JP2842106B2 - Transmission method of acoustic signal - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sound signal transmission method for compressing and transmitting the information amount of a sound signal.

[0002]

2. Description of the Related Art In recent years, analog signals have been frequently transmitted (or recorded and reproduced) as digital signals, for example, PCM (pulse code modulation) signals.
Since the PCM signal has a large amount of information, a wide transmission band is required for transmission (or recording / reproduction) of the PCM signal. So in transmission equipment (recording and playback equipment) that performs signal processing of digital signals, and other various equipment,
Conventionally, digital signals have been efficiently processed with a small amount of information. As a high-efficiency coding method capable of coding a signal efficiently with a smaller amount of information, a signal is predicted, and only a component (residual component) that is deviated from a predicted value is recorded and transmitted. Various high-efficiency coding schemes, and certain types of transforms (generally, orthogonal transforms) are applied to signals to extract signal features, and the characteristic portions of the signals or the human visual or auditory sense The amount of information (bits) per sample is high, such as by using the property that it is sensitive to changes in parts with little change, but difficult to detect even if there is some error in parts where the signal changes rapidly. It is well known that various high-efficiency coding schemes and the like that reduce the number have been conventionally proposed.

[0003] However, for example, in a well-known high-efficiency coding system configured by applying linear prediction that is practically used for compressing the amount of information when transmitting a telephone voice signal, a numerator of a prediction system ( Although zero) and the denominator (pole) are predicted, their prediction ability is not so good and has little effect on the transmission of music signals. In addition, there are other methods for performing linear prediction, such as the Perco method, which also have a performance limit. As described above, there has been no prediction method suitable for compressing the amount of information related to the transmission (recording) of a music signal. The bit compression method that has been used is to perform bit compression by using a masking effect between adjacent frequencies by, for example, orthogonal transform, and for example, a period having a predetermined number of sample points from an acoustic signal is defined by a window function. After multiplying the joints of the respective frames and cutting them out as successive one frame periods in a state of being gently connected to each other, and performing orthogonal transform by fast Fourier transform (FFT) for each frame, The operation of a masking curve is performed on the data obtained by the fast Fourier transform of FIG. (Recorded) to the spectrum having smaller amplitude than the masking curve does not transmit (record), and so on, spectrum not hear masked by the main spectrum, record it, it is to avoid transmission. Note that DFT, DCT, and others can be used as the orthogonal transform.

[0004] By the way, the reason why the data of an acoustic signal can be reduced as described above is that, when a sound of a certain frequency component is strongly radiated, the vicinity of the frequency component is a characteristic of human hearing. This is because the detection capability for the frequency of the signal decreases, and a small number of bits are allocated to the frequency component of the portion where the detection capability is reduced, and the signal amount of the small amplitude is not transmitted at all. Can be realized. Although the signal accuracy is considerably reduced due to the decrease in the amount of data as described above, the auditory masking effect often makes it impossible to distinguish the original signal from the case where the original signal is heard. As described above, conventionally, high-efficiency coding has been performed by predicting a signal or orthogonally transforming a signal. However, both a signal prediction technique and an orthogonal transform technique are used. A better fusion can be achieved by predicting the orthogonally transformed signal of the next frame from the orthogonally transformed data, but this has not been done before. .

[0005] In order to solve the above-mentioned problem, the applicant company has previously set a predetermined time length according to Japanese Patent Application No. 4-30076 "Method for predicting the phase of an acoustic signal". The first and second Fourier transform frames in each of the sequential Fourier transform frames extracted from the audio signal are discretely Fourier-transformed using the same window function, and the first and second Fourier transforms described above are performed. By the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each frame, for each of the first and second Fourier transform frames, phase information for each discrete frequency is obtained, In each of the first and second Fourier transform frames, the aspect of the change of the phase information for each of the same discrete frequencies corresponding to each other is obtained, and the state of the change of the phase information for each of the individual discrete frequencies is determined. Is constant on the time axis, and a predetermined number in the Fourier transform frame of a third time position existing at a time position that is an integral multiple of the time difference between the first time position and the second time position described above. A method for determining individual phase information of discrete frequencies of the above, and predicting phase information of a Fourier transform frame at a third time position. The encoding facilitated the transmission (recording) of the acoustic signal and the configuration of the sound source of the musical instrument.

FIG. 9 is a block diagram of an encoder that can be used in an application of the phase information prediction method already proposed by the applicant company, and FIG. 10 is a block diagram of a decoder. In FIG. 11, the waveform indicated by AUDIO in the uppermost part is an audio signal targeted for high-efficiency encoding, and the audio signal is Fr
.. are cut out so as to have a certain time length, respectively, and are sequentially Fourier-transformed frames, such as 0, 1, 2,... A state in which each of the above-mentioned Fourier transform frames is multiplied by a window function so as to have a period having 1024 sample points from the acoustic signal, and successive frames are successively overlapped with each other to be loosely connected. Are extracted as one successive frame period. The above-described Fourier transform frame for each frame is subjected to an orthogonal transform by, for example, Fourier transform (DFT) or fast Fourier transform (FFT) of a discrete finite series. In the following description, it is described that the orthogonal transform is performed by a fast Fourier transform (FFT).

When an FFT operation is performed on a Fourier transform frame for each frame, the number of data (the number of samples) in the Fourier transform frame for each frame is N, and the sampling frequency is fs. , F
= Fs / N The FFT operation result for each discrete frequency (total N frequencies) for each frequency interval of f indicated by fs / N is obtained, but the FFT operation result is a real number for each discrete frequency. Part (Real) amplitude and the imaginary part (Ima
g) amplitude. Using the real part (Real) amplitude and the imaginary part (Imag) amplitude of each discrete frequency described above, by the following equations 1 and 2, for each discrete frequency described above, The synthetic amplitude term (Am
p) and a phase term (Phase) are obtained. By the way, FF
In N real part (Real) amplitudes and N imaginary part (Imag) amplitudes obtained for each discrete frequency as a result of the T operation, the N real part (Real) amplitudes have the same value. Two appear, and N imaginary parts (I
mag) Since two amplitudes having the same value appear in absolute value, the number of effective combined amplitude terms (Amp) becomes の of the total number, and the effective phase term (Phase) term Since the number is also と of the total number, the number of FFT output data is
It can be equal to the number of real input data of FT.

[0008]

(Equation 1)

(Equation 2)

Now, for a continuous Fourier transform frame (frame) that is continuous on the time axis, the real part (Real) amplitude and the imaginary part (Ima) of each discrete frequency are provided.
g) By using the amplitude and the equations (1) and (2), the composite amplitude term (Am
When p) and the phase term (Phase) are obtained, it is natural to predict that two adjacent frames on the time axis will have the same amplitude in the most common sense. Yes, and actually the sampling frequency is 44.1K
Assuming that the number of samples is 1024 points in Hz and the frame length (the frame length is about 1/40 second), the spectrum obtained by performing the FFT operation on the signal of the sound of the piano, which is obtained by performing the FFT operation, has a certain value. Comparing the 512 amplitudes in one frame with the 512 amplitudes in the next frame of that frame or the 512 amplitudes in the next frame of the next frame, It has been confirmed that the amount of change between spectra is extremely small.

On the other hand, it is difficult to predict the phase in successive frames that are continuous on the time axis. The reason is that there is no relation between the repetition time of the successive frames and the frequency of the signal. It is clear from the beginning and end of the frame regardless of the frame, and it has been conventionally difficult to predict the signal by orthogonal transform. Then, assuming that the sampling frequency is 44.1 KHz and the number of samples is 1024 as a frame length,
Even if the distribution of 512 phases in one frame based on the FFT operation result obtained by actually performing the FFT operation on the signal of the piano sound is seen, the distribution of the phases is random. 5 in the frame of
It turns out that the distribution of the 12 phases cannot be predicted.

The previously proposed inventor of the present applicant, Takahashi, predicts the phase information of another frame by using the phase information of one frame in a sequential frame on the time axis. Although it is not possible, for two frames, the variation of the phase information between the same predetermined number of discrete frequency phase information obtained as a result of the Fourier transform for each Fourier transform frame is constant on the time axis. Paying attention to the fact that if it exists, it is possible to predict the phase information of another frame by using the relationship, and as described above, "for two frames, as a result of the Fourier transform for each Fourier transform frame, The change state of the obtained phase information among the same predetermined number of discrete frequencies for each discrete frequency is constant on the time axis. "Bridge's hypothesis), and actually use various signals such as a single frequency sine wave signal, a composite signal of multiple frequency sine wave signals, a musical instrument (piano) sound signal, etc. When conducting experiments, correct prediction results were obtained to the extent that the phase of the frame predicted according to the above hypothesis and the phase of the actual frame were practically consistent, and Takahashi's hypothesis was The fact that it holds in practice is supported by various experimental results.

According to Takahashi's hypothesis, the data θi (1) of N / 2 phase terms obtained for each discrete frequency in, for example, frame 1 shown in FIG. N / 2 phase term data θi (2) obtained for each discrete frequency and N / 2 phase term data θi (3) obtained for each discrete frequency in frame 3, for example. And N / 2 phase term data θi (4) obtained for each discrete frequency in frame 4, for phase term data at the same frequency value, The data is, for example, θ1
If the data of the phase term in frame 2 is, for example, θ2, the data of the phase term in frame 3 is, for example, θ3, and the data of the phase term in frame 4 is, for example, θ4, θ2
−θ1 ≒ θ3−θ2 ≒ θ4−θ3 ≒ Δθa, the amount of phase change between the frames is substantially the same, so if this hypothesis holds, then for each of the two frames Is different from the above two frames if the phase difference between the data of the phase terms of all the frequency values corresponding to each other in the N / 2 phase terms obtained for each discrete frequency of It is possible to predict the phase of another frame. Specifically, as in the above-described example, the data of the phase term of a specific frequency value fa in frame 1 is θ1 and that in frame 2 is If the data of the phase term of the specific frequency value fa is θ2, the data θ3 of the phase term of the specific frequency value fa in the frame next to the frame 2 is converted to θ3
≒ θ2 + (θ2−θ1) = 2θ2−θ1... (A), and the above phase prediction is individually performed for each of the discrete frequencies in frames 1 and 2. By doing so, the phase of the signal of frame 3 can be predicted.

According to the Takahashi hypothesis described above, even if the phase information of one frame, for example, the phase information of only frame 1 is known, it is impossible to predict the phase information of another frame using the phase information. However, if the phase information of two frames is known, it is possible to predict the phase information of another frame, and it is possible to predict the phase information of two adjacent frames, for example, the phase information of frame 1 and the phase information of frame 2. This indicates that it is possible to predict the phase information of frames other than the two frames described above, and it is 2 K times apart from the time length of one frame.
If the phase information of one frame, for example, frame 1 and the phase information of frame 4 are known, the phase information of another frame, for example, frame 7, which is separated from frame 4 by K times the time length of one frame is predicted. It is also possible to express the above-mentioned Takahashi's hypothesis in general, as follows: “First, first, second, and third in each sequential Fourier transform frame cut out from the audio signal so as to have a predetermined time length.
The same predetermined number of discrete numbers obtained as a result of the Fourier transform for each of the first and second Fourier transform frames by performing a Fourier transform discretely on each of the second Fourier transform frames using the same window function. From the data for each frequency, phase information for each discrete frequency is obtained for each of the first and second Fourier transform frames, and the phase information is obtained for each of the first and second Fourier transform frames. The manner of change of the phase information for each identical discrete frequency is determined, and the manner of change of the phase information for each discrete frequency is assumed to be constant on the time axis, and the first time position and the The individual phase information of a predetermined number of discrete frequencies in the Fourier transform frame at the third time position existing at a time position that is an integer multiple of the time difference from the second time position is determined. Fourier It can be to be able to predict the phase information of the conversion frame ".

The block diagram shown in FIG. 9 is obtained by applying the above-described acoustic signal phase prediction technology based on the Takahashi hypothesis.
When performing recording and transmission by compressing the information amount of a signal to be recorded and transmitted, the residual signal Ai (m) -Ai (m-1) of the amplitude and the residual of the phase for each frame. Difference signal Δθi
FIG. 10 shows a configuration example of an encoder configured to record and transmit (m) = θi (m) − {2θi (m−1) −θi (m−2)}. 2 shows a configuration example of a decoder. Each of the residual signals becomes zero if the prediction is successful, but usually a slight deviation from the predicted value occurs.
The residual signal rarely becomes zero, but the information amount of the residual signal is much smaller than the information amount of the original signal.

In FIG. 9, reference numeral 1 denotes an input terminal of a digital audio signal to be recorded and transmitted. The digital audio signal supplied to the input terminal 1 of the digital audio signal is converted by the overlap unit 2 into an input terminal. The sequential Fourier transform frames cut out so as to have a predetermined fixed time length in an overlapped state are stored in the window function processing unit 3 for each period having, for example, N sample points. Is multiplied by a window function to form a continuous one frame period in which the seams of the respective frames are overlapped with each other so as to be gently connected, and then the fast Fourier transform operation (FFT) is performed in block 4. Operation) is performed. As a result of the FFT operation, for each Fourier transform frame, the same constant frequency interval f {where 1
The number of data in the Fourier transform frame for each frame The number of samples is N, the sampling frequency is fs, and f = fs /
N} for each of the N discrete frequencies
l) The data of the FFT operation result including the amplitude and the imaginary part (Imag) amplitude is obtained.

