JPH0764544A - Device and method for generating tempo data - Google Patents

Abstract

PURPOSE:To generate tempo data from an acoustic signal in a live performance without providing the tempo data. CONSTITUTION:When the acoustic signal is inputted, the levels of specific frequency are extracted by peak detection processing (step S3), and peak values over a prescribed level out of them and a generating time are stored in the peak value area PA of memory. The mean value of the peak values is calculated in mean value calculation processing (step S4), and the peak value over a calculated mean value out of the peak values in the peak value area PA is stored in the mean value area AA of the memory. Two peak values in the mean value area AA are selected by tempo calculation processing (step S5), and it is retrieved whether or not another peak value exists in the peak value area PA of the memory every generating interval. When another peak value exists, the tempo data can be generated based on the generating interval, and when no another peak value exists, another peak value is selected, and it is retrieved similarly. Found tempo data is outputted to external equipment (step S6).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a tempo data generating device and a tempo data generating method, and more particularly to a tempo data generating device and tempo data generating device for detecting tempo from an acoustic signal of an actual performance. Regarding the generation method.

[0002]

2. Description of the Related Art Recently, with the widespread use of electronic musical instruments, playing styles have been diversified, and a plurality of devices are playing in synchronization with each other.

In this case, in order to synchronize the performance, it is necessary to detect the tempo of the performance, generate tempo data, and synchronize with the tempo data.

Therefore, conventionally, out of the performance data composed of a plurality of parts, the tempo data contained in the control track is taken out, and an external device such as a sequencer is synchronized with this tempo data.

[0005]

However, in such a conventional tempo data generation device, since only the tempo data of the performance data is taken out, the acoustic sound of a live performance such as when a person plays is performed. There is a problem in that the tempo cannot be accurately detected based on the signal to generate the tempo data, and the acoustic signal of the live performance cannot synchronize with the external device.

Therefore, it is an object of the present invention to provide a tempo data generating device and a tempo data generating method capable of generating tempo data even for a live performance acoustic signal having no tempo data. .

[0007]

The tempo data generation apparatus of the present invention is an extraction means for extracting the level of a specific frequency band of an input acoustic signal, a peak value of the level extracted by the extraction means, and an occurrence time. Of the peak values of the levels stored in the storage means, the calculation means for calculating the average value of the peak values of the levels stored in the storage means, and the calculation means of the peak values of the levels stored in the storage means. Selecting means for selecting two peak value levels exceeding the average value;
Search means for searching whether or not another level stored in the storage means exists at every occurrence interval of the two levels selected by the selection means; and the search means has another level for the occurrence interval. When it is detected that each level exists, when the tempo data generation unit that generates tempo data based on the occurrence interval, and the search unit detects that another level does not exist at each occurrence interval, Reselection means for causing the selection means to select another level from the storage means,
With the provision of the above, the above object is achieved.

In this case, the storage means stores only the peak value of a level having a peak value of a specific value or more among the levels extracted by the extraction means, for example. May be

The tempo data generation method of the present invention is
Extraction processing for extracting the level of a specific frequency band of the input acoustic signal, storage processing for sequentially storing the peak value and generation time of the level extracted by the extraction processing, and peak of the level stored by the storage processing A calculation process for calculating an average value of the values, a selection process for selecting two peak value levels exceeding the average value calculated by the calculation process among the peak values of the levels stored by the storage process, A search process for searching whether or not another level stored by the storage process exists at each occurrence interval of the two levels selected by the selection process; and another level in the search process at each occurrence interval. When it is detected that the tempo data is present, the tempo data generation process for generating tempo data based on the occurrence interval and another level in the search process for each occurrence interval. When it is detected not with, by providing a re-selection process for selecting the other level stored in the storing process by the selection process, we have achieved the above objects.

In this case, for example, as described in claim 4, the storage processing stores only the peak value of a level having a peak value of a specific value or more among the levels extracted by the extraction processing. Good.

[0011]

According to the tempo data generation device and the tempo data generation method of the present invention, the level of a specific frequency band of the input acoustic signal is extracted, and the peak value of this extracted level and the generation time are sequentially stored. The average value of the peak values of the stored levels is calculated, and the peak value level exceeding the calculated average value among the peak values of the stored levels is set to 2
Select one. Then, at each occurrence interval of the two selected levels, it is searched whether or not there is another stored level, and when another level exists at each occurrence interval, the occurrence interval is determined. Generate tempo data based on
When another level does not exist at each occurrence interval, the stored other level is selected and similarly searched.

