JPH0261760B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to an input device for an electronic musical instrument that extracts the pitch of an input waveform signal and generates a musical tone of a corresponding pitch.

[Prior art]

Advances in electronic technology have made it possible to generate musical sound waveforms electronically and to generate musical tones through speakers. This aforementioned device is generally called an electronic musical instrument. In this electronic musical instrument, the musical tones to be generated are generally selected using the keys, and furthermore, the tones of the piano or other musical instruments can be specified and generated using a lever switch or the like.

[Problems with conventional technology]

As mentioned above, this electronic musical instrument uses keys to select musical tones, so in order to play it, it is necessary to practice piano skills and operate the keys. In other words, it was not possible to simply provide accompaniment to, for example, a person's song, and only those who had the skills to operate electronic musical instruments could do so.
That is, conventional electronic musical instruments operated by keys have a problem in that they cannot be easily played by anyone.

[Purpose of the invention]

The present invention solves the above-mentioned problems, and makes it possible to play music even if one does not have the skills to play an electronic keyboard instrument. An input device for an electronic musical instrument that can generate musical tones with good responsiveness, and that can prevent unnatural pitch changes caused by noise and fluctuations in extracted pitch data. The purpose is to provide.

[Structure of the invention]

In order to achieve the above object, according to the present invention, the processing means performs different processing on pitch data extracted from an input waveform signal depending on the time from the start of input of the input waveform signal, and the processed pitch data is The key point is that the pitch determining means determines the frequency of the generated musical tone in accordance with the processed pitch data.

In other words, optimal processing is performed in accordance with the time from the start of input of the input waveform signal, so that optimal pitch data can be obtained.

Specifically, in the initial stage, the extracted pitch data is used as is to quickly determine the pitch of the generated musical note, and in the later stage, the pitch data is averaged multiple times, such as by moving average calculation or with hysteresis. By performing processing such as moving average calculation,
To prevent an unnatural pitch change caused by noise or fluctuation of pitch data from being caused to a musical tone being output.

Furthermore, which pitch data to use as the pitch data may be determined based on the amount of change between the previous pitch data and the current pitch data.

[Embodiments of the invention]

FIG. 1 is a circuit configuration diagram of an embodiment of the present invention,
This figure shows the configuration of an electronic musical instrument that generates musical tones in response to audio input. The output of the microphone 1, which converts audio into an electrical signal, is applied to a preprocessing section 2. The output of the preprocessing section 2 is connected to the pitch extraction section 3. The output of the pitch extractor 3 is applied to a processor (CPU) 5 via a latch 4. The output of the processor 5 is connected to a pitch extraction section 3, a storage section 6, an error removal section 7, a moving average calculation section 8, and a flag creation section 9.
The output of the storage section 6 is applied to the error removal section 7 and the pitch data control section 10, and the output of the error removal section 7 is applied to the moving average calculation section 8 and the pitch data control section 10, respectively. The output of the moving average calculation section 8 is connected to a pitch data control section 10 and a code generator 11. The output of the pitch data control section 10 is also connected to the code generator 11. The output of the chord generator 11 is connected to a musical tone generator 12, and the output of the musical tone generator 12 is connected to a speaker 13 that converts an electrical signal into sound.

The audio signal converted into an electrical signal by the microphone 1 is applied to a preprocessing section 2, where it is preprocessed. This preprocessing first involves removing out-of-band signals using, for example, a low-pass filter (LPF). This out-of-band removal is done to prevent malfunctions in pitch extraction performed in the next stage. Furthermore, the second step of the preprocessing is to amplify the input audio signal from which out-of-band components have been removed by an automatic gain control circuit to a specific amplitude value.

Third, the input audio signal having the above-mentioned specific amplitude value is converted into digital data by the analog/digital conversion circuit. The aforementioned amplification to achieve the specific amplitude value is performed in order to make the number of output bits of this analog/digital conversion circuit effective.