As described above, the data at each of the N discrete frequencies obtained as a result of the FFT operation is subjected to signal processing by a different signal processing device for each of the discrete frequency data. However, FIG. 9 illustrates only the configuration of one of the N signal processing devices. In FIG. 9, the signal processing device is a component between the orthogonal coordinate → polar coordinate converter 6 and the multiplexer 18. Data of a specific discrete frequency consisting of a real part and an imaginary part for each of N discrete frequencies obtained as a result of the FFT operation is subjected to polar coordinate conversion in a rectangular coordinate → polar coordinate conversion unit 6 to obtain an amplitude term and After being separated into the phase term, the calculation of the amplitude according to the above-described equation 1 and the calculation of the phase according to the above-described equation 2 are performed. For each frequency,
The composite amplitude term Ai (m) and the phase term θi (m) are obtained. The synthesized amplitude term Ai of a specific discrete frequency obtained as a result of the calculation performed by the above-described rectangular coordinate → polar coordinate conversion unit 6
(M) is supplied to the latch circuit 7, the subtractor 8, and the data selector 9. The data selector 9 is connected to the terminal 34
The composite amplitude term A
One of the setting data based on i (m) and the residual data output from the subtracter 8 is selected and output to the multiplexer 18. The switching operation of the data selector 9 is performed in conjunction with the switching operation of the data selector 16 described later.

Further, the above-described rectangular coordinate → polar coordinate conversion unit 6
The phase term θi (m) of a specific discrete frequency obtained as a result of the calculation is supplied to the latch circuit 10, the subtractor 14, and the data selector 16. Data of the phase θi (m) output from the rectangular coordinate → polar coordinate converter 6 as a calculation result of the phase of a specific discrete frequency (supposed to be fa) in the m-th frame is held in the latch circuit 10. The phase data held in the latch circuit 10 before the operation, that is, the phase θi output from the orthogonal coordinate → polar coordinate conversion unit 6 as a calculation result of the phase of the specific discrete frequency fa in the (m−1) -th frame. The data of (m-1) is held in the latch circuit 12 in the phase prediction unit PFC. In the illustrated configuration example, the phase prediction unit PFC includes a latch circuit 12, an amplifier 11 having a gain of 2, and a subtractor 13.
And is constituted by. The data of the phase held in the latch circuit 12 before the data of the phase θi (m−1) is held in the latch circuit 12, that is, the data of the specific discrete frequency fa in the (m−2) th frame The data of the phase θi (m−2) output from the orthogonal coordinate → polar coordinate conversion unit 6 as the calculation result of the phase is supplied to the subtractor 13 as a subtraction signal. Since what is supplied as the minuend signal is the output from the amplifier 11 having the above-mentioned gain of 2, the data of the predicted phase output from the subtractor 13, that is,
The prediction phase data output from the phase prediction unit PFC is 2
θi (m−1) −θi (m−2).

A phase θi (m) of a specific discrete frequency obtained as a result of the calculation performed by the rectangular coordinate → polar coordinate converter 6
Are a latch circuit 10, a subtractor 13, and a data selector 16.
And supplied to. The phase θi output from the rectangular coordinate → polar coordinate converter 6 as a calculation result of the phase of a specific discrete frequency (here, supposed to be fa) in the m-th frame
The data of the phase held in the latch circuit 10 before the data of (m) is held in the latch circuit 10, that is, m
The data of the phase θi (m−1) output from the rectangular coordinate → polar coordinate conversion unit 6 as the calculation result of the phase of the specific discrete frequency fa in the −1st frame is the latch circuit 1
2 and a latch circuit 12 in a phase prediction unit PFC constituted by an amplifier 11 having a gain of 2 and a subtractor 13.
Is held. The terminal 5 is a supply terminal for the shift clock signal. The data of the phase held in the latch circuit 12 before the data of the phase θi (m−1) is held in the latch circuit 12, that is, the data of the specific discrete frequency fa in the (m−2) th frame The phase θ output from the rectangular coordinate to polar coordinate conversion unit 6 as the phase calculation result
The data of i (m−2) is supplied to the subtractor 13 as a decrement signal, and the data of i (m−2) is supplied to the subtractor 13 as a subtrahend signal. , The data of the predicted phase output from the subtractor 13, that is, the phase prediction unit PF
The predicted phase data output from C is 2θi (m-1) -θ
i (m-2).

By subtracting the predicted phase data 2θi (m−1) −θi (m−2) output from the phase prediction unit PFC from the actual phase data θi (m) in the subtractor 14, , The phase residual signal Δθi (m) = θi (m) − {2θi (m−1) −θi (m−2)} from the subtractor 14.
Is output to the quantization scaling unit 17 via the data selector 16, where the quantization size is set according to the frequency, and then the multiplexer 1
8 is supplied. The multiplexer 18 combines the amplitude residual signal ΔAi (m) having a specific discrete frequency (hereinafter, supposed to be fa) with the phase residual signal Δθi (m) having a specific discrete frequency fa. To the output terminal 19. Note that the output terminal 19 has at least once an original amplitude component Ai (m) and a phase component θi
(m) Needless to say, it is supplied. The multiplexer 18 has an amplitude residual signal ΔAi (m) for each of a predetermined number of discrete frequencies in the Fourier transform frame.
And output data from all the other signal processing circuits that generate the phase residual signal Δθi (m), the multiplexer 18 outputs the compressed audio signal of the compressed state. The data is output and sent to the transmission line via the output terminal 19.

Next, an example configuration of the receiving side of the transmission system will be described with reference to FIG. In FIG. 10, reference numeral 20 denotes an input terminal of a decoder provided on the receiving side.
(Not shown), the data of the acoustic signal in a state where the amount of information transmitted to the receiving side is compressed, that is, the amplitude residual signal ΔAi (m ) And a phase residual signal, an amplitude residual signal ΔAi (m) for each specific discrete frequency, separated by a demultiplexer (not shown), Residual signal Δθi (m)
Are provided. The signal including the amplitude residual signal ΔAi (m) and the phase residual signal Δθi (m) for each specific discrete frequency supplied to the input terminal 20 is used for data of a specific discrete frequency. Is subjected to predetermined signal processing by a signal processing device that performs the above signal processing. FIG. 10 representatively shows one signal processing device that performs signal processing on data of a specific discrete frequency.

The input terminal 20 includes an original amplitude component Ai (m) and phase component θi (m) of a specific discrete frequency, an amplitude residual signal ΔAi (m), a phase residual signal Δθi (m), and the like. In the signal processing circuit supplied with the signal constituted by
(m), the phase component θi (m), and the amplitude residual signal ΔAi
(m) and the phase residual signal Δθi (m) are separated and supplied to the amplitude component signal processing circuit and the phase component processing circuit.
In FIG. 10, the data supplied from the latch circuit 24 to the adder 25 is the amplitude residual signal ΔAi (m) supplied from the demultiplexer 21 to the adder 25.
Is the amplitude residual signal ΔAi (m) in the m-th frame.
In the case of, since the data is the data Ai (m-1) of the combined amplitude term of the (m-1) th frame, the adder 25 calculates the data Ai (m-1) of the combined amplitude term of the (m-1) th frame. And the amplitude residual signal ΔAi (m) in the m-th frame are added, and the adder 25 outputs data Ai (m) of the synthesized amplitude term of the m-th frame, which is
4 and supplied to the polar → rectangular coordinate converter 27 via the data selector 26.

As described above, the original amplitude component Ai (m) and the phase component θ of a specific discrete frequency (hereinafter, supposed to be fa) separated by the demultiplexer 21 are described.
i (m) and amplitude residual signal ΔAi (m) and phase residual signal Δ
The phase component θi (m) and the phase residual signal Δθi (m) with respect to θi (m) are supplied to the adder 29 and the data selector 30 after being requantized by the requantizer 22. The data output from the adder 29 is
The data θi (m) of the phase term of the specific discrete frequency in the m-th frame is determined by the adder 29 described above.
Are added to the phase residual signal Δθi (m) supplied to the adder 29 from the demultiplexer 21.
Is the phase residual signal Δθi (m) in the m-th frame.
And 2θi (m−1) −θ output from the phase prediction unit PFC
This is because i (m−2). The data θi (m) of the phase term of a specific discrete frequency in the m-th frame output from the adder 29 is
Are supplied to a polar coordinate → rectangular coordinate conversion unit 27 via a latch circuit 28. Terminal 2
3 is supplied with a shift clock signal.

The phase θ output from the adder 29
Before the data of i (m) is held in the latch circuit 28,
The data of the phase held in the latch circuit 28, that is, the data of the phase θi (m−1) of the specific discrete frequency fa in the (m−1) -th frame is divided into the latch circuit 12 and the amplifier 11 having the gain of 2. And a subtractor 13, which is held in the latch circuit 12 in the phase prediction unit PFC. The data of the phase held in the latch circuit 12 before the data of the phase θi (m-1) is held in the latch circuit 12, that is, the data of the specific discrete frequency fa in the (m-2) th frame The data of the phase θi (m−2) is supplied to the subtractor 13 as a subtraction signal, and the data of the phase θi (m−2) is supplied to the subtractor 13 as a subtrahend signal. 11, the data of the predicted phase output from the subtractor 13, that is, the data θi (m) of the predicted phase output from the phase prediction unit PFC is 2θi (m−1) −θi ( m-
2). Therefore, the prediction phase data θi (m) = 2θi (m−1) − output from the phase prediction unit PFC.
θi (m−2) and the phase residual signal Δθi of the m-th frame
When (m) = θi (m) − {2θi (m−1) −θi (m−2)} is added by the adder 29, the adder 29 outputs a specific discrete signal in the m-th frame. The data θi (m) of the phase term of the frequency fa is output.

As described above, m output from the adder 25
The data Ai (m) of the combined amplitude term of the n-th frame and the data θi (m) of the phase term of a specific discrete frequency in the m-th frame output from the adder 29 are predetermined data, respectively. Polar coordinates via selectors 26 and 30 →
When supplied to the rectangular coordinate converter 27, the polar coordinates →
The Cartesian coordinate converter 27 converts the data Ai (m) of the composite amplitude term of the m-th frame and the data θi (m) of the phase term of a specific discrete frequency in the m-th frame output from the adder 29. m), the real part (Real) amplitude and the imaginary part (Imag) amplitude at the above-mentioned specific discrete frequency fa are calculated and output, and are supplied to the inverse FFT operation part 31. Since the output data from all the signal processing circuits provided for each of a predetermined number of discrete frequencies in the Fourier transform frame is supplied to the inverse FFT operation unit 31, the inverse FFT operation unit 31
From the original digital audio signal, the data of the original audio signal is restored, and the window function processing unit 32 and the overlap addition unit 33 perform multiplication of the window function and overlap addition. Output terminal 3
5 will be sent.

[0025]

If the sound signal is transmitted (recorded and reproduced) by applying the above-mentioned "method of predicting the phase of the sound signal" proposed by the present applicant, the sound signal is increased. The desired effect can be obtained because the data can be transmitted (recorded and reproduced) with efficient coding. However, the amount of data can be reduced by using even higher efficiency coding, and the sound signal can be transmitted in a high quality state. There has been a demand for a method of transmitting an acoustic signal that can be (recorded, reproduced). In the present applicant's company, in Japanese Patent Application No. 4-265340, the same window function is applied to the signal of each successive Fourier transform frame cut out from the acoustic signal so as to have a predetermined fixed time length. Fourier transform is performed discretely by using the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each Fourier transform frame, and obtained for each Fourier transform frame described above. In a method of transmitting an acoustic signal to be transmitted after highly efficient encoding of an amplitude component and a phase component of each discrete frequency, the same predetermined number of discrete components obtained as a result of Fourier transform for each successive Fourier transform frame Among the phase components obtained from the data for each frequency, the phase component in the Fourier transform frame to be transmitted first is transmitted as the initial setting data, and is determined in advance. In the Fourier transform frames existing between the two Fourier transform frames, an arithmetic progression having a slope determined by the phase component in the two Fourier transform frames selected at a predetermined interval as a tolerance is used. The prediction residual of the phase component obtained by performing the prediction on the phase component, and the transmission method of the acoustic signal configured to transmit the above-described tolerance data, and the Fourier transform obtained for each sequential Fourier transform frame as a result of the Fourier transform Among the phase components obtained from the same predetermined number of discrete frequency data, the phase component of the first Fourier transform frame in N (where N is a natural number of 3 or more) Fourier transform frames is set as initial setting data. And transmitting the phase components of the first Fourier transform frame in the N Fourier transform frames. And an arithmetic progression having a slope determined by the phase component of the Nth Fourier transform frame as a tolerance, the difference between the first Fourier transform frame and the Nth Fourier transform frame in the N number of Fourier transform frames described above. The present invention proposes a sound signal transmission method for transmitting a prediction residual of a phase component obtained by performing prediction of a phase component in each existing Fourier transform frame and the above-described tolerance data, and a transmission method of another sound signal. ing.

