JPH071429B2 - Chord generator - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a chord generating device that extracts a pitch of an input waveform signal and generates a chord according to the pitch data.

[Prior art]

Advances in electronic technology have made it possible to generate musical tone waveforms electronically and to produce musical tones using speakers. The above-mentioned device is generally called an electronic musical instrument. This electronic musical instrument generally uses a key to select a musical tone to be generated, and a lever switch or the like can be used to specify a piano sound or another musical instrument sound.

[Problems of conventional technology]

Thus, playing a conventional electronic keyboard instrument requires a considerable amount of technology. In particular, playing chords that require pressing a plurality of keys on the keyboard is extremely difficult for a beginner. Therefore, as another performance mode, a pitch of an input waveform signal such as a voice is extracted, and a chord constituent tone is generated based on the extracted pitch and a designated arbitrary chord type, thereby performing a chord performance. It was suggested to be able to do it.

According to such a method, the performer does not need to operate the keyboard, and even a beginner can easily perform a chord performance.

However, the pitch extracted from the input waveform signal is
Since it has noise and fluctuations, unnatural pitch changes are likely to occur. Moreover, it is known that such noises and fluctuations change their generation modes with the lapse of time after the start of extraction.

[Object of the Invention]

An object of the present invention is, in a chord generating device capable of playing a chord based on the extracted pitch and the type of a specified chord, removes the influence of noise and fluctuation included in the extracted pitch, and makes it unnatural. The purpose of this is to prevent changes in pitch.

[Structure of Invention]

In order to achieve the above-mentioned object, the present invention differs from the extraction means for extracting the pitch data of the input waveform signal, and the pitch data extracted by the extraction means, based on the change in the extracted pitch data. Processing means for performing processing to obtain processing pitch data, input operation means for designating an arbitrary chord type from a plurality of chord types, and a sound represented by the processing pitch data from the processing means. A root note designating means for designating to generate a root note of the chord at a high pitch, and the root note based on the type of the chord specified by the input operation means, and a plurality of other constituent notes of the chord. Output means for outputting a plurality of pitch difference data corresponding to the pitch difference, and the pitch difference data output from the output means and the processing pitch data from the processing means to calculate the corresponding chord Calculating means for obtaining pitch data of each of the synthesized sounds, and constituent sound designating means for designating to generate the other plurality of constituent sounds at pitches represented by the respective pitch data obtained by the calculating means. And is provided.

Example of Invention

FIG. 1 is a circuit configuration diagram of an embodiment of the present invention, showing a configuration of an electronic musical instrument which generates a musical sound in response to a voice input. The output of the microphone 1 for converting voice into an electric signal is applied to the preprocessing unit 2. The output of the preprocessing unit 2 is connected to the pitch extraction unit 3. The output of the pitch extraction unit 3 is applied to the processor (CPU) 5 via the latch 4. The output of the processor 5 is connected to the pitch extraction unit 3, the storage unit 6, the error removal unit 7, the moving average calculation unit 8, and the flag creation unit 9. The output of the storage unit 6 is sent to the error removal unit 7 and the pitch data control unit 10, and the output of the error removal unit 7 is sent to the movement calculation control unit 10 and the error removal unit 7
Is applied to the movement calculation unit 8 and the pitch data control unit 10, respectively.

The output of the movement calculation unit 8 is connected to the pitch data control unit 10.

The output of the pitch data control unit 10 is connected to the external control unit 14. The output of the flag creating section 9 connects the pitch data control section 10 to the external control section 14. The output of the key input unit 5 is connected to the external control unit 14, and the output of the external control unit 14 is connected to the code generator 11.

The output of the code generator 11 is connected to the musical sound generating unit 12, and the output of the musical sound generating unit 12 is connected to the speaker 13 which converts an electric signal into a sound.

The voice signal converted into an electric signal by the microphone 1 is added to the pre-processing unit 2 and pre-processed. This pretreatment is the first
For example, a signal outside the band is removed by a low pass filter (LPF) or the like. The removal out of the band is performed in order to prevent a malfunction of pitch extraction performed by the next stage. Further, in the preprocessing, secondly, the input audio signal from which the out-of-band is removed by the automatic gain control circuit is amplified to have a specific amplitude value.

Thirdly, the input audio signal having the above-mentioned specific amplitude value is converted into digital data by the analog / digital conversion circuit. The amplification to obtain the above-mentioned specific amplitude value is made to make the number of output bits of this analog / data conversion circuit effective.