The audio signal processed by the preprocessing section 2, that is, the digital audio signal, is applied to the pitch extraction section 3. The pitch extraction section 3 extracts the pitch of the fundamental frequency of the input audio signal under the control of the processor (CPU) 5. The pitch extraction unit 3 includes an audio data quantization circuit, a scale extraction circuit, and the like, although details are described in, for example, Japanese Patent Application No. 58-31284, Japanese Patent Application No. 58-31285, and Japanese Patent Application No. 58-31286. It has a circuit and a ternary correlation processing circuit, and these circuits perform pitch extraction. The data quantization circuit sequentially obtains the maximum and minimum values of the audio digital data over a specific period of time, and uses the maximum and minimum values to determine respective threshold levels for, for example, ternarization. Then, the input signal is ternarized using the threshold. The scale extraction circuit uses the above-mentioned audio digital data to
This is a circuit that delays digitized audio ternary data in accordance with the scale to be extracted, and multiplies the delayed data by input data. This circuit provides a time correlation. This data simply represents the temporal correlation between the delayed data and the input data, and is further processed to extract the pitch. The above-mentioned ternary correlation processing circuit performs this extraction. The ternary correlation processing circuit is a circuit that performs window processing and the like for pitch extraction. That is, since a plurality of pieces of data are outputted from the above-mentioned scale extracting circuit corresponding to the scale to be extracted, the plurality of data are subjected to window processing in this ternary correlation processing circuit, in other words, weighting is performed in relation to the scale. and then repeat it for a specific time (1
frame) to accumulate. Then, find the maximum value from the accumulated values (accumulated values corresponding to the delay time),
The pitch is determined from the delay time corresponding to the maximum value and output. This output is data related to the pitch of the fundamental wave of the input audio (including harmonics).

The pitch data outputted by the pitch extraction section 3 described above is stored in the storage section 6 via the latch 4 and the processor 5. Thereafter, this pitch data is transferred to the pitch data control section 10 and also to the error removal section 7. The error removing unit 7 uses the data stored in the storage unit 6 to detect a change in specific pitch data, especially a change of one octave or more in only one frame, and when that change is detected, it detects a change in pitch data before changing the data. This is a circuit to restore the data. It should be noted that although there may be sudden changes in the pitch, the error removal unit 7 does not perform the above operation if the pitch data changes suddenly for two or more frames in a row. is designed to change. The data from which one frame error has been removed in the error removal section 7 is transferred to the pitch data control section 10 and averaged at the moving average calculation section 8. The output obtained by performing a plurality of averaging processes is output. That is, the first output is the result of averaging the pitch data inputted from the error removing section 7 over the past four frames including the current one.
Further, as a second output, the past six frames including the current one and the previous average output are appropriately weighted and averaged for each pitch data. For example, the past 6 frames and the 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. These processes are performed by a shift register and an accumulator that sequentially store pitch data input for each frame.

The data averaged in this way is outputted from the moving average calculation section 8, and is outputted from the pitch data control section 10.
In addition, processing for generating musical tones is performed.
In other words, the error removal unit 7 described above is a circuit that detects and removes malfunctions in only one frame of pitch extraction due to external noise, etc., and eliminates large step-like changes in data due to malfunctions of pitch extraction. Some things need to be removed. The moving average calculation unit 8 is a circuit that appropriately stabilizes changes in pitch data by performing a plurality of averaging processes in response to subtle changes in data due to erroneous pitch extraction due to subtle pitch changes in input audio, etc. In particular, the extracted data for each frame is
This prevents the sound from fluctuating finely with a width similar to that of a musical scale.

The human voice is not limited to singing; it also has consonants and vowels. Vowels contain many fundamental pitches, but consonants have very few pitches. For this reason, while only consonants are being produced (for example, “SA”)
(while generating "S"), the pitch of the fundamental wave may be erroneously extracted.
The pitch data control 10 eliminates malfunctions of musical tones due to these erroneous extractions and ensures that the tones match the scale of the input voice. Then, this control section selects and outputs the data stored in the storage section 6, the output data of the error removal section 7, or the output data of the moving average calculation section 8, using data such as audio power and audio input detection. .

