JP3525482B2 - Sound source device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention divides tone waveform data into time-series frames and expresses the frequency spectrum in each frame as frame data consisting of a predetermined number of peak data (frequency data, intensity data). The present invention relates to a sound source device of an analysis and synthesis system that generates a musical tone signal by generating peak sounds represented in time-series frame data and performing additive synthesis.

[0002]

2. Description of the Related Art The applicant of the present invention has previously filed Japanese Patent Application No. 3-27.
No. 1183 and Japanese Patent Application No. 3-282988 propose sound sources for analysis and synthesis. This analysis / synthesis sound source has the good reproducibility of the original sound, which is the characteristic of the waveform memory type sound source, and the good waveform processability, which is the characteristic of the fundamental wave synthesis type sound source such as the FM sound source. It has both. That is, almost the same waveform as the original sound can be reproduced by generating basic waveform data (sin waves) corresponding to a plurality of peaks (about 128) in the frequency spectrum of the audio signal and adding and synthesizing the basic waveform data. In addition, the amount of data can be reduced as compared with the waveform memory type sound source. Further, since the features are extracted as peak data (frequency data, intensity data) from the original sound waveform, the tone color can be easily processed by processing the features.

[0003]

In such a sound source of the analysis and synthesis system, when an intermediate tone color of a plurality of tone colors (voice data) is generated or when a tone color being sounded is changed to another tone color, Frequency data and intensity data are linearly interpolated between different timbres to generate an intermediate value, and a tone waveform is formed from the data. However, such a linear interpolation method is not optimal as a method for generating an intermediate tone color by such a sound source of the analysis and synthesis system. For example, when linear interpolation is performed in a part with a non-linear character such as the attack part, or between a clear tone with little temporal variation and a noise-like tone with many temporal parameter variations. When linear interpolation is performed, the intermediate tone color may be a muddy tone color that is completely different from the original tone, and the inharmonicity may increase.

Further, a huge amount of calculation is required to perform linear interpolation on all peak data, and a large-scale calculation is required to perform this in real time every time frame data is output in time series. I needed a section.

The present invention relates to a sound source of analysis and synthesis system,
Frequency data that constitutes frame data of multiple tones,
By combining intensity data, it is an object to provide a sound source device that can generate a natural intermediate tone color. Furthermore, by gradually replacing frequency data and intensity data that form frame data of a plurality of tone colors, It is an object of the present invention to provide a sound source device capable of smooth transition of tone colors and having simple processing.

[0006]

SUMMARY OF THE INVENTION The invention of claim 1 of this application, the frequency spectrum of the tone waveform data of a predetermined number
Is divided into channels to extract the peak for each channel, the frequency data of the peak of each said respective channels, the intensity data and frame data, arranges the frame data in the time series to mute the sound of the musical tone waveform data Synthesizing to form a tone signal by sequentially inputting frame data and a storage means storing a plurality of constituted voice data, and generating and synthesizing a fundamental wave indicated by frequency data and intensity data of each peak data. means and reads out the respective frame data from a plurality of voice data stored in the storage means at the same time, the frequency data of <br/> plurality of frame data for each channel, or intensity data
Select one of the Luo optionally, characterized by comprising an intermediate tone generating means for input to the combining means to configure a new one-frame data by combining this,. This out
According to the invention of claim 2, the storage means is externally input.
Data for adjusting the weight of the manipulated variable data
Store coefficient data that changes with time for each channel.
The synthesizing means uses the input operation amount data as the correlation
Intensity data of each channel with values weighted by several data
It is characterized by controlling.

According to the third aspect of the present invention, a predetermined number of peaks are extracted from the frequency spectrum of the musical tone waveform data, and the frequency data and intensity data of the predetermined number of peaks are used as frame data. A storage means for storing a plurality of voice data arranged in time series from sounding to muffling of waveform data and frame data are sequentially input to generate a fundamental wave indicated by frequency data and intensity data of each peak data. In a sound source device including a synthesizing unit for generating and synthesizing a tone signal, a frame for simultaneously reading frame data from two voice data stored in the storing unit and outputting the frame data to the synthesizing unit. The contents of the frequency data and intensity data that make up the data are changed from those of the first frame data to the second Characterized in that a tone color changing means Yuku replaced gradually for each frame to that of the frame data.

