JPH02137896A - Generating method for sound source data - Google Patents

Abstract

PURPOSE:To enable excellent repetitive reproduction by interpolating waveform data consisting of existent samples and generating interpolated samples, and using the interpolated sample which is closest to the existent sample as the connection sample of repetitive waveform. CONSTITUTION:The specific-cycle waveform data consisting of the existent samples (marked with 'o' in figure) with a sampling period T3 is interpolated to generate the interpolated samples ('x' in figure) and the point where the existent sample and interpolated sample are closest in value (crest value) is regarded as the connection sample. Thus, the connection point (looping start point and looping end point) of the repetitive waveform are obtained including the interpolated samples obtained by interpolating the existent samples of the waveform data, so the continuity at the connection point of the waveform is improved, so no looping noises are generated.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a method of generating sound source data such as, for example, an electronic musical instrument, and in particular, a method of generating sound source data by digitally processing a musical tone signal or the like. It is related to.

[Summary of the invention]

The present invention generates a plurality of interpolated samples by interpolating waveform data consisting of real samples, and connects the repeating waveform at the point of the interpolated sample closest to the real sample, thereby eliminating discontinuity at the repeating points. The present invention provides a method of generating sound source data that enables good repeated playback.

[Conventional technology]

-i, the sound sources used in electronic musical instruments, TV game machines, etc. are analog sound sources such as VCO, VCA, VCF, etc.
It is broadly classified into digital sound sources such as PSG (programmable sound generator) and waveform ROM readout type. In recent years, as a type of digital sound source, sampler sound sources have become widely known, in which sound source data that is sampled from live musical instruments and digitally processed is stored in a memory or the like (for example, Publication No. 62-264099, JP-A No. 62-2
(See Publication No. 67798).

This sampler sound source requires a large capacity of memory for storing sound source data when fishing on a boat, so various methods have been proposed to save memory, such as looping processing that uses the periodicity of musical sound waves A typical example of this is bit compression processing using nonlinear quantization. Note that the looping process described above is also a method for continuing to output a sound for a longer time than the original duration of the sampled musical tone. In other words, when considering a musical tone signal waveform, for example, immediately after the start of sound generation, a formant portion containing so-called non-pitch components such as piano key-press noise or wind instrument press noise occurs, and the periodicity of the waveform is unclear, and then, The same waveform appears repeatedly with a fundamental period corresponding to the musical interval (pitch, pitch) of the musical tone. By setting n cycles (n is an integer) of this repetitive waveform as a looping section and repeatedly reproducing the sound as needed, a long-lasting sound can be obtained with a small memory capacity.

As mentioned above, sampling a musical tone and looping its pitch components means connecting the looping start point and looping end point of the looping section and looping repeatedly. It is required that the values of are approximately equal. This is because if the values at the two points are not approximately equal, that is, if the connection points are discontinuous, noise will occur.

However, it is difficult to select substantially equal values for the looping start point and the looping end point due to the sampling period, etc., and no effective solution has been found in conventional looping processing methods.

The present invention has been proposed in view of the above-mentioned circumstances, and aims to provide a method for generating sound source data that eliminates discontinuity in repeating points and does not generate looping noise.

(Problem to be Solved by the Invention) [Means for Solving the Problem] A sound source data generation method according to the present invention has been proposed to achieve the above-mentioned object, and is shown in FIG. As shown in FIG. 2, the waveform data of a predetermined period consisting of multiple real samples (marked with ○ in the figure) with sampling period T is interpolated to generate multiple interpolated samples (marked with x in the figure, etc.), and the real samples and interpolated samples are It is characterized in that the point with the closest value (peak value) among them is used as a connected sample of the repeating waveform. For example, the looping start point LP in the looping section LP corresponding to the repeating waveform is an actual sample, and this start The looping end point LPE closest to the peak value of point LP is the interpolation sample. This is the conventional looping section LP' consisting of only real samples.
The peak value of the above interpolated sample is higher than the peak value of the actual sample at the looping end point L P E " of the starting point L
This is done in consideration of the fact that it is close to the peak value of P3.

In order to connect the starting point LP and the ending point LP, interpolation processing is used to prevent the sampling period from being disrupted when returning from the ending point to the starting point during looping processing, as shown in Figure 2. each interpolation sample (x mark in the figure)
, and form a signal waveform as shown by the broken line.

[For production]

According to the present invention, interpolated samples obtained by interpolating a plurality of real samples of waveform data are also used as connecting points (looping start point and looping end point) of repeated waveforms, so as shown in the broken line in FIG. It is possible to improve the continuity of the waveform at the connection point of the waveform.

〔Example〕

An embodiment to which the present invention is applied will be described below with reference to the drawings. It goes without saying that the present invention is not limited to the following examples.

First, the waveform diagram of the looping section LP shown in FIG. 1 will be explained in more detail.

In Fig. 1, looping corresponds to a repetitive waveform obtained using only each sample (actual sample, marked with ○ in the figure) actually obtained by sampling a signal waveform of a predetermined period at a constant sampling frequency MT s. Section L
P' has actual samples at the looping start point LPff and the looping end point LP,', but when connecting the start point LP3 and the end point [, Pi', the start point LPs and the end point Pi' are connected. The peak values of ' are rarely close to each other, and as shown in the connection waveform shown by the solid line in Figure 2, the starting point LPs and the ending point LP! ' are very often discontinuous. Therefore, interpolation samples are generated by interpolating the waveform data made up of the actual samples, and the looping start point LPi and end point LP, which have the closest peak values to each other including the interpolated samples, are determined, and the connection point of the waveforms is determined. It is used as a.

Here, in a normal sampler sound source device, etc., an interpolation method such as oversampling is used as a pitch conversion method to obtain a musical sound waveform of the pitch of a pressed key from sound source data of a constant pitch stored in a memory etc. In this case, there is no need to increase hardware to obtain the interpolated samples, and the present invention can be easily applied. FIG. 3 is a diagram for explaining a pitch conversion method using interpolation. For example, pitch conversion is performed by changing the sampling interval of each interpolated sample obtained by four times oversampling. In the example shown in Fig. 3, samples are sampled at a period of, for example, 5 samples from all samples, including interpolated samples (x marks) obtained by oversampling each real sample (O mark) of the original waveform A by 4 times. The pitch-converted waveform B is obtained by extracting and arranging it at the original sampling period T.