The code generator 11 is a circuit that converts the output of the pitch control section 10 described above into the code of the musical tone to be generated by the musical tone generating section 12, that is, musical tone data.

The pitch extraction section 3, storage section 6, error removal section 7, and moving average calculation section 8 described above are controlled by the processor 5. Also, the pitch data control section 1
0, the code generator 11 is controlled by the processor 5 via the flag generator 9. In other words, the flag creation section 9 is created by the processor 5.
A flag and the like are set, and the above-mentioned pitch data control section 10 and code generator 11 operate according to the flag.

The output of the code generator 11, ie, the converted data, is applied to the musical tone generator 12, and specifies the musical tone to be output by the speaker 13. In other words, the musical tone generator 12 uses the chord generator 1
A musical tone electric signal (analog) specified by the conversion data of No. 1 is generated and applied to the speaker 13. As a result, the speaker 13 generates a musical tone corresponding to the input voice.

The circuit shown in Figure 1 represents the configuration of an electronic musical instrument that generates musical tones in response to audio input.
This electronic musical instrument is made possible by the circuitry described in more detail below.

The invention particularly relates to the pitch data control section 10 of the embodiment of the invention shown in FIG.

The pitch data control section 10 will be explained in detail below.

At the beginning of a human voice's pronunciation, the pitch is often unstable even if an attempt is made to pronounce it at a specific pitch. Moreover, as mentioned above, in consonant parts, the pitch is unclear, so erroneous extraction is likely to occur.

On the other hand, the moving average output with hysteresis in the moving average calculation unit 8 is effective for stabilizing pitch data for a steady scale, but is not effective for a portion with large changes at the beginning.
Furthermore, the output of the error removing section 7 can appropriately remove errors from one frame of pitch data and can follow sudden changes in musical scales, but cannot follow irregular changes at the beginning of the audio. Furthermore, the 4-frame moving average output from the moving average calculating section 8 has a property intermediate between the above two outputs. Therefore, after the vocalization begins, for example, the first
In frames 3 to 3, the unprocessed pitch data from the storage section 6 is output to the code generator 11 as is. For the next 4 to 7 frames, pitch data is outputted through the error removing section 7.

Furthermore, for the next 8 to 11 frames, a 4-frame moving average output from the moving average calculating section 8 is output. From the 12th frame onward, the 8-frame moving average output with hysteresis of the calculation unit 8 is output. By doing this, suitably processed pitch data is selectively output to the code generator 11 as the vocalization progresses from the beginning to the steady portion.

In addition, in human speech, for example, when pronouncing "do-so", the constant part of "do" is changed to "so".
When moving to , the pitch data may be interrupted at that part because there is a consonant at the beginning of "so". However, when actually producing a musical tone as "do-so", it is desirable that "do" and "g" connect smoothly without interruption. Therefore, when a consonant part is detected, the pitch data immediately before the consonant part is successively output for that part and the vowel parts of the following 1 to 3 frames. After that 4~
At the transition to the 7th frame, the pitch may be interrupted at the consonant part and the change in pitch may be large, so
Pitch data from each processing section is selectively outputted to the chord generator 11 in the same way as the processing for the beginning part of the pronunciation.

In order to realize the above operation by the pitch data control section 10, various flags are created in the flag creation section 9 (FIG. 1). The various flags created include a power flag PF, a correlation value flag CF, a key-on flag KOF, a key attack flag KATF, and a key-off flag KOFF. First the power flag
Explain about PF.

The pitch extractor 3 sends the pitch data to the processor 5 via the latch 4, in addition to the pitch data, the power of the ternary quantized data for each frame (accumulated value within one frame of the square value of the ternary quantized data). is output. When this power value exceeds a certain threshold,
Power flag PF is high level (hereinafter referred to as H)
becomes. This power flag PF is an indicator of the presence of speech regardless of whether it is a consonant or a vowel.

