JP2794561B2 - Waveform data generator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a waveform data generator for use in, for example, an electronic semester sound source device, and more particularly to a device for storing different musical tones in a waveform memory and simultaneously reading a plurality of waveforms to generate a synthesized tone. I understand.

[0002]

2. Description of the Related Art It has been known that a plurality of waveforms are stored in a waveform memory and read out at the same time to change the mixing ratio in order to change the timbre of a musical tone in accordance with a volume or a tone range. Such techniques, for example, in keyboard instruments,
JP-A-54-39615 and JP-A-54-39616, and for rhythm instruments, JP-A-61-205997.
Each is disclosed. However, in these devices, the memory capacity was increased, and a very large tone change was not obtained.

[0003]

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and has a waveform memory means for storing a plurality of waveform data sampled at mutually different sampling frequencies, and an address generator. Reading means for reading the plurality of waveform data from the waveform memory in accordance with the generated address; performing interpolation on at least one waveform data read from the reading means; Interpolating means for obtaining a plurality of waveform data having the same frequency, and mixing means for mixing a plurality of waveform data including the interpolated waveform data in accordance with a mixing ratio, wherein sampling is performed at different sampling frequencies. It is characterized by generating waveform data that mixes multiple waveform data. That.

[0004]

FIG. 1 shows a circuit configuration of a tone generator of an electronic musical instrument according to an embodiment of the present invention, which generates a musical tone under the control of a microcomputer 1 (CPU). The CPU 1 operates in accordance with a program stored in the ROM 3. The CPU 1 stores tone color parameters set by the operation panel 4 in the RAM 2, and when performance information is generated, the address generators 5 and 8 store the tone data in the waveform memories 6 and 9. At the same time as instructing the generation of an address for reading the waveform, the envelope circuits 7 and 10 also generate the envelope waveform in order to add a predetermined envelope waveform to the tone waveform read from the waveform memories 6 and 9, respectively. Control the envelope circuits 7, 10.
It is known that the performance information is generated from a keyboard, a sequencer, an automatic rhythm performance device or the like (not shown) to drive a sound source having the configuration shown in FIG. Various types of envelope circuits are also known as the envelope circuit, and a detailed description thereof will be omitted.

[0005] These enveloped waveforms are:
The signals are synthesized by an adder 11, converted into an analog signal by a D / A converter 12, and output.

The waveform memories 6 and 9 store musical sounds obtained by dividing musical tones for each frequency band and sampling waveforms. As a method of band division, as shown in FIG. 2, there is a method in which an analog signal is filtered and divided, and the output of each filter is A / D converted and stored in each waveform memory. As shown in FIG. 3, there is also a method in which a musical tone is A / D converted once, stored in a waveform memory, digitally filtered, and stored in a waveform memory for each band.

[0007] As can be seen from the sampling theorem, it is necessary to increase the sampling frequency when digitizing a high frequency band. However, in the case of a low frequency band, even if the sampling frequency is low, no alias (aliasing noise) is generated, so that the memory can be reduced.

[0008] In the case of ordinary musical tones, since harmonics are included up to about 20 KHz, the sampling frequency is 40
KHz or more is required. On the other hand, the frequency at which a musical sound is divided into a plurality of frequency bands differs depending on the timbre, but if the musical sound is divided at 2 KHz, it is sufficient to sample the sound at a low frequency band at 5 KHz, and the memory capacity is significantly reduced.

In the case of a piano sound or a drum sound, such as a piano sound or a drum sound, whose amplitude rapidly increases at the start of sound generation and then gradually attenuates, many harmonics are generated at the start of sound generation. Broadband sounds disappear quickly as they decay. For example, when schematically shown in FIG. 4, (a) shows the envelope of the entire waveform, (b) shows the high-band envelope obtained by frequency division, and (c) shows the low-band envelope. Therefore, the storage time at a high sampling frequency is short.

In order to simultaneously read and synthesize waveform data having different sampling frequencies from the waveform memory, it is necessary to convert the sampling frequencies so that they match. Also, by changing the speed at which the musical tone waveform stored in the waveform memory is read out, the pitch of the musical tone can be changed. For such a purpose, conversion of a sampling frequency is performed. In Japanese Patent Publication No. 59-17838,
One such method is disclosed.

In this method, the sampling frequency to be read is kept constant, and the step width of the address to be repeatedly read is changed. The outline is shown below. FIG.
As shown in (b), the waveform memory stores amplitude values Y _n and Y _{n + 1} corresponding to the sample points X _n and X _{n + 1} ,
Assuming that the point f between them (the interval between X _n and X _{n + 1} is 1, f
Is obtained by interpolation from the preceding and following sampling values.