Therefore, the tempo can be accurately detected from the live performance acoustic signal of the musical instrument or the like to generate the tempo data, and the external equipment such as the sequencer can be synchronized based on the tempo data. .

[0013]

EXAMPLES The present invention will be specifically described below based on examples.

1 to 16 are views showing an embodiment of the tempo data generating apparatus and tempo data generating method of the present invention.

FIG. 1 is a block diagram of the tempo data generator 1.

The tempo data generator 1 is a CPU (Cent
Ral Processing Unit) 2, a frequency filter 3, a volume level filter 4, a memory 5, a data memory 6, a clock sequence composer 7, a clock generation unit 8 and the like, and the tempo data generation device 1 includes a player 9 that generates an acoustic signal. For example, an external device 10 that synchronizes with a tempo data output from the tempo data generation device 1 with a playback device or the like that plays back a musical instrument or a played sound, for example,
A sequencer etc. is connected.

A player 9 is connected to the frequency filter 3, and the frequency filter 3 allows only a sound signal in a specific frequency band among sound signals input from the player 9 to pass therethrough and outputs the sound signal to the volume level filter 4.

The volume level filter 4 transmits only the acoustic signal of a predetermined level or higher among the acoustic signals input from the frequency filter 3 and outputs it to the memory 5. That is, the volume level filter 4 extracts the peak value of the acoustic signal that has passed through the frequency filter 3 and outputs it to the memory 5.

The memory 5 is composed of a RAM or the like, and has a peak value area PA for storing the peak value of the acoustic signal that has passed through the volume level filter 4 and the time of occurrence of the peak value. It has a peak value at a level exceeding the average value and an average value area AA for storing the time of occurrence of the peak value.

The peak value stored in the memory 5 is read by the clock sequence composer 7 and processed by the clock sequence composer 7.

That is, the clock sequence composer 7 calculates the average value of the peak values written in the peak value area PA, writes the calculated average value in the average value area AA of the memory 5, and exceeds the average value. Two peak values are selected, and it is searched whether or not there is a peak value other than the selected peak value in the peak value area PA in the occurrence time interval of the two selected peak values. Data is generated and written in the data memory 6. Also, the clock sequence composer 7
When there is no peak value other than the selected peak value in the time interval of occurrence of the two selected peak values, the peak value is selected again, and similarly processed to generate tempo data.

The data memory 6 is composed of a RAM or the like,
The data memory 6 includes a clock sequence composer 7
The tempo data generated and the occurrence time are written.

The clock generator 8 operates under the control of the CPU 2 and generates a clock signal of a tempo corresponding to the tempo data written in the data memory 6 to generate the external device 1
Output to 0.

The CPU 2 controls each part of the tempo data generation device 1 according to the program stored in the internal ROM, and causes the tempo data generation device 1 to perform processing.

Next, the operation will be described.

As shown in FIG. 1, the tempo data generating apparatus 1 uses a player 9 for generating an acoustic signal of a musical instrument or a reproducing apparatus for reproducing a played sound, and a clock signal output from the tempo data generating apparatus 1. It is connected to an external device 10 such as a sequencer for synchronization, detects the tempo of the audio signal generated by the player 9, generates tempo data for that tempo, outputs it to the external device 10, and synchronizes the external device 10. take.

The tempo data generation process will be described below.

<Main Processing> Tempo Data Generation Device 1
When the power is turned on, first, as shown in FIG. 2, first, initialization processing such as setting initial values to various registers and the like is performed (step S1), and when the initialization processing is completed, a sound signal from the player 9 is transmitted. It is checked whether or not the song has started depending on whether or not is input (step S2).

When the music starts, a peak detection process (step S3), an average value calculation process (step S4), and a tempo calculation process (step S5), which will be described later, are sequentially performed on the acoustic signal from the player 9 to obtain a tempo. Data is generated and the generated tempo data is output to the external device 10 (step S6).

Hereinafter, peak detection processing, average value calculation processing,
The tempo calculation process will be described in detail.

