit 17 calculates the number of single notes N2 between the J1 semi-note length note string data and the most similar note length note string data l[) sample based on the individual note length matching discrimination data LDeq, Data D representing this - number of tone cylinders
(N2) is output.

The pitch similar tone sequence number detection circuit 18 detects each -several tone pitch data PDeq- outputted from the similar tone sequence extraction circuit 7.
Based on 1 to -k, the number N of sets of similar sound ^ sound string partial data
3 is obtained, and data D (N3) representing this number N3 is output.

The note length similar note string number detection circuit 19 uses each matching note length note number data LDe output from the similar note length note string extraction circuit 10.
Based on q-1 to -k, the number N4 of sets of similar note length tone string partial data is determined, and data D(N4) representing this number N4 is output.

The reference tone number construction circuit 20 uses the reference tone treble string data PDli output from the reference tone treble string data generation circuit 1.
ne (ref) k: base tweet, xii single note its The number N5 of constituent notes of the note string data is determined, and data D (N5) representing this number N5 is output.

The reference note length number generation circuit 21 uses the reference note length tone string data L D 1i output from the reference note length tone string generation circuit 12.
ne (ref), find the number N6 of constituent notes of the reference note length tone sequence data, and calculate the data D representing this number (lIN6).
(N6) is output.

The performance time/time difference detection circuit 22 uses the individual tone ^ coincidence discrimination data p output from the individual pitch coincidence discrimination circuit 14.
o eq-bat and the matching length part valve type combination circuit 1
Most similar note length string data l-[)sa- output from 1
〇: e, reference note length note string data LDline<ref) output from the reference note length note string generation circuit 12, and individual note length coincidence discrimination data LDeQ-bit output from the individual note length coincidence discrimination circuit 16. Demonstration based on *1
IIIT and performance error time ΔT are determined, and data D(T) and T(6 pieces) representing these are output.

Next, the similarity calculation circuit 23 calculates the -sound ^ number data D (Nl), the matching note length number data D (N2), the pitch similar note string set number data D (N3), and the note length similar note string set. Number data D
(N4), reference pitch number data D (N5), 1! Based on the quasi-note length number data D (N6), the model performance time data D (T), and the performance time difference data D (ΔT), similarity score data [)SCore is obtained, and this is passed to the similarity discrimination circuit 2.
Promote to 4.

The calculation formula for determining the similarity score X based on each of the above data is as follows.

×-([100×(N1/>N5)×(N2/N6)×
(D-1ΔTl)/T]-Y)÷5×5Here, N1
;-Number of note lengths 17- N2; Number of matching note lengths N3; Number of similar note sequence sets N4; Number of note length similar note sequence sets N5; Reference number of pitches N6; Reference number of note lengths ■: Model performance time ΔT; Performance time difference Y; N3 or N4 In this way, in this similarity score calculation method, when sifting the similarity scores, the main scoring i elements are the monotone pitch N1 and the number of matching note lengths N2. Even if there are errors in the input melody that are resung during the singing, these will not greatly affect the final score, and the similarity score is very close to the music teacher's feeling. You can get it.

The similarity discrimination circuit 24 discriminates the similarity score data [)SCore based on a predetermined standard score (for example, 80 points), and determines mra only when the value of this similarity score data exceeds the standard 80 points. Confirmation signal 18--Outputs "1" as SOk.

Door opening/closing control circuit 26. The lighting control circuit 27 and the bath water filling control circuit 28 respectively open and close the door.

It controls lighting control and bath water injection.
Each of these circuits receives the aforementioned melody designation signal 35eie.
It is alternatively enabled depending on the contents of ct-1 to 35elect-4.

Then, in the state where each control circuit is enabled, only when the melody recognition confirmation signal SOk is output, each control circuit executes the required operation, and upon completion of the execution, outputs the completion signal 5OVer.

As a result, as described above, for example, when a melody corresponding to a predetermined door opening/closing command is sung from the microphone 4, the sung melody is stored in the reference pitch tone string generation circuit 1 and the reference note length tone string generation circuit 12. If it is recognized as sufficiently similar to each sound string data,
The door opening/closing control circuit executes the required operations, thereby automatically opening/closing the door.

In response, the control circuit 25 sends a melody designation signal 5s.
In the state where select-3 or 5select-4 is output, if a melody corresponding to a lighting control command or bath water injection control command is input from the microphone 4 or an electronic musical instrument, a melody confirmation signal 3ok"1 is output.
In response to this, the lighting control circuit 27 or the bath water filling control circuit 28 performs the respective required operations, whereby lighting control of the lighting or water injection control for the bath is automatically performed.

Next, the basic operation flow of the present apparatus described above will be explained in detail with reference to the drawings from FIG. 2 onwards.

Since the operation of this device is controlled by various control signals output from the control circuit 25, the detailed configuration of the control circuit 25 will be explained first with reference to FIGS. 20 to 24.

As shown in FIG. 20, the control circuit 25 is configured to selectively designate any one of first to second similar pitch tone sequence detection circuits or first to second similar note length tone sequence detection circuits, which will be described later. A similar film designation signal generation circuit 2510 that outputs a similar stage designation signal 3ssl -1 to -(n+1) and a cyst register in the designated similar pitch tone string detection circuit and similar note length tone string detection circuit are incremented MIl. Shift signal 5shif for
The input pitch tone string data and the input note length tone string data are latched by the shift signal generation circuit 2520 that outputs t and the shift registers in the specified similar pitch tone string detection circuit and similar note length tone string detection circuit. latch signal S for
In response to each rising edge of each similar stage designation signal 55m1-1 to -(gate+1) outputted from the latch signal generation circuit 2530 that outputs 1ch and the similar film designation signal generation circuit 2510, a minute width "1" is generated. A load signal generation circuit 2540 that outputs a pulse, and the shift signal generation circuit 25
a selector for switching and supplying the shift signal 5shift outputted from the similar film designation signal generation circuit 2510 to each shift register in the similar pitch tone string detection circuit and the similar note length tone string detection circuit of the similar stage designated by the similar film designation signal generation circuit 2510; 25
50 and the latch signal S1atch outputted from the latch signal generation circuits 221-530 to the similar pitch tone string detection circuit and similar note length tone string detection circuit of each similar stage designated by the similar film designation signal generation circuit 2510. and a selector 2560 for switching and supplying.

The counter 2580 is reset every time the operation end signal 3 over arrives, and is controlled every time the determination end signal arrives. Designated signal 85elect -1~S
5elect -4.