In the present invention, frame data is simultaneously read from a plurality of voice data, and frequency data and intensity data of the plurality of frame data are combined to form one new frame data. The combination method is arbitrary, but a method of comparing specific peak data (a set of frequency data and intensity data corresponding to one peak) between a plurality of frame data and selecting the peak data with the larger intensity data. Alternatively, there is a method of alternately selecting peak data from a plurality of frame data.
The new frame data thus configured is input to the synthesizing means. Since the peak of the frequency spectrum of the original sound is included as data in the new frame data, the timbre is similar and the processing is easy.

Further , according to the present invention, the frame data is simultaneously read from the two voice data, but first, the first frame data is directly input to the synthesizing means. In the next frame, part of the first frame data is replaced with that of the second frame data. Further, in the next frame, more data is replaced with that of the second frame data. In this way, finally, all of them become the second frame data. As a result, the voice data (tone color) can be smoothly and easily transferred. The replacement method may be any method as in the first aspect of the invention. For example, there is a method of replacing higher harmonic overtones or a method of alternating higher and lower harmonics.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIGS. 6 to 9 are views for explaining voice data used in a sound source device according to an embodiment of the present invention. Figure 6
FIG. 4 is a diagram for explaining a method of dividing one musical tone waveform into frames. FIG. 7 is a diagram illustrating a filter configuration for analyzing the frequency characteristic of each frame. FIG. 8 is a diagram for explaining the method of data extraction in each channel. FIG. 9 is a diagram showing a structure of voice data generated by the above method.

The musical tone signal is 32 kHz from sound generation to mute.
Are sampled at the frequency of and are made into digital musical tone waveform data. Consecutive 204 from this tone waveform data
Eight samples are extracted to make one frame. This one frame of musical tone waveform data is subjected to Fourier analysis to obtain a frequency spectrum, which is divided into 128 channel frequency bands by a filter array as shown in FIG. For each filter of the filter array, the peak frequency data and its intensity data in that frequency band are extracted (FIG. 8).
reference). In this way, 128 peak data (frequency data, intensity data) are extracted per frame. This is frame data. The frame is cut out from the musical tone waveform data by overlapping every 6 samples (2 ms) from the beginning (during sound generation) to the end (silence) of the musical tone waveform data as shown in FIG. Assuming that the number of frames cut out from the sounding of the tone waveform data to the mute is n, the voice data is data consisting of n frame data.

When synthesizing a tone signal using such voice data, the frame data is read out and the basic waveform data of the frequency indicated by the frequency data of each peak data is converted into the intensity indicated by the intensity data. It is generated and these are added and synthesized. 6 to read frame data
The sampling may be performed at every sampling timing (2 ms), and each data may be linearly interpolated during 5 sampling timings in between. Since this interpolation is processing within the same voice data, it does not become anharmonic. By doing so, it is possible to synthesize a musical tone signal having a quality close to that of the sampling sound source.

FIG. 1 is a block diagram of an electronic musical instrument to which the sound source device of the present invention is applied. Further, FIG. 2 is a timing chart of various data and signals transmitted and received in the electronic musical instrument. This electronic musical instrument forms a musical tone signal using the voice data. The analysis unit 6 extracts the voice data, and the extracted voice data is stored in the voice data memory 10 of the sound source unit 5.

This electronic musical instrument is provided with a performance operator 1, an auxiliary operator 2, and a tone color designating operator 3 as operators operated by the player. As the performance operator 1, for example, a keyboard such as a keyboard with a touch strength sensor that can specify a pitch and its strength is used. The auxiliary operator 2 is, for example, an operator such as a modulation wheel,
It controls the effect added to the tone signal and the change in tone color. The tone color designating operator 3 is, for example, a tone color selection switch or the like, and selects and designates a tone color of a musical tone generated by the sound source unit 5. This tone designating operator 3
The tone color that can be designated is not only a tone color obtained by directly combining the voice data stored in the voice data memory 10 but also an intermediate tone color of a plurality of voice data.

The operation contents of the performance operator 1, the auxiliary operator 2, and the tone color selection operator 3 are input to the control section 4. Control unit 4
Generates various data according to the operation contents of these operators and outputs it to the sound source unit 5. The data output from the control unit 4 to the sound source unit 5 includes a key-on pulse KON, a key code KC, touch intensity data TCH, and tone color designation data T.
C, the operation amount data DCONTT of the auxiliary operator 2, and the count value FRMCN of the frame counter.
Various timing signals including T are output.