In this case, the frequency is converted to 5/4 times. Although the pitch conversion shown in FIG. 3 is an example in which the pitch is raised, the pitch can also be lowered by setting the sampling period to three or less. Furthermore, accuracy can be improved by increasing the oversampling multiple, and in practical terms, oversampling of about 256 times is considered, for example.

Next, FIG. 4 shows a circuit configuration for performing pitch conversion by the interpolation process shown in FIG. 3 above.

In this FIG. 4, the RO where the sound source data is stored is
The memory 1 such as M outputs sound source data based on the address data of the address generation circuit 2. Pitch information that determines the pitch conversion ratio when performing the pitch conversion process is sent to the address generation circuit 2 from the pitch information generation circuit 3, and further includes an attached information register 7a, a loop start address register 7b, and a loop end address. - Each data from register 7C is also supplied. Based on these data, the address data for accessing the memory I is generated. Further, the attached information register 7a, the loop start address register 7b, and the loop end address register 7C are supplied with respective information from the sound source data.

The attached information register 7a is for taking in block-by-block header information FR shown, for example, in FIG. and each address of the looping end point LPt.

These attached information registers and loop start address
The output data of the register 7b and the loop end address register 7C are sent to the address generation circuit 2 and also to the coefficient ROM address conversion circuit 6. The pitch information from the pitch information generation circuit 3 is also input to the coefficient ROM address conversion circuit 6, and based on these output data, the coefficient ROM 5 sends pre-stored coefficients to the interpolation filter 4. Determine the filter characteristics of 4. Here, the interpolation filter 4 includes, for example, n delay devices DLL=DLn, multipliers M1 to Mn, and adders P1 to Pn, and receives sound source data from the ROMI. As mentioned above, the sound source data input to the interpolation filter 4 undergoes pitch conversion by the interpolation filter 4, is D/A converted by the D/A converter 8, and is output as a sound source signal from the output terminal 9. .

FIG. 5 shows an example of the loop start address and loop end address taken into each of the registers 7b and 7c, assuming a case where a block is formed for each of a plurality of samples.

Here, the loop start point (the above looping start point L
If the sample of P s ) is always placed at the beginning of the block, the block number alone is sufficient as the loop start address, but when looping starts from any sample within the block, the sample number within the block ( (see broken line in the figure) is required. Furthermore, the loop end address includes not only the sample number within the block but also information indicating the interpolation point between samples.

As a result, high-precision looping processing including interpolation samples can be performed using a method similar to the pitch conversion described above. For example, in the case of 256 times oversampling, the looping section LP is set to 1/256 of the sampling period T.
It can be set with an accuracy of 1/2, and the accuracy of waveform connection is also 1/2
It can be increased to about 56. Furthermore, the present invention can be applied to sound source devices that employ pitch conversion using an oversampling method, with almost no additional hardware, and looping noise can be significantly reduced by smooth waveform connections. can be reduced to

Here, the normal musical tone signal waveform is sampled and transmitted,
Alternatively, looping processing during recording and reproduction will be explained with reference to FIG.

Generally, immediately after the start of sound production, a formant part FR, which is a part where the periodicity of the waveform is unclear, occurs due to the inclusion of non-pitch components such as piano key tapping noise and press noise of wind instruments, and after that, the pitch of the musical note The same waveform will appear repeatedly with a fundamental period corresponding to (pitch, pitch). N cycles (n is an integer) of this repetitive waveform are defined as a looping section LP, and this looping section LP is expressed between the looping points of the looping start point LP and the looping end point LPt.

And the above formant part FR and looping section LP
is recorded on a storage medium, and during playback, the formant part F
Following the reproduction of R, a looping section t.

By repeatedly reproducing P, musical tones can be generated over an arbitrary long period of time.

FIG. 7 shows a specific example of each function from sampling a human-powered musical tone signal to recording it on a storage medium, in order to explain another specific embodiment of the sound source data generation method according to the present invention. It is a block diagram.

In this case, the input musical tone signal supplied to the input terminal 10 may be a signal directly picked up by a microphone, or a signal obtained by reproducing a digital audio signal recording medium, etc., in the form of an analog signal or a digital signal. It can be used in

First, in the sampling processing function block 11 of FIG. 7, the input musical tone signal is sampled at a frequency of 38 kHz, for example, and extracted as 16-bit sample digital data.

This sampling process corresponds to A/D conversion process when the input musical tone signal is an analog signal,
Furthermore, when the input signal is a digital signal, it corresponds to processing of sampling rate conversion and bit number conversion.

Next, in the pitch detection function block 12, the frequency (
fundamental frequency) f, that is, pitch information is detected.

The detection principle in this pinch detection function block 12 will be explained. Here, the frequency of the fundamental tone of the musical tone signal serving as the sampling sound source is often considerably lower than the sampling frequency fs, and it is difficult to identify the pitch with high accuracy just by detecting the peak of the musical tone on the frequency axis. Therefore, it is necessary to use some means to utilize the spectrum of overtone components of musical tones.

First, let f (t) be the waveform of a musical tone signal whose pitch is to be detected, then convert this musical waveform fct> to the amplitude a(
ω) and phase φ(ω), the musical sound waveform f(t)
is a Fourier-expanded formula, f(t)=Σa (6))CO3(ωt +φ(ω
) ) ... can be expressed as ■. Here, if the phase shift φ(ω) of each harmonic is set to zero, f(t) = Σa(ω)cosωt ・ ・ ・ ・
・ ・ ・It can be expressed by the formula of ■. In this way, the points that are integral multiples of the period of all harmonics that the phase alignment has and t
・It is a point of 0. This is nothing but the period of the fundamental tone.

Based on this principle, the pitch detection procedure will be explained using the functional block diagram shown in FIG.