Details of the similar film designation signal generation circuit 2510 are shown in FIG. As shown in the figure, a similar film designation signal generation circuit 2510
is set by the judgment enable signal 3 judge output from the input pitch sequence forming circuit 6 and reset by the n+2 bit output of the decoder 2515 described later, and the latch signal generation circuit 2530. A D-type flip 70 knob 2512 is forcibly reset by the output latch signal 31stch and configured to take in the d output in response to the rise of the output "1" of the 22-AND gate 2513, which will be described later. and the RS flip-flop 251
1 and the Q output of the above-mentioned flip-flop 2512.
an AND gate 2513 whose opening/closing is controlled by the output and which passes the clock signal Sφ output from the shift signal generation circuit 2520; and this AND gate 251.
A counter 2514 counts the pulses output from 3 and is reset by the n+, 2 pit output of a decoder 2515, which will be described later.
and a decoder 2515 for decoding the count output of , and each pit output of this decoder 2515 becomes a load signal 31°ad-1 to -(n+1), respectively.

Details of shift signal generation circuit 2520 are shown in FIG.

As shown in the figure, the shift signal generation circuit 2520 is connected to a decoder 2515 in the similar film designation signal generation circuit 2510.
OR gate 25 for taking the logical sum of the outputs of each pit of
21, a monomulti 2522 that outputs a minute width ``1'' pulse in response to the rise of the output ``1'' of this OR gate 2521, and a minute width ``1'' pulse output from this monomulti 2522 as a minute @@dt. a delay circuit 2523 for delaying the latch signal S I set by the "1" pulse output from the delay circuit 2523 and output from the latch signal generation circuit 2530.
The RS flip-flop 2524 is reset by atCh, and the opening r11 is controlled by the Q output of this RS flip 70 knob 2524, and the clock generator 25
and an AND gate 2526 for controlling the opening/closing of the clock signal Sφ output from the clock signal S5hif.
It becomes t.

Details of the latch signal generation circuit 2530 are shown in FIG.

As shown in the figure, the latch signal generation circuit 2530 receives the shift signal 5 output from the shift signal generation circuit 2520.
In addition to counting the number of shifts, a shift number counter 2531 that is reset by the monomer specific force Seq of a coincidence determination circuit 2533, which will be described later, and the reliable key press signal 3 push outputted from the input pitch string formation circuit 6 are counted. , the key press number counter 2532 that is reset by the determination end signal 3 end output from the similar film designation signal generation circuit 2510, the count value of the shift count counter 2531, and the count value of the key press number counter 2532. A match determination circuit 2533 is configured to determine a match, and a match signal from this match determination circuit 2533 is output as a latch signal S1atch.

As a result, when each of the circuits described above operates, the determination enable signal 3 is output as shown in the time chart of FIG.
When the "1" pulse arrives at the judge, one "1" pulse is first output to the load signal 5load-1, and then the delay time dt determined by the delay circuit 2523 is output.
After a delay of 1, the number of "1" pulses corresponding to the number of key presses is outputted to the shift signal 5shirt-1, and the load signals 5load-2 to -4 and the shift signal 5 are output in the same manner.
25- similar pulses are output from shift -2 to -4, and the final shift signal 3sh
After a predetermined number of pulse trains are output to ift-4, when the time required for a certain degree of similarity calculation has elapsed, a "1" pulse is finally output to the determination end signal 3 end.

Note that the time chart in FIG. 24 shows the case where n-3 is used in FIG. 19.

Next, details of each circuit shown in FIG. 1 will be explained along the flow of data processing while referring to each control signal explained above.

First, details of the sound detection circuit 5 are shown in FIG.

In the figure, a fundamental wave detection circuit 501 detects a fundamental wave included in an input audio signal or musical tone signal, converts it into a corresponding analog voltage, and outputs the same.

The A/D conversion circuit 2602 is the fundamental wave detection circuit 501.
Converts the analog voltage output from the converter into a digital signal.

The tone reference data generation circuit 503 outputs a series of reference pitch data consisting of C1, CI26-#, D1...Cn in parallel.

Permissible discrimination circuits 504-1 to 504-p compare the input pitch data outputted from the A/D conversion circuit 502 and each reference pitch data outputted from the pitch reference data generation circuit 503. 1'' is output to each terminal EQ only when the pitch difference corresponding to the comparison result falls within a predetermined tolerance range.

The gate circuits 505-1 to 505-p are controlled to open and close by the minus number specific power of each of the permissible discrimination circuits 504-1 to 504-1), thereby passing pitch data corresponding to 01 to Cn, respectively.

The OR gate 506 includes each of the gate circuits 505-1 to 505-5.
05-1), and that is, this OR gate 506 selectively outputs tone sequence data corresponding to the input audio signal.

The O gate 507 calculates the logical sum of each bit of the OR gate 506, and thereby detects that some tone string data has been input.

The monomulti 507 and 508 output a minute width 111 I+ in response to the output “1” of the OR gate 507,
This "1"' pulse becomes the key press timing signal Skon- shown in FIG. 1, and the output of the OR gate 506 becomes the input pitch data PD(in) shown in FIG.

Next, details of the input sound string forming circuit 6 are shown in FIG.

In the same figure, the RS flip 70 knob 6°1 is set by a minute width 111 II pulse output from a differentiating circuit 609, which will be described later, and reset by a determination end signal S end output from the control circuit 25.

The AND gate 602 includes the RS flip 70 knob 60.
Opening/closing is controlled by the Q output of No. 1, and a key press timing signal 5kon- outputted from the key press detection circuit 5 is passed.

The gate circuit 603 is the Q of the RS flip 70 tube 601.
Opening/closing is controlled by the output, which causes the key press detection circuit 5
The input pitch data PQ(in) output from the input pitch data PQ(in) is passed through.

The match determination circuit 604 includes a shift register 607, which will be described later.
The input pitch data PD (
in) and the input pitch data PD(in) output from the gate circuit 603. If it is determined that the two match, the EQ ratio output is "°1'". Ru.

The OR gate 605 takes the logical sum of the output of the AND gate 602 and the output of the coincidence determination circuit 604, and the output of this OR gate 605 is the reliable key press signal 3.
Push.

The monomulti 606 is output from the OR gate 605.
It outputs a very small width "1" pulse in response to the rising edge of "1", and the output of this monomulti 606 becomes the inverted pitch data acquisition signal "3 kon".

The shift register 607 responds to the inverted data acquisition signal 5kon'' output from the monomulti 606, inputs the input pitch data that has passed through the gate circuit 603 into its first stage, and sequentially inputs the input pitch data through each of the 29 stages. It is configured to shift to the center right.

The monomulti 608 is connected to 411 of the monomulti 606.
It is configured so that it can be triggered repeatedly in response to the rise of the #l output, and there is a relatively long “1” in the trigger trace.
'' pulse (for example, 2.5 seconds), and the output of this monomulti 608 becomes the playing signal 5play.

The differentiating circuit 609 outputs the output 111 of the monomulti 608 $
The differential circuit 609 is configured to output a minute width "1" pulse in response to a rise of "1", and the output of this differentiating circuit 609 becomes the determination enable signal Sjudge.

The OR gate 610 takes the logical sum of the judgment end signal 3 end and the initial clear signal 3 ic, and the output of this OR gate 610 determines the logical sum of the judgment end signal 3 end and the initial clear signal 3 ic.
All stages of 07 are cleared, and at the same time, the output of this OR gate 610 is output as a clear signal 5clr.