The voice signal processed by the preprocessing unit 2, that is, the digital voice signal is added to the pitch extraction unit 3. The pitch extraction unit 3 extracts the pitch of the fundamental frequency of the input voice signal under the control of the processor (CPU) 5. Details of the pitch extraction unit 3 are Japanese Patent Application No. Sho 58.
-31284, Japanese Patent Application No. 58-31285, Japanese Patent Application No. 58-31286 have a voice data quantization circuit, a scale extraction circuit, and a three-value correlation processing circuit, and pitch extraction is performed by these circuits. . The data quantization circuit sequentially finds the maximum and minimum values of audio digital data over a specific time, and then determines the maximum value,
Using the minimum value, for example, each threshold level for ternarization is determined. Then, the input signal is ternarized, for example, by the threshold. The scale extraction circuit is a circuit for delaying the tone ternary data that has been ternarized using the above-mentioned voice digital data in correspondence with the tone scale from which it is extracted, and multiplying the delay data and the input data. This circuit correlates over time. This data simply represents the time-related correlation between the delay data and the input data, and this data is further processed to extract the pitch. The above-described ternary correlation processing circuit performs this extraction. The ternary correlation processing circuit is a circuit that performs window processing for pitch extraction. That is,
Since a plurality of data are extracted from the above-mentioned scale extraction circuit corresponding to the scale to be extracted, this ternary correlation processing circuit performs window processing on the plurality of data, in other words, performs weighting in relation to the scale, Further, it is accumulated for a specific time (1 frame). Then, the maximum value is obtained from the accumulated value (accumulated value corresponding to the delay time), and the pitch is obtained and output from the delay time corresponding to the maximum value. This output is data related to the pitch of the fundamental wave of the input voice (including harmonics).

The pitch data output from the pitch extraction unit 3 is stored in the storage unit 6 via the latch 4 and the processor 5. After this, this pitch data is transferred to the pitch data control unit 10 and also to the error removing unit 7. The error removing unit 7 uses the data stored in the recording unit 6 to detect a change in specific pitch data, particularly a change of, for example, one octave or more only in one frame, and changes the data when the change is detected. This is a circuit for returning to the previous data. Although there may be a sudden change in the scale, the error removing section 7 does not perform the above operation when the pitch data changes rapidly for two or more consecutive frames. At this time, there is a delay of one frame. The data is changing. The data from which one frame error has been removed by the error removing unit 7 is transferred to the Pichi data control unit 10 and averaged by the moving average calculation unit 8. Then, as its output, the one obtained by performing a plurality of averaging processes is output. That is, as the first output, a value obtained by performing arithmetic mean on each of the pitch data input from the error removing unit 7 over the past four frames including the present is output. Further, as the second output, the past 6 frames including the present for each pitch data and the addition average output immediately before are appropriately weighted and added and averaged. For example, the past 6 frames and a value obtained by doubling the previous addition average output value are accumulated and averaged, and this is an 8-frame moving average output with apparent hysteresis. Then, these processes are performed by a shift register and an accumulator which sequentially store the pitch data input for each frame.

The data averaged in this way is used as the moving average calculator 8
Then, the pitch data control unit 10 is added to perform processing for generating a musical tone. In other words, the above-mentioned error removing unit 7 is a circuit for detecting and eliminating a malfunction of only one frame of pitch extraction due to external noise or the like, and eliminating a large stepwise change in data due to the result of the malfunction of pitch extraction. There are also things. The moving average calculator 8 is a circuit that appropriately stabilizes a change in pitch data by a plurality of averaging processes against a subtle change in pitch mis-extraction data due to a subtle pitch change in an input voice or the like.
In particular, it is possible to prevent the extracted data for each frame from fluctuating in a width of about one scale in the vicinity of a specific scale that is constantly extracted by the moving average with 8 frames with hysteresis.

Human voices have not only songs but consonants and vowels. Although the pitch of the fundamental wave of vowels is large, there are very few consonants. Therefore, the pitch of the fundamental wave may be erroneously extracted while only the consonant is being generated (for example, when "SA" is being generated, "S" is being generated). Pitch data control is to eliminate the erroneous operation of musical tones caused by these erroneous extractions so that the scale of the input voice is achieved.
Is 10. Then, this control unit uses data such as voice power and voice input detection output from the flag generation unit 9,
The data stored in the recording unit 6, the output data of the error removal unit 7 and the output data of the moving average calculation unit 8 are selected and output.

The external control unit 14 is a circuit that outputs the pitch data output from the pitch data control unit 10 to the code generator 11 based on the data output from the key input unit 15 and the control data from the flag creating unit 9. For example, the key input unit 15
Generates data according to the chord selected in step 1, that is, the chord data, and outputs it to the chord generator 11. In other words, the rhythm timing for generating musical tones and chord chords such as major chords, minor chords, and seventh chords are manually selected by the key input unit 15, and the external control unit 14 performs control according to the manual operation. Is.