<Peak Detection Processing> In the peak detection processing,
As shown in FIG. 3, first, "0" is set to the pointer I (step P1), and it is checked whether data (acoustic signal) is input (step P2).

In step P2, when the data is not input, since the data for peak detection is not input, the peak detection processing is ended as it is.

At step P2, when data is input,
A band pass filter (BPF) process is performed to perform an envelope process for taking in only data of a specific frequency band (step P3), and the detected envelope value is stored in the register E0 (step P4).

That is, in the BPF processing, when the data (acoustic signal) as shown in FIG. 4 is input, only the data in the specific frequency band of this data is taken in and the envelope as shown in FIG. 5 is extracted.

It is checked whether the envelope value stored in the register E0 is smaller than the envelope value detected in the previous process and stored in the register E1 (step P5). If NO, that is, the value of the register E1 is registered. When the value is smaller than or equal to the value of E0, it is determined that the envelope value tends to increase, and the peak flag is set to "0" (step P13).
The value of 0 is stored in the register E1 (step P1
1).

Then, it is checked whether peak detection (extraction) has been completed for all data (step P1).
2) If not completed, the process returns to step P2, waits for data input, and performs BPF processing for the next data (step P3).

Similarly, the detected envelope values are sequentially stored in the register E0 (step P4), and the register E0 is stored.
As compared with the value of 1 (step P5), as long as the determination in step P5 is NO, it is determined that the envelope value tends to increase, and the same processing as above is performed.

When the determination in step P5 is YES, the envelope value (E0) detected at this time is
Since the envelope value (E1) detected last time is larger, it is checked whether the peak flag is "1" (step P6).

Now, in the peak flag, step P13
Since "0" is set in the processing of step 1, it is judged that the envelope value E1 detected last time is the peak value, "1" is set to the peak flag (step P7), and the value of the register E1 is preset. It is checked whether or not it is equal to or more than the specified value Eth (step P8).

If NO at step P8, it is determined that the sound is not the tempo detection target sound, and the register E0 is selected.
Is stored in the register E1 (step P11), and it is checked whether the extraction process is completed (step P1).
2).

If YES in step P8, it is determined that the sound is the target of tempo detection, and the register E1 is selected.
And the value of the time counter TM are stored in the peak value area PA of the memory 5 (step P9), the pointer I is incremented by "1" (step P10), and the value of the register E0 is stored in the register E1 ( Step P11).

The time counter TM is incremented by the clock interrupt process which is performed every predetermined time.

That is, the clock interrupt processing is
As shown in FIG. 6, when an interrupt occurs at every predetermined time, "1" is added to the value of the time counter TM, and the addition result is stored in the time counter TM (step T
1).

The above peak detection processing is sequentially repeated,
When all the input data are processed, step T12
The determination is YES, and the process ends.

That is, in the peak detection process, when the acoustic signal as shown in FIG. 4 is input, the BPF process is performed to obtain the envelope as shown in FIG. And
As shown in FIG. 7, the peak value of this envelope is sequentially detected, and as shown in FIG. 8, the detected peak value is compared with the specific value Eth, and a peak value larger than the specific value Eth (see FIG. The peak value E) indicated by the solid line and the generation time at that time are sequentially stored in the peak value area PA of the memory 5.

<Average Value Calculation Processing> When the peak detection processing is completed, next, as shown in FIG. 2, average value calculation processing is performed.

In the average value calculation processing, as shown in FIG. 9, first, the pointer I, the register SE for storing the accumulated peak value and the register M for storing the average peak value are set to "0". Then, the peak value E is read from the peak value area PA of the memory 5 by using the pointer I as an address (step Q2) (step Q2).

The read peak value E is stored in the register SE.
Stored in register SE (step Q
3) The pointer I is incremented by "1" (step Q4), and it is checked whether or not the addition processing has been completed for all the data in the peak value area PA of the memory 5 (step Q5).

When the addition process is not completed for all the data in step Q5, the process returns to step Q2, and the incremented pointer I is used as an address to read the peak value E from the peak value area PA of the memory 5 (step Q2). , Register SE value (step Q
3) Increment pointer I (step Q
4).

That is, all the peak values E detected in the peak detection process are read out from the memory 5, accumulated, and the accumulated result is stored in the register SE.