As a result, when each of the circuits explained above operates, the monomulti 60
6 is a key press signal 3 pu output from the OR gate 605
The shift register 607 receives the input pitch data output from the gate circuit 603 in response to the rise of the data acquisition signal inverted 5kon'' output from the monomulti 606. As shown in Figure 26 (A> and Figure 27 (A)),
If there is incorrect pitch data due to jitter etc. in the input original pitch data, these input pitch data will not be taken into the shift register 607, and as shown in FIG. As shown in (B), incorrect input pitch data is removed from each stage of the shift register 607, and only pitch data corresponding to a note length that is long enough as a musical element is taken in. It is. Then, the pitch data taken into each stage is output in parallel, and this parallel output is the input pitch string data P D
1ine (full).

Next, details of the input note length string forming circuit 9 are shown in FIG.

In the figure, an OR gate 901 receives a determination end signal S
Outputs the logical sum of end and initial clear signal 3ic.

The OR gate 902 outputs the logical sum of the initial clear signal Sic and the determination enable signal 3 judQf3.

The differentiating circuit 903 outputs a minute width "1" pulse in response to the rise of the data acquisition signal 9kon" output from the pitch string forming circuit 6.

The AND gate 904 is controlled to open and close by the playing signal 81) laV, and allows the minute width 111 II pulse output from the differentiation circuit 903 to pass therethrough.

RS flip 70 knob 905 is data acquisition signal 3k
It is set at the rising edge of "On" and reset at the output of the OR gate 902.

The tempo clock oscillator 906 has a predetermined period (for example, 1
00es, 500μS) tempo clock is output.

In this example, the frequency is configured to be variably controllable.

A counter 907 is enabled by the Q output of the RS flip 70 knob 905 and counts the tempo clock TCL output from the tempo clock oscillator 906. Furthermore, it is repeatedly reset by the minute width "1" output of the AND gate 904 which is delayed via the loop circuit 908.

The shift register 908 is shift-controlled by the minute width "1" pulse output from the AND gate 904, and is configured to input the counting output of the counter 907 as note length data to the first stage. , and this shift register 908 is connected to the OR gate 90.
Cleared by outputting 1.

When the above-described circuits operate, each stage of the shift register 908 stores data in each stage of the shift register 607 in the input pitch string forming circuit 6, as shown in FIG. The note length data corresponding to the note pitch data are sequentially imported, and the parallel output data of each stage is the input note length note string data LDli.
αe (full).

33- Next, details of the similar sound sequence extraction port 117 are shown in FIG. As shown in the figure, the similar sound roar sequence extraction circuit 7 inputs input pitch sequence data P D li, ne (full
) and reference tone ^ tone sequence data P [) 1ine (ref
) are supplied in parallel.
kiil similar sound^ sound sequence detection circuit 700-1~
700-k.

As described above, the first to kll-like high pitch string detection circuits 700-1 to 700-- have input pitch string data PDline (
full) and reference sound roar sequence data p[) 1ine
(ref) existing in the same time zone while shifting the reference points of each other's casting axes, as if trains were passing each other. This method detects the pitch tone string data portions that are similar to the 2nd to klith pitches, and each of these detected similar tone string portions is 1
As shown in FIGS. 127 (C) to (J), first to second similar sound roar sequence partial data PD1inesml-1 to PD1inesml-1 representing the number of shifts to the input pitch sequence data are output.

Further, each similar pitch tone sequence detection circuit 700-1 to . From 734-00-k, a number tone indicating how many tones that match the reference pitch tone string data are present in each detected similar tone tone string partial data p Q 1ine 5g1l -1 to -k. This means that the high school data PDeq-1 to -k are output.

Next, the details of the first similar sound string detection circuit 700-1 are shown in FIG. In the same figure, the shift register 701 is
In response to the rise of the load signal 5load-1, input pitch string data PDline (full) output from the input pitch string forming circuit 6 is loaded. Further, in response to the "1" pulse included in the shift signal 5shift-1, the shift control is performed to the left in the figure, and the contents of all stages are simultaneously cleared by the clear signal 3clr.

The parallel pitch coincidence determination circuit 702 is configured to perform the shift register 7.
The pitch string partial data p□+the(1 to 7) output in parallel from the first to seventh stages of 01 and the reference tone treble string data P output from the reference tone treble string generation circuit 1.
D 1ine (ref), and 1° +, i + □ corresponding to each determination result.
Output signals to terminals EQI to EQ7.

The sound^-politics/religion detection circuit 703 detects the number of coincidences between the two based on the respective -number ratios EQ1 to EQ7 of the parallel sound^ coincidence determination circuit 702, and outputs the corresponding sound^-politics/religion data.

The counter 704 is reset by the determination enable signal Sjudge, counts the number of "1" pulses included in the shift signal 3shift, and thereby outputs shift number data.

The match number comparison circuit 705 compares the pitch match number data latched by a latch circuit 707, which will be described later, with the pitch match number data output from the pitch-political religion detection circuit 703, and compares the pitch match number data output from the pitch-political religion detection circuit 703. Only when the pitch matching number data output from the detection circuit 703 is larger than the pitch matching number data latched by the latch circuit 707, "1" is output.

The AND gate 706 is controlled to open and close by the output of the -politics-religion comparison circuit 705, and thereby the shift signal 3sh
Ift -1 is passed.

The latch circuit 707 receives a judgment enable signal Sjudge.
and the AND gate 70
In response to the "1" pulse output from the tone ^-political/religious detection circuit 703, the pitch matching number data output from the tone ^-political/religious detection circuit 703 is latched.

Similarly, the latch circuit 708 receives the determination enable signal Sju
The counter 7 is reset by dge, and each time a “1” pulse is output from the AND gate 706
Latch the shift number data output from 04.

The latch circuit 709 is reset by the clear signal 5clr, and the latch circuit 709 is reset by the clear signal 5clr.
Every time a 1'' pulse arrives, the pitch matching number data latched in the latch circuit 707 is latched.

The latch circuit 710 is similarly reset by the clear signal S ctr, and is also reset by the latch signal 5latch.
Each time a 1° pulse arrives in the latch circuit 7,
The shift count data latched at 08 is latched.

37- Then, the output of the latch circuit 709 becomes the -sankyo data PDeq-1, and the output of the latch circuit 710 becomes the shift number data p[) 1ine ssl-1.

As a result, when each of the circuits described above operates normally, the shift register 701 operates as shown in FIGS.
), the shift control is performed sequentially, and in the parallel sound matching determination circuit 702, the data PDline (ref) and the data p[
) line (1 to 7) is determined.

In the example of FIG. 28, the pitch string partial data p□1ine (1 to 7) corresponding to 0 shifts is detected as the most similar pitch string partial data, and as a result, the data output from the latch circuit 710 P[) 1ine ssl
The content of -1 becomes rOJ, and the data PDeq-1
The content of is "3".