The sound source unit 5 includes a voice data memory 10, a frame data read control unit 11, and a frame data processing unit 1.
2, a frame data interpolating unit 13, a synthesizing unit 14, and a waveform processing unit 15. The voice data memory 10 stores a plurality of voice data shown in FIG. The voice data is read by the frame data read control unit 11. Tone color designation data TC is input to the frame data read control unit 11, and the frame data read control unit 11 determines voice data to be read from the voice data memory 10 based on the tone color designation data TC. When the tone color designated by the tone color designation data TC is an intermediate tone color, the frame data read control unit 11 simultaneously reads out a plurality of voice data for generating the tone color and designates the tone color during the tone generation. When the data TC is switched, two voice data are read at the same time in order to change the tone color by overlapping.

In the frame data processing unit 12, the frame data read control unit 11 uses the frame data Va (FRQa, i, MAGa, i) and Vb (FRQb, FRQb,
i, MAGb, i) (i = 0 to 127) are read out at the same time, new frame data (FRQi, MAG) is created by combining these frame data.
i) is generated and output to the next frame data interpolating unit 13. When an intermediate tone color is generated, any peak data is selected from both frame data for each channel of 0 to 127 based on the operation amount data of the auxiliary operator 2, and this is combined to generate new frame data. To do. Also, when changing the timbre by overlapping it, replace the peak data of the frame data of the original timbre with the peak data of the frame data of the new timbre,
When all the peak data have been replaced with the tone color of the change destination, the reading of the tone color of the change source is stopped and the tone color change is completed.

The frame data interpolating unit 13 supplies the frame data (FRQi, MAG) every 6 sampling clocks.
i) is input. The frame data interpolating unit 13 performs a linear interpolation calculation between the newly input frame data and the immediately preceding frame data to generate frame data at 5 sampling timings between the frame data and the frame data. Therefore, the frame data interpolating unit 13 starts outputting the interpolated frame data (IFRQi, IMAGi) at each sampling timing when the frame data of two frames is input from the start of sound generation (see FIG. 2).

The synthesizing section 14 is a frame data interpolating section 13
Frame data (IFRQi, IMAG
It is a circuit for synthesizing tone waveform data based on i).
The frame data is composed of 128 pieces of peak data, but the basic waveform data of the frequency specified by the frequency data of each peak data is generated with the strength specified by the strength data of the peak data. The generated 128 basic waveform data are added and synthesized to form musical tone waveform data WAVE. The musical tone waveform data WAVE synthesized by the synthesizer 14 is input to the waveform processor 15. The waveform processing unit 15 gives various effects such as reverberation to the waveform data WAVE. The type and degree of the effect to be given are determined based on the key code TC, the touch intensity data TCH, and the tone color designation data TC input from the control unit 4. Waveform processing unit 1
The tone waveform data to which the effect is added in 5 is output data OU
It is output as T. The output data OUT that has been output is analogized, amplified, and emitted from a speaker or the like.

FIG. 3 is a block diagram of the frame data processing unit 12. The frame data processing unit 12 receives the frame data Va (FRQa, MAGa) of two voice data which are simultaneously input from the frame data read control unit 11,
This is a circuit that forms one new frame data from Vb (FRQb, MAGb). The configuration of new frame data is performed by selecting one of the two peak data for each channel of 0 to 127. The peak data is selected by the intensity data MAGa, i, MAGb, i (i = 0 to 1) of both frame data.
27) is multiplied by a predetermined value, and the peak data having the larger value of the multiplication result is selected. The predetermined value means the coefficients MCa, i, MCb, which are read from the intensity data memory 20 in the operation amount data DCONT of the auxiliary operator 2.
i (i = 0 to 127).

Frequency data FRQ of frame data Va
a is input to one input terminal 0 of the selector 25. The intensity data MAGa of the frame data Va is input to the multiplier 23. A coefficient value obtained by multiplying the operation amount data DCONT of the auxiliary operator 2 by the intensity coefficient MCa read from the intensity coefficient memory 20 is input to the multiplier 23.
This multiplication is performed by the multiplier 21. The intensity data MAGa ′ multiplied by the coefficient value by the multiplier 23 is the comparator 26.
Input terminal A and input terminal 0 of the selector 27.

