JPS62182800A - Acoustic signal analyzer/synthesizer - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application The present invention synthesizes a sound i9 signal inputted from the outside by attaching a subtone scale obtained by Fourier vector 1 to 1 to Le analysis.

The present invention relates to an 8-basic analysis and synthesis device that generates musical tones, and is particularly concerned with improving the square root calculation circuit that performs the square root calculation required when calculating the amplitude coefficient of a frequency spectrum.

B. Summary of the Invention The present invention provides a device for creating musical tones by adding a scale to an input sound using Fourier spectrum analysis and harmonic addition synthesis.
In particular, J3 is used in the square root calculation circuit necessary to calculate the amplitude coefficient of the frequency spectrum, and extracts the exponent of the input data X = (M/2N)2 expressed as bit data and the bit data corresponding to M2. A 2-bit shift converter, a ROM table that provides bit data of the square root M corresponding to a predetermined high-order effective bit of the bit data corresponding to M2, and a ROM table that provides bit data of the square root M by -NN pit shift 1-shi/
Since a 1-bit shift converter that provides 2N bit data is provided, square root PI4SS for obtaining M/2N can be executed at high speed and with a small storage capacity I11.

C. Conventional technology Conventionally, acoustic signals input from the outside (for example, the sound of a bell or the sound of an animal) are subjected to Fourier spectrum analysis, the main spectral components are extracted, a scale is attached to this, and overtones are added and synthesized to produce musical sounds. The device was known. By the way, Fuli.”
The Fourier class expansion formula revealed by spectral analysis is (1)
It is expressed by the formula.

Q (-t) =zaO+Σ2a cos2 π
nt/T 1-, ;bo s+n2 πn1:/T
・(1)ll n However, the king is the period.

For the sake of later spectral synthesis (1)
It is necessary to convert the expression into the expression (2>).

a (-) ) = ao +, 5A, cos (2
πnt/T-〇,)...(2)A
is the amplitude coefficient of the frequency spectrum. To find this A, the square root of X=a2+b2 ffn
It is necessary to calculate n n. Note that the phase component θ. In some cases, constants may be introduced to simplify the calculation without performing any calculations.

Below, a square root q circuit is provided for this purpose.

D Problems to be solved by the invention However, the conventional square root calculation circuit executes *a using an approximation formula that gives the square root, so it is complicated CP
This placed a burden on tJ and was an obstacle to high-speed processing. There is also a calculation using a square root conversion table, but the table requires a large size, which is an obstacle to speeding up and reducing costs.

E. Means for Solving the Problems The present invention aims to solve the above-mentioned conventional problems. For this purpose, we obtained the following configuration means.

In other words, is it the spectral data of the main components obtained by Fourier spectrum analysis of the acoustic signal?
The bit data corresponding to the season when the number of sea urchins increases is 2 to the right.
Shift bit by bit. Therefore, we provided a 2-bit shift converter that conversely shifts the bit data of X to the left by 2 bits and reads the number of shifts N when the most significant bit reaches the left end. At the same time, bit data corresponding to M2 is extracted. Next, R gives bit data of the square root M corresponding to a predetermined number of upper effective bits of the bit data of M2.
I set up an OM table. Roughly calculated numerical value ff=M/2
To obtain N, a 1-bit shift converter is provided to shift the pit data of square root M by -N bits to the right, and M÷2N is executed.

``Examples'' Preferred embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram of an acoustic signal analysis and synthesis device.

This will be explained in conjunction with FIG. 2 showing the output signal diagram of each circuit block. Reference numeral 1 denotes an A/D conversion circuit which converts an externally input acoustic analog signal S1 (for example, the sound of a bell or the sound of an animal picked up by a microphone) into digital waveform data S2. 2 is a sambling data memory for temporarily storing sampled digital waveform data. Reference numeral 3 denotes a fast Fourier transform (FFT) calculation circuit, which reads 82 from the zumbling data memory 2 in predetermined lofts and performs spectrum analysis. 4 is a spectrum extraction circuit which extracts a predetermined number of spectra with particularly large amplitude intensity among the spectra input from the FF calculation circuit 3, and extracts the obtained data into 8
3. Show me how. The spectral group with large amplitude intensity is the main one, and it is possible to reproduce the source waveform with sufficient fidelity, and by narrowing down the number of subextensions, the speed of synthesis can be increased. S3 is the fundamental tone component t'o and the harmonic component child,
f2° includes f3 and fA. The overtone components are not necessarily limited to the integer A-8 tones. Depending on the characteristics of the externally input sound, non-integer overtones may also be included.

5 is a square root calculation circuit and a spectrum extraction circuit 4.
It is connected to the. The square root calculation circuit 5 is necessary when calculating the amplitude coefficient of the extracted frequency scale I. As a result, the following overtone addition synthesis is performed using the spectrum of equation (2) instead of equation (1), so that a total of t11 times is simplified. The details will be explained in Shun.

6 is a harmonic addition sound source circuit, which synthesizes sine waves of a predetermined amplitude and frequency based on the spectrum data inputted from the spectrum extraction circuit 4, and synthesizes musical tone data into one.
Do 5. At this time, the predetermined amplitude is 1 determined in proportion to the amplitude intensity If value of each spectrum shown in S3, and the frequency is S3.
While maintaining the relative ratio of To to fA shown in Figure 3, the fundamental frequency 1'O is set to match a predetermined scale. Note that the value of r- calculated by the square root arithmetic circuit 5 is used as shown in S3 for the amplitude strength 1!u ia t of each spectrum l''o-fA.When performing 3rd overtone addition, the sine wave of equation (1) It is sufficient to add only one COS wave in equation (2) instead of the two COS waves.