When the addition process is completed for all the data in step Q5, the value of the register SE is divided by the value of the pointer I, and the division result (SE / I) is stored in the register M (step Q6).

That is, the average value of the peak values E is calculated by dividing the accumulated value of the peak values E by the number of data,
The average value is stored in the register M.

Next, the pointers I and J are set to "0" (step Q7), and the peak value E is read from the peak value area PA of the memory 5 with the pointer I as an address (step Q8).

Then, it is checked whether the read peak value E is equal to or less than the value of the register M, that is, the average value (step Q9). If NO, it is judged that the peak value E is larger than the average value, and the pointer is set. Using the value of J as an address, the peak value E and the time of occurrence of the peak value E are written in the average value area AA of the memory 5 (step Q1).
0).

Next, the pointer J is incremented by "1" and it is checked whether the processing is completed for all the data (step Q12). When the processing is not completed for all the data, the pointer I is set to "1". Incremented only (step Q13) and the process returns to step Q8.

Similarly, using the pointer I as an address, the peak value E is read from the peak value area PA of the memory 5 (step Q8), and it is checked whether the peak value E exceeds the average value (step Q9).

When the average value is not exceeded, the memory 5
It is checked whether or not all the data is completed without writing the peak value E to the average value area AA (step Q12), and when the processing of all the data is not completed, the pointer I is incremented (step S12). Q1
3) Similarly, the next peak value E is read from the memory 5 with the pointer I incremented similarly as an address, and the same processing is performed (steps Q8 to Q13).

Then, in step Q12, when the processing of all data is completed, the average value calculation processing is completed.

That is, in the peak value area PA of the memory 5, as shown in FIG. 10, the detected peak value E and the value (occurrence time) of the time counter TM are stored with the pointer I as an address. In the average value calculation process,
The average value of the peak values E detected in this peak detection process is calculated, and the peak value E exceeding the calculated average value is calculated as shown in FIG.
As shown in 1, the pointer J is sequentially written in the average value area AA of the memory 5 as an address.

<Tempo Calculation Process> When the average value calculation process is completed, next, as shown in FIG. 2, the tempo calculation process is performed.

In the tempo calculation process, as shown in FIG. 12, first, the pointer I, the pointer J, the pointer G, and the pointer Q are set to "0" (step R).
1) Using the pointer J as an address, the generation time is read from the average value area AA of the memory 5 and stored in the register A (step R2).

Further, using the value obtained by adding "1" to the pointer J as an address, the generation time is read from the average value area AA of the memory 5 and stored in the register B (step R).
3).

Then, the generation time of the register A is subtracted from the generation time of the register B, and the generation time interval of the subtraction result (BA) is stored in the register C0 (step R
4) The value of the register C0 is stored in the register C (step R5).

Next, "60" is divided by the generation time interval of the register C, the division result (60 / C) is stored in the register D (step R6), and the value of the register D and the value of the register B are stored. The generation time is stored in the data memory 6 using the pointer G as an address (step R7).

That is, in the present embodiment, since the tempo is the number of quarter notes in one minute, the generation time of the peak value E exceeding the average value stored in the average value area AA of the memory 5 in the above average value calculation processing. The interval is calculated (step R4), and "60" is divided by this occurrence time interval to calculate the tempo (step R6). The calculated tempo and the generation time (value of the register B) which is the basis of this tempo are stored in the data memory 6 (step R7).

Then, "1" is set in the register K (step R8), the value of the register C0 is added to the value of the register C, and the addition result is stored in the register C (step R9).

Next, using the value of the pointer I as an address,
The occurrence time is read from the peak value area PA of the memory 5, stored in the register H (step R10), and the value of the register C is obtained by dividing "15" by the value of the register D and subtracting the value from the value of the register H. It is checked whether it is between the value (H-15 / D) and the value (H + 15 / D) obtained by adding "15" divided by the value of the register D to the value of the register H (step R11). .

That is, in an ordinary performance, the tempo is not exactly constant and is somewhat disturbed, but the disturbance of the tempo during the performance is usually around 15 / tempo. Therefore, in step R11, it is checked whether or not the next peak value E is within the permissible range of the disturbance of the tempo.