The code generator 11 is a circuit that converts the output of the external control unit 14 into a code of a musical tone to be generated in the musical tone generating unit 12, that is, musical tone data.

The pitch extraction unit 3, recording unit 6, error removal unit 7, and moving average calculation unit 8 described above are controlled by the processor 5.

The output of the code generator 11, that is, the converted data, is added to the musical tone generating section 12 and the musical tone is designated so as to be output by the speaker 13. In other words, the tone generator 12 causes the code generator 11 to generate a tone electrical signal (analog) specified by the conversion data, and adds it to the speaker 13.

As described above, the speaker 13 generates musical tones corresponding to the pitch of the input voice input from the microphone 1, the chord data of the key input in the key input unit 15, and the rhythm timing. That is, the pitch is input by the input voice, and the generation timing of the musical tone and the selection of the chord are input by the key, thereby making it possible to create a complicated sound with a few operations.

The circuit shown in FIG. 1 shows the structure of an electronic signal for generating a musical tone in response to a voice input, and this electronic musical instrument can be realized by the circuit described in more detail below.

FIG. 2 is a detailed circuit configuration diagram of the external control unit 14 and the key input unit 15 according to the present invention. The output of the pitch data control unit 10 is connected to one input of the adder circuit 17 via the latch 16. The key input section 15 is an attack key 30, a major code 31,
It has a minor code key 32 and a seventh code key 33, the output of which is passed through the encoder 18 to the flag register 19
Connected to. The parallel output of the flag register 19 becomes an address input to the read-on memory 20, and the OR circuit 23
Will be input to. The output of the read-only memory 20 becomes the start address input to the read-only memory 22 for transposing tone generation via the counter 21 controlled by the flake clock φf. The output of the read-only memory 22 is the adder circuit 17
Connected to the other input of. The output of the OR circuit 23 is connected to the latch 24 controlled by the frame clock φf. The output of the latch 24 is connected to the latch 25 controlled by the frame clock φf, connected to the AND circuit 26, and also connected to the AND circuit 27 via the inverter 28. Similarly, the output of the latch 25 is connected to the AND circuit 27 and also connected to the AND circuit 26 via the inverter 29. The output of the OR circuit 23 is connected to the latch 24 controlled by the frame clock φf. The key-on signal KON output from the AND circuit 26 is connected to the control terminal of the latch 16 and the code generator 11. The key-off signal KOFF output from the AND circuit 27 is connected to the code generator 11. Further, the output of the adding circuit 17 is also connected to the musical sound generating section 12 via the code generator 11.

First, the case where the attack key 30 is pressed in the key input unit 15 in the above configuration will be described. The attack key 30 does not generate chords in the musical tone generator 12,
This is a key for taking rhythm timing when only one scale corresponding to the pitch of the voice input from the microphone 1 (Fig. 1) is generated. As an operation, the performer gives the timing of generation with the attack key 30 while giving the pitch of the musical sound to be generated by voice. As a result, first microphone 1 (first
Pitch is extracted from the voice input from FIG. 1), and after each process is performed, pitch data control unit 10 outputs pitch data. Note that these extraction processes are performed for each frame (for example, 12mesc). If the performer does not press the attack key at this time, no data is output from the key input unit 15, and the output of the OR circuit 23 remains low. Therefore, the contents of the latches 24 and 25 are low level, and the key-on signal KON output from the AND circuit 26 remains low level. Therefore, no pitch data is input to the latch 16 and no musical sound is generated. Next is the attack key 30
When is pressed, the output signal from the key input section 15
Are encoded by and output to the flag register 19. This content is coded to a value that is not all "0", and the OR circuit
23 output goes high. Then, this signal is taken into the latch 24 by the frame clock φf (clock of 12 msec if one frame is 12 mesc). At this time, the latch 25 takes in the contents of the latch 24 one frame before. Therefore, when the attack key 30 is pressed, first, the contents of the latch 24 become high level and the contents of the latch 25 become low level. As a result, the AND circuit 26 is turned on and the key-on signal KON becomes high level. As a result, the pitch data from the pitch data control unit 10 becomes the key-on signal.
It is taken into the latch 16 at the rising edge of KON, and the output appears at the rising edge. In other words, when the attack key 30 is pressed, the content of the latch 24 becomes high level and the content of the latch 25 becomes low level, whereby the pitch data is first input by the key-on signal KON. That is, the key-on signal KON is a signal indicating that a key has been pressed,
This has the same effect as pressing a key on a conventional keyboard-type electric musical instrument.