On the other hand, the frequency data FRQb of the frame data Vb is input to the input terminal 1 of the selector 25. The intensity data MAGb of the frame data Vb is input to the multiplier 24. A coefficient value obtained by multiplying the operation amount data DCONT of the auxiliary operator 2 by the intensity coefficient MCb read from the intensity coefficient memory 20 is input to the multiplier 24.
This multiplication is performed by the multiplier 22. The intensity data MAGb ′ multiplied by the coefficient value by the multiplier 24 is the comparator 26.
Input terminal B and input terminal 1 of the selector 27.

Here, the comparator 26 outputs "1" when the data at the input terminal B is larger than the data at the input terminal A. This signal is supplied to the selectors 25 and 27 as a select signal. When "1" is input to the select terminals of the selectors 25 and 27, the data on the input terminal 1 side is output from the output terminal. When "0" is input to the select terminals of the selectors 25 and 27, the data on the input terminal 0 side is output from the output terminal. Since the data of the frame data Va is input to the input terminal 0 and the data of the frame data Vb is input to the input terminal 1 of both the selectors 25 and 27, the frame data Va is selected and output when MAGa ′ ≧ MAGb ′. And then MAGa ′ <MAG
When it is b ', the frame data Vb is selectively output.

This circuit is provided in parallel for 128 channels from channel 0 to channel 127, and the comparison / selection process is performed independently for each channel. Further, the DCONTT is set so that 1 is output when the auxiliary operator 2 is in the initial state, and the multipliers 23, 24 are no matter how the auxiliary operator 2 is operated.
The value input to is never 0.

With this circuit, the frame data with the larger MAG is selected in real time for each channel, an intermediate tone color can be generated, and the tone color can be adjusted by the controller.

The coefficients MCa, i, MCb, i stored in the intensity coefficient memory 20 are data for adjusting the action of the operation amount data DCONTT of the auxiliary operator 2.
The value is stored for each frame and each channel.
The coefficient value changes in time series, changes for each channel, and is different for each voice data Va, Vb.
For example, if the coefficient value at the time of attack is increased, the tone color at the time of attack can be greatly controlled by the auxiliary operator 2. In addition, by increasing the coefficient value of the high-order overtone, it can function as an equalizer or a filter.

FIG. 4 is a block diagram showing another embodiment of the frame data processing section. The frame data processing unit of this embodiment uses FRQai and FRQbi (i = 0 to 127).
Is provided and a selector 32 to which MAGa and MAGb are input, and the selection combination control unit 30 supplies a select signal to these selectors 31 and 32. The selection combination control unit 30 includes tone color designation data TC and operation amount data DC of the auxiliary operator 2.
Count value FRMCN of ONT and frame counter
T is input. The selection combination control unit 30
Based on these data, a select signal is output for each frame to the selector of each channel. Selector 3
It is assumed that 1 and 32 are provided in parallel for 128 channels. Further, as the intensity data MAGi, it is possible to select data MAGa / b, j (j ≠ i) of another channel. That is, the intensity data of 127 channels is set to 0
Processing such as outputting as channel intensity data is possible. The selection combination control unit 30 instructs the selector 32 to perform such selection. This instruction data is input to the selector 32 via the PORT terminal.

The selection combination control unit 30 stores a plurality of types / procedures for creating one new frame data from two input frame data and a plurality of peak data selection rules.

An example thereof is shown in FIG. FIG. 9A shows a procedure for changing the timbre by gradually changing the frame data Va to the frame data Vb. When the timbre change data TC instructed to change the timbre, the data of the frame data Va was output on all the channels until then. However, the frame data Vb is sequentially output from the higher harmonic overtone channel (channel 127) each time one frame is advanced. I am trying to replace it with the peak data of. As a result, the frame data output in 127 frames (= 254 ms) is completely switched from the frame data Va to the frame data Vb. Even in this case, since it is not necessary to perform interpolation calculation or the like on each peak data, the processing is easy, and since an intermediate tone color completely different from both tone colors a and b is not generated, it becomes anharmonic. There is no.

In the example of FIG. 3B, the frame data Va
Shows a procedure for changing only the intensity data to that of the frame data Vb to generate a new timbre while producing a sound. Further, in this example, the intensity data of the high-order overtone of the frame data Vb is replaced with the intensity data of the low-order overtone. That is, first MAGb, 127 is MA
Set as MAG0 instead of Ga, 0, then MAGb, 12
Set 6 as MAG1, and finally set MAGb, 0 to MAG
It is set as 127. In this way, a new musical tone can be created by diverting the intensity data of the high-order overtone as the intensity data of the low-order overtone.