In FIG. 8, musical tone data is supplied from a real part data input terminal 31 and "0°" is supplied from an imaginary part data input terminal 32 to a fast Fourier transform (FFT) functional block 33.

Here, in the fast Fourier transform performed by the fast Fourier transform function block 33, the musical tone signal whose pitch is estimated is x(t), and the overtone component included in the musical tone signal x(t) is aacos(2πf1% L+θ)・・・・・・■, then x(t) is Oo Rewriting this in complex representation, 0゜X (t) = (1/2)Σa, exp(jθ,,)
exp(jωnt) ・・■n=-ω However, cosθ=(exp(jθ) + exp(-jθ))/2
- ・■ was used. When this formula is Fourier transformed, O x ((1)) = S x (t)exp(-3(1)
L) dt-Oo ・Σa exp(jθ,,)δ(ω-ω+1) ・
・■n;- where δ(ω-ω, 1) is a delta function.

In the next functional block 34, the norm (absolute value, that is, the square root of the sum of the squares of the real part and the imaginary part) of the data after the fast Fourier transform is calculated.

That is, when we take the absolute value Y(ω) of X(ω), the phase component is canceled, and Y(ω)=[X(ω)X(ω)]l/! =(1/2)a
, 1δ(ω-ω7) ・ ・ ・ ・■ This is done to match the phases of all the high frequency components of the musical tone data, and by setting the imaginary part to zero,
Phase components can be aligned.

Next, this calculated norm is supplied as real part data to the fast Fourier transform (corresponding to inverse FFT in this case) function block 36, and "°0" is supplied to the imaginary part data input terminal 35 to perform inverse FFT. and restore the musical sound data. In other words, the above inverse Fourier transform is 0koy(t)=(1/2πN Y(ω)exp(-jωt
) dt-Oo = 5 a1%co! (dnj + HHHH - [phase]. The restored musical tone data after this inverse Fourier transform is extracted as a waveform that can be expressed by a combination of cosine waves in which all high frequency components are in phase.

Thereafter, a peak detection function block 37 detects the peak of the restored sound source data. Here, the peak is the point where the extreme values (peaks) of all the high frequency components of the musical tone data coincide, and in the next function block 38, the detected peak values are sorted by 1114 (sorting) from the largest value. do. By measuring the period of the detected peak, the pitch of the musical tone signal can be determined.

FIG. 9 is for explaining the configuration for detecting the local maximum value (peak) of musical tone data in the peak detection function block 37 of FIG. 8.

In this case, the musical tone data has many peaks (extreme values) of different values, and by finding the maximum value of the musical tone data and detecting its period, the pitch of the musical tone can be determined.

That is, in FIG. 9, the musical tone data string after the inverse Fourier transform is supplied to the N+1 stage shift register 42 via the input terminal 41, and each stage of the shift register 42 has registers a-N/2...ao.・・a M/! are sent to the output terminal 43 via the following sequentially. This N+1 stage shift register 42 acts as a window with a width of N+1 samples for the musical tone data string, and N+1 samples of the musical tone data string are sent to the maximum value detection circuit 44 via the window. That is, the above musical tone data is first input to registers a-N/□, and then sequentially transmitted to registers aN/□, and is then input to each register a□7. ...ao...
- The above-mentioned musical tone data of N+1 samples from aN/□ are sent to the maximum value detection circuit 44.

This maximum value detection circuit 44 detects, for example, the value of the central register a0 in the shift register 42 when the value of the register a0 becomes the maximum among the N+1 samples of data. data is detected as a peak value, and output terminal 45
It outputs more. In addition, the width N of the above window
+1 can be set arbitrarily.

Returning to FIG. 7, the envelope detection function block 13 obtains the so-called envelope waveform of the musical tone signal by performing envelope detection processing using the pitch information on the digital musical tone signal after the sampling process described above. ing.

For example, the waveform shown in Figure 10B is obtained by sequentially connecting the peak points of the musical tone signal waveform shown in Figure 10A, and the level (or volume) changes over time from immediately after the sound is produced. It represents change. This envelope waveform is generally expressed by parameters such as ADSR (attack time/date time/sustain level/release time). Here, as a specific example of a musical sound signal, when considering a piano sound etc. that is produced in response to a keystroke operation, the above attack time TA is the target for the volume to gradually increase when a key is pressed (key-on) on the keyboard. It represents the time it takes to reach the volume, and the above Decay Time T,
is the next volume from the volume reached at the above attack time TS (
For example, the sustain level L represents the volume of the sustained sound that is maintained until the key is released and the key is turned off, and the release time TR is This represents the time from when the key is turned off until the sound disappears. Note that each of the above times TA, To
, To may also indicate the slope or rate of volume change. Furthermore, more enherobe parameters may be used in addition to these four parameters.

Here, in the envelope detection function block 13, envelope waveform information represented by each parameter such as ADSR (attack time TA/decay time T./sustain level L, release time T.) as described above is simultaneously detected. In order to extract the aforementioned formant portion with the attack waveform remaining, information indicating the overall decay rate of the signal waveform is obtained.

This decay rate information is, for example, as shown in FIG.
The waveform takes a reference value "1" between , and then monotonically decays.

Here, an example of the configuration of the envelope detection function block 13 shown in FIG. 7 will be explained with reference to the functional block diagram shown in FIG. 12.

The principle of envelope detection is similar to envelope detection of so-called AM (amplitude modulation) signals. That is, the envelope is detected by considering the pitch of the musical tone signal as the frequency of the carrier of the AM signal. The envelope information is used when reproducing a musical tone, and the musical tone is formed based on the envelope information and pitch information.

For the musical tone data supplied to the input terminal 51 in FIG. 12, an absolute value of the peak value data of the musical tone is determined in an absolute value output function block 52. FI this absolute value data
It is sent to a functional block 55 of an R (finite impulse response) type digital filter.

Here, the FIR filter function function blank acts as a low-pass filter, and the function block 5 is activated based on the pitch information supplied to the input terminal 53 in advance.
By supplying the filter coefficients formed in step 4 to the FIR filter function block 55, the cutoff characteristic of the low-pass filter is determined.