For example, as shown in FIG. 13, in the peak value area PA of the memory 5, the peak values p1, p2,
... is stored, and p1, p2, p3, p6, p8, p9 ... Are stored in the average value area AA of the memory 5 as peak values exceeding the average value. 5, the peak value p1 and the peak value p stored at the first address and the next address of the average value area AA
2 occurrence time is read out and these two peak values p1,
The tempo obtained from the value of the register C0, which is the generation time interval between p2, is stored in the register D, and the tempo and the generation time of the peak value p2 are stored in the data memory 6 (steps R1 to R6). The K to be set is set to "1" (step R8), and the time obtained by adding the time interval between the peak value p1 and the peak value p2 to the time interval between the peak value p1 and the peak value p2 is stored in the memory using the pointer I as an address. It is checked whether or not the peak value read from the peak value area PA of No. 5, for example, the peak value p1 is within the above-mentioned permissible range centered on the generation time (step R11).

If NO at step R11, the pointer I is incremented by "1" (step R1
2), the value of register C is the value of register H added with the value of "15" divided by the value of register D (H + 15 /
D), that is, the above-mentioned peak value p
It is checked whether the time interval obtained by subtracting the peak value p1 from 2 is longer than the generation time of the peak value p1 read from the peak value area PA of the memory 5 (step R13),
If YES, it is determined that the peak value (for example, the peak value p1) read from the peak value area PA of the memory 5 is not the peak value to be searched, and step R
Return to 10.

Similarly, using the incremented pointer I as an address, the time of occurrence of the next peak value is read from the peak value area PA of the memory 5 and stored in the register H (step R10), and the read peak value is read in the same manner as above. It is checked whether the occurrence time of is within the allowable time as the tempo (step R11).

If YES in step R11, it is determined that the next peak value is within the permissible range of tempo disorder, the value in register K is multiplied by "60", and the multiplication result is stored in register C. A value obtained by subtracting the value of the register C0 from the value is divided [60 × K (C−C0)] to calculate the tempo, and the calculated tempo is stored in the register D (step R23).

Next, the pointer G is incremented by "1" (step R24), and the tempo and generation time of the register D are stored in the data memory 6 with the incremented pointer G as an address (step R2).
5).

Then, it is checked whether or not the tempo calculation processing has been completed for all the data (step R26). If the processing has not been completed, the pointer I is incremented by "1" (step R27) and step R8.
Return to.

Similarly, the pointer K is set to "1" (step R8), and the result of adding the value of the register C0 to the value of the register C is stored in the register C (step R8).
9) The generation time is read from the peak value area PA of the memory 5 and stored in the register H using the incremented pointer I as an address (step R10).

Similarly to the above, the peak value area PA of the memory 5
If the occurrence time of the peak value read from is within the allowable range of the tempo, similarly, the tempo is calculated and stored in the register D (step R23), the pointer G is incremented (step R24), and then the pointer G is set. The calculated tempo and occurrence time in the register D are stored in the data memory 6 as an address (step R25).

If all the data are not completed in step R26, the pointer I is incremented and similarly, it is checked whether or not the time when the next peak value occurs is within the tempo allowable time (step R8). ~ Step R1
1).

If NO at step R11, the pointer I is incremented in the same manner as described above (step R11).
12), the value of register C is set to the value of register H by “15”
Value obtained by dividing the value by the value of register D (H + 15
/ D) is checked (step R)
13).

If NO in step R13, the pointer K is incremented by "1" to increase the note value (step R14), and it is checked whether or not the value of the pointer K exceeds "4" (step R15). .

If NO in step R15, the pointer I is returned to "0" (step R16), and the process returns to step R9.

Then, the value of the register C is added to the register C0 in which the peak value generation time interval is stored (step R9), and the generation time of the next peak value is stored in the memory 5
Read from the peak value area PA (step R1
0), similarly, it is checked whether the tempo is within the allowable generation range (step R11).

If NO in step R11, the pointer I is incremented (step R12), and similarly to the above, the size of the generation time of the register H is checked (step R13). If NO in step R13, the pointer I is set. Increment K (step R1
4) It is checked whether the value of the pointer K exceeds "4" (step R15).

In step R15, when the value of the pointer K exceeds "4", it is determined that the peak value generation interval exceeds the time interval limit of the note, and the pointer J is incremented by "1" ( At step R17), it is checked whether the value of the pointer J exceeds the total number of data stored in the average value area AA of the memory 5 (step R18).