Next, details of the similar pitch sequence detection circuit 700- are shown in FIG. In the figure, the shift register 751 loads the 38-input pitch string data P D 1ine (full) in response to the load signal 5load-k, and also loads the "11"' pulse included in the shift signal 5shift-k. Shift controlled in response,
Furthermore, the contents of all stages are simultaneously cleared by the clear signal 5clr. Also, the first shift register 751
~Each data in the seventh stage is configured to be individually resettable.

Next, parallel pitch matching determination circuit 752. Pitch matching number detection circuit 753. - Politics and Religion Comparison Circuit 755. and gate 756
．． Counter 754. Latch circuit 759, latch circuit 76
Each operation of 0 is completely the same as that of the corresponding circuit in the first similar pitch sequence detection circuit 700-1, and will not be repeatedly described here.

Matching determination circuits 761-1 to 761- (k-1> are the shift number data output from the counter 754, respectively, and similar pitch sequence detection circuits 700-1 to 700 at the previous stage)
-(k-1>) shift number data P [
) 1ine s*I -i 〜PD 1iness
1-(k-1), and outputs "1" only when it is determined that both data match.

The OR gate 762 includes each of the matching determination circuits 761-1 to 761-1.
761- Outputs the logical sum of the outputs of (k-1).

The gate circuit 763 is controlled to open and close by the output from the OR gate 762, and thereby transfers each negative number ratio power of the parallel pitch matching determination circuit 752 to the first one of the shift register 751.
~Supplied to the reset terminal R of the seventh stage.

AND gate 764 is controlled to open and close by the output of OR gate 762 which is inverted by inverter 765, thereby inhibiting the output of AND gate 756.

The latch circuit 757 receives a judgment enable signal Sjudge.
and the AND gate 7
In response to the "1" pulse passing through 64, the pitch-religion data output from the tone-religion detection circuit 753 is latched.

The latch circuit 758 similarly receives the determination enable signal 3 J
In response to the "1" pulse reset by the AND gate 764 and passed through the AND gate 764, the shift number data output from the counter 754 is latched.

As a result, when each of the above circuits operates, Fig. 27(B)
As shown in (J), the input pitch data stored in each stage of the shift register 751 is sequentially shifted to the left in the figure in response to the 111 I+ pulse of the shift signal 5shift-, and at the same time the input pitch data is shifted to the left in the figure in response to the 111 I+ pulse of the shift signal 5shift-, and at the same time, In the coincidence determination circuit 752, as shown in FIGS. 29 and 30, a process of determining coincidence between the partial pitch string data for each number of shifts and the reference pitch string data is performed.

Here, FIGS. 28, 29, and 30 are the first
．． Second. Third similar high pitch string detection circuit 700-1.700
-2.Each parallel pitch matching determination circuit 70 in 700-3
2.752 operation.

As is clear from these figures, in each of the parallel pitch matching circuits 41 to 752 in the second to second similar tone high pitch string detection circuits, the similar pitch matching circuits 752 that have already been detected in the previous stage similar tone roar string detection circuit Sound row partial data p0
When the shift timing for 1ine (1 to 7) arrives,
Reference pitch sequence data included in these data P D
1ine (ref >) are individually reset by the output of the gate circuit 763.

As a result, the portion that matches the reference pitch string data included in the already detected similar pitch tone string partial data will not be recognized as another similar tone roar string partial data again.

Furthermore, when the shift timing of the similar tone pitch string partial data already detected in each preceding stage similar tone pitch string detection circuit arrives, the AND gate 764 is inhibited by the output of the OR gate 762, and therefore the similar tone string In the detection circuit 700-k, the similarity is always one step lower than the similar tone string partial data PDline (1 to 7) that has already been detected in the first to eleventh similar pitch tone string detection circuits. 4 because the data will be detected.
2-.

Next, the details of the -number part valve type matching circuit 8 are shown in FIG. In the same figure, a shift register 801 receives a load signal 5.
Load the input pitch string data P D 1ine (full) in response to load −(k +1 >).

Then, the shift signal 5shift −(k + 1 >
In response to the 1" pulse included in the 1" pulse, the stages are shifted downward in the figure, and the contents of each stage are simultaneously cleared by the clear signal 3clr.

The parallel pitch coincidence determination circuit 802 receives pitch string partial data pDline (1 to 7) output in parallel from the first to seventh stages of the shift register and the pitch string partial data pDline (1 to 7) output from the reference tone pitch string generation circuit 1. Reference pitch pitch sequence data P D 1
ine (ref) for each stage, and outputs the result of the determination to terminals EQI to EQ7 as a 1-bit signal for each stage.

The counter 803 is reset by the judgment enable signal Sjudge and also by the shift signal 5s.
shift - (k + 1) is counted, thereby outputting shift count data corresponding to the shift count of the shift register 801.

The match determination circuits 804-1 to 804- each receive the shift number data output from the counter 803 and the shift number data P D 1ine output from each of the similar tone pitch sequence detection circuits 700-1 to 700-k. ssl
-1 to P D 1ine ssl -k, and only when it is determined that they match, it returns 1"
Output.

The OR gate 805 includes each of the match determination circuits 804-1 to 804-1.
The output of OR gate 804- is logically summed, and the output of this OR gate 805 controls data superimposition processing, which will be described later.

The AND gates 806-1 to 806-7 are controlled to open and close by the output of the OR gate 805, and each output of the parallel pitch matching determination circuit passes through them.

The latch circuits 807-1 to 807-7 are latch-controlled by the respective minus number ratio powers of the parallel pitch coincidence determination circuit 802 supplied via AND gates 806-1 to 806-7, respectively, and thereby the shift Among the outputs from the first stage to the seventh stage outputted from the register 801, the reference pitch pitch sequence data PQ 1ine (raf
) is latched at each shift. And these latch circuits 807-1 to 807
Each series of pitch data latched to -7 constitutes the most similar sound roar sequence data p[) sample.

As a result, when each of the circuits described above operates, as shown in FIG.
From each similar tone string data detected in 1 to 700-k, only the parts that match the reference pitch tone string data are extracted, and by superimposing them on each other,
The most similar sound roar sequence data p[) 6ample is formed.

Next, the details of the similar note length string extracting circuit 10 are shown in FIG.

In the same figure, the first to kw4th analog long tone string detection circuit 10
The configuration of 00-1 to 1000- includes the first to second similar pitch series detection circuits 700-1 to 700 shown in FIG.
The configuration is almost the same as that of 45-, that is, each similar note length tone string detecting circuit calculates the number of shifts corresponding to a plurality of similar note length tone string partial data selected in order from the one with the highest degree of similarity. Data l Q 1
ine ssl -1~L D 1ine ssl -
k, and note length-political/religious data LDeQ-1 to LDeq-k are output.