Here, the filter characteristic is, for example, the characteristic shown in FIG. 13, and has zero points at the fundamental tone (frequency r0) of the musical tone signal and the frequencies of its overtones. for example,
From the musical tone signal shown in FIG.
Envelope information as shown in Figure 0B is detected. Note that the characteristics of the filter coefficients described above are shown by the following equation.

H(f) = k ・(sin(πf/fo))/
f...■fo in this formula represents the fundamental frequency (pitch) of the musical tone signal.

Next, from the peak value data (sampling data) of the sampled musical tone signal mentioned above, the peak value data of the formant part FH signal shown in FIG. 6 and the peak value data of the signal in the looping section LP (loop data)
The process of generating .

In the first function block 14 for generating the loop data, the peak value data of the sampled musical tone signal is divided (or multiplied by the reciprocal) by the data of the previously detected envelope waveform [Fig. 10B]. By performing envelope correction, peak value data of a signal with a constant amplitude waveform as shown in FIG. 14 is obtained. By filtering (the peak value data of) this envelope-corrected signal, (the peak value data of) a signal in which components other than pitch components are attenuated or pitch components are relatively emphasized is obtained. Here, the pitch component is a frequency component that is an integral multiple of the fundamental frequency f0.

Specifically, an HPF is used to remove low frequency components such as vibrato included in the envelope-corrected signal.
(bypass filter), and then through a comb filter having frequency characteristics as shown in the dashed line in FIG. Only the pitch components included in the above HPF output signal are passed through, other non-pitched components and noise components are attenuated, and if necessary, the LP
By passing through F (low pass filter), noise components superimposed on the signal after passing through the comb filter are removed.

That is, when considering a musical sound signal such as the sound of an instrument as the input signal, this musical sound signal usually has a constant pitch (pitch,
As shown by the solid line in Fig. 15, the frequency spectrum contains energy in the vicinity of the fundamental frequency f0 corresponding to the pitch of the musical note itself and in the vicinity of frequencies that are integral multiples thereof. A distribution is obtained in which .

On the other hand, it is known that general noise components have a −-like frequency distribution. Therefore, by passing the input musical tone signal through a comb-shaped filter having frequency characteristics as shown in the dashed line in FIG. 15, the fundamental frequency f0 of the musical tone signal is
Only frequency components (so-called pitch components) that are integral multiples of are passed through or emphasized, while other components (non-pitch components and part of noise) are attenuated, and as a result, the S/N ratio can be improved. Here, the frequency characteristic of the comb-shaped filter shown by the dashed line in FIG. 15 is expressed by the following formula. In this equation, fo is the fundamental frequency of the input signal (the frequency of the fundamental tone corresponding to the pitch), and N is the number of stages of the comb filter.

The musical tone signal whose noise components have been reduced in this way is sent to the repetitive waveform extraction circuit, and this repetitive waveform extraction circuit extracts an appropriate repeated waveform section such as the looping section LP in FIG. 6 described above. After that, it is sent to a storage medium such as a semiconductor memory and recorded. Since the musical tone signal data recorded on this storage medium has non-pitch components and some noise components attenuated, it is possible to reduce the so-called looping noise, which is the noise that occurs when the above-mentioned repetitive waveform section is repeatedly reproduced. can.

Note that the frequency characteristics of the HPFS comb filter and LPF are set based on the fundamental frequency f0, which is pitch information detected by the pitch detection function block 12 in advance.

Next, the loop section detection function block 16 in FIG. 7 detects an appropriate repeating waveform section for the musical tone signal whose components other than pitch components have been attenuated by the above-described filtering process, thereby determining the looping starting point LP and the looping ending point. Set a looping point with point LP.

That is, the loop section detection function block 16 selects two looping points that are relatively separated by (an integral multiple of) the repetition period corresponding to the pitch (interval) of the musical tone signal, and the selection principle is explained below. Explain.

When performing looping processing on musical tone data, the looping interval must be an integral multiple of the fundamental period (the reciprocal of the frequency of the fundamental tone) of the musical tone signal. Therefore, if the pitch of the musical note is accurately identified, it can be determined easily.

That is, a looping interval is determined in advance, two points separated by the interval are extracted, and a looping point is set by evaluating the correlation or similarity of signal waveforms in the vicinity of the two points. As an example of this evaluation function, one that uses convolution (synthetic product, convolution) of samples of signal waveforms near each of the two points will be described. In other words, the above convolution operation is sequentially applied to all sets of points to evaluate the correlation or type (other than) of the signal waveform. The musical tone data inputted sequentially and taken in by each register are sent to a DSP, which will be described later, for example.
(digital signal processing device), which calculates and outputs the above convolution. The set of two points resulting in the maximum convolution thus obtained is defined as a looping start point LP and a looping end point LPE.

That is, in the 16th time, the looping starting point LP,
Let the candidate point of the looping end point LP be ao, and let the candidate point of the looping end point LP be bo, then the candidate point a of the looping start point LP,
Each wave height value data of a plurality of points around 0, for example 2N+1 points, is a-N"+a-2+a-1, respectively.
+a6+al+ all ” all sThe same number of points (2N
+1 point) each wave height value data is bl・・・, b−z
, b-1, bo+bl, bz, . . . b, then the evaluation function E(an, bo) can be determined by the following equation. This phase [phase] equation is an equation for finding convolution centered on the points ao and be. Then, the set of candidate points ao and bo is sequentially changed to find the value of the evaluation function E for all looping point candidates, and the value is the largest among all the evaluation functions E obtained. The point where , is the looping point.

Further, in addition to the method of finding the looping point from convolution as described above, it is also possible to find the looping point from the method of least squares of errors. That is, the candidate point ao of the looping point is determined by the least squares method.

bo is t (ao, bo) = Σ (ak - b) z...
It can be expressed by the second part of [phase]. In this case, aOJ6 with the minimum value of the evaluation function ε may be found.

Further, in the loop section detection function block 16 described above, the looping start point LP is determined as necessary.