If NO at step R18, the pointer J is decremented by "1" (step R1).
9) Returning to step R3, the occurrence time is read from the average value area AA of the memory 5 using "J + 1" as an address and stored in the register B (step R3).

The occurrence time stored in the register A is subtracted from the occurrence time stored in the register B, and the result is stored in the register C0 (step R4) and stored again in the register C (step R5). At the position corresponding to the time interval of occurrence of C, the memory 5
It is checked whether or not there is a peak value stored in the peak value area PA.

Then, when there is a corresponding peak value, the tempo calculated from the generation time interval of the register C is stored in the data memory 6 (steps R22 to R25), the pointer I is incremented (step R27), and sequentially. Similarly, the peak value in the peak value area PA of the memory 5 is searched (steps R8 to R11).

When there is no corresponding peak value,
The processing from step R12 to step R19 is sequentially performed.

When the value of the pointer J reaches the total number of data in step R18, the pointer Q is incremented by "1" (step R20) and the pointer J is reset to "0" (step R21).

The value of the pointer J and the value of the pointer Q are added and set to the pointer J (step R22). Then, the process returns to step R2 and the added value of the pointer J is used as an address, and the average value area of the memory 5 is set. The occurrence time interval is read from AA and stored in register A (step R
2).

That is, the tempo is calculated based on the time interval of the peak value generation time stored at the start address and the next address of the average value area AA of the memory 5, and the memory 5 is stored at the generation time corresponding to this tempo. When the peak value is stored in the peak value area PA, the tempo and the time of occurrence are stored in the data memory 6.

If the peak value is not stored in the peak value area PA of the memory 5 at the time of occurrence corresponding to this tempo, it is checked whether the peak value is stored at the time of occurrence of different notes. If there is, the occurrence time and tempo are stored in the data memory 6.

Further, when the search for the time of occurrence of different notes is completed, the same processing is performed based on the time interval of the time of occurrence stored at the start address and the next next address, and this process is fixed at the start address. Then, the same processing is performed with respect to the time interval with the time of occurrence of the peak value of the address which is sequentially added so that the time interval of occurrence is widened.

When the address addition proceeds to the address of the last peak value stored in the average value area AA of the memory 5, the head address is incremented by one and the peak value of this head address and the next address is generated. The time is processed in the same manner as above.

The processing is carried out while carrying out the carry-over of the leading address in sequence, and when the processing of all the data is completed in step R26, the tempo calculation processing is ended.

That is, in the tempo calculation process, for example,
As shown in FIG. 13, regarding the tempo corresponding to the time interval between the peak value p1 and the peak value p2 stored in the average value area AA of the memory 5, the peak value area P of the memory 5
The occurrence time of the peak value p1 to the peak value p9 stored in A is searched for within the allowable range, and if there is a peak value of the corresponding occurrence time, the occurrence time and tempo are stored in the data memory 6. To do.

When all the data stored in the peak value area PA of the memory 5 are processed, the time interval between the peak value p1 and the peak value p3 stored in the average value area AA of the memory 5 is determined. Do the same,
When this process ends, the same process is performed for the time interval between the peak value p1 and the peak value p4.

When the above processing is performed up to the final data stored in the average value area AA of the memory 5, next, the same processing is performed for the time interval between the peak value p2 and the peak value p3.

In this way, the average value area A of the memory 5 is
When all the data stored in A are sequentially processed and the processing of all the data is completed, as shown in FIG.
The data memory 6 stores data on the relationship between the tempo indicated by the horizontal solid line and the time of occurrence thereof.

In this way, the clock signal corresponding to the tempo can be output from the clock generator 8 to the external device 10 based on the tempo data stored in the memory 6, and the external device 10 can be synchronized.

As described above, according to the tempo data generation device 1, the level of the specific frequency band of the audio signal input from the player 9 is extracted by the frequency filter 3, and the audio signal of a predetermined level or higher is extracted from the extracted levels. The peak value and the generation time are sequentially stored in the peak value area PA of the memory 5. The average value of the stored peak values is calculated by the clock sequence composer 7, and among the peak values of the levels stored in the peak value area PA of the memory 5, the peak value exceeding the calculated average value is the average value of the memory 5. Store in area AA.
Then, two peak values in the average value area AA of the memory 5 are selected in the order of addresses, and there is another level stored in the peak value area PA of the memory 5 at every occurrence interval of the selected two levels. Search whether or not. When a peak value of another level exists at each occurrence interval, tempo data is generated based on the occurrence interval, and when no other level exists at each occurrence interval, the peak value area PA of the memory 5 is generated. Select another level stored in and search in the same manner.