Details of the 1111th similar note length string detection circuit 1000-1 are shown in FIG. In the same figure, shift register 1oo
i, note length-political/religious detection circuit 1003. Counter 1004.
- Politics and Religion Comparison Circuit 1005. ANDGATE 1006. Latch circuit 1007. latch circuit i oos, latch circuit 1
009. The configuration of the latch circuit 1010 is the same as the first similar sound roar sequence detection circuit 700-1 shown in FIG. 6 above, except that the data handled is simply changed from pitch data to note length data. Therefore, the explanation will not be repeated here.

On the other hand, regarding the configuration of the parallel note length match determination circuit 1002, the parallel note length match determination circuit 70 shown in FIG.
The configuration is slightly different from 2. In other words, as shown in FIG. 25(C), the reference note length tone string data L D 1i output from the reference note length tone string generation circuit 10
For each constituent note of ne (ref), exactly an eighth note, a quarter note, and a scored quarter note, respectively.

It has a fixed standard length, such as a half note.

On the other hand, the note length string partial data output from the first to seventh stages of the shift register 701 have errors Δ1 to Δ9, respectively, as shown in FIGS. 31(B) to (J). Here, for convenience of explanation, the values of errors Δ1 to Δ9 attached to each note are assumed to be errors within a predetermined tolerance range that can be recognized as each corresponding note.

Therefore, when the parallel pitch match determining circuit 702 accurately determines the match between the corresponding stages in both tone string data, it is almost impossible for both data to completely match.

Therefore, in the parallel note length match determination circuit 1002, when comparing the note sequence data L D 1ine (ref) and the note sequence data PDline (1 to 7), the error with the reference note length data is determined for each stage by a predetermined value. If it is within the allowable range, it is considered to be a match. In other words, terminals EQI to EQ7 have terminals A1 to EQ7, respectively.
A matching output is obtained only when the difference between A7 and 81 to B7 is within a predetermined tolerance range.

As a result, when each circuit shown in FIG. 10 operates, in the shift register 1001, FIGS.
), each stage is shifted, and at the same time, in the parallel note length coincidence determination circuit 1002, as shown in FIG.
e (ref) and note length string partial data pi) line
(1 to 7) are compared in magnitude, and it is determined for each stage whether the length of each note is within a predetermined allowable range.

Then, the note length/religion detection circuit 1003 outputs numerical data corresponding to the number of each -sound, and at the same time the counter 10
04, the number of shifts is outputted, and finally the latch circuit 1009 and the latch circuit 1010 output the number of matching notes and the number of shifts corresponding to the note length note string partial data with the largest number of matching notes, DeQ -1, L D 1ine
5ill -1 is the output sale.

Next, details of the similar note length tone sequence detection circuit 1000- are shown in FIG. In the figure, shift register 105
1. Note length-religion detection circuit 1053° counter 1054.
- Politics and Religion Comparison Circuit 1055. ANDGATE 1056. Latch circuit 1057. Latch circuit 1058. Latch times vs1
059. Latch circuit 1060. Match determination circuit 1061-
1-1061-(k-1), ORGATE 1062. Gate circuit 1063. ANDGATE 1064. Inverter 1
The configuration of 065 is the same as that of the similar pitch sequence detection circuit 700- shown in FIG. 7, except that the type of data handled has changed from pitch data to note length data. The configuration of the note length match determination circuit 1052 is the same as that of the first similar note length tone sequence detection circuit 1000-1, so a repeated explanation will be avoided here.

As a result, when each of the circuits described above operates, 49-, for example, the second . Third similar note length string detection circuit 1000-2.
In 1000-3, as shown in FIGS. 33 and 34, a match determination operation is performed between the reference note length string data LDline (ref) and the note length string partial data PDline (1 to 7). Similar to the case of the pitch data described above, the already selected similar tone string data is excluded from the matching determination target.

Then, the second similar note length string detection circuit 1000-2 extracts the note length string partial data with the second largest number of matching notes, and the third similar note length string detection circuit 1000-3 extracts the third note length string partial data. Similar note length note string partial data is extracted.

Then, each extracted note length note string partial data is converted into input note length note string data Q 1ine (full)
LDlines
al -1 to Dline sat -k, and the number of matching tones in each tone string data is determined by LDeQ
-1 to LDeq-.

Next, details of the matching length part valve type matching circuit 11 are shown in FIG. 150-2. In the figure, shift register 1101.
Counter 1103. Match determination circuits 1104-1 to 110
4-, Aogate 1105. ANDGATE 1106-
1 to 1106-7, latch circuits 1107-1 to 1107
-7 is exactly the same as the single-note high-valve combination circuit 8 shown in FIG. 8 above, except that the handled data is changed from pitch data to note length data, and parallel Since the configuration of the coincidence determination circuit 1102 is the same as the corresponding circuit in each of the similar note length string detection circuits 1000-1 to 1000-, repeated explanation will be avoided here.

As a result, when the circuit described above operates, in the latch circuits 1107-1 to 1107-7, as shown in FIG. Second. Each -several note in the note length string partial data detected by the third similar note length string detection circuit is superimposed on each other, so that the most similar note length string data L
D sample will be formed.

Next, details of the performance time/time difference detection circuit 22 are shown in FIG. In the same figure, a selector 2201 selects between the most similar pitch tone sequence data P D sagiple output from the individual pitch coincidence determination circuit 14 and the reference tone tone sequence data P D
1ine (ref) corresponding to the number of notes that match the pitch-political data PDeq, each stage can be individually switched and controlled, and therefore the output terminals 0UT1 to
The contents of each stage of either data p[) Sample or data P D 1ine (rer) are output to 0UT7.

The gate circuit 2204 is controlled to open and close by a latch signal 31atch, and determines the number of notes in which the most similar note length data L D sample outputted from the individual note length coincidence discrimination circuit 16 and the reference note length note string data L D 1ine (ret) match. The note length-political and religious data LDec+ corresponding to is passed.

The or gates 2202-1 to 2202-7 produce the above sound ^-
It outputs the logical sum of political and religious data PDeq and note length-political and religious data LDeQ for each stage, and the latch circuits 2203-1 to 2203-1 to be described later are output from these OR gates.
2203-7 is latch controlled.

The latch circuits 2203-1 to 2203-7 are specially latch-controlled by the outputs of the OR gates 2202-1 to 2202-7.
The note length data output from each output 0UT1 to 0UT7 of 2 is latched.

The adder circuit 2205 includes the latch circuits 22o3-1 to 22o3-2.
Add all the note length data latched to 203-7,
Output these sums.

The addition circuit 2206 adds the reference note length tone string data L D
All the note length data constituting 1ine (ref) are added, and the total sum is output. The output of this adder circuit 2206 becomes the aforementioned performance time data D (T).

A subtraction absolute value circuit 2207 calculates the difference between the addition data output from the addition circuit 2205 and the addition data output from the addition circuit 2206, and further outputs the absolute value, and this absolute value is the performance time difference. Data will be D (6 guns).