According to the transmission method of the audio signal proposed by the present applicant, the audio signal can be encoded and transmitted with high efficiency. In the method, a phase component in each of the Fourier transform frames existing between the two Fourier transform frames is obtained by an arithmetic progression having a slope determined by a re-phase component in the two predetermined Fourier transform frames. Is performed to obtain the prediction residual of the phase component. If the selection of the two Fourier transform frames is not appropriate, for example, (a) in FIG.
(B) The problem is that the residual represented by the distance between the solid line shown by the solid line and the phase component in each Fourier transform frame becomes large, and a solution to that problem is required. Was done.

[0027]

SUMMARY OF THE INVENTION The present invention discretely uses the same window function for each successive Fourier transform frame signal extracted from an audio signal so as to have a predetermined fixed time length. Fourier transform, and using the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each Fourier transform frame, the amplitude for each discrete frequency obtained for each Fourier transform frame described above. The component and the phase component are obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame in the transmission method of the acoustic signal transmitted after high-efficiency encoding. Within the phase component,
A phase component in a Fourier transform frame to be transmitted first is transmitted as initial setting data, and three or more continuous Fourier transform frames are sequentially set as individual Fourier transform frame groups. For each, a phase component in the Fourier transform frame at one end of the Fourier transform frame group,
A method for transmitting an acoustic signal in which an arithmetic average of inclination values individually determined by phase components in other individual Fourier transform frames in the group is transmitted as tolerance data of each Fourier transform frame group. And the above-described method of transmitting an acoustic signal, among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, the Fourier to be transmitted first. A phase component in the transform frame is transmitted as initial setting data, and a continuous Fourier transform frame group in which three or more consecutive Fourier transform frames are set as individual groups is converted into a Fourier transform frame group for each Fourier transform frame group. Using the phase component in each Fourier transform frame constituting the group Method of transmitting an acoustic signal to the value of the slope which is determined by application of the least squares method, transmitted as the tolerance data for each Fourier transform frame group,
In the method of transmitting an acoustic signal, among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each sequential Fourier transform frame, the Fourier transform to be transmitted first is used. A phase component in a frame is transmitted as initial setting data, and a continuous Fourier transform frame group in which three or more continuous Fourier transform frames are set as individual groups is provided. The arithmetic mean value of the slope values individually determined by the phase component in the Fourier transform frame at one end and the phase components in the other individual Fourier transform frames in the group are calculated by using the respective Fourier transform frame groups. The above-mentioned individual Fourier transform equations are given by an arithmetic progression with a tolerance of A method for transmitting an acoustic signal configured to transmit a prediction residual of a phase component obtained by performing prediction on a phase component in each Fourier transform frame belonging to a frame group and the tolerance data, and the acoustic signal In the transmission method, among the phase components obtained from the data of the same predetermined number of discrete frequencies obtained as a result of the Fourier transform for each successive Fourier transform frame, the phase component in the Fourier transform frame to be transmitted first is While transmitting as the initial setting data, for the sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as an individual group, the Fourier transform frame group is configured for each Fourier transform frame group. Determined by applying the least-squares method using the phase component in each Fourier transform frame The prediction residual of the phase component obtained by performing prediction on the phase component in each Fourier transform frame belonging to the individual Fourier transform frame group by an arithmetic progression as a tolerance of each Fourier transform frame group, In the method of transmitting an acoustic signal and the method of transmitting an acoustic signal described above, the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame. The phase components of the first Fourier transform frame in N (where N is a natural number not less than 3) Fourier transform frames are transmitted as initialization data among the phase components obtained from Of the first Fourier transform frame and the other N-1 Fourier transform frames in the Fourier transform frame A method of transmitting an acoustic signal, wherein the arithmetic average of the values of the gradients individually determined by the individual phase components of the audio signal are transmitted as tolerance data of the respective Fourier transform frame groups, and In the transmission method, N (where N is a natural number of 3 or more) N out of phase components obtained from data of the same predetermined number of discrete frequencies obtained as a result of Fourier transform for each successive Fourier transform frame The phase component of the first Fourier transform frame in the Fourier transform frame is transmitted as initialization data, and the slope value determined by applying the least squares method using the phase components in the N Fourier transform frames described above. In the transmission method of the audio signal in which the audio signal is transmitted as the tolerance data, and in the transmission method of the audio signal, Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each of the Fourier transform frames, the first among N (where N is a natural number of 3 or more) Fourier transform frames Is transmitted as initialization data, and the phase components of the first Fourier transform frame in the N number of Fourier transform frames and the individual components of the other N-1 Fourier transform frames are individually transmitted. The first Fourier transform frame in the N number of Fourier transform frames described above is calculated by an arithmetic progression using the arithmetic mean value of the slope values individually determined by the phase components as the tolerance data of the respective Fourier transform frame groups. Phase component in each Fourier transform frame existing between the Fourier transform frame and the Nth Fourier transform frame In the acoustic signal transmission method for transmitting the prediction residual of the phase component obtained by performing the prediction and the above-described tolerance data, and the acoustic signal transmission method, the Fourier transform is performed for each successive Fourier transform frame. Of the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of
(Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, and the phase component of the N Fourier transform frames is minimized using the phase component. By using an arithmetic progression in which the value of the gradient determined by the application of the square method is used as the tolerance data, each of the N Fourier transform frames existing between the first Fourier transform frame and the N-th Fourier transform frame is used A method for transmitting an acoustic signal is provided which transmits a prediction residual of a phase component obtained by predicting a phase component in a Fourier transform frame and the above-described tolerance data.

[0028]

The discrete Fourier transform is performed on the signals of the respective sequential Fourier transform frames cut out from the audio signal so as to have a predetermined fixed time length using the same window function. The data to be transmitted first among the phase components for each discrete frequency obtained for each Fourier transform frame described above, using the same predetermined number of data for each discrete frequency obtained as a result of the Fourier transform for each A phase component in the Fourier transform frame is transmitted as initial setting data, and a continuous Fourier transform frame set as three or more consecutive Fourier transform frames is set as an individual group. For each Fourier transform frame group, a Fourier transform is performed. The phase component in the Fourier transform frame at one end of the group of frames and the other individual Fourier transforms in the group The arithmetic mean value of the values of the slope, each is determined separately by the phase component in the frame, transmitted as tolerance data for each Fourier transform frame group. Fourier transform is performed discretely using the same window function on the signal of each sequential Fourier transform frame cut out from the acoustic signal so as to have a predetermined time length, and the Fourier transform for each of the above-described Fourier transform frames is performed. Using the same predetermined number of discrete frequency data obtained as a result of the conversion, the Fourier transform frame to be transmitted first among the phase components for each discrete frequency obtained for each Fourier transform frame described above. Are transmitted as initial setting data, and a sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as an individual group, for each Fourier transform frame group described above, a Fourier transform frame Determined by applying the least-squares method using the phase component in each Fourier transform frame constituting the group The value of the slope, is transmitted as the tolerance data for each Fourier transform frame group. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data. In addition, with respect to a sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as individual groups, the Fourier transform at one end of the Fourier transform frame group is performed for each of the Fourier transform frame groups described above. The arithmetic mean of the slope values individually determined by the phase components in the frames and the phase components in the other individual Fourier transform frames in the group is the arithmetic sequence that is the tolerance of each Fourier transform frame group. , Each Fourier transform frame belonging to a group of individual Fourier transform frames A prediction residual of the phase component obtained by performing prediction of the phase component that transmits and tolerance data described above. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data. In addition, with respect to a sequential Fourier transform frame group in which three or more continuous Fourier transform frames are set as individual groups, each Fourier transform frame group constituting the Fourier transform frame group is provided for each of the above-described Fourier transform frame groups. The phase value in each Fourier transform frame belonging to each individual Fourier transform frame group is calculated by an arithmetic progression using the value of the gradient determined by applying the least squares method using the phase component in the frame as the tolerance of each Fourier transform frame group. Prediction residual of the phase component obtained by performing prediction on the component, To transmit and tolerance data. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, N (where N is a natural number of 3 or more) Fourier transform frames The phase component of the first Fourier transform frame is transmitted as initial setting data, and the phase components of the first Fourier transform frame in the N Fourier transform frames and the individual components of the other N-1 Fourier transform frames are individually transmitted. The arithmetic average of the slope values individually determined by the phase components is transmitted as tolerance data of each Fourier transform frame group. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, N (where N is a natural number of 3 or more) Fourier transform frames 1
The phase component of the fourth Fourier transform frame is transmitted as initial setting data, and the value of the gradient determined by applying the least square method using the phase components in the N Fourier transform frames is transmitted as tolerance data. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, N (where N is a natural number of 3 or more) Fourier transform frames The phase component of the first Fourier transform frame is transmitted as initial setting data, and the phase components of the first Fourier transform frame in the N Fourier transform frames and the individual components of the other N-1 Fourier transform frames are individually transmitted. The first Fourier transform frame in the N number of Fourier transform frames is calculated by an arithmetic progression using the arithmetic mean value of the slope values individually determined by the phase components as the tolerance data of the respective Fourier transform frame groups. And the phase in each Fourier transform frame existing between the Nth Fourier transform frame Min and prediction residual of the phase component obtained by performing prediction of a tolerance data described above is transmitted. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, N (where N is a natural number of 3 or more) Fourier transform frames The phase component of the first Fourier transform frame is transmitted as initial setting data, and the value of the gradient determined by applying the least square method using the phase components of the N Fourier transform frames is used as tolerance data. According to the difference sequence, N
Prediction residual of a phase component obtained by performing prediction of a phase component in each Fourier transform frame existing between the first Fourier transform frame and the Nth Fourier transform frame in the Fourier transform frames, and the above-described tolerance. And data transmission.

[0029]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an audio signal transmission method according to the present invention. FIG.
3 is a block diagram showing a configuration example of an encoder used in the audio signal transmission method of the present invention, and FIG. 2 is a block diagram showing a configuration example of a decoder used in the audio signal transmission method of the present invention. FIGS. 4 to 8 are diagrams used for explaining the audio signal transmission method of the present invention, FIG. 9 is a block diagram showing an example of the configuration of an encoder used for the already proposed audio signal transmission method, and FIG. FIG. 11 is a block diagram showing a configuration example of a decoder used in the audio signal transmission method of FIG. 1, and FIG. 11 is a diagram used for explaining a previously proposed audio signal transmission method. FIGS. 1 and 3 show configuration examples of an encoder used in the audio signal transmission method according to the present invention. Corresponding components use the same reference numerals as those used in the encoder shown in FIG. 9, and show an example of the configuration of a decoder used in the audio signal transmission method of the present invention. 2 shown in FIG.
The same reference numerals as those used in the encoder shown in FIG. 10 denote the same components as those in the encoder used in the proposed method of transmitting an acoustic signal described above with reference to FIG. You are using

In FIGS. 1 and 3 showing examples of the configuration of an encoder used in the audio signal transmission method of the present invention, reference numeral 1 denotes an input terminal of a digital audio signal to be recorded and transmitted; A sequential Fourier transform frame cut out so that the digital audio signal supplied to the input terminal 1 of the digital audio signal is overlapped by the overlap unit 2 and has a predetermined fixed time length. In the window function processing unit 3, a window function is multiplied for each period having, for example, N sample points, so that the joints of the respective frames are sequentially overlapped with each other and connected gradually. After one frame period, a fast Fourier transform operation (FFT operation) is performed in the fast Fourier transform operation unit 4 shown in block 4.
Then, F in the fast Fourier transform operation unit 4 described above is obtained.
As a result of the FT operation, the same constant frequency interval f 、 for each Fourier transform frame, where N is the number of data samples in the Fourier transform frame for each frame, fs is the sampling frequency, and f = fs For each of the N discrete frequencies having / N}, data of an FFT operation result including a real part (Real) amplitude and an imaginary part (Imag) amplitude is obtained.

As described above, the data for each of the N discrete frequencies obtained as a result of the FFT operation is subjected to signal processing by a different signal processor for each data of each discrete frequency. 1 and 3, only the configuration of one of the N signal processors is shown between the orthogonal coordinate → polar coordinate converter 6 and the multiplexer 18 in the figures. ing. Data of a specific discrete frequency consisting of a real part and an imaginary part of each of N discrete frequencies obtained as a result of the FFT operation is subjected to polar coordinate conversion by a rectangular coordinate → polar coordinate conversion unit 6 to obtain an amplitude term and After being separated into the phase term, the calculation of the amplitude by the above-described equation 1 and the calculation of the phase by the above-described equation 2 are performed, so that for each of the discrete frequencies described above for the sequential frames, A composite amplitude term Ai (m) and a phase term θi (m) are obtained.