Therefore, the tempo can be accurately detected from the acoustic signal of the live performance of the player 9 such as a musical instrument, and tempo data can be generated. Based on this tempo data, the external device 10 such as a sequencer can be synchronized. Can be taken.

[0102]

According to the present invention, it is possible to accurately detect the tempo from an acoustic signal of a live performance of a musical instrument or the like and generate tempo data, and based on the tempo data, an external device such as a sequencer can be used. Can be synchronized.

[Brief description of drawings]

FIG. 1 is a block configuration diagram of an embodiment of a tempo data generation device of the present invention.

FIG. 2 is a flowchart showing main processing of the tempo data generation device of FIG.

FIG. 3 is a flowchart showing detailed processing of peak detection processing of FIG.

FIG. 4 is a waveform diagram of an input acoustic signal.

5 is a waveform diagram of the envelope signal of the acoustic signal of FIG.

6 is a flowchart showing a clock interrupt process by the tempo data generation device of FIG.

FIG. 7 is a diagram illustrating a process of detecting a peak value of the envelope signal of FIG.

FIG. 8 is a diagram illustrating a process of extracting a peak value that exceeds a specific value Eth among the peak values in FIG. 7.

FIG. 9 is a flowchart showing detailed processing of the average value calculation processing of FIG.

10 is a diagram showing a storage state of a peak value E and an occurrence time in a peak value area PA of the memory of FIG.

11 is a peak value E of the average value area AA of the memory of FIG.
The figure which shows the storage state of the occurrence time.

FIG. 12 is a flowchart showing detailed processing of the tempo calculation processing of FIG.

13 is a diagram showing a relationship between a peak value E stored in an average value area AA of the memory 5 of FIG. 1, an average value and an occurrence time.

14 is a diagram showing the relationship between the tempo calculated from the peak value E of FIG. 13 and the time of occurrence.

[Explanation of symbols]

 1 Tempo Data Generator 2 CPU 3 Frequency Filter 4 Volume Level Filter 5 Memory 6 Data Memory 7 Clock Sequence Composer 8 Clock Generator 9 External Equipment

Claims (4)

[Claims] 1. Extraction means for extracting a level of a specific frequency band of an input acoustic signal, storage means for sequentially storing a peak value and an occurrence time of the level extracted by the extraction means, and the storage means Calculating means for calculating an average value of the peak values of the level thus selected, and two peak value levels exceeding the average value calculated by the calculating means are selected from among the peak values of the levels stored in the storage means. Selecting means, searching means for searching whether or not another level stored in the storing means exists at every occurrence interval of the two levels selected by the selecting means; and the searching means for another level When it is detected that each of the occurrence intervals is present, tempo data generating means for generating tempo data based on the occurrence interval; And without upon detection, tempo data generating apparatus being characterized in that and a re-selecting means for selecting the other levels from the storage means to the selection means. 2. The tempo data generation according to claim 1, wherein the storage means stores only a peak value of a level having a peak value of a specific value or more among the levels extracted by the extraction means. apparatus. 3. An extraction process for extracting a level of a specific frequency band of an input acoustic signal, a storage process for sequentially storing a peak value and an occurrence time of the level extracted by the extraction process, and a storage process for the storage process. A calculation process for calculating the average value of the peak values of the selected levels, and two peak value levels exceeding the average value calculated by the calculation process are selected from the peak values of the levels stored by the storage process. A selection process; a search process for searching whether or not another level stored by the storage process exists at every two-level occurrence interval selected by the selection process; and another level in the search process. When it is detected that each is present at each of the occurrence intervals, tempo data generation processing for generating tempo data based on the occurrence interval; When it is detected that does not exist in every generation interval, the tempo data generating method, characterized in that it and a re-selecting process to select another level stored in the storing process by the selection process. 4. The tempo data generation according to claim 3, wherein the storage processing stores only the peak value of a level having a peak value of a specific value or more among the levels extracted by the extraction processing. Method.