As the tone color changing method, various methods other than the method shown in FIG. 5 can be adopted. In the above two examples, one voice data (tone color) is gradually changed to another voice data in time series, but the tone color change may be stopped at an intermediate tone color. Alternatively, the preset intermediate tone color may be immediately changed. Further, as in the frame data control circuit of FIG.
It may be changed in real time according to the input of CONT.

In the above embodiment, the case where the intermediate tone color is generated or the tone color is changed between completely different tone colors has been described, but the present invention can be similarly applied to the multi-sampling tone color. That is, the present invention can be applied to tone color change between two frame data having the same tone color but different tone pitch and touch strength, and generation of an intermediate tone.

In the above embodiment, the sampling frequency is 32 kHz and the number of frequency division channels is 128.
However, the present invention is not limited to this specification.

Further, the sound source system shown in the embodiment is
It can be realized by hardware or by a configuration including a computer including program processing or a DSP. Of course, a hybrid system in which these are appropriately combined may be used.

[0035]

As described above, according to the present invention, since the intermediate tone color is generated by combining the data from the frame data of the plurality of voice data, the process is easier than the interpolation calculation. , It is possible to generate a harmonious timbre having the characteristics of both timbres.

Further, since the timbre is changed by gradually replacing the data from the first frame data to the second frame data, the processing is facilitated and unnatural intermediate sounds are passed through. And there is no new tone.

[Brief description of drawings]

FIG. 1 is a block diagram of an electronic musical instrument that is an embodiment of the present invention.

FIG. 2 is a timing chart of various signals of the electronic musical instrument.

FIG. 3 is a diagram showing a configuration of a frame data processing unit of the electronic musical instrument.

FIG. 4 is a diagram showing another embodiment of the frame data processing unit.

FIG. 5 is a diagram showing an example of a frame data changing procedure of the frame data processing unit.

FIG. 6 is a diagram for explaining a structure of voice data used in the electronic musical instrument, and is a diagram for explaining a method of dividing one musical tone waveform into frames.

FIG. 7 is a diagram illustrating a filter configuration for analyzing frequency characteristics of each frame.

FIG. 8 is a diagram illustrating a data extraction method in each channel.

FIG. 9 is a diagram showing a structure of voice data generated by the above method.

[Explanation of symbols]

10-Voice data memory 12-frame data processing unit

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G10H 1/00-7/12

Claims (3)

(57) [Claims] 1. A musical tone waveform own the frequency spectrum of the data <br/> divided into channels of a constant number to extract a peak for each channel, the peak of the frequency data for each said respective channels,
The intensity data is used as frame data, and a plurality of storage means for storing a plurality of voice data configured by arranging the frame data in time series from sounding of the musical tone waveform data to muffling, and the frame data are sequentially input, and each peak data is stored. read frequency data, a synthesizing means for forming a tone signal by combining by generating a fundamental wave indicated by intensity data, each frame data from a plurality of voice data stored in the storage means at the same time, each Channel
Frequency data of the plurality of frame data in each Le, select one from the intensity data of an arbitrary, an intermediate tone generating means for input to the combining means to configure a new one-frame data by combining this, A sound source device characterized by being provided . 2. The operation amount stored in the storage means is input externally.
Data for adjusting the weight of the data, and for each channel
Separately, the coefficient data that changes with time is stored, and the synthesizing unit converts the input operation amount data into the coefficient data.
The intensity data of each channel is controlled by the value weighted by the data.
The sound source device according to claim 1. 3. A predetermined number of peaks are extracted from the frequency spectrum of the musical tone waveform data, and frequency data and intensity data of the predetermined number of peaks are used as frame data. A storage means for storing a plurality of voice data arranged in series and frame data are sequentially input, and a fundamental wave indicated by the frequency data and intensity data of each peak data is generated and synthesized to generate a musical tone. In a sound source device comprising a synthesizing unit for forming a signal, frequency data constituting frame data to be simultaneously read out from the two voice data stored in the storage unit and output to the synthesizing unit,
A sound source device comprising a tone color changing means for gradually replacing the content of intensity data from that of the first frame data to that of the second frame data for each frame.