The synthesized amplitude term Ai (m) of a specific discrete frequency obtained as a result of the calculation performed by the orthogonal coordinate → polar coordinate conversion unit 6 is supplied to and stored in a storage device (memory) 36. The phase term θi (m) of a specific discrete frequency obtained as a result of the calculation performed by the orthogonal coordinate → polar coordinate converter 6 is supplied to and stored in a storage device (memory) 37. As the memory 36, a memory having a function of storing at least information on the current frame and information on the immediately preceding frame is used. As the memory 37, signal processing as described later is used. One having a function of storing information on a plurality of frames of a predetermined number of frames required for performing signal processing on the phase component in the embodiment is used.

The memory 36, which stores the sequential data of the synthesized amplitude terms of the specific discrete frequencies obtained as a result of the calculation performed by the above-described orthogonal coordinate → polar coordinate converter 6, is stored in a control device (not shown). Under the control, the data Ai (m) and Ai (m-1) of the two combined amplitude terms stored in the memory 36 adjacent to each other on the time axis are read. Said 2
The data Ai (m) of one of the combined amplitude terms Ai (m) and Ai (m−1) is supplied to the subtractor 8 as a minuend signal and to the data selector 9. The data Ai (m-1) of the other combined amplitude term of the two combined amplitude term data Ai (m) and Ai (m-1) is supplied to the subtractor 8. Then, the data selector 9 described above uses a switching signal output from a control device (not shown) to switch between the setting data based on the synthesized amplitude term Ai (m) and the residual data output from the subtractor 8. One of them is selected and output to the multiplexer 18. In the case of the encoder shown in FIG. 3, the switching operation of the data selector 9 is performed in conjunction with the switching operation of the data selector 16 described later. The masking method is applied to the data stored in the memory 36, that is, the sequential data of the synthetic amplitude term of a specific discrete frequency obtained as a result of the calculation performed by the rectangular coordinate → polar coordinate conversion unit 6. By applying adaptive quantization by various methods such as data thinning, etc.
Ensuring that even higher efficiency encoding is performed is a preferred embodiment.

In the encoder shown in FIG. 1, sequential data of phase terms θi (m) of specific discrete frequencies obtained as a result of calculation by the above-described orthogonal coordinate → polar coordinate converter 6 are stored. The memory 37 is connected to respective components such as a tolerance change point determination unit 38, a tolerance calculation unit 39, an adder 40, a quantization scaling unit 17, and the like.
The above rectangular coordinates in the encoder shown in →
The memory 37 in which the sequential data of the phase term θi (m) of the specific discrete frequency obtained as the calculation result of the polar coordinate conversion unit 6 is stored, a tolerance change point determination unit 38, a tolerance calculation unit 39, Adder 40, subtractor 14 and data selector 16
Are connected.

In the encoder of FIG. 1 shown as an encoder used in the method of transmitting an acoustic signal of the present invention,
Using the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each Fourier transform frame by the signal processing in the memory 37 and the components described above, the above-described Fourier transforms are performed. Among the phase components for each discrete frequency obtained for each frame, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data, and three or more consecutive Fourier transform frames are formed as individual groups. For the set sequential Fourier transform frame group, for each Fourier transform frame group, a phase component in a Fourier transform frame at one end of the Fourier transform frame group and a phase component in another individual Fourier transform frame in the group. And the arithmetic mean of the slope values, each determined individually by The phase component in the Fourier transform frame to be transmitted first among the phase components for each discrete frequency obtained for each Fourier transform frame is transmitted as the tolerance data of the Fourier transform frame group. While transmitting as setting data,
With respect to a sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as individual groups, the phase in each Fourier transform frame constituting the Fourier transform frame group is determined for each of the Fourier transform frame groups described above. The value of the gradient determined by applying the least squares method using the components is transmitted as tolerance data of each Fourier transform frame group, or the same value obtained as a result of Fourier transform for each successive Fourier transform frame is used. The phase component of the first Fourier transform frame in N (where N is a natural number of 3 or more) Fourier transform frames among the phase components obtained from the data for each of the predetermined number of discrete frequencies is transmitted as the initial setting data. And the first Fourier transform frame in the N Fourier transform frames described above. And transmitting the arithmetic mean value of the slope values individually determined by the phase components of each of the other N-1 Fourier transform frames as tolerance data of each Fourier transform frame group, Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each successive Fourier transform frame, N (where N is a natural number of 3 or more) Fourier transform frames The phase component of the first Fourier transform frame is transmitted as initial setting data, and the value of the gradient determined by applying the least squares method is transmitted as tolerance data using the phase components of the N Fourier transform frames. Or perform an action.

In the encoder of FIG. 3, which is shown as an encoder used in the sound signal transmission method of the present invention, the signal processing in the memory 37 and each of the above-mentioned components makes it possible to sequentially perform the Fourier transform frame by frame. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform, the phase component in the Fourier transform frame to be transmitted first is transmitted as the initial setting data, and the continuous 3 For a sequential Fourier transform frame group set as an individual group with four or more Fourier transform frames, for each of the Fourier transform frame groups described above, a phase component in a Fourier transform frame at one end of the Fourier transform frame group, With the phase components in other individual Fourier transform frames in the group. The arithmetic mean of the values of the slopes determined individually is set to be a tolerance of each of the Fourier transform frame groups, and the prediction of the phase component in each Fourier transform frame belonging to each Fourier transform frame group is performed by an arithmetic progression. The phase difference obtained by transmitting the predicted residual of the phase component obtained by performing the above and the above-mentioned tolerance data, or the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame Among the components, a phase component in a Fourier transform frame to be transmitted first is transmitted as initial setting data, and three or more consecutive Fourier transform frames are sequentially set as a group of successive Fourier transform frames. For each Fourier transform frame group described above, each Fourier transform frame constituting the Fourier transform frame group is used. In the Fourier transform frames belonging to the individual Fourier transform frame groups, the slope value determined by applying the least squares method using the phase components in the Fourier transform frame groups is used as the tolerance of each Fourier transform frame group. For each of the same predetermined number of discrete frequencies obtained by transmitting the prediction residual of the phase component obtained by performing the prediction on the phase component and the above-mentioned tolerance data, or as a result of the Fourier transform for each successive Fourier transform frame, Of the phase components obtained from the data of
The phase component of the first Fourier transform frame in N (where N is a natural number of 3 or more) Fourier transform frames is transmitted as initial setting data, and the first Fourier transform frame in the N Fourier transform frames described above is transmitted. An arithmetic progression in which the arithmetic mean value of the slope values individually determined by the phase components of the above and the other N-1 Fourier transform frames is used as tolerance data of the respective Fourier transform frame groups The prediction residual of the phase component obtained by performing the prediction of the phase component in each Fourier transform frame existing between the first Fourier transform frame and the Nth Fourier transform frame in the N Fourier transform frames described above And the above-mentioned tolerance data, or a Fourier transform for each successive Fourier transform frame. Among the phase component obtained from the data of each discrete frequency resulting identical predetermined number of determined the conversion, N
(Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, and the N
The first Fourier transform frame in the N Fourier transform frames and the first Fourier transform frame in the N Fourier transform frames are obtained by an arithmetic progression in which the slope value determined by applying the least square method using the phase components in the four Fourier transform frames is used as tolerance data. An operation of transmitting a prediction residual of a phase component obtained by predicting a phase component in each Fourier transform frame existing between the fourth Fourier transform frame and the above-described tolerance data is performed.

In the encoder of FIGS. 1 and 3 used in the above-described audio signal transmission method according to the present invention, high efficiency encoding by signal processing performed on phase components is shown.
In the above aspect, in the case of the above-described high efficiency coding mode, in the case of the sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as individual groups, For each group, the arithmetic mean of the slope values individually determined by the phase components in the Fourier transform frame at one end of the group of Fourier transform frames and the phase components in the other individual Fourier transform frames of the group. Is the tolerance of each of the Fourier transform frame groups, and in the case of the above-described high efficiency coding mode, successive Fourier transform frames in which three or more consecutive Fourier transform frames are set as individual groups. For each of the Fourier transform frame groups described above, each Fourier transform frame group The value of the slope which is determined by application of the least squares method using the phase component in the converted frame, and the tolerance of each Fourier transform frame group.

Here, a method of obtaining the tolerance in the above-described Fourier transform frame group will be described with reference to FIGS. FIG. 4 shows a group of Fourier transform frames composed of three or more consecutive Fourier transform frames as in the case of the above-described high efficiency coding mode.
The arithmetic mean of the slope values individually determined by the phase component in the Fourier transform frame at one end in each Fourier transform frame group and the phase components in the other individual Fourier transform frames in the group is calculated by the Fourier transform. FIG. 5 and FIG. 6 are diagrams for explaining the case where the average inclination of the transform frame group is used, and FIGS. 5 and 6 show the average inclination of the Fourier transform frame group shown in FIG. FIG. 7 is a diagram comparing the previously proposed Fourier transform frame group with the gradient, and FIGS. 7 and 8 further show three consecutive frames in the case of the high efficiency coding mode described above.
FIG. 7 is a diagram for explaining a gradient determined by applying the least-squares method using a phase component in an individual Fourier transform frame of a group of Fourier transform frames by a plurality of Fourier transform frames, and FIG. FIG. 8 is a diagram used to explain the case where the above-described gradient is obtained by applying the least squares method, with the phase component in the end Fourier transform frame in the group of Fourier transform frames as the origin, and FIG. It is a figure used for description at the time of calculating the above-mentioned inclination by applying a general least squares method which does not make a phase component in a Fourier transform frame of an end in a frame group the origin.

In FIG. 4 to FIG. 6, 0, 1, 2,.
Are the numbers indicating the sequential Fourier transform frames, and the vertical lines at the respective positions of 0, 1, 2,... .Theta.i (m) described above, where m is a number 0, 1, 2,... Indicating sequential Fourier transform frames, that is, the same predetermined number obtained as a result of Fourier transform for each successive Fourier transform frame Are obtained as examples of phase components θi (0), θi (1), θi (2),... Obtained from the data for each discrete frequency. 4 to 6 show the same predetermined number obtained as a result of the Fourier transform for each successive Fourier transform frame (indicated by numbers 0, 1, 2,... On the horizontal axis). Of the phase components {θi (0), θi (1), θi (2)...} Obtained from the data for each discrete frequency of the Fourier transform frame in which the phase components are transmitted as initialization data, This is an example where the Fourier transform frame to which the phase component is to be transmitted first is the Fourier transform frame 0. First, with reference to FIG. 4, a method of obtaining the average inclination in the Fourier transform frame group in the case of the high efficiency coding mode described above will be described as follows. In FIG. 4, each phase component θi (0), obtained from data of the same predetermined number of discrete frequencies obtained as a result of Fourier transform for each successive Fourier transform frame,
θi (1), θi (2)... data coordinates (x0, y0), (x
With the point of the coordinates x0, y0 of the initial value in (1, y), (x2, y2)... As the starting point (origin), the coordinates (x0, y0) of the aforementioned origin
And the inclination of the straight line connecting the coordinates (x1, y), the inclination of the straight line connecting the coordinates (x0, y0) of the origin and the coordinates (x2, y2),
The coordinates of the origin (x0, y0) in the group of Fourier transform frames, such as the slope of a straight line connecting the coordinates (x0, y0) and the coordinates (x3, y3) of the origin, All coordinates (x1, y1) other than coordinates (x0, y0),
(x2, y2)... are obtained, and the slope value of the straight line La shown by the broken line in FIG. 4 having the slope of the arithmetic average of the slope values of the straight lines individually determined as described above. To the tolerance k
Used as The specific calculation of the inclination of the straight line La indicated by the broken line in FIG. 4 is obtained according to the following “Equation 3”. In FIG. 4, a straight line Lr indicated by a solid line is a straight line directly connecting the coordinates of both ends in a group of Fourier transform frames, as in the above-described method of transmitting an acoustic signal.

[0040]

(Equation 3)

In both cases (a) and (b) of FIG. 4, the total amount of the residuals of each group of Fourier transform frames is
The method of transmitting an acoustic signal according to the present invention is less than the method of transmitting an acoustic signal already proposed. In FIG. 4, a vertical arrow indicated by a broken line indicates a residual amount in the case of the audio signal transmission method of the present invention, and a vertical arrow indicated by a solid line indicates the residual amount in the case of the proposed acoustic signal transmission method. The difference is shown. FIG. 5 is exemplified in FIG. 1 described above with reference to FIG. 1, where the average inclination of each successive Fourier transform frame group obtained by the method of obtaining the average inclination of the Fourier transform frame group described above with reference to a tolerance k. This is an example in which high efficiency coding is performed in a high efficiency coding mode (a mode in which no residual is transmitted) using an encoder having a certain configuration. Is represented. FIG. 6 shows the average of each successive Fourier transform frame group obtained by the method of obtaining the average inclination of the Fourier transform frame group described above with reference to FIG. 4 using the encoder having the configuration illustrated in FIG. Is a case where high efficiency coding is performed in a high efficiency coding mode (a mode in which a residual is also transmitted) using an encoder having the configuration illustrated in FIG. And
The tolerance k in this case is represented by “Equation 5”.