At this time, the read-only memories 20 and 22 are accessed while the attack key 30 is being pressed.
The content of 20 is loaded into the counter 21, and the count value changes with the frame clock. However, since the content of the read-only memory 22 designated by the data at this time is “0”, the latch 16
The output pitch data of is output from the adding circuit 17 to the code generator 11 as it is.

As a result, a tone code corresponding to the pitch data is generated, and the tone generator 12 generates a predetermined tone. Next, if the attack key 30 is still pressed, the contents of the flag register 19 become the same data, and the output of the OR circuit 23 remains high level. As a result, the contents of the latches 24 and 25, which are rewritten by the frame clock φf, both become high level. Therefore, the AND circuit 26 is turned off, and the key-on signal KO
N becomes low level. Therefore, new pitch data is not input to the latch 16 and the pitch data taken in when the key-on signal KON becomes high level is output as it is, so that the chord generator 11 continues to generate the same tone data. In other words, while the attack key 30 is being pressed, data is continuously generated with the same musical sound.

Then, when the attack key 30 is released next time, the flag register
The content of 19 becomes "0", and the output of the OR circuit 23 becomes low level. This causes the contents of latch 24 to go low. At this time, the content of the latch 25 is high because it is the content of the latch 24 one frame before. This is the AND circuit 27
Turns on and the key-off signal KOFF goes high.
This signal is sent to the code generator 11, and the generation of the musical sound is completed. That is, the key-off signal KOFF is the attack key
This is a signal indicating that 30 has been released. As described above, the pitch of the musical tone is given by the pitch data, and its generation timing is given by the attack key 30 to generate the corresponding musical tone. This place love pitch extraction unit 3 (first
The pitch data extracted according to the drawing) is extracted from the pitch data control unit 70 by the error removal unit 7 and the moving average calculation unit 7 regardless of whether the attack key 30 is pressed or not. To be done. In this state, it is determined whether the attack key 30 is pressed or not, and it is determined whether the key is on, the key is off, or the key is still pressed, and a tone code is generated at that timing. Therefore, the player can generate various musical tones by properly matching the timing of producing a sound and the timing of pressing the attack key 30.

Next, a case where the major chord key 31, the minor chord key 32, the seventh chord key 33, or the like is pressed will be described. These keys exhibit exactly the same operation as the attack key 30 in that they generate the timing of musical tone generation. However, at the same time, a major chord, a minor chord, or a tone code of a chord of the 7th chord can be created. Now, a case where the major chord key 31 is pressed will be described. Whether the key is key-on, key-off, or kept pressed is determined by the OR circuit 23, the latches 24, 25, the AND circuits 26, 27, etc., as in the case of the attack key 30. At this time, the flag register 19 stores the encoded contents of the major chord key 31. Then, this content becomes an address input of the read-only memory 20. Then, the content of the read-only memory 20 at that address becomes the head address where the major chord creation data of the transpose sound creation read-only memory 22 is stored. For example, read-only memory 22 has 100
It is called address "0", address 101 "+8", address 102 "+14", address 103 "+24".

Stores data for creating major chords. now,
When the major code key 31 is pressed, the code register 19
"100" is output from the read-only memory 20 depending on the contents of. Then, the value is loaded into the counter 21 by the frame clock φf. As a result, the content “0” at address 100 is read from the read-only memory 22 and output to the adder circuit 17. Since the key-on signal KON is now at the high level, the pitch data is output from the latch 16, and the data and the output of the read-only memory 22 are added by the adder circuit 17. Since the output of the read only memory 22 is "0", the pitch data from the latch 16 is output to the code generator 11 as it is. Next, time division clock φt in the same frame (for example, 12 msec)
The counter 21 is incremented by 1 and its content is changed to "10.
1 ". This means 101 from read only memory 22.
The content "+8" of the address is read and output to the adder circuit 17. At this time, the same pitch data is still output from the latch 16, so the value obtained by adding +8 to the pitch data is output to the coat generator 11. In this way, the contents of addresses 102 and 103 are time-division clock φ in the same frame.
The data is sequentially read at t, added to the pitch data, and output to the code generator 11 in a time division manner. That is, in addition to the input pitch data, +8,
Data generated by +14 and +24 is code generator 1
Output to 1 in a time-sharing manner. In this case, the input pitch data is 1 octave, not as frequency information.
It is divided into 24 steps and given as a number within this one octave. That is, pitch data increases by +2 when the scale increases by a semitone. Therefore, for example, when the data "0" representing the scale "C" is given as the pitch data, "+8", "+1" in addition to the pitch data "0" in the same frame.
The data "4", "+24" is output to the chord generator 11 in a time division manner. These data are the scale "C",
Represents "E", "G", "C" (1 octave up). And these represent chords of major chords with "C" as the reference note. In other words, data representing a plurality of chords having a major chord relationship with the input pitch data is created in the adder circuit 17 in a time division manner and given to the chord generator 11. As a result, a chord code is generated from the chord generator 11, and a chord musical tone is generated in the musical tone generating section 12. At this time, the timing at which the chord is generated is determined by the timing at which the major chord key 31 is turned on or off, just as in the case of the attack key 30. Naturally, the pitch data of the reference sound of the chord is given by voice from the microphone 1 (FIG. 1).