Next, the pitch extractor 3 outputs the maximum correlation value at the delay time corresponding to the pitch to the processor 5 via the latch 44. When this maximum correlation value exceeds a certain threshold, the correlation value flag CF
becomes H. This correlation value flag CF indicates the degree of presence of pitch, and in vowels etc., CF
becomes H, but when there is a consonant or silence, CF becomes a low level (hereinafter referred to as L).

Next, the key-on flag KOF is a flag indicating that the instrument is in a sounding state, and becomes H when the power flag PF and correlation value flag CF are both H, and becomes L when both are L.
No change at other times. key attack flag
KATF becomes H only for the first clock when the key-on flag KOF becomes H. Also the key off flag
KOFF becomes H only for the first clock when the key-on flag KOF becomes L.

These two flags serve as indicators indicating the beginning and end of pronunciation. It should be noted that since the above flag creation operation is performed for each frame, they are all synchronized by the frame clock φ _f having a time width equal to the frame interval.

The above five flags are used to determine the pronunciation start frame, consonant frame, and pronunciation end frame. In other words, the start frame of pronunciation is the frame where the key attack flag KATF is H,
At the end frame of pronunciation, the key-off flag KOFF is H.
This is the frame immediately before . Further, the consonant frame is a frame in which the power flag PF and key-on flag KOF are H, and the correlation value flag CF is L.
In order to realize the pitch data control operation based on the flag, the pitch control section 10 performs the following steps.
Two counters are used: a counter ANC14 that counts from the pronunciation start frame to the 11th frame, and a counter SNC15 that counts from the frame following the consonant frame to the 11th frame.

Thus, raw pitch data _N1 from storage 6 (FIG. 1) is connected to latch 22 via latch 18 and gate 26. Error removal section 7
Pitch data _N2 from (FIG. 1) is connected to latch 22 via latch 19 and gate 27.
The 4-frame moving average output pitch data N ₄ from the moving average calculating section 8 is connected to the latch 22 via the latch 20 and the gate 28 . Similarly, the moving average output pitch data _N8 with hysteresis from the moving average calculating section 8 is connected to the latch 22 via the latch 21 and the gate 29. The output of latch 22 is output to code generator 11 (FIG. 1) and is also fed back to its own input via latch 23 and gate 30. Flag creation section 9
The key-on flag KOF from (FIG. 1) is input to the counter 14 via AND circuits 31 and 32, respectively.
and count enable input EN of counter 15
Enter. Furthermore, this signal KOF is input to the OR circuit 44 via an inverter 46. Power flag PF and correlation value flag from flag creation unit 9
Both CF are input to a NAND circuit 45, the outputs of which are input to the count reset inputs R of counters 14 and 15, respectively, via OR circuits 36 and 39, and to the OR circuit 44. In addition, the key attack flag KATF from the flag creation section 9,
Each bit output of the counter 14 via the OR circuit 37 is sent to the count enable input EN of the counter 14 via the OR circuit 35 and the AND circuit 31.
Enter. On the other hand, the correlation value flag CF from the flag generator 9 is connected to a latch 24 operated by the frame clock φ _f , and its output is connected to a latch 25. Output of latch 24 and inverter 48
The output of latch 25 is input to AND circuit 34.

The output of the AND circuit 34 and each bit output of the counter 15 via the OR circuit 40 are input to the count enable input EN of the counter 15 via the OR circuit 38 and the AND circuit 32. counter 1
4 and counter 15 perform counting in accordance with frame clock φ _f , and their respective bit outputs are input to logic circuits 16 and 17, respectively. 16 and 1
Outputs 49 and 54, which become H when each count value is 0 in 7, both control the gate 29 via the OR circuit 41 and the AND circuit 33. At this time, the other input of the AND circuit 33 is connected to the OR circuit 44.
The output is inputted via the inverter 47.