As a result, when each circuit described above operates, in the 53-latch circuits 2203-1 to 2203-7,
As shown in FIG. 37, the most similar pitch sequence data P D s
If the note length is incorrect in the part where the note ^ is incorrect in the sample, and the note length is also incorrect in the most similar note length note string data L This will be corrected using the correct note length data present at the corresponding location in the column data.

In other words, pitch is particularly important in melody recognition, and therefore, the note length of a note pressed with an incorrect pitch is considered to be the correct note length in the note length judgment, and thereby the note length of a note pressed with the wrong pitch is treated as the correct note length. This means that if the note length is determined only for key tones, the recognition will be more accurate and similar to that of a music teacher.

Next, details of the pitch similar tone sequence number detection circuit 18 are shown in FIG. In the figure, a zero data generation circuit 1801 outputs two pieces of zero data in parallel.

The coincidence determination circuit 1802 is configured to match the zero data generation circuit 180.
It is determined individually whether the zero data outputted in parallel from 1 and the 54-tone ^-political/religious data PDeq-1 to PDeQ- outputted from the similar pitch sequence extraction circuit 7 is matched, and each The determination result is output to terminals EQI to EQk as a 1-pit signal.

The mismatch number detection circuit 1803 includes the match determination circuit 180.
Detect the number of "0" signals output from EQ1 to EQk for each pair of pitches EQ1 to EQk, and use this as pitch similar sound sequence set number data D (N3).
Output as .

Next, the details of the note length similar tone sequence detection circuit 19 are shown in FIG. In the figure, a zero data generation circuit 1901 outputs two pieces of zero data in parallel.

The coincidence determination circuit 1902 is configured to match the zero data generation circuit 190.
zero data output from 1 in parallel, and note length - output in parallel from the similar note length note string extraction circuit 10.
It individually determines whether the political/religious data LDeQ-1~eq- match, and outputs each determination result to terminals EQ1~EQk as a 1-bit signal.

'The mismatch number detection circuit 1903 is the match determining circuit 1.
The number of signals "O" outputted from 902 is detected and outputted as note length similar sound sequence set number data D (N4).

Next, the details of the reference tone antinomial detection circuit 20 are shown in FIG.

In the figure, a zero data generation circuit 2001 outputs 14 pieces of zero data in parallel.

The coincidence determination circuit 2002 is a match determination circuit 2002
14 zero data output from 1 and the reference tone treble string data P D output from the reference tone treble string generation circuit 1.
1ine (ref) for each stage individually, and the determination results are sent to terminals EQ1 to EQ14.
Outputs in parallel depending on pit signal.

In the mismatch number detection circuit 2003, the match determination circuit 20
The number of °°0'' is detected from each output outputted from 02, and this is output as the reference tone ^ number data D (N5).

Next, the details of the reference code length number detection circuit 21 are shown in FIG. In the figure, a zero data generation circuit 2101 outputs 14 pieces of zero data in parallel.

The coincidence determination circuit 2102 is connected to the zero data generation circuit 210.
14 zero data output from 1, and the reference note long tone string data L output from the reference note long tone string generation circuit 12.
D1ine (ref) is individually determined for each stage, and the determination results are sent to terminals EQ1 to EQI4.
1-bit signal is output in parallel.

The mismatch number detection circuit 21 determines whether "
Detect the number of 0'' and use it as the reference mark length number data D (N6)
Output as .

Next, in the similarity calculation circuit 23, the one-tone ^ number data D (Nl) obtained in this way, the matching note length song data D (
N2), pitch similar note string set number data D (N3), note length similar note string phase number data D (N4).

Standard note number data D (N5), standard note length number data D (
N6), the similarity score for the melody singing performed by the singer is determined based on the performance time difference data D (ΔT) and the performance opening data D (T) using the following calculation formula.

Score X - ([100x (N1/N5)x (N2/N
6) ;Performance time difference Y: Number of pitch-similar tone string sets N3 or note length-similar note string groups N4 Next, details of the door opening/closing control circuit 26 are shown in FIGS. 18 and 19.

AND gate 2608 is the melody designation signal 3 mentioned above.
Opening/closing is controlled by select-1, thereby allowing the melody recognition confirmation signal Sok to pass.

ANDGATE 2609 is the melody specified message @
Opening/closing is controlled by S5elect-2, thereby allowing the melody recognition confirmation signal Sok to pass through.

On the other hand, as shown in FIG. 19, the rotational force of motor M is
The rotational speed is transmitted to the driven gear 2602 via the gear ratio 2601, so that the screw ring 2610 identified in the driven gear 26o2 is supported between the bearings 2604 and 2605 and rotates.

Then, the boss 2603 that meshes with the screw ring 2610 performs a linear movement in the vertical direction in the figure according to the rotational direction of the motor M, and thereby the door 2607 is moved via the lever 2606.
will be driven to open and close.

In addition, the moving distance of the boss 2603 is determined by the limit switch L.
The boss 2603 is regulated by S1 or S2, and the boss 2603 reciprocates between these limit switches.

In the above configuration, if a voice melody corresponding to the sound of opening the door is input to microphone 4, first the melody designation signal sse+ect-i is "1".
'', and as a result, the AND gate 2608 becomes open.

In this state, if the input voice melody is sufficiently similar to the melody corresponding to opening a predetermined door, the melody recognition confirmation signal Sok becomes °゛1'' and the ant game t-2608 The AND condition is satisfied,
Flip 70 knobs FF1 and FF2 are set. As a result, relay R1 is driven and switch R1-1 is turned on, relay R1 is driven and switches R2-1 and R2 are turned on.
-2 is switched to the normal rotation terminal side, so the motor M starts rotating in the normal rotation mode, and the switch LS2 is turned off.

When the silence has finished opening, limit switch LS1 is turned on,
Since the flip-flop FF1 is reset, the switch R1-1 is turned off and the motor is stopped.

Also, due to the limit switch LSI being turned on,
Flip-flop FF2 is also reset and the motor enters the reverse rotation mode.

Next, when the timer T triggered by turning on the switch LSI reaches a predetermined time, it generates an output pulse, thereby setting the flip 70-tube FFI. As a result, switch R1-1 turns on and the motor starts rotating in reverse mode. Switch LS1 is turned off.

Next, when the cap is tightened, the limit switch LS2
is turned on, flip-flop FF1 is reset, switch R1-1 is turned off, and the motor is stopped. stop signal 5
over is output.

Note that when "Stop in a quiet manner with the gate open" is recognized and the AND condition of the AND gate 2609 is satisfied, the one shot is driven, the switch R1-1 is turned off for a certain period of time, and the motor is stopped.

Further, the lighting control circuit 27. The operation within the m8 water injection circuit 28 is also substantially the same, so a repeated explanation will be avoided.