Next, when the minor code key 32 is pressed, the content of the flag register 19 becomes a code represented as the minor code key 32. As a result, the start address (for example, address 104) of the read-only memory 22 in which the minor code creation data is stored is read from the read-only memory 20, and the value is counted by the frame clock φf in the counter 21.
Loaded in. Then, as in the case of the major code key 31, the minor code creation data is sequentially read from the read-only memory 22 in a time division manner in the same frame,
A plurality of chords having a minor chord relationship with the pitch data of the reference tone are created in the adding circuit 17, and are output to the chord generator 11 in a time division manner. When the Seventh code key 33 is pressed, the Seventh code is created in exactly the same manner.

When the chord key representing each chord is pressed as described above, each chord code is created with the pitch data input as voice as the reference note, and a tone of the chord is generated at the timing of turning on or off with each chord code key. It becomes possible.

In addition, in the embodiment of the present invention, as an input key, an attack key, a major chord key, a minor chord key,
Although limited to the case of the 7th code key, other octave conversion keys (simply +24), transposing keys, etc. can be considered, and various data can be stored in the read-only memory by storing the corresponding data. It is possible to have a function.

〔The invention's effect〕

According to the present invention, for the extracted pitch data, one of a plurality of different processing operations is selected and applied based on the change of the pitch data, and based on the processed pitch data. By generating a chord whose pitch is the root of the chord, it is possible to remove noise and fluctuation included in the extracted pitch data that cause an unnatural pitch change. In particular, such noise and fluctuations, the generation mode thereof changes with the lapse of time from the start of pitch extraction, but in the present invention, by selecting and applying a processing process according to the change from different processing processes, Noise or fluctuation can be removed more accurately.

Moreover, other constituent tones other than the root tones that compose the chord are calculated based on the pitch difference data representing the pitch difference from the root note obtained based on the specified chord type and the processing pitch data. Therefore, various chords can be easily generated. Moreover, the root note can be changed by changing the pitch of the input waveform signal, and the performance effect can be improved. Moreover, the data processing can be performed by numerical calculation, and the configuration can be simplified.

[Brief description of drawings]

FIG. 1 is an overall circuit configuration diagram of an embodiment according to the present invention, and FIG. 2 is a detailed circuit configuration diagram of an external control unit and a key input unit according to the present invention. 14 …… External control part, 15 …… Key input part, 16,24,25 …… Latch, 17 …… Adding circuit, 18 …… Encoder, 19 …… Flag register, 20,22 …… Read only memory, 27 ……counter,
30 …… Attack key, 31 …… Major code key, 32 ……
Nainer code key, 33 …… Seventh code key.

Claims (2)

[Claims] 1. Extraction means for extracting pitch data of an input waveform signal, and selectively processing different pitches on the pitch data extracted by the extraction means based on changes in the extracted pitch data. Processing means for performing processing to obtain processing pitch data, input operation means for designating an arbitrary chord type from a plurality of chord types, and a pitch indicated by the processing pitch data from the processing processing means. Pitch of each of the root note designating means for designating to generate a root note of the chord, and the root note based on the type of the chord specified by the input operation means Output means for outputting a plurality of pitch difference data corresponding to the difference, and the pitch difference data output from the output means and the processing pitch data from the processing means to calculate each of the constituent sounds of the chord. A calculation means for obtaining the pitch data; and a constituent sound designating means for designating to generate the other plurality of constituent sounds at the pitches represented by the respective pitch data obtained by the calculation means. A chord generating device characterized in that 2. The processing means, as different processing operations, removes abrupt changes in the pitch data from the extraction means and averages the past extracted pitch data and the currently extracted pitch data. The chord generating device according to claim 1, characterized in that the processing for setting the value as the current pitch data is performed based on the change of the extracted pitch data.