An output 50 that becomes H when the count value is 1 to 3 in the logic circuit 16 controls the gate 26, and an output 55 that becomes H when the count value is 1 to 3 in the logic circuit 17 is input to the OR circuit 44. do. The count value in logic circuits 16 and 17 is 4 to 7.
Outputs 51 and 56 which become H when
The gate 27 is controlled via the gate 27. Furthermore, outputs 52 and 57 which become H when the count value is 8 to 11 in the logic circuits 16 and 17 control the gate 28 via the OR circuit 43. In addition, outputs 53 and 58 which become H when the count value reaches 12 in the logic circuits 16 and 17 are output from the OR circuit 36, respectively.
and 39 to the reset inputs R of the counters 14 and 15. On the other hand, the output of the OR circuit 44 controls the gate 30.

The operation for performing the pitch data control by the pitch control section 10 having the above configuration will be explained with reference to FIG. 3.

FIG. 3 shows each flag and two counters 14 and 1.
5 and pitch data output from the pitch data control unit 10 to the code generator 11. First, PF becomes H when voice is input. CF becomes H with a delay of two frames (this delay is due to the circuit configuration). Then KOF with a delay of 2 frames
becomes H, and at the same time KATF becomes H for only one frame.
becomes. Then KATF becomes H, KOF, PF,
Counter 14 on the condition that CF are both H.
starts counting. This condition is OR circuit 35
This is established when the count enable input EN becomes H via the AND circuit 31. Also, KATF that starts counting becomes L, but at least 1 bit of the 4 bits output from the counter becomes H, so the output of the OR circuit 37 becomes H, and the OR circuit 3
5 and the AND circuit 31, the count enable input EN becomes H and continues counting. Extracted data is continuously selected and output by this counting operation.

Next, when the count value of this counter 14 is between 1 and 3, the gate 26 is opened and the unprocessed pitch data _N1 from the storage section 6 (FIG. 1) inputted to the latch 18 is passed through the latch 22. It is selectively output to the code generator 11. This operation is realized by the output 50 of the logic circuit 16 becoming H. Next, when the count value is 4 to 7, logic circuit 1
The output 51 of 6 becomes H and is applied to the control terminal of the gate 27 via the OR circuit 42, so the gate 27 opens and the error remover 7 input to the latch 19
selectively outputs pitch data _N2 from the latch 22 to the latch 22. Further, when the count value is between 8 and 11, the gate 28 is opened, and the 4-frame moving average output pitch data _N4 from the moving average calculating section 8 input to the latch 20 is selectively output to the latch 22. This operation is performed when the output 52 of the logic circuit 16 becomes H and H is applied to the control terminal of the gate 28 via the OR circuit 43.

Then, when the count value of the counter 14 reaches 12, the output 53 becomes H and the counter 14 is cleared via the OR circuit 36 to end the counting.
After that, the gate 29 opens on the condition that PF, CF, and KOF are all H, and the 8 with hysteresis from the moving average calculation unit 8 input to the latch 21
Frame moving average output pitch data _N8 is selectively output to latch 22. This operation is OR circuit 4
The output 49 through 1, the NAND circuit 45, the OR circuit 44, and the inverter 47 become H, which causes the AND circuit 33 to go H, and this is realized by applying it to the control terminal of the gate 29. Ru.

Next, if a consonant frame is detected in the key-on state (KOF is H), that is, if PF is H,
When CF becomes L, the gate 30 opens and selectively outputs the pitch data BN of the immediately previous frame (N ₈ in the case of FIG. 2) to the latch 22. This operation is realized by the NAND circuit 45 and the OR circuit 44 becoming H and being applied to the control terminal of the gate 30. Then, when the consonant frame ends and the next pronunciation frame begins, the counter 15 starts counting. That is, when CF rises to H while KOF is at H, the counter 15 starts counting. This operation is realized by inputting a pulse corresponding to one frame to the count enable input EN of the counter 15 via the AND circuit 34 and the OR circuit 38 by the latch 25 via the latch 24 and the inverter 48.

When the count starts, the OR circuit 4
0, 48 and the AND circuit 32, the count enable input EN becomes H, and counting continues.

Next, when the count value is 1 to 3, the pitch data BN for the consonant frame is selectively output to the latch 22. In this operation, the output 55 of the logic circuit 15 becomes H and is passed through the OR circuit 44 to the gate 30.
This is achieved by opening.