Next, the least squares method is applied by using the phase components in the individual Fourier transform frames of a group of Fourier transform frames by three or more consecutive Fourier transform frames in the case of the high-efficiency encoding mode. 7 and 8 for explaining the gradient determined by the following equation, the above-mentioned gradient is obtained by applying the least squares method with the origin being the phase component in the Fourier transform frame at the end of the group of Fourier transform frames. In FIG. 7 used for the description of FIG. 7, the points of the coordinates x0, y0 of the initial values in the coordinates (x0, y0), (x1, y), (x2, y2). Assuming that a straight line indicated by a dashed line as the start end (origin) is y = ax + b, the straight line passes through the origin, and this line is represented by y−y0 = a (x−x0)
Can be expressed as Now, assuming that the error between the coordinates of the respective data and the straight line is Δi (i = 1, 2, 3,...), The error Δi = y−y0 = a (x−x0) is minimized. In order to obtain a straight line by the least squares method, the value obtained by partially differentiating "Equation 6" with "a" becomes 0, and "Equation 8" and "Equation 9" are obtained by solving "Equation 7".

[0043]

(Equation 6)

(Equation 7)

(Equation 8)

(Equation 9)

FIG. 1 shows a method of transmitting an acoustic signal according to the present invention.
In the case of using the encoder having the configuration as exemplified in (1) and not performing the transmission of the residual of the phase component, the variables in the above (Equation 9) are set to the variables in the description of the acoustic signal transmission method of the present invention. Replace with the variable used in "Formula 10"
Then, the "Equation 10" is arranged, and the tolerance is calculated by "Equation 11". Note that, when the transmission method of the acoustic signal of the present invention is implemented in a mode in which the residual of the phase component is transmitted by using the encoder having the configuration mode illustrated in FIG. Is set to k-1 to obtain a tolerance.

Next, application of the least squares method using the phase components in the individual Fourier transform frames of a group of Fourier transform frames by three or more consecutive Fourier transform frames in the case of the highly efficient encoding mode In FIG. 8 for explaining the slope determined by the following equation, the slope is obtained by applying a general least-squares method, which does not use the phase component in the end Fourier transform frame in the group of Fourier transform frames as the origin. In FIG. 8 used for explaining the case, the points of the coordinates x0, y0 of the initial values in the coordinates (x0, y0), (x1, y1), (x2, y2). Is a starting line (origin), a straight line indicated by a broken line is represented by y = ax + b, and an error between the coordinates of the respective data and the straight line is represented by Δi (i = 1, 2, 3,...). yy In order to obtain a straight line that minimizes 0 = a (x−x0) by the least squares method, the value obtained by partially differentiating “Equation 12” with a becomes 0, and “Equation 1”
By solving "3", "Equation 14" and "Equation 15" are obtained. Also,
What is necessary is that the value obtained by partially differentiating “Equation 12” with b becomes 0,
By substituting b of “Equation 17” obtained by solving “Equation 16” into “Equation 14”, “Equation 18” is obtained, and “a” of “Equation 19” is obtained. Further, a of “Equation 19” is substituted into “Equation 17” to obtain “Equation 20”.
Get. Also in this case, the transmission method of the audio signal of the present invention is implemented by using the encoder having the configuration illustrated in FIG. 1 and not transmitting the residual of the phase component,
Alternatively, the audio signal transmission method of the present invention can be implemented as an embodiment for transmitting the residual of the phase component using the encoder having the configuration illustrated in FIG.

[0046]

(Equation 12)

(Equation 13)

[Equation 14]

(Equation 15)

(Equation 16)

[Equation 17]

(Equation 18)

[Equation 19]

(Equation 20)

In the encoder shown in FIG. 1, the sequential data of the phase term θi (m) of a specific discrete frequency obtained as a result of the calculation performed by the rectangular coordinate → polar coordinate converter 6 is stored. The signal processing for the phase component in each component such as the memory 37 and the tolerance change point determination unit 38, the tolerance calculation unit 39, the adder 40, the quantization scaling unit 17, etc. High-efficiency coding is performed in any of the high-efficiency coding modes. When the encoder shown in FIG. 1 performs high-efficiency encoding in the high-efficiency encoding mode described above,
Using the same predetermined number of data for each discrete frequency obtained as a result of the Fourier transform for each Fourier transform frame, among the phase components for each discrete frequency obtained for each Fourier transform frame, , The phase component in the Fourier transform frame to be transmitted first, which is read from the memory 37, is given to the quantization scaling unit 17 by the operation of a control unit (not shown), and the quantization scaling unit 17 outputs the predetermined number of bits. From the initial setting data, the tolerance k is supplied to the multiplexer 18, and the tolerance k in the sequential Fourier transform frame group set as an individual group by the tolerance change point determination unit 38 for three or more consecutive Fourier transform frames is shown in the figure. Each data read from the memory 37 and supplied to the tolerance calculator 39 under the operation of Using the data of the phase component for each of the Fourier transform frame groups, the phase component in the Fourier transform frame at one end of the Fourier transform frame group is obtained by an operation performed by the tolerance calculator 39 in accordance with the above-described “Equation 3”. The tolerance data is supplied to the multiplexer 18 with the arithmetic mean of the slope values individually determined by the phase components in the other individual Fourier transform frames of the group as the tolerance.

When the encoder shown in FIG. 1 performs high-efficiency encoding in the above-described high-efficiency encoding mode, it is obtained as a result of Fourier transform for each Fourier transform frame described above. Using the same predetermined number of data for each discrete frequency, the memory 3 stores the phase components for each discrete frequency obtained for each Fourier transform frame.
7, the phase component in the Fourier transform frame to be transmitted first to be transmitted is provided to the quantization scaling unit 17 by the operation of a control unit (not shown), and the quantization scaling unit 17 initializes the predetermined number of bits. After being set as setting data, it is supplied to the multiplexer 18, and
With respect to three or more consecutive Fourier transform frames, the tolerance k in the sequential Fourier transform frame group set as an individual group by the tolerance change point determining unit 38 is obtained from the memory 37 under the operation of a control unit (not shown). Using the data of the phase component for each Fourier transform frame group read out and supplied to the tolerance calculation unit 39, the above-described “Equation 8” is used.
Alternatively, a value of the slope determined by applying the least squares method by an operation performed by the tolerance calculation unit 39 according to “Equation 19” or “Equation 20” is obtained as tolerance data,
The tolerance data is supplied to the multiplexer 18.

Further, when the encoder shown in FIG. 1 performs high-efficiency encoding in the above-described high-efficiency encoding mode, the encoder obtains the result of the above-described Fourier transform for each Fourier transform frame. Using the same predetermined number of data for each discrete frequency, among the phase components for each discrete frequency obtained for each Fourier transform frame, the N The phase component in the Fourier transform frame to be transmitted first is given to the quantization scaling unit 17 by the operation of a control unit (not shown), and the quantization scaling unit 17 Is supplied to the multiplexer 18, and the tolerance change point determination unit 3 is used for the N Fourier transform frames.
8, the tolerance k in the N sequential Fourier transform frame groups set as individual groups is read out from the memory 37 under the operation of a control unit (not shown) and supplied to the tolerance calculation unit 39. Using the data of the phase component for each of the transformed frame groups, one of the Fourier transformed frame groups is obtained by an operation performed by the tolerance calculation unit 39 in accordance with the above-described "Equation 4" (j in Eq. 4 is N). The arithmetic mean value of the slope values individually determined by the phase component in the Fourier transform frame at the end of the and the phase component in the other individual Fourier transform frames in the group is obtained as a tolerance, and the tolerance data is obtained. The signal is supplied to the multiplexer 18.

Furthermore, when the encoder shown in FIG. 1 performs high-efficiency encoding in the above-described high-efficiency encoding mode, the result of the above-described Fourier transform for each Fourier transform frame is as follows. Using the same predetermined number of data obtained for each discrete frequency, among the phase components for each discrete frequency obtained for each Fourier transform frame described above,
The phase component in the Fourier transform frame to be transmitted first among the N Fourier transform frames read out from the memory 37 is given to the quantization scaling unit 17 by the operation of a control unit (not shown), and the quantized scaling unit 17 The data is initialized to a predetermined number of bits by the scaling unit 17 and then supplied to the multiplexer 18.
For each of the Fourier transform frames, the tolerance k in the sequential Fourier transform frame group set as an individual group by the tolerance change point determination unit 38 is read from the memory 37 under the operation of a control unit (not shown) and Using the data of the phase component for each Fourier transform frame group supplied to the calculation unit 39, the above-described “Equation 8” or “Equation 1”
The value of the slope determined by the application of the least squares method is obtained as the tolerance data by the operation performed by the tolerance calculator 39 according to “9” or “Equation 20”, and the tolerance data is supplied to the multiplexer 18.

On the other hand, in the encoder shown in FIG. 3 which is shown as an encoder used in the sound signal transmission method of the present invention, the above-described rectangular coordinate → polar coordinate converter 6 is used.
A memory 37 in which sequential data of phase terms θi (m) of specific discrete frequencies obtained as a result of the calculation are stored.
And the tolerance change point determination unit 38, the tolerance calculation unit 39, and the adder 4
0, the signal processing for the phase component in each component such as the subtractor 14 and the data selector 16, etc.
The high-efficiency coding is performed in any of the high-efficiency coding modes. When the encoder shown in FIG. 3 performs high-efficiency encoding in the above-described high-efficiency encoding mode, the same predetermined code obtained as a result of the Fourier transform for each Fourier transform frame described above is used. Using the data for each discrete frequency of the number, among the phase components for each discrete frequency obtained for each of the above-mentioned Fourier transform frames, the phase component in the first Fourier transform frame to be transmitted and read out from the memory 37 is determined. Is supplied to the quantization scaling unit 17 via the data selector 16 by the operation of the control unit (not shown), and the quantization scaling unit 17 converts the data into initialization data of a predetermined number of bits.
The signal is supplied to the multiplexer 18, and the tolerance change point determination unit 38 is used for three or more consecutive Fourier transform frames.
The tolerance k in the sequential Fourier transform frame group set as an individual group is read from the memory 37 under the operation of a control unit (not shown), and the tolerance calculation unit 39
Using the data of the phase component for each Fourier transform frame group supplied to the above, it is obtained by an operation performed by the tolerance calculation unit 39 in accordance with the above-described “Equation 3”, and the Fourier transform of one end of the Fourier transform frame group is performed. Providing, as a tolerance, the arithmetic mean of the slope values individually determined by the phase component in the frame and the phase components in the other individual Fourier transform frames in the group, and providing the tolerance data to the multiplexer 18; The aforementioned tolerance k and the phase component θi (m−1) read from the memory 37.
Is added by the adder 40 to obtain a predicted value of the phase component, which is given to the subtractor 14 as a minuend signal, and the phase component θi read out from the memory 37 to the subtractor 14 is obtained.
By supplying (m) as the subtrahend signal, the subtractor 14 outputs the residual component Δθ (m) of the phase component, and supplies it to the quantization scaling unit 17 via the data selector 16, and outputs it to the quantization scaling unit 17. The multiplexer 1 converts the data to the initial setting data of a predetermined number of bits in the
8

Further, the encoder shown in FIG.
When performing high-efficiency encoding in the high-efficiency encoding mode described above, the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each Fourier transform frame described above is used. Among the phase components for each discrete frequency obtained for each Fourier transform frame, the phase component in the first Fourier transform frame to be transmitted among the Fourier transform frames read from the memory 37 is as follows:
The data selector 1 is operated by a control unit (not shown).
6, the data is supplied to a quantization scaling unit 17, the data is converted into a predetermined number of bits of initial setting data by the quantization scaling unit 17, and then supplied to a multiplexer 18.
In addition, for three or more consecutive Fourier transform frames, the tolerance k in the sequential Fourier transform frame group set as an individual group by the tolerance change point determination unit 38.
Using the phase component data for each Fourier transform frame group read from the memory 37 and supplied to the tolerance calculation unit 39 under the operation of a control unit (not shown), The value of the slope determined by applying the least squares method by the calculation performed by the tolerance calculation unit 39 according to “Equation 19” or “Equation 20” is defined as tolerance data.
The tolerance data is supplied to the multiplexer 18, and the above-described tolerance k and the phase component θ read from the memory 37 are supplied.
The data of i (m-1) is added to the adder 40 to obtain a predicted value of the phase component, which is provided to the subtractor 14 as a minuend signal, and the phase component read from the memory 37 is supplied to the subtractor 14. By supplying θi (m) as the minuend signal, the subtractor 14 outputs the residual component Δθ (m) of the phase component, and supplies it to the quantization scaling unit 17 via the data selector 16. Quantization scaling unit 1
At 7, the data is supplied to the multiplexer 18 as initialization data of a predetermined number of bits.