FIG. 1A is a circuit configuration diagram for performing interpolation.
The CPU 1 supplies a step address corresponding to the sampling frequency and pitch at the time of writing to the waveform memory to be read to the step address register 11. The increment address is a decimal value, and is accumulated by the accumulator 12 at each timing corresponding to the read sampling frequency.
This accumulated value is the address of the waveform to be read, but contains the decimal value.
An interpolation operation is performed from the waveform amplitude value near the address indicated by the integer value. The amplitude value interpolated in this way is added with other waveform amplitude values after the envelope is given, and D / D
A conversion is performed.

As an application example of this sound source device, for example,
The two piano sounds are divided into bands, the high-frequency waveform and the low-frequency waveform are stored respectively, and the mixing ratio (the respective envelope waveforms) is changed according to the speed at which the key is pressed, resulting in different pitches As for the key, by changing the readout speed and the mixture ratio, the timbre can be changed according to the pitch and volume.

[0014] Further, by combining different musical instruments having different frequency bands, it is possible to obtain a musical tone which has not been obtained in the past. For example, in the case of a percussion instrument sound, a low range of a bongo sound and a high range of a snare drum sound are synthesized, and if the pitch is changed, a completely new percussion sound can be synthesized.

In the case of a percussion instrument such as a tom-tom, which has substantially the same timbre and a different pitch depending on the caliber of the tom, storing one tom in the waveform memory and changing the reading speed will result in an attack. Change. That is, if reading is performed late, the pitch becomes lower, and at the same time, the attack is impaired. However, as in the present invention, the tom sound is divided by a frequency band, and the sound in the low band changes the readout speed according to the caliber (pitch) of the tom, while the sound in the high band is
If the reading speed is slightly changed in accordance with the pitch and the two sounds are combined, the pitch will change and the attack feeling will not be impaired.

[0016]

As described above, according to the present invention, a plurality of waveform data sampled at different sampling frequencies are converted into a plurality of waveform data having the same sampling frequency, and the plurality of waveform data having the same sampling frequency are obtained. Can be generated, so that the sampling frequency of at least one waveform data can be lower than the sampling frequency of the other waveform data. Therefore, the capacity of the waveform memory means for storing a plurality of waveform data can be reduced without deteriorating the sound quality of the reproduced waveform.

[Brief description of the drawings]

FIG. 1 is a circuit configuration diagram of a waveform data generation device according to the present invention.

FIG. 2 is a circuit configuration diagram in which an analog signal is frequency-divided and stored in a waveform memory.

FIG. 3 is a circuit configuration diagram in a case where an analog signal is converted into a digital signal and then frequency-divided.

FIG. 4 is an envelope of each band when frequency division is performed.

FIG. 5 is a circuit configuration diagram when a waveform is read.

[Explanation of symbols]

 5 8 Address generator 6 9 Waveform memory 7 10 Envelope circuit 11 Adder 13 Interpolation circuit

Claims (8)

(57) [Claims] 1. A waveform memory means for storing a plurality of waveform data sampled at different sampling frequencies, and an address generating means for generating an address, wherein the plurality of waveform data are stored in the waveform memory according to the generated address. Reading means for reading the data; interpolating means for performing an interpolation operation on at least one waveform data read by the reading means to obtain a plurality of waveform data having the same sampling frequency; and a waveform subjected to the interpolation operation Mixing means for mixing a plurality of waveform data including data in accordance with a mixing ratio, and generating waveform data in which a plurality of waveform data sampled at different sampling frequencies are mixed. 2. A waveform data generating apparatus according to claim 1, wherein said interpolation means performs an interpolation operation in accordance with the waveform data read by said reading means and an address generated by said address generation means. apparatus. 3. The waveform data generating device according to claim 2, wherein said reading means includes an address generating means for generating an address which changes at a step address corresponding to a sampling frequency at the time of writing in said waveform memory. A waveform data generator for reading out the waveform data from the waveform memory in accordance with the generated address. 4. The waveform data generating device according to claim 3, wherein said address generating means of said reading means generates an address including an integer value and a decimal value. 5. The waveform data generating device according to claim 3, wherein said reading means reads out waveform data near an address generated by said address generating means. 6. A waveform data generating apparatus according to claim 5, wherein sample values before and after an address generated by said address generating means are read. 7. The waveform data generating apparatus according to claim 6, wherein said interpolation means obtains an amplitude value between the sampling values before and after read by said reading means. 8. The waveform data generating device according to claim 4, wherein said reading means reads sample values before and after an address generated by said address generating means, and said interpolation means reads said sample values by said reading means. A waveform data generating apparatus for interpolating an amplitude value corresponding to a decimal value generated by the address generating means from sampled values before and after the output.