The pitch conversion ratio is calculated based on the looping end point LP, and the looping end point LP. This pitch conversion ratio is determined by the following function block 17.
It is used as time axis correction value data during the time axis correction process. This time axis correction process is performed in order to align the pitches of various sound source data when actually recording the various sound source data in a storage means such as memory, and uses pitch detection instead of the above pitch conversion ratio. The pitch information detected in the functional block 12 may be used.

The pitch normalization operation in the time axis correction function block 17 will be explained with reference to FIG. 17.

FIG. 17A shows a tone signal waveform before time axis correction processing (mainly time axis companding process), and FIG. 17B shows a corrected waveform after the companding process. These Figures 17A and B
The time axis of is marked with a scale in units of blocks during quasi-instantaneous bit compression encoding processing, which will be described later.

In the waveform A before time axis correction, the looping section LP is normally unrelated to the above block, but the 17th
As shown in FIG. B, the time axis companding process is performed so that the looping section LP is an integral multiple (m times) of the block length (block period), and the block boundary position is the looping start point LP, and shifted in the time axis direction to match the looping end point LPt. That is, the starting point LPs and the ending point LP of the looping section LP.

Time axis correction (
time axis companding and shifting), an integer number (m
The looping process can be performed on a block-by-block basis, making it possible to normalize the pitch of sound source data during recording.

Here, the deviation ΔT from the block boundary that occurs at the beginning of the musical tone signal waveform due to the above-mentioned time shift may be filled with 0 as the peak value data.

FIG. 18 shows the block structure when the peak value data of the waveform after time axis correction is divided into blocks for the bit compression encoding process described below. (number of words, number of words) is set to h. In this case, the above pitch normalization means -1
Time axis companding processing is performed so that the number of words in the looping section LP for n cycles of the musical tone signal waveform with a constant period Tw shown in the figure is an integral multiple (m times) of the number of words in the block. More preferably, time axis processing (
shift processing). In this way, each point LPs
, t, p coincide with the block boundary position, it is possible to reduce errors caused by block switching during decoding in the bit compression encoding system.

Here, the words wt, ps, and WLP in the shaded portion in one block of FIG. 18A are the looping start point LPs and the looping end point LP! of the corrected waveform in the figure.
(To be exact, the point immediately before point LPI) is a word indicating a sample. Note that if the above shift processing is not performed, the looping start point LP.

Since the terminal point LPE and the terminal point LPE do not necessarily coincide with the block boundary, as shown in FIG. 18B, the word wLps
, WLP, are set at any position within the block. However, from the above word WL P, to the word W
LP! The number of words up to this point is an integral multiple (m times) of the number of words in l block, and the pitch is normalized.

Here, as mentioned above, various methods can be considered for time-axis companding processing of the musical tone signal waveform in order to make the number of words in the looping section LP an integral multiple of the number of words in one block.
For example, this can be realized by interpolating the peak value data of a sampled waveform, and as a specific example, a filter configuration for oversampling processing, etc. can be used.

(Left below) By the way, the looping period of an actual musical sound waveform has a fraction of the sampling period unit, and the looping start point L
The sampling peak value at P and the looping termination point LPE
If there is a deviation from the sampling peak value at
By interpolation processing using oversampling, etc., a peak value that matches the sampling peak value of the looping start point LP is obtained at a position near the looping end point LP (a position at a distance shorter than the sampling period). ,
Non-integer multiple of the sampling period including interpolated samples (
It is conceivable to realize a looping period (with a fraction). Such a looping period that is a non-integer multiple of the sampling period can also be made an integer multiple of the block period by the above-mentioned time axis correction processing. By correction processing, the block period can be set to an integral multiple of the above block period. For example, when performing time axis companding processing using 256 times oversampling, the looping start point LP,
By reducing the error in the peak value between the LPI and the end point LPI to 1/256, smoother looping reproduction can be realized.

In the next functional block 21, the waveform whose looping section LP is determined and subjected to time axis correction (companding) processing as described above is converted into a looping section L as shown in FIG.
Loop data is generated by connecting P in front and behind each other. In other words, FIG. 19 shows the musical sound waveform (the first
A loop data waveform is shown in which only the looping section LP is cut out from Figure 7B) and a plurality of these looping sections LP are arranged.
and the other looping start point LPs are connected and arranged in sequence. This loop data waveform is generated by the loop data generation function block 21.

Since this loop data is formed by connecting looping sections LP many times, a word WLP corresponding to each looping start point LP of the connected loop data waveform is
Immediately before the start block containing l, the looping end point L
The data of the end block including the word WLP corresponding to PE (more precisely, the point immediately before point LPt) will be placed as is. In principle, in order to perform bit compression encoding processing, it is sufficient that at least the end block exists immediately before the start block of the looping section LP to be stored. Furthermore-
To put it simply, at the time of bit compression encoding on a block-by-block basis, the parameters of the starting block (information on focus compression encoding for each compressed block, such as range information and filter selection information to be described later) are the same as the starting block. It may be formed based on the data of the end block. This is a technique that can also be applied when the sound source is a musical tone signal consisting only of loop data without a formant part, which will be described later.

By doing this, during the above encoding, the looping start point L
Regarding Ps and the terminal point LPt, the same data is lined up over a plurality of samples before and after each of them. Therefore, the parameters for bit compression encoding for each block immediately before each point LPs and LPE are the same, and errors (noise) during looping playback during decoding processing can be reduced. can.

That is, the musical tone data that is looped and reproduced becomes stable without any connection noise. In this embodiment, the number of samples of data in the looping section LP placed immediately before the start block is approximately 500 samples.

Next, in the process of generating data for the signal of the formant portion FR, envelope correction processing is first performed in the functional block 18, similar to the functional block 14 in generating the loop data.

However, in this case, the envelope correction is performed by dividing the sampled musical tone signal by the envelope correction of only the decay rate information (Fig. 11) described above, resulting in a waveform as shown in Fig. 20. The signal (peak value data) is obtained.

That is, in the output signal of FIG. 20, the envelope of the attack portion (between time TAO) remains, and the other portions have a constant amplitude.