7 is a scale generator, which is composed of, for example, a keyboard;
For example, if the scale name is A4, by pressing the V- key of A4, the scale data 44011z@杙音加咋B is sent to the B source circuit 6 to determine the fundamental tone frequency.

8 is a D/A converter, which converts the digital waveform synthesized by the overtone addition sound source circuit 6 into analog waveform data as shown at 84. S4 is amplified by the amplifier 9 and then output to the speaker 101. :J, is converted to a musical tone. 1
A control circuit 1 performs synchronization and timing control of each circuit described above.

Next, a specific example of the square root calculation circuit, which is a feature of the present invention, will be explained with reference to FIG. 3 and FIG. 4. 12 is 2
It is an N-hit shift converter that converts manual bit data 2 bits at a time and outputs the number of shifts. The input 16-bit data is obtained from the spectrum extraction circuit 4 as X=a 2+
There is n corresponding to b 2. Now X = (M/2')2=M2X2-2'
Let's consider how it corresponds to bit data in the format.

If N=O, the most significant effective bit of the bit data will be at the left end of the 16-bit data. Thereafter, each time N increases by one, the most significant effective bit is shifted two digits to the right. Considering this in reverse, the 2N pit shift converter 12 has input pitches 1 to 1.
It shifts the data (indicated by number 101 in Figure 4) to the left by two bits at a time until the most significant bit is on the left side, and N is known by the number of shifts in -. In the example, N=1. At the same time, the 2N bit shift converter is extracting bit data D+ (indicated by number 102 in FIG. 4) corresponding to M2.

Reference numeral 13 is a square root conversion ROM table, which provides bit data D3 (indicated by number 103 in FIG. 4) of M2 (corresponding to bit data 102 in M2). At this time, the eight upper significant bits of the bit data 102 of M2 are used as the address of the ROM table 13. In other words, this calculation does not require such precision (even rough approximation is sufficient to synthesize the original 1?J waveform), but rather high-speed processing is required, and for that purpose, the effective pitch 1~
Limited number. This result 1 (0M table 13 content d
28=256 is sufficient. Next, 14 is a -N pit shift converter, which shifts the M bit data 103 (8 digits) obtained from the ROM table 13 to the right by -N bits. This is equivalent to executing M/2N operations T and F, and obtains the corresponding P1-data (indicated as number 104 in FIG. 4). For the right 7 bits 1- of J3 pin 1 to data 104, 0 is inserted to supplement the missing data. As described above, the square root data M/2N of the input X=(M/2N)2 can be obtained.

G. Effects of the Invention The present invention comprises a square root 1iFH2 circuit for determining the amplitude coefficient of a Fourier frequency spectrum, consisting of a 2-bit shift register, a 1-bit shift register, and a ROM table that provides the square root M of M2 with a limited effective pitch. Therefore, the square root part τ can be executed at high speed without putting a burden on CPtJ, and the storage capacity of the ROM table can be kept to a minimum.

[Brief explanation of drawings]

Figure 1 is a configuration diagram of the acoustic signal analysis and synthesis device, Figure 2 is an output signal diagram of each block, Figure 3 is a square root disease n circuit diagram, and Figure 4 is a diagram of the output signal of each block.
The figure is a square root replacement flow diagram. 1...A/D converter 2...Sampling data memory 3...FFT calculation circuit 4...Spectrum extraction i11 circuit 5...Square root calculation circuit 6...Overtone addition circuit 7...Tone scale Generator 8...D/A converter 9...Amplifier 10...Speaker 11...Control circuit 12...2N bit shift converter 13...Square root conversion Fi I'l< OM table 14
...-N bit shift] Converter Applicant Seiko Electronic Industries Co., Ltd. Agent Patent Attorney Tsutomu Mogami (1 other person) I'-41 Figure 32 DB showing the output of each folotsf = FC Square I calculation circuit ☆chi 3 four

Claims (2)

[Claims] (1) An A/D converter that converts an analog acoustic signal into digital waveform data, a sampling data memory that stores the digital waveform data, a Fourier analysis calculation circuit that spectrally analyzes the stored digital waveform data, and a main a spectrum extraction circuit that extracts spectral components;
A square root calculation circuit connected to the spectrum extraction circuit and performing the square root calculation necessary for calculating the amplitude coefficient of the extracted spectrum, a scale generator that inputs desired scale data, and the extracted spectrum data and scale data. an acoustic signal analysis/synthesis device consisting of an overtone addition sound source circuit that receives and performs additive synthesis of spectra based on this, a D/A converter that converts the resulting synthesized digital waveform into an analog acoustic signal, and an electroacoustic transducer. In the above, the square root calculation circuit calculates input data represented by bit data X=(
a 2-bit shift converter that extracts the exponent N of M/2^N)^2 and bit data corresponding to M^2;
From a ROM table that provides M bit data corresponding to a predetermined upper effective bit of bit data corresponding to , and a 1-bit shift converter that shifts M bit data by -N bits and provides M/2^N bit data. An acoustic signal analysis and synthesis device characterized by: (2) The acoustic signal analysis and synthesis apparatus according to claim 1, wherein said ROM table provides square root data using upper bits of said predetermined number of digits as an address.