Thus, according to the melody mll device shown in this embodiment, a predetermined melody, which has been determined in advance in accordance with each required operation, is transmitted from the microphone 4 or the input terminal 2 via the output signal of the electronic keyboard instrument, etc. When you enter it,
Only when the similarity with the corresponding melody is sufficiently similar, the door opening/closing, lighting control, bath water filling control, etc. are automatically performed, and as a result, the voice or musical instrument, etc. It is possible to remotely control various control objects 61- through.

In particular, in this embodiment, the premise for determining whether or not an input melody is similar to a reference melody is that the input melody is sufficiently similar to the reference melody. Since the phrases that are considered to be recognized are extracted, for example, even if there is an incorrect part in the middle of the input melody, it can be corrected by re-singing it. In this case, the erroneous part is removed from the similarity score, and it is possible to obtain a scoring result that is extremely close to the senses of a music teacher (human).

Therefore, even if the melody is not necessarily input accurately through the microphone 4, as long as the melody is input within a certain range of accuracy, the desired movement can be performed reliably. Even people who cannot pronounce strictly accurate melodies, such as people with disabilities, can reliably perform remote control using this type of voice or musical instrument.

Furthermore, in the above embodiments, the key press detection circuit 5 and the pitch data detection circuit 26 are shown as converting each input note into two-dimensional data consisting of pitch and note length. , the application of the present invention is not limited to this, but for example to the pitches described above.

Of course, in addition to the note length, detection may be performed by converting the accent dimension, timbre dimension, or volume dimension, etc.

For example, to give a concrete explanation using tones as an example, the first
The tone sequence pattern of the reference tone high pitch sequence generating circuit 1 shown in the figure and the tone sequence pattern sequentially input to the shift register 607 shown in FIG. , piano, violin, trumpet, etc., may be configured in that timbre sequence pattern in that order.

That is, what is input from the shift register 607 is digital pitch information, but the formant data that each constituent note of the tone string has is inputted from the shift register 607.
It is only necessary to configure it so that it is input to 07,
More specifically, each constituent tone of the tone sequence is sequentially filtered through a multiband filter (for example, with a center frequency of 500 Hz).
, 1000Hz, 30001-1z) and convert the analog voltage level output from each bandpass filter into digital information, and if the values are r4J, r3J, r2J, for example, All that is required is to input the digital data into the shift register sequentially for each constituent sound.

A device configured in this manner can be used, for example, for training in timing alignment of each musical instrument sound by an orchestra.

Furthermore, using the same idea as this type of thing, we extracted the relatively leading 7 ormantos of cat's cry, loud cry, small bird's cry, etc., and analyzed them by tone. Pattern extraction can be easily realized in a similar manner.

Furthermore, regarding the volume, each volume level of each constituent sound of the tone sequence is converted into digital information.)
This can be easily realized by configuring the data to be converted and input to the shift register 607.

With the device configured in this way, as the song progresses,
It is suitable for use in training that requires emotion.

Furthermore, when configuring a harmony or rhythm practice device, it may be done as follows. In other words, each detected performance information is filtered through a plurality of bandpass filters similar to those described above, each of which is numbered in parallel, and the constituent note data of each chord is thereby converted into digital data. All you have to do is input the digital data into the shift register.

Furthermore, when constructing an accent practice device, it is sufficient to similarly detect dynamics data from inputted performance data, convert this into digital data, and then input it to the shift register.

In this way, in addition to the above-mentioned timbre, it is suitable for use in training that requires a greater sense of emotion.

65- Furthermore, in the above embodiment, for notes whose pitch is already incorrect in the two-time difference detection circuit 22 during performance,
The note length is considered to be correct, and the score is based on the unique characteristics of the melody, i.e., the note ^.On the other hand, for notes with the wrong note length,
Of course, if the music is configured so that the sound ^ is regarded as correct, it is possible to obtain performance scoring results that place emphasis on rhythm.

As is clear from the description of the embodiments above, according to the first and second inventions related to this application, information inputted by the sound of a musical instrument or a voice can be recognized with a sense similar to that of a music teacher. It is possible to remotely control a desired object based on the method, and its configuration is relatively simple compared to conventional voice recognition methods that control objects through spoken words, and this type of remote control is widely used. This makes it possible to apply it to control.
, 5shi-4, Brief explanation of the drawings
' □Figure 1 is a block diagram showing the overall configuration of the melody recognition @ stomach including the first and second inventions related to this application, Figure 2 is a block diagram showing details of the pitch detection circuit, and Figure 2 is a block diagram showing details of the pitch detection circuit. Figure 3 is a block diagram showing details of the input pitch string forming circuit, Figure 4 is a block diagram showing details of the input note length string forming circuit, and Figure 5 is a block diagram showing details of the similar tone string extraction circuit. , Fig. 6 is a block diagram showing details of the 111th similar tone high tone sequence detection circuit, Fig. 7 is a block diagram showing details of the @similar tone high tone sequence detection circuit, and Fig. 8 is a block diagram showing details of the 111th similar sound high tone sequence detection circuit. FIG. 9 is a block diagram showing details of the similar note duration string extraction circuit; FIG. 10 is a block diagram showing details of the 1111i similar note length string detection circuit; FIG. 11 is a block diagram showing details of the similar note length string detection circuit. Figure 12 is a block diagram showing details of the matching note length part valve type matching circuit; Figure 13 is a block diagram showing details of the performance time/time difference detection circuit. , 14!11 is a block diagram showing the details of the circuit for detecting the number of note length similar sound sequences, Fig. 15 is a block diagram showing the details of the circuit for detecting the number of note length similar sound sequences, and Fig. 16 is the reference note ^ number structure. A block diagram showing the details of the circuit, Fig. 17 is a block diagram showing the details of the reference code length number composition circuit, Fig. 18 is a sequence circuit diagram showing the details of the door opening/closing control circuit, and Fig. 19 is a door opening/closing drive device. FIG. 20 is a block diagram showing details of the control circuit, FIG. 2111 is a block diagram showing details of the similar film designation signal generation circuit, FIG. 22 is a block diagram showing details of the shift signal generation circuit, and FIG. FIG. 23 is a block diagram showing details of the latch signal generation circuit, FIG. 24 is a time chart showing the status of each control signal output from the control circuit 25, and FIG. 25 is an example of each tone sequence data corresponding to the reference melody. FIG. 26 is a time chart showing the states of various timing signals corresponding to the input melody, and FIGS. 27 to 30! l
is the input sound roar column forming circuit.

Explanatory diagrams showing details of tone sequence data processing performed by the similar pitch tone sequence extraction circuit, Figures 31 to 341%! 1 is an explanatory diagram showing the flow of tone sequence data processing performed in the input note length forming circuit and the similar note length note sequence extraction circuit, and FIG. 36 is an explanatory diagram showing the flow of note string data processing performed in the matched note length part valve type matching circuit, and FIG. 37 is an explanatory diagram showing the note length data correction performed in the performance time/time difference detection circuit. It is an explanatory diagram showing a flow of processing.