This envelope-corrected signal is subjected to filter processing in the functional block 19 as necessary, so that
The non-pitch component of the musical tone signal becomes a signal that is relatively emphasized. For the filter processing in the functional block 19, a comb-shaped filter having a frequency characteristic as shown by the broken line in FIG. 15, for example, which is symmetrical to the functional block 15, is used. That is, this comb-shaped filter has a frequency characteristic that attenuates frequency band components that are integral multiples of the fundamental frequency f0 corresponding to the pitch and relatively emphasizes non-pitch components,
The frequency characteristics of this comb filter are also set based on the pitch information (fundamental frequency re) detected by the pitch detection function block 12. Such a signal in which non-pitch components are relatively emphasized is used to generate signal data of a formant part in sound source data that is finally recorded in a storage medium such as a memory.

In the next functional block 20, the above functional block 1
The same time axis correction as in step 7 is also performed on the formant part generation signal. This is the function block 1 above.
This is to equalize (normalize) the pitches of each sound source by compressing and expanding the time axis based on the pitch conversion ratio obtained in step 6 or the pitch information detected in the functional block 12.

Next, in a functional block 22, the loop data and formant part generation data, both of which have been time-base corrected using the same pitch conversion ratio or pitch information, are mixed. In this mixing, a Hamming window is applied to the formant part generation signal from the functional block 20 to form a fade-out signal that attenuates over time in the part to be mixed with the loop data. On the other hand, a similar Hamming window is applied to the loop data from the functional block 20, and in this case, a fade-in type signal that increases with time is formed in the portion to be mixed with the formant signal, and these Mix the signals of (
(crossfading) to obtain musical tone signals that ultimately become sound source data. Here, as the loop data to be recorded in a storage medium such as a memory, the data of one looping section that has been cut to a certain extent from the crossfade section is taken out and the noise during looping playback is eliminated.
looping noise) can be reduced. In this way, peak value data of the sound source signal is obtained, which consists of the formant portion FR, which is a waveform portion including non-pitch components from the time of sound generation, and the looping section LP, which is a repeating waveform portion containing only pitch components.

In addition, a process of cutting and connecting each part such that the start point of the loop data signal is connected to the position of the looping start point in the formant part generation signal may also be considered.

By the way, when actually performing loop section detection, looping processing, and even mixing loop data and formant parts, a rough mix is performed by repeating trial and error trial listening manually, and at this time, loop points (looping start point LP, looping end point L
Pt) Processing with higher precision is performed based on information etc.

That is, prior to detecting a loop section with high accuracy in the function block 16, the loop section detection and the above-mentioned mixing etc. are manually performed while repeatedly listening to the sample in the procedure shown in the flowchart of FIG. 21, and then the above-mentioned Highly accurate processing (from step 326 onward) is performed.

In FIG. 21, in the first step 321, the loop point is detected with relatively rough accuracy by using, for example, the zero crossing point of the signal waveform or visually checking the display of the signal waveform, and in step S22, the loop point is detected. The waveform between the loop points is processed and reproduced repeatedly, and in the next step S23, a human listens to the sample and determines whether it is good or not.

If it is defective, the process returns to the first step 321 and the loop point is detected again. If a good listening result is obtained by repeating this process, the process proceeds to the next step S24, where it is mixed with the above-mentioned formant part signal by cross-fading, etc. In the next step S23, a human listens to the signal from the formant part to the looping part. Determine whether the transition is good or not. If the mixture is defective, the process returns to step S24 and the above mixing is repeated.

Thereafter, the process proceeds to step S26, where the loop section detection function block 16 performs highly accurate loop section detection. Specifically, loop section detection including the above-mentioned interpolated samples is performed, for example, in the case of 256 times oversampling, loop section detection is performed with an accuracy of 1/256 of the sampling period.
In the next step, step S27, a pitch conversion ratio for the pitch normalization is calculated. Based on this pitch conversion ratio, time axis correction processing is performed in the functional block 17.2o in the next step 32B, loop data generation is performed in the functional block 21 in the next step S29, and
In step 330, the mixing process in the functional block 22 is performed. In the processing after step 326, the loop point information etc. obtained in steps S21 to S25 are used. Note that steps S21 to 325 may be omitted to fully automate the looping process and the like.

The formant part F obtained by such a mixing process
The peak value data of the signal consisting of R and the looping section LP is subjected to bit compression encoding processing in the next functional block 23.

Various bit compression encoding methods can be considered as the above-mentioned bit compression encoding method, but here, the present applicant first proposed Japanese Patent Application Laid-Open No. 61-0
The quasi-instantaneous companding type proposed in Publication No. 08629 and Japanese Patent Application Laid-open No. 62003516, etc. is highly efficient, in which the peak value data is divided into blocks for each fixed number of words (h samples) and focus compression is performed in units of blocks. This high-efficiency bit compression encoding method will be briefly described with reference to FIG. 22.

In FIG. 22, the high-efficiency bit compression encoding system is composed of an encoder 70 on the recording side and a decoder 90 on the reproduction side. data x (n)
is supplied.

This input signal (peak value data) x(n) is obtained by the predictor 7
The predicted signal (peak value data) 7(n) from the predictor 72 is supplied to the FIR (finite impulse response) digital filter 74, which is composed of the adder 2 and the adder 73, and the predicted signal (peak value data) 7(n) from the predictor 72 is sent to the adder 73 as a subtraction signal. being sent. The adder 73 subtracts the prediction signal x(n) from the input signal x(n), thereby outputting a prediction error signal or a broadly defined difference output d(n). Predictor 72 generally uses two past inputs x(n-p), x(
The predicted value Ma(n) is calculated by a one-component combination of n-p+1), . . . x(n-1). Note that the FIR filter 74 is hereinafter referred to as an encode filter.