Further, the encoder shown in FIG.
When performing high-efficiency encoding in the high-efficiency encoding mode described above, the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each Fourier transform frame described above is used. Of the phase components for each discrete frequency obtained for each Fourier transform frame described above, the phase component in the first Fourier transform frame to be transmitted among the N Fourier transform frames read from the memory 37 is The data is supplied to a quantization scaling unit 17 via a data selector 16 by an operation of a control unit (not shown), and the quantization scaling unit 17 supplies the data as initial setting data having a predetermined number of bits to a multiplexer 18.
Further, regarding the above-mentioned N Fourier transform frames,
The tolerance k in the N sequential Fourier transform frame groups set as individual groups by the tolerance change point determination unit 38
Is calculated using the phase component data for each Fourier transform frame group read from the memory 37 and supplied to the tolerance calculation unit 39 under the operation of a control unit (not shown). (J in Equation 4 is N) in accordance with the above-described operation, the phase component in the Fourier transform frame at one end of the group of Fourier transform frames and the other individual Fourier transform in the group are obtained. With the arithmetic mean of the slope values individually determined by the phase components in the frame as the tolerance, the tolerance data is supplied to the multiplexer 18, and the above-described tolerance k and the phase component θi read from the memory 37 are further provided. (m-
The data of 1) is added by the adder 40 to obtain a predicted value of the phase component, and the result is given to the subtractor 14 as a minuend signal.
The phase component θi (m) read out from the memory 37 is supplied to the subtractor 14 as a minuend signal, so that the subtractor 14 outputs the residual component Δθ (m) of the phase component, and the data is output to the data selector. The data is supplied to a quantization scaling unit 17 through 16, and is supplied to a multiplexer 18 as initialization data having a predetermined number of bits by the quantization scaling unit 17.

Further, when the encoder shown in FIG. 3 performs the high-efficiency encoding in the above-described high-efficiency encoding mode, as a result of the Fourier transform for each Fourier transform frame described above, Using the same predetermined number of data obtained for each discrete frequency, among the phase components for each discrete frequency obtained for each Fourier transform frame described above,
The phase component in the Fourier transform frame to be transmitted first among the N Fourier transform frames read from the memory 37 is given to the quantization scaling unit 17 via the data selector 16 by the operation of a control unit (not shown). After the data is converted into initial setting data of a predetermined number of bits by the quantization scaling unit 17, the multiplexer 18
, And for the N consecutive Fourier transform frames, the tolerance k in the sequential Fourier transform frame group set as an individual group by the tolerance change point determination unit 38 is determined by the operation of a control unit (not shown). Memory 3
7 using the data of the phase component for each Fourier transform frame group read from 7 and supplied to the tolerance calculation unit 39, in accordance with the above-described “Equation 8”, “Equation 19” or “Equation 20”. The value of the slope determined by the application of the least squares method is obtained as the tolerance data by the operation to be performed, the tolerance data is supplied to the multiplexer 18, and the tolerance k and the phase component θi ( The data of m-1) are added by the adder 40 to obtain a predicted value of the phase component, which is given to the subtractor 14 as a minuend signal, and the phase component θi ( m) as the minuend signal, the subtractor 14 outputs the residual component Δ
θ (m) is output and supplied to the quantization scaling unit 17 via the data selector 16. The quantization scaling unit 17 supplies the data as initialization data of a predetermined number of bits, and then supplies the data to the multiplexer 18.

When the transmission of the sound signal data by the sound signal transmission method of the present invention is performed as shown by a broken line La in FIG. 5, that is, the result of the Fourier transform for each successive Fourier transform frame Among the phase components obtained from the same predetermined number of discrete frequency data obtained as
After transmitting the phase component in the Fourier transform frame to be transmitted first as the initial setting data, three or more consecutive Fourier transform frames are sequentially set for each of the Fourier transform frame groups set as individual groups. For each group, the arithmetic mean of the slope values individually determined by the phase components in the Fourier transform frame at one end of the group of Fourier transform frames and the phase components in the other individual Fourier transform frames of the group. Is transmitted only as the tolerance data of each Fourier transform frame group, and when the residual is not sent, the phase component in the Fourier transform frame group after the first Fourier transform frame group and thereafter. Data differs from the true phase component data. However, since the residual amount of the phase component is not transmitted, the compression rate of the transmitted data amount becomes large, so that even if the data amount of the phase component is largely compressed as described above, The present invention can be applied favorably to applications that do not cause a major problem, for example, to store a musical sound signal of a musical instrument sound source with a small amount of data.

An example of the data compression amount in the case where the transmission of the audio signal data by the audio signal transmission method of the present invention as described above is performed as shown by a broken line La in FIG. It is as follows. Now, when the number of Fourier transform frames constituting the Fourier transform frames constituting each group of Fourier transform frames is set to 10, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data. Also, since only the tolerance is sent as the data of the phase components in the remaining nine Fourier transform frames, the data amount for transmitting the phase components in the group of Fourier transform frames is reduced to two-ninths. Become. It goes without saying that the compression rate of the data amount increases as the number of Fourier transform frames constituting the Fourier transform frames constituting each group of Fourier transform frames increases. Note that in each Fourier transform frame group subsequent to the group of Fourier transform frames that transmitted the phase component in the Fourier transform frame to be transmitted first as the initial setting data, only the tolerance data needs to be transmitted. Therefore, the compression rate of the data amount is 1/9.

Since the actual data has a certain length of time (for example, the sound of a piano is about 10 seconds), the data amount is reduced to about 1/9 as a whole. Also, for the amplitude component, for example, the data amount is reduced to 1/4 to 5 by application of a masking processing method or the like.
Since the audio signal is compressed by a factor of 1, the compression ratio of the audio signal data is about 1/50 in total. On the other hand, a conventional compression method of the data amount of an audio signal using a masking processing method or the like generally has a compression ratio of 1/4 to 1/5. Therefore, according to the acoustic signal transmission method of the present invention,
Except for the overlap due to the window function processing, a large amount of data can be compressed as compared with the conventional method.

When the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by a broken line La in FIG. 5, the value of the tolerance itself is transmitted when the tolerance value is transmitted. Is likely to be affected by the cumulative error component caused by the rounding, but when the sequential Fourier transform frame groups set as individual groups are implemented as being constituted by a large number of Fourier transform frames, It is preferable to increase the number of digits of the tolerance value (increase the precision) in accordance with the number of Fourier transform frames constituting. In the case where the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by a broken line La in FIG. 5, the Fourier transform frame transmitted with the phase component as the initial setting data is used. The origin to be used for calculating the average value of the gradient of the phase component in the next group of Fourier transform frames after the group of Fourier transform frames has the first Fourier transform frame of the next group of Fourier transform frames described above. Coordinates indicating phase components {(x, y ') are used instead of (x, y) in FIG.

When the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by a broken line La in FIG. 5, the gradient of the phase component in the group of Fourier transform frames is set. The phase may become discontinuous between the average values of the above, but the human hearing becomes less sensitive to the phase as the frequency becomes higher. Will be
It depends on the frequency value of the frequency spectrum. Therefore, the degree of phase discontinuity that would cause a problem in the sense of hearing is determined in advance by obtaining the phase detection limit frequency characteristic, tabulating it, and determining the frequency between the fourier transform frame groups of each group. By comparing the discontinuity amount of the phase and the value on the table, it is determined whether the discontinuity amount is equal to or less than an allowable value, and the discontinuity amount is equal to or less than the allowable value. In this case, the signal is transmitted as it is, and when the discontinuity is equal to or larger than the allowable value, even if the residual amount of the phase slightly increases, the discontinuity of the phase is improved.
The inclinations of the phase components are made closer to each other. In addition,
If the number of Fourier transform frames constituting the Fourier transform frames constituting a group of Fourier transform frames such as a preset sound source may be arbitrarily changed, the degree of the discontinuity of the phase is equal to or less than the allowable limit value. For each group of Fourier transform frames, the number of Fourier transform frames constituting each Fourier transform frame may be changed such that

Further, when the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by the broken line La in FIG. 5, and the residual of the phase component is also transmitted, that is, Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each sequential Fourier transform frame, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data. After that, with respect to a sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as individual groups, for each Fourier transform frame group, a Fourier transform frame at one end of the Fourier transform frame group Determined separately by the phase component at, and the phase components at other individual Fourier transform frames in the group When the arithmetic mean of the slope value is transmitted as the tolerance data of each Fourier transform frame group, and the residual of the phase component is also transmitted, the data amount increases by the data amount of the residual of the phase component. However, the sound signal can be transmitted with high quality in a state where the amount of data to be transmitted is significantly reduced as compared with the conventional method.

Next, when the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by a broken line La in FIG. 6, that is, for each successive Fourier transform frame Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform, the first Fourier transform frame in N (where N is a natural number of 3 or more) Fourier transform frames Is transmitted as initialization data, and the phase components of the first Fourier transform frame in the N Fourier transform frames and the other N−1
If the arithmetic mean value of the slope values individually determined by the individual phase components of the Fourier transform frames is transmitted as tolerance data of the respective Fourier transform frame groups, each group of Fourier transform frames Although the amount of data increases by the amount of transmitting the phase component of the first Fourier transform frame as the initial setting data for each transform frame, the problem of phase discontinuity as described with reference to FIG. Can be transmitted with good quality.

Next, when the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by the broken line La in FIG. 6, and the residual of the phase component is also transmitted, ie, sequentially. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each Fourier transform frame, N (where N is 3
The phase component of the first Fourier transform frame in the (natural number) Fourier transform frames described above is transmitted as initial setting data. And transmitting the arithmetic mean value of the slope values individually determined by the individual phase components of the N-1 Fourier transform frames as the tolerance data of the respective Fourier transform frame groups, 5 is transmitted, the data amount of the phase component and the data amount of the phase component of the first Fourier transform frame increase for each group of Fourier transform frames.
As described above, the acoustic signal can be transmitted with good quality in a state where the amount of data to be transmitted is significantly reduced as compared with the conventional method, without causing the phase discontinuity problem as described above. It can be said that the above-described transmission mode of the acoustic signal is suitable for the transmission of an acoustic signal including various frequency components, unlike the case of the musical instrument sound. It goes without saying that the number of bits to be allocated to the residual component of the phase may be smaller (for example, about 程度) than the number of bits to be allocated to the initial setting data. Further, when the transmission of the audio signal data by the audio signal transmission method of the present invention is performed as shown by a broken line La in FIG. 6, there is an advantage that the reproduction can be performed halfway.

Next, the transmission of the audio signal data by the audio signal transmission method of the present invention is obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame. Among the phase components, a phase component in a Fourier transform frame to be transmitted first is transmitted as initial setting data, and three or more consecutive Fourier transform frames are sequentially set as a group of sequential Fourier transform frames. For each Fourier transform frame group, the slope value determined by applying the least squares method using the phase component in each Fourier transform frame that constitutes the Fourier transform frame group is calculated using the tolerance of each Fourier transform frame group. Data, or Fourier transform for each successive Fourier transform frame. Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data, and three or more consecutive For the sequential Fourier transform frames set as the Fourier transform frames as individual groups, for each Fourier transform frame, the least squares method is used using the phase component in each Fourier transform frame constituting the Fourier transform frame. A phase component obtained by predicting a phase component in each Fourier transform frame belonging to each Fourier transform frame group by an arithmetic progression in which a value of the slope determined by application is set as a tolerance of each Fourier transform frame group. Of the prediction residual and the above-mentioned tolerance data, or Among the phase component obtained from the data of each discrete frequency of the same predetermined number of determined as a result of the Fourier transform for each Fourier transform frame, N
(Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, and the phase component in the N Fourier transform frames is used. The first Fourier transform frame of the N Fourier transform frames and N are calculated by an arithmetic progression using the slope value determined by applying the least square method as the tolerance data.
The transmission of the acoustic signal by transmitting the prediction residual of the phase component obtained by performing the prediction of the phase component in each Fourier transform frame existing between the fourth Fourier transform frame and the above-described tolerance data. Is implemented, the slope of the phase component in each group of Fourier transform frames is obtained by the least squares method of one of the two methods described above with reference to FIGS. This is the same as described above with reference to FIGS. 5 and 6 except that the component tolerance is used.

Next, an example configuration of the decoder will be described with reference to FIG. In FIG. 2, reference numeral 20 denotes an input terminal of a decoder provided on the receiving side. The input terminal 20 is connected to the receiving side via a transmission path (not shown) from the transmitting side described with reference to FIG. Data of an acoustic signal in a state where the amount of information transmitted to the original is compressed, that is, the original amplitude component Ai (m) and the amplitude residual signal ΔAi (m) of the specific discrete frequency, and the original of the specific discrete frequency. For example, the phase component of the first Fourier transform frame 0 selected to transmit the initialization data, or the phase component of the first Fourier transform frame of the N Fourier transform frames} and The information on the tolerance k of the arithmetic progression determined by the phase component of the Fourier transform frame, and the prediction of the phase component in each Fourier transform frame is performed using the tolerance k. The original amplitude component Ai (m of a specific discrete frequency separated by a demultiplexer (not shown) from the audio signal data including the prediction residual Δθi (m) of the phase component obtained as described above. ), Amplitude residual signal ΔAi (m),
A predetermined one of an original phase component θi (m) of a specific discrete frequency, information on the tolerance k, a predicted residual Δθi (m) of the phase component, and the like are supplied. FIG. 2 representatively shows only one signal processing unit that performs signal processing on data of a specific discrete frequency.