1.......Reference tone high tone string generation circuit 2...
...... Input terminal 4 ...... Microphone 5 ...... Pitch detection circuit 6 ...... Input pitch pitch string formation four circuits 7・・・・・・
...Similar pitch sequence detection circuit 8...
- Discrete high part valve type matching circuit 9...Input note length string formation circuit 10...Similar note length string extraction path 1
1... Matching code length part ■ Matching circuit 12...
・Reference note length tone string generation circuit 13...Performance result table holo-14...Individual pitch matching discrimination circuit 15...
・-Number sequence number detection circuit 16...Individual note length match discrimination circuit 17...
- Matching note length song detection circuit 18... Pitch similar note sequence number detection circuit 19...
...Note length order similar sound sequence number detection circuit 70-20...Reference note sequence number detection circuit 21...Reference note sequence number detection circuit 22...Open during performance/time difference detection Circuit 23...
... Similarity calculation circuit 24 ... Similarity discrimination circuit 25 ... Control circuit 26 ... Door opening/closing control circuit 27 ... Lighting control circuit 28. ...Bath water injection control circuit PD (in) ...Input pitch data PCI
ine (full)... Input sound ^ tone string data P D 1ine (ref)...
・Reference pitch pitch sequence data PD (ref>・・・・・・・・・・
・・・・・・Standard pitch data p[)line sm
l −l ~p[)line sg+l−k ・
-・-・-Shift number data PDeq-1 to PDeq-k corresponding to the first to second similar high pitch string partial data, respectively......First to
Pitch-politico-religious data P D sample corresponding to the similar pitch series partial data
...Most similar pitch sequence data PDeq...
・・・・・・・・・・・・・・・Pitch - Politics and Religion Data L
D 1ine (full)... Input note length string data L D 1ine (ref)...
・・・・・・Reference note long tone string data L D 1ine
5g11-1 ~LD 1ine ss
+l−k ・・・・・−・−−Shift number data LDeq−1 to 5DeQ−k ・・・・・・・・・・1st to
The note length-political/religious data L D sample corresponding to the similar note length string partial data L D sample ...... Most similar note length string data LDeQ ......・Note length - Political and religious data Dscore・・・・・・・・・Score data P D 1ine (1, ~7)・・・・・・
...Input pitch string partial data LD (ref) ...Reference note length data LD (in
)...Input note length data D(N1)...
......-Kanon Kokyo Data D (N2)...
... Matching code length number data D (N3) ......
Pitch similar tone sequence set number data D (N4)...
Note length similar note sequence set number data D (N5)...
Reference pitch number data D (N6)......Reference note length number data 5kOn...Key press signal 5kOn'... Key press timing signal 3k
on”・・・・・・Data acquisition signal inversion 5kO
n''...data acquisition signal s play...
...Playing signal 3 judge...
・Judgment enable signal Sφ・・・・・・・・・・・・
・・・Clock signal 3 end・・・・・・・・・・・・
・Judgment end signal 5load・・・・・・・・・Load signal 3shift・・・・・・・・・Shift signal 3
1atch・・・・・・Latch signal 5 push
・・・・・・・・・Reliable key press signal 35m1...
......Similar film designation signal Seq...
......Concordance signal 5C1r...
・・・・Clear signal 3ic・・・・・・・・・・・・・
...Initial clear signal 5CO5...
...Switching signal S 5elect...Melody designation signal 73-3ok...Melody recognition confirmation signal 3 over... ...Operation end signal Patent applicant Nippon Musical Instruments Manufacturing Co., Ltd. 74- Figure 28, Figure 29, Figure 30, Figure 33.

Claims (6)

[Claims] (1) Convert performance or singing information into one or two or more dimensional tone sequence data having at least the pitch dimension; then, convert the tone sequence data and the tone sequence data corresponding to the reference tone sequence into a tone sequence pattern. Compare predetermined features with; Based on the comparison result, extract the phrase most similar to the reference sound string from the input sound string; Use the sound string data corresponding to this extracted phrase as the sound string corresponding to the reference sound string. A method for recognizing a musical performance or singing characterized by recognizing an input musical performance or singing by comparing it with data. (2) Each constituent sound generated sequentially by playing or singing,
A single note data detection means that sequentially converts and detects single note data of one or two or more dimensions having at least the pitch dimension; and stores the detected single note data for each dimension and in the order of occurrence, and performs or sings it. an input sound string data forming means for forming input sound string data corresponding to; a sound string data registration memory in which reference sound string data of each dimension corresponding to one or more performances or singing is registered; and: the sound string. The tone sequence data of each dimension corresponding to one performance or singing read from the data registration memory and the input tone sequence data of the corresponding dimension are transferred to the same time zone while shifting the reference points of both time axes from each other. Similar sound string part extracting means for comparing existing sound string parts and extracting one or more similar sound string parts from the input sound string data, selected in descending order of similarity; a data superimposition means for extracting matching portions from each similar sound string portion with the registered sound string data of the corresponding dimension and superimposing them on each other to form the most similar sound string data corresponding to the most similar phrase; Among each constituent sound data of the registered sound sequence data, calculate the ratio of the total number of sound dates included in the formed most similar sound sequence data to the total number of each constituent sound data of the registered sound sequence data. similarity score calculating means for determining the dimensions and calculating a similarity score for the input performance or singing using at least the value of the ratio as a scoring element; Recognition of performance or singing melody 11@lf, characterized by comprising an input information discriminating means for discriminating. (3) The input tone string data and the reference tone string data each consist of two-dimensional data consisting of a note and a note length,
In addition, the data superimposition means extracts from each extracted similar sound string part only the portion that matches the reference sound string data of the corresponding dimension, and the data value of one dimension set in advance is set in the corresponding dimension. The patent claim is characterized in that data in the other dimension corresponding to a sound that does not match the value of the registered data in the corresponding dimension is assumed to match the value in the registered data in the corresponding dimension, and the matching portion is extracted. Range 2nd
A performance or singing recognition device as described in Section 1. (4) The reference tone string data and the input tone string data are each composed of two-dimensional data consisting of pitch and note length,
In addition to the ratio between the total number of matching data notes and the total number of all registered data, the similarity score calculation means calculates the difference between the model performance time determined by the standard note length note string data and the excess performance time determined by the most similar note length note string data. 3. The performance or singing recognition apparatus according to claim 2, wherein a similarity score is calculated using a ratio between the model performance time and the model performance time as a scoring element. (5) The reference tone string data and the input tone string data each consist of two-dimensional data consisting of a note and a note length,
In addition to the ratio of the total number of matching data sounds to the total number of all reference data sounds, the similarity score calculation means uses the number of similar sound string sets of at least one dimension extracted by the similar sound string extraction means as a scoring element. The performance or singing recognition method according to claim 2, characterized in that the similarity score is calculated as follows.
l device. (6) The input note string data forming means removes, from the detected single note data, single note data of a certain standard or less that cannot be a musical element. The performance or singing recognition device described in .