In the above-mentioned high-efficiency bit compression code system, data within a certain period of time of the above-mentioned sound source data, that is, input data of a certain number of words, are divided into blocks, and the above-mentioned encoding filter 74 having characteristics suitable for each block is applied. I try to choose. In this method, multiple (for example, four) encoding filters with different characteristics are provided in advance, and the filter with the optimal characteristics, that is, the one that can obtain the highest compression rate, is selected. This can be achieved by doing so. However, in the configuration of a general digital filter, a plurality of sets of coefficients Mi (for example, 4 sets) of the coefficients of the predictor 72 of one encode filter 74 shown in FIG. 22 are stored in a coefficient memory or the like. By time-divisionally switching and selecting a set of coefficients, an operation substantially equivalent to selecting one of the plurality of encode filters described above is often performed.

Next, the difference output d (n) as the prediction error is sent via an adder 81 to a bit compressor consisting of a shifter 75 with a gain G and a quantizer 76. Compression processing or ranging processing is performed such that the exponent part in the display form corresponds to the gain G and the mantissa part corresponds to the output from the quantizer 76. That is, the shifter 75 converts the input data according to the gain G. The quantizer 76 performs requantization to extract a fixed number of bits from the bit-shifted data.

Here, the noise shaping circuit (noise shaper) 77 uses an adder 78 to obtain a so-called quantization error corresponding to the error between the output and the input of the quantizer 76, and applies this quantization error to a shifter with a gain of Q −1. 79 to the predictor 80, so-called error feedback is performed in which the predicted signal of the quantization error is fed back to the adder 81 as a subtraction signal.

In this way, requantization by the quantizer 76 and error feedback by the noise shaping circuit 77 are performed, and an output 2(n) is taken out from the output terminal 82.

By the way, the output d''(n) from the adder 81 is obtained by subtracting the predicted signal τ(n> of the quantization error from the noise shaper 77) from the difference output d(n), and the gain G The output d'' (n) from the shifter 75 is the gain G and the output d' (n) from the output adder 81.
is multiplied by Also, the output from the quantizer 16? (n) is the quantization error e in the quantization process
(n) and the output d '' (n) from the shifter 75, and the quantization error e (n) is extracted in the adder 78 of the noise shaper 77. This quantization error e ( n> is the above gain Q −1
The predicted signal 7(n) of the quantization error is obtained by passing through the chic 79 and the predictor 80 which takes a 1-free sum of one past input.

The sound source data is subjected to the encoding process as described above, and is outputted from the quantizer 76 as an output ↑(n) through the output terminal 82.

Next, the prediction/range adaptation circuit 84 outputs mode selection information as optimal filter selection information and sends it to the predictor 72 and output terminal 87 of the encoding filter 74, and also outputs the gain G and the gain G-1 or the range information for determining the above bit shift amount is outputted to each shifter 75, 79 and output terminal 8.
It has been sent to 6.

Next, the output ↑(n) from the output terminal 82 of the encoder 70 is input to the input terminal 91 of the decoder 90 on the playback side.
F, a signal obtained by being sent, recorded, or reproduced, a-1(n), is supplied. This input signal 71 (n) is sent to an adder 93 via a shifter 92 with a gain of Q-1. Output x' from adder 93
(n) is sent to the predictor 94 and is the predicted signal? '(n)
So, this predicted signal? '(n) is the adder 93
Output from the shifter 92 91 (n)
is added. This addition output is the decoded output ↑°(n)
The signal is outputted from the output terminal 95 as a signal.

Further, each output terminal 86 and 87 of the encoder 70
The range information and mode selection signal output, transmitted, recorded, and reproduced are input to input terminals 96 and 97 of the decoder 90, respectively. The range information from the input terminal 96 is sent to the shifter 92.
The mode selection information from the input terminal 97 is sent to the predictor 94 to determine the prediction characteristic. The prediction characteristic of this predictor 94 is selected to be equal to the characteristic of the predictor 72 of the encoder 70.

In the decoder 90 having such a configuration, the thick 9
The output ↑''(n) from 2 is the product of the input signal ↑''(n) and the gain G-1. In addition, the adder 9
The output 9'(n) of 3 is the sum of the output ↑''(n) from the shifter 92 and the prediction signal ?'(n).

Next, FIG. 23 shows the bit compression encoding encoder 7.
This shows an example of the above one block of output data from 0. This one block of data consists of 1 byte of header information (parameter information) RF and 8 bytes of sample data DAO to D1. . Above header information R
F includes the above-mentioned range information of 4 bits, the above-mentioned filter selection information of 2 bits, and two flag information of 1 bit each, for example, information Ll indicating the presence or absence of a loop, and information indicating whether there is a terminal block (end block) of the waveform. It is made up of the information El shown. Here, the peak value data for one sample is
The data is compressed and expressed in 4 bits, and the data DA to D0 includes 4-bit data DAOH-D131 for 16 samples.

Next, FIG. 24 shows each block of the quasi-instantaneous (blocked) bit compression encoded peak value data corresponding to the leading portion of the musical tone signal waveform as shown in FIG.

In this Fig. 24, the header is omitted and only the peak value data is shown, and for convenience of illustration, one block is divided into 8
Although this is a sample, it goes without saying that it can be set arbitrarily, such as 16 samples per block. this is,
The same applies to the case shown in FIG. 18 above.

Here, the above-mentioned quasi-instantaneous bit compression code system has a mode in which the above-mentioned input musical tone signal is directly output, that is, a straight P
Select the mode in which a signal with the highest compression rate is obtained from the CM mode and the mode in which the musical tone signal is output through a filter, that is, the primary or secondary differential filter mode, and transmit the musical tone data that is the output signal. It was designed to do so.

When sampling a musical tone and recording it in a storage medium such as a memory, the musical tone signal waveform of the musical tone starts to be captured at the sound generation start point KS. When a filter mode that requires an initial value, such as the 1st or 2nd order differential filter mode, is selected, it is necessary to prepare this initial value in advance. It is desirable that Therefore, after adding a pseudo input signal that selects the straight PCM mode (a mode in which the input musical tone signal is directly output) in the period preceding the sound generation start point KS, signal processing including that input signal is performed. I try to do that.