Next, when the count value is 4 to 7, the counter 14
Open gate 27 in the same way as when
N ₂ is selectively output to latch 22. This is realized by the output 56 becoming H and the gate 27 opening.

Furthermore, when the count value is between 8 and 11, the gate 28 similarly opens and the pitch data _N4 is selectively output to the latch 22. This is realized by the output 57 becoming H and the gate 28 opening. And in frames where the count value exceeds 11, the output is 5 as well.
8 clears the output of the counter 15 and ends counting. Thereafter, gate 29 opens and selectively outputs pitch data _N8 to latch 22. In this operation, the output 54 becomes H and the gate 29
This is achieved by opening.

Finally, following the CF in the last part of the pronunciation
There is a possibility that the sound continues to be unstable until PF becomes L and KOF becomes L with a delay of two frames, so the gate 30 opens and the pitch data BN of the previous frame (N in the case of Fig. 2) is generated. ₈ ) is selectively output to the latch 22.

This operation is realized by opening the gate 30 via the NAND circuit 45 and the OR circuit 44.

FIG. 4 summarizes the relationship between the counter values of counters 14 and 15 and the type of pitch data to be selected. When the count value of each counter is 1 to 3, it is the beginning of sound generation in counter 14, so the unprocessed pitch data
N ₁ is selected. Furthermore, since the frame follows the consonant, the counter 15 selects the pitch data BN of the frame immediately before the consonant.

Furthermore, the count value of the beginning part of the pronunciation is 4~
7, 8 to 11, unprocessed pitch data 1
Pitch data 4 with frame error removed
Pitch data subjected to a moving average of frames and pitch data subjected to an 8-frame moving average with hysteresis are selectively adopted in this order, so that both irregular changes and steady parts can be treated.
Pitch data with less erroneous extraction can be supplied to the code generator 11. Furthermore, in the consonant part and the end part of pronunciation, the pitch data is connected smoothly without interruption, so that the change in the scale created by the chord generator 11 becomes natural. In the embodiment of the present invention, the correlation value flag CF off is set for the consonant part, but this is not limited to the consonant part, but also corresponds to changes in scales generated by musical instruments, for example.

〔Effect of the invention〕

As explained above, the present invention determines the pitch of an output musical sound by performing optimal processing on pitch data in accordance with the time from the start of input of an input waveform signal to obtain processed pitch data. This makes it possible to prevent unnatural changes in pitch due to fluctuations in pitch data, noise during quantization, and the like.

[Brief explanation of the drawing]

FIG. 1 is an overall circuit diagram of an embodiment according to the present invention, FIG. 2 is a detailed circuit diagram of the pitch data control section of FIG. 1, FIG. 3 is an operation timing chart of the present invention, and FIG. Counter ANC14, SNC
15 is a relationship diagram between count values of No. 15 and types of selected pitch data. FIG. 10... Pitch data control section, 14, 15...
Counter, 16, 17...Logic circuit, 18, 1
9, 20, 21, 22, 23, 24, 25... Latch, 26, 27, 28, 29, 30... Gate.

Claims (1)

[Scope of Claims] 1. Pitch data extraction means for extracting pitch data of an input waveform signal, detection means for detecting the start of input of the input waveform signal, and detecting the pitch data extracted by the pitch data extraction means. processing means for obtaining processing pitch data by selectively performing different processing processes corresponding to the time from the start of input of the input waveform signal detected by the processing means; and the processing pitch data obtained by the processing means. An input device for an electronic musical instrument, comprising: pitch determining means for determining the frequency of a musical tone generated according to the following. 2. The processing means directly outputs the pitch data from the pitch data extraction means as the processed pitch data without performing any processing at the time of input start, and at a later point in time outputs the pitch data extracted previously and the pitch data extracted this time. Claim 1, characterized in that the machining pitch data is obtained by performing an average calculation of
An input device for an electronic musical instrument as described in Section 1.