The input terminal 20 receives an original amplitude component Ai (m) of a specific discrete frequency and an amplitude residual signal Δ
Ai (m), an original phase component θi (m) of a specific discrete frequency, information on a tolerance k, a predicted residual Δ of the phase component
In a signal processing circuit to which a signal including a predetermined one of θi (m) and the like is supplied, the demultiplexer 21 generates an original amplitude component Ai of a specific discrete frequency (hereinafter, supposed to be fa). (m), the phase component θi (m),
And the amplitude residual signal ΔAi (m) and the phase residual signal Δθi
(m), information on the tolerance k, etc., are separated, and the original amplitude component Ai (m), phase component θi (m), amplitude residual signal ΔAi (m), and phase residual signal Δθi (m ) Are transmitted via a bus 41 to a signal processing circuit for an amplitude component and a requantization unit 4 provided in a processing circuit for a phase component.
The information on the tolerance k is supplied to a latch circuit 54 provided in a phase component processing circuit. That is, of the respective signals separated by the demultiplexer 21, the original amplitude component Ai (m) is transmitted to the requantization unit 43 provided in the amplitude component signal processing circuit via the bus 41. The phase component θi (m) is supplied to the requantization unit 44 provided in the phase component processing circuit via the bus 41, and the amplitude residual signal ΔAi (m) is supplied to the bus 41. Is supplied to the requantization unit 42 provided in the signal processing circuit for the amplitude component, and the phase residual signal Δθi (m) is further provided to the processing circuit for the phase component via the bus 41. The information on the tolerance k is supplied to the latch circuit 54 provided in the phase component processing circuit via the bus 41.

In the decoder shown in FIG. 2, the original amplitude component Ai (m) requantized by the requantizer 43
Is supplied to the data selector 26 provided in the signal processing circuit for the amplitude component, and the amplitude residual signal ΔAi (m) requantized by the requantization unit 42 is used for signal processing of the amplitude component. It is supplied to an adder 50 provided in the circuit. In the adder 50, the amplitude residual signal Δ
Ai (m) is added to the required amplitude component data supplied to the adder 50 via the bus 49, and the output signal Ai (m) = Ai (m-1) + ΔAi (m) is converted to data. This is given to the selector 26. That is, the amplitude residual signal ΔAi supplied from the requantizer 42 to the adder 50.
When (m) is, for example, the amplitude residual signal ΔAi (m) in the m-th frame, the data of the combined amplitude term read from the memory 47 and supplied to the adder 50 via the bus 47 is Since it is the data Ai (m-1) of the combined amplitude term of the (m-1) th frame, the adder 50 sets the data Ai (m-1) of the combined amplitude term of the (m-1) th frame and m
The amplitude residual signal ΔAi (m) in the n-th frame is added, and the adder 50 outputs data Ai (m) of the composite amplitude term of the m-th frame.
4 is supplied to and stored in a memory 47 via a data selector 26 and a bus 49 which perform a switching operation at a predetermined timing according to a switching signal supplied to the polar signal → rectangular coordinate via the data selector 26. The data is supplied to the conversion unit 27.

Also, information on the phase term data θi (m) and the tolerance k of the specific discrete frequency in the m-th frame output from the requantization unit 44 provided in the phase component processing circuit is supplied. Data selector 30
Switches the data θi (m) described above and the data of the phase term output from the adder 29 described later by a switching signal supplied to the input terminal 34 and outputs the data. The data output from the data selector 30 is supplied to and stored in the memory 46 via the bus 48, and the data output from the requantization unit 44 in the data output from the data selector 30 described above. The data θi (m) of the phase term of the specific discrete frequency in the m-th frame, the data of the phase term output from the adder 29 described later, and the like are supplied to the polar coordinate → rectangular coordinate conversion unit 27. An operation of writing data to the memory 46 provided in the phase component processing circuit, an operation of reading data from the memory 46, and a signal processing operation are performed under the control of a control device (not shown). However, in the decoder illustrated in FIG. 2, signal processing in different signal processing modes is performed corresponding to the difference in the signal processing modes of the high efficiency coding applied in the encoder described above with reference to FIG. 1. Done.

The phase term information in the audio signal transmitted from the encoder shown in FIG. 1 to the decoder shown in FIG. 2 according to the audio signal transmission method of the present invention is shown in FIG. Which has been generated by applying the high-efficiency coding in the signal processing mode described above with respect to the encoder described above, that is, as a result of the Fourier transform described with reference to FIG. Phase component ｛θi obtained from the same predetermined number of data for each discrete frequency.
(0), θi (1), θi (2)...}, The first Fourier transform frame 0 selected to transmit the initialization data
Using the phase component θi (0) and the tolerance k due to the average inclination determined by the phase component of each frame transform frame of the Fourier transform frame group, or the phase component of each frame transform frame of the Fourier transform frame group. When the tolerance data k based on the value of the slope determined by the application of the multiplication is transmitted from the encoder side to the decoder side, the control device (not shown) on the decoder side transmits the data from the encoder side. Storing the phase component θi (0) in the first Fourier transform frame m = 0 selected to transmit the incoming initialization data in the memory 46 via the data selector 30 and the bus 48;
Further, the phase component θi (0) at the first Fourier transform frame m = 0 stored in the memory 46 is read and given to the adder 53, and the adder 53 reads the phase component θi (0) at the first Fourier transform frame m = 0. Data obtained by adding the phase component θi (0) and the tolerance k latched by the latch circuit 54 described above is output. The phase component θi (m) of the predetermined Fourier transform frame sequentially output from the adder 53 is supplied to the polar coordinate → rectangular coordinate conversion unit 27 via the adder 29 and the data selector 30, and the bus 48 A control operation for performing signal processing to be stored in the memory 46 via the memory is performed.

As described above, the data Ai (m) of the combined amplitude term of the m-th frame output from the adder 8 and the phase of the specific discrete frequency in the m-th frame output from the adder 29 The term data θi (m) is supplied to the polar coordinate → rectangular coordinate conversion unit 27 via the predetermined data selectors 9 and 30, respectively. And the data Ai (m) of the combined amplitude term of
The real part (Real) amplitude and the imaginary part at the specific discrete frequency fa described above are obtained from the data θi (m) of the specific discrete frequency phase term in the frame.
(Imag) The amplitude is calculated and output, and is stored in the memory 51. The data read from the memory 51 is supplied to the inverse FFT operation unit 31. The inverse FFT operation unit 31 includes all signal processing provided for each of a predetermined number of discrete frequencies in the Fourier transform frame. Since the output data from the circuit is supplied, the original acoustic signal data is restored from the inverse FFT operation unit 31 and stored in the memory 52. The restoration data of the audio signal stored in the memory 52 is subjected to the window function multiplication shown in block 32 and the overlap addition shown in block 33 to obtain the original digital data. The sound signal is sent to the output terminal 35.

Next, information on the phase term in the audio signal transmitted from the encoder shown in FIG. 1 to the decoder shown in FIG. 2 according to the audio signal transmission method of the present invention is shown in FIG. As mentioned above,
Generated by applying the high-efficiency coding in the signal processing mode, that is, for each of the same predetermined number of discrete frequencies determined as a result of the Fourier transform described with reference to FIG. The phase components {θi (0), θi (1),
θi (2)...}, only the phase component in the first Fourier transform frame selected to transmit the initial setting data is initialized for each group of Fourier transform frames each including N Fourier transform frames. While transmitting as data, the least squares method is applied using a tolerance k due to an average inclination determined by a phase component of each frame transformed frame of the Fourier transformed frame group, or a phase component of each frame transformed frame of the Fourier transformed frame group. When the tolerance data k based on the slope value determined by the above is transmitted from the encoder side to the decoder side, the signal processing performed on the decoder side is already performed for the encoder shown in FIG. The detailed description is omitted because it can be easily inferred from the description of the data processing in the case of .

Next, from the encoder shown in FIG. 3 to the decoder shown in FIG. 2, the information of the phase term in the audio signal transmitted according to the audio signal transmission method of the present invention is The encoder shown in FIG. 3 has been generated by applying the high-efficiency encoding in the signal processing mode described above, that is, the same predetermined number of encoders obtained as a result of the Fourier transform. Phase component ｛θi (0), θi (1), θi obtained from data for each discrete frequency
(2) The phase component θ in the first Fourier transform frame 0 selected to transmit the initial setting data within｝
Application of the least squares method using i (0) and a tolerance k due to an average inclination determined by the phase component of each frame transform frame of the Fourier transform frame group, or the phase component of each frame transform frame of the Fourier transform frame group. In the case where the tolerance data k and the residual of the phase component due to the value of the slope determined by the above are transmitted from the encoder side to the decoder side, the control device (not shown) on the decoder side The phase component θi (0) in the first Fourier transform frame m = 0 selected to transmit the initialization data transmitted from the memory 46 via the data selector 30 and the bus 48,
Further, the phase component θi (0) in the first Fourier transform frame m = 0 stored in the memory 46 is read out and given to the adder 53, and the first Fourier transform frame m =
0 and the latch circuit 54 described above.
, And the output data of the adder 53 is supplied to the adder 29. The adder 29 then calculates the predicted residual Δθi (m) of the phase component and the adder 5 described above.
3, and the phase component θi (m) of the predetermined Fourier transform frame is sequentially supplied from the adder 29 to the polar coordinate → rectangular coordinate converter 27 via the data selector 30. , A control operation for performing signal processing to be stored in memory 46 via bus 48 is performed.

As described above, the data Ai (m) of the synthesized amplitude term of the m-th frame output from the adder 8 and the phase of the specific discrete frequency in the m-th frame output from the adder 29 The term data θi (m) is supplied to the polar coordinate → rectangular coordinate conversion unit 27 via the predetermined data selectors 9 and 30, respectively. And the data Ai (m) of the combined amplitude term of
The real part (Real) amplitude and the imaginary part at the specific discrete frequency fa described above are obtained from the data θi (m) of the specific discrete frequency phase term in the frame.
(Imag) The amplitude is calculated and output, and is stored in the memory 51. The data read from the memory 51 is supplied to the inverse FFT operation unit 31. The inverse FFT operation unit 31 includes all signal processing provided for each of a predetermined number of discrete frequencies in the Fourier transform frame. Since the output data from the circuit is supplied, the original acoustic signal data is restored from the inverse FFT operation unit 31 and stored in the memory 52. The restoration data of the audio signal stored in the memory 52 is multiplied by the window function shown in block 32 and the overlap addition shown in block 33 to obtain the original digital data. The signal is restored to an acoustic signal and sent to the output terminal 35.

Next, the information of the phase term in the audio signal transmitted from the encoder shown in FIG. 3 to the decoder shown in FIG. 2 according to the audio signal transmission method of the present invention is shown in FIG. As described above for the third encoder,
That is generated by applying the high-efficiency coding in the signal processing mode, that is, for each of the same predetermined number of discrete frequencies determined as a result of the Fourier transform described with reference to FIG. Phase component ｛θi (0),
Of the θi (1), θi (2)...｝, only the phase component in the first Fourier transform frame selected to transmit the initialization data is converted into a group of Fourier transforms each including N Fourier transform frames. While transmitting as initial setting data for each frame group, using the tolerance k due to the average inclination determined by the phase component of each frame conversion frame of the Fourier transform frame group, or the phase component of each frame conversion frame of the Fourier transform frame group Data k based on the value of the slope determined by applying the least squares method
Is transmitted from the encoder side to the decoder side, the signal processing performed on the decoder side can be easily analogized from the description about the data processing in the case of the encoder described above with reference to FIG. So you can
The detailed description is omitted.

[0074]

As is apparent from the above description, the method of transmitting an acoustic signal according to the present invention is characterized in that each of the Fourier segments sequentially cut out from the acoustic signal so as to have a predetermined fixed time length. Using the same window function to discretely Fourier-transform the signal of the transformed frame, and using the same predetermined number of discrete frequency-based data obtained as a result of the Fourier transform for each Fourier transformed frame described above, Among the phase components for each discrete frequency obtained for each Fourier transform frame, the phase component in the Fourier transform frame to be transmitted first is transmitted as initial setting data, and three or more consecutive Fourier transform frames are individually transmitted. For each of the Fourier transform frame groups, the one end of the Fourier transform frame group is set. The arithmetic mean of the slope values individually determined by the phase components in the Fourier transform frames of the group and the phase components in the other individual Fourier transform frames of the group are transmitted as tolerance data of the respective Fourier transform frame groups. Or Fourier transform discretely using the same window function on the signals of successive Fourier transform frames cut out from the audio signal so as to have a predetermined fixed time length. Using the same predetermined number of discrete frequency data obtained as a result of the Fourier transform, among the phase components for each discrete frequency obtained for each Fourier transform frame described above, the Fourier to be transmitted first The phase component in the transformed frame is transmitted as initial setting data, and three or more continuous Fourier transforms are transmitted. For each of the above-described Fourier transform frame groups, for the sequential Fourier transform frame groups set as frames as individual groups, the least squares method using the phase component in each Fourier transform frame constituting the Fourier transform frame group The slope value determined by the application is transmitted as tolerance data of each Fourier transform frame group, or obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame. Of the phase components
A phase component in a Fourier transform frame to be transmitted first is transmitted as initial setting data, and each of the Fourier transform frames described above is used for a sequential Fourier transform frame group in which three or more consecutive Fourier transform frames are set as individual groups. For each frame group, the arithmetic mean of the slope values individually determined by the phase components in the Fourier transform frame at one end of the Fourier transform frame group and the phase components in other individual Fourier transform frames in the group. The prediction residual of the phase component obtained by performing the prediction on the phase component in each Fourier transform frame belonging to the individual Fourier transform frame group by an arithmetic progression having a value as a tolerance of each Fourier transform frame group, Or the Fourier Transform Among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of the Fourier transform for each system, the phase component in the Fourier transform frame to be transmitted first is transmitted as the initial setting data, and the continuous The phase component in each Fourier transform frame constituting the Fourier transform frame group is determined for each of the Fourier transform frame groups described above with respect to the sequential Fourier transform frame groups in which three or more Fourier transform frames are set as individual groups. The prediction of the phase component in each Fourier transform frame belonging to each individual Fourier transform frame group by an arithmetic progression in which the value of the gradient determined by applying the least square method using the And the prediction residual of the phase component obtained by performing the N (where N is a natural number of 3 or more) Fourier transforms out of phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame The phase component of the first Fourier transform frame in the frame is transmitted as initial setting data, and the phase component of the first Fourier transform frame in the N number of Fourier transform frames and the other N-1 Fourier transform frames are transmitted. The arithmetic average value of the slope values individually determined by the individual phase components is transmitted as tolerance data of each Fourier transform frame group,
N (where N is a natural number of 3 or more) among phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame
The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, and the slope value determined by applying the least squares method using the phase components in the N Fourier transform frames described above. Among the phase components obtained from the same predetermined number of discrete frequencies obtained as a result of the Fourier transform for each successive Fourier transform frame, N (where N is 3 or more) The phase component of the first Fourier transform frame in the (natural number) Fourier transform frames is transmitted as initial setting data, and the phase component of the first Fourier transform frame in the N Fourier transform frames and the other N−1 And the individual phase components of the Fourier transform frames. The arithmetic mean of the values of the above is calculated by an arithmetic progression as the tolerance data of each Fourier transform frame group, and the first Fourier transform frame and the Nth Fourier transform frame in the N number of Fourier transform frames are used. The prediction residual of the phase component obtained by performing the prediction of the phase component in each existing Fourier transform frame,
N (where N is 3) among the phase components obtained from the same predetermined number of discrete frequencies obtained by transmitting the above-mentioned tolerance data and the same predetermined number of discrete frequencies obtained as a result of Fourier transform for each successive Fourier transform frame. The phase component of the first Fourier transform frame in the (natural number) Fourier transform frames is transmitted as initial setting data, and is determined by applying the least squares method using the phase components in the N Fourier transform frames. By the arithmetic progression using the slope value as the tolerance data, the phase component of each Fourier transform frame existing between the first Fourier transform frame and the Nth Fourier transform frame in the N Fourier transform frames described above is obtained. Because the prediction residual of the phase component obtained by performing the prediction and the above-described tolerance data are transmitted. In the method of transmitting an acoustic signal of the present invention, as compared with the transmission method of a conventional acoustic signal in the state of performing the highly compressed data for transmission of audio signals. For example, in an electronic musical instrument, it is not necessary for the original sound to be the same as the original sound in terms of the sense of hearing, and the sound finally output may be of high quality as a musical instrument sound. The discrete Fourier transform is performed discretely using the same window function on the signals of the sequential Fourier transform frames cut out from the acoustic signal so as to have a predetermined time length, and the above-described Fourier transform frames Using the same predetermined number of discrete frequency data obtained as a result of the Fourier transform, the Fourier transform to be transmitted first among the phase components for each discrete frequency obtained for each Fourier transform frame described above. The phase component in the frame is transmitted as initialization data, and the average gradient determined by the phase component of each frame transform frame of the Fourier transform frame group is used. In a state where the data compression ratio is extremely high, only the tolerance data k based on the value of the slope determined by applying the least squares method using the tolerance k or the phase component of each frame transform frame of the Fourier transform frame group is transmitted. Can be implemented. In other words, in the case of electronic musical instruments, it is necessary to record a large amount of acoustic data in the sound source of the musical instrument sound in order to perform various kinds of music, and it is necessary to produce sound without delay immediately after the performance. Therefore, it is required that the data of the instrument sounds be stored in a solid-state memory with a high readout speed.However, in order to record several hundred types of instrument sounds with the required high sound quality, the storage capacity is still limited. Therefore, it is necessary to provide a large number of memory elements having a certain size. Therefore, it is necessary to reduce the recorded sound data to the limit by transmitting the residual data of the phase component as little as possible, that is, by transmitting the quantization scale of the residual data of the phase component as coarsely as possible. However, in such a case, according to the present invention, by applying the phase prediction method based on the average gradient as described above, the variation in the residual of the phase component is suppressed as much as possible, and the data amount is greatly reduced. Even if compression is possible, the effect on the sound when the residuals of the phase components are not sent can be averaged as a whole by reducing the absolute amount of the residuals of the individual phase components. Recording). Also, if the phase prediction is continued with only the phase difference between the sequential Fourier transform frames as a tolerance as already proposed without sending the residual of the phase component, the phase prediction gradually deviates. In the case of natural music, the phase component itself fluctuates,
Also, it is difficult to predict the whole due to the phase difference between two Fourier transform frames due to the presence of noise at the time of data capture and the presence of quantization errors. Considering the realization, practical cost, and the like, the calculation accuracy at the time of calculating the initial tolerance and the quantization scaling accuracy at the time of transmission cannot be made very high. If the phase prediction shifts, a noise-like sound may be generated, or the frequency of the window may be heard audibly, and the timbre changes extremely when the next original frame is reproduced. Therefore, the position of the original frame can be clearly recognized by hearing and becomes far from practical, but when applying the acoustic signal transmission method of the present invention to a sound source, it is necessary to send the residual of the phase component without sending it. When the data is significantly compressed, the above-mentioned problems can be eliminated. As described above, in the acoustic signal transmission method of the present invention, discrete Fourier transform frames are sequentially extracted from an acoustic signal so as to have a predetermined time length using the same window function. Each discrete frequency obtained for each of the above-mentioned Fourier transform frames by using the same predetermined number of data for each of the discrete frequencies obtained as a result of the Fourier transform for each of the above-mentioned Fourier transform frames. Among the phase components, the phase component in the Fourier transform frame to be transmitted first is transmitted as the initial setting data, and the tolerance k due to the average inclination determined by the phase component of each frame transform frame in the Fourier transform frame group is k, Or the slope determined by applying the least squares method using the phase component of each frame transform frame of the Fourier transform frame group In addition to transmitting the tolerance data k by value, the residual of the phase signal is also transmitted, or the phase component in the Fourier transform frame to be transmitted first for each Fourier transform frame group is transmitted as initial setting data, Transmission (recording) of sound signals with good sound quality in a state where a large amount of data is compressed compared to the sound signal transmission method according to the conventional method
Can be easily performed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of an encoder used in a method for transmitting an acoustic signal according to the present invention.

FIG. 2 is a block diagram illustrating a configuration example of a decoder used in the audio signal transmission method according to the present invention.

FIG. 3 is a block diagram illustrating a configuration example of an encoder used in the sound signal transmission method of the present invention.

FIG. 4 is a diagram used for explaining a sound signal transmission method of the present invention.

FIG. 5 is a diagram used for describing an acoustic signal transmission method of the present invention.

FIG. 6 is a diagram used for explaining a sound signal transmission method according to the present invention.

FIG. 7 is a diagram used for explaining a sound signal transmission method according to the present invention.

FIG. 8 is a diagram used for explaining the sound signal transmission method of the present invention.

FIG. 9 is a block diagram illustrating a configuration example of an encoder used in a previously proposed sound signal transmission method.

FIG. 10 is a block diagram illustrating a configuration example of a decoder used in a proposed transmission method of an audio signal.

FIG. 11 is a diagram used to explain a previously proposed method of transmitting an acoustic signal.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Input terminal of digital audio signal targeted for recording and transmission, 2 ... Overlap section, 3,32 ... Window function processing section, 4 ... Fast Fourier transform operation (FFT operation) section, 6 ...
Cartesian coordinate → polar coordinate converter, 7, 10, 12, 24, 28
... Latch circuit, 8, 13, 14 ... Subtractor, 9, 16, 2
6, 30 ... data selector, 11 ... amplifier with gain of 2,
17: quantization scaling unit, 18: multiplexer,
20 ... input terminals of a decoder provided on the receiving side, 21 ...
Demultiplexer, 22 requantizer, 25, 29 adder, 27 polar → rectangular coordinate converter, 31 inverse FFT
Calculation unit, 36, 37 memory, 38 tolerance change point determination unit, 39 tolerance calculation unit, 40 adder,

(Equation 4)

(Equation 5)

(Equation 10)

[Equation 11]

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-1-221800 (JP, A) JP-A-3-144700 (JP, A) JP-A-5-216482 (JP, A) (58) Field (Int.Cl. ⁶ , DB name) G10L 3/00-9/18 H03M 7/30 H04B 14/00 JICST file (JOIS)

Claims (8)

(57) [Claims] 1. A discrete Fourier transform is applied to signals of successive Fourier transform frames cut out from an audio signal so as to have a predetermined constant time length using the same window function, and the above-mentioned Fourier transform is performed. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a transmission method of an acoustic signal transmitted after high-efficiency encoding, among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame, first, The phase component in the Fourier transform frame to be transmitted is transmitted as initial setting data, and three or more consecutive Fourier transform frames are grouped into individual groups. For the set of sequential Fourier transform frames set as follows, for each Fourier transform frame group,
The arithmetic mean of the slope values individually determined by the phase component in the Fourier transform frame at one end of the group of Fourier transform frames and the phase component in the other individual Fourier transform frames in the group is calculated as A method of transmitting an acoustic signal which is transmitted as tolerance data of a Fourier transform frame group. 2. Fourier transform discretely using the same window function on signals of successive Fourier transform frames cut out from an audio signal so as to have a predetermined fixed time length, and each of the Fourier transforms described above is performed. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a transmission method of an acoustic signal transmitted after high-efficiency encoding, among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame, first, The phase component of the Fourier transform frame to be transmitted is transmitted as initial setting data, and three or more consecutive Fourier transform frames are divided into individual groups. For the set of sequential Fourier transform frames set as follows, for each Fourier transform frame group,
An acoustic signal that transmits the value of the gradient determined by applying the least squares method using the phase component in each Fourier transform frame constituting the Fourier transform frame group as tolerance data of each Fourier transform frame group Transmission method. 3. A discrete Fourier transform is performed discretely using the same window function on signals of respective sequential Fourier transform frames cut out from an audio signal so as to have a predetermined fixed time length. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a transmission method of an acoustic signal transmitted after high-efficiency encoding, among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame, first, The phase component in the Fourier transform frame to be transmitted is transmitted as initial setting data, and three or more consecutive Fourier transform frames are grouped into individual groups. For the set of sequential Fourier transform frames set as follows, for each Fourier transform frame group,
The arithmetic mean of the slope values individually determined by the phase component in the Fourier transform frame at one end of the group of Fourier transform frames and the phase component in the other individual Fourier transform frames in the group is calculated as By the arithmetic progression as the tolerance of the Fourier transform frame group, the prediction residual of the phase component obtained by performing prediction on the phase component in each Fourier transform frame belonging to each Fourier transform frame group, and the tolerance data described above. A method of transmitting an acoustic signal to be transmitted. 4. A discrete Fourier transform is applied to signals of successive Fourier transform frames cut out from an audio signal so as to have a predetermined fixed time length by using the same window function, and the above-mentioned Fourier transform is performed. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a transmission method of an acoustic signal transmitted after high-efficiency encoding, among the phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame, first, The phase component in the Fourier transform frame to be transmitted is transmitted as initial setting data, and three or more consecutive Fourier transform frames are grouped into individual groups. For the set of sequential Fourier transform frames set as follows, for each Fourier transform frame group,
Each of the Fourier transform frames constituting the Fourier transform frame group is calculated by using a phase component in each of the Fourier transform frames to determine a slope value determined by applying the least squares