Specifically, in FIG. 24, a block in which all data is "0" is placed as the pseudo input signal prior to the sound generation start point KS, and all data from the beginning of this block is "0". The data is bit-compressed and imported as sampled peak value data. This can be done, for example, by creating a block whose data is all "0" in advance and storing it in a memory etc., or by creating a block in which the data of the l block is all "0" and storing it in a memory etc. This can be obtained by starting sampling from an input signal where all data are '0' (that is, a silent part before the start of sound generation). Note that the number of blocks of the pseudo input signal is at least l blocks.

The musical tone data including the pseudo input signal formed as described above is subjected to signal compression processing using a high-efficiency bit compression encoding system as shown in FIG. , reproduces this compressed signal.

Therefore, when playing musical tone data including the above pseudo input signal, the straight PCM mode is selected for the filter at the start of playback (block portion of the pseudo input signal), so the initial value of the first or second difference filter is There is no need to configure settings in advance.

Here, there is a concern about a delay in the sound generation start time due to the pseudo input signal (there is no sound because all the data is on) at the start of playback. However, for example, when the sampling frequency is 32 kHz and one block has 16 samples, the delay in the sound generation time is approximately 0.5 m5ec, which is not a delay that can be discerned audibly and does not pose a problem.

Nowadays, the above-mentioned bit compression encoding processing and other digital signal processing for generating sound source data are often implemented in software using a digital signal processing bag Z (OSP). A software configuration using a DSP is often employed to reproduce recorded sound source data as well. As an example, FIG. 25 shows an example of the overall configuration of a system including an audio processing unit (APU) 107 as a sound source unit that handles sound source data and its surroundings.

In this 25th episode, a host computer 104 installed in, for example, a general personal computer device, a digital electronic musical instrument, a TV game machine, etc. is connected to the APU 107 as the sound source unit, and The sound source data etc. are APU
107. This APU
I07 is configured to include at least a CPU (Central Processing Unit) 103 such as a microprocessor, a DSP (Digital Signal Processing Unit) 101, and a memory 102 in which the above-mentioned sound source data etc. are stored. It is. That is, this memory 102 stores at least sound source data, and the DSP 101 performs various processing including readout control of the sound source data, such as looping processing, bit expansion (restoration) processing, pitch conversion processing, addition of an envelope, and echo processing. (Reverberation) processing etc. are applied. The memory 102 is also used as a buffer memory for these various processes. The CPU 103 is a DSP
It controls the operations and contents of these various processes of the IOI.

Furthermore, D for the above sound source data from memory +02.
The digital musical tone data finally obtained by performing the various processing described above by SP 101 is converted into digital/analog C
The signal is converted into an analog signal by a D/A converter 105 and supplied to a speaker 106.

Note that the present invention is not limited to the embodiments described above; for example, in the embodiments described above, the formant part and the looping section were connected to form the sound source data, but the sound source data consists only of the looping section. It can also be easily applied to the case of forming sound source data. Further, the present invention can be applied not only to sound source data but also to other uses such as speech synthesis. Furthermore, the generation of sound source data in the present invention includes the possibilities of data generation on the reproduction side and data generation on the recording side.

〔Effect of the invention〕

According to the present invention, since the waveform is repeatedly interpolated and the interpolated samples are used as connection samples, the continuity of the waveform at the connection point is improved.

Therefore, it is possible to obtain a method of generating sound source data that does not generate looping noise due to discontinuity of repetition points. Further, when an interpolation method is used for pitch conversion of sound source data, the present invention can be applied without increasing hardware.

[Brief explanation of the drawing]

FIG. 1 is a waveform diagram for explaining the principle of the sound source data generation method of the present invention, FIG. 2 is a waveform diagram showing the connection state of waveforms, FIG. 3 is a waveform diagram for explaining pitch conversion, and FIG. Figure 4 is a block circuit diagram for explaining a specific example of interpolation processing,
FIG. 5 is a schematic diagram for explaining the relationship between connection points and blocks in looping processing, FIG. 6 is a musical tone signal waveform diagram, and FIG.
8 is a functional block diagram for explaining another specific example of the present invention, FIG. 8 is a functional block diagram for explaining pitch detection operation, FIG. 9 is a block diagram for explaining peak detection operation, and FIG. Figure 10 is a waveform diagram of the musical tone signal and envelope, Figure 11 is a waveform diagram of decay rate information of the musical tone signal,
FIG. 12 is a functional block diagram for explaining envelope detection operation, and FIG. 13 is F! Characteristic diagram of R filter, 1st
Figure 4 is a waveform diagram showing peak value data after envelope correction of the musical tone signal, Figure 15 is a characteristic diagram of the comb filter,
Figure 16 is a waveform diagram for explaining the operation of setting the optimal looping point, Figure 17 is a waveform diagram showing musical tone signals before and after time axis correction, and Figure 18 is a quasi-instantaneous waveform diagram for wave height data after time axis correction. A schematic diagram showing the structure of a block for bit compression. Figure 19 is a waveform diagram showing loop data obtained by repeatedly connecting waveforms in the looping section. Figure 20 is a formant part after envelope correction based on decay rate information. FIG. 21 is a flowchart for explaining operations before and after actual looping processing, FIG. 22 is a block circuit diagram showing the schematic configuration of a quasi-instantaneous bit compression encoding system, and FIG. 23 is a waveform diagram showing generation data. A schematic diagram showing a specific example of one block of data obtained by quasi-instantaneous bit compression encoding. Figure 24 is a schematic diagram showing the contents of the processing at the beginning of a musical tone signal. Figure 25 is an audio processing diagram. - It is a block diagram showing a configuration example of a system including a unit and its surroundings. Pitch 12 Comprehensive drawings to rule the 12 marquis 3rd figure + L-7" start? F Renu 5th figure P/+ Kyoshichi creation 8th figure! 11 Figure 7 Oh! Zata for Shimarehi partial tribes Figure 20 Envelope → Exit 1 Figure 12 Characteristics of the FIR filter Figure 13 Figure 14 -Pinigudebi 145 etc. Who B's Figure 21

Claims (1)

[Claims] A sound source data generation method in which a plurality of interpolated samples are also generated by interpolating waveform data of a predetermined period consisting of a plurality of real samples, and a point with the closest value between the real sample and the interpolated sample is repeatedly used as a connected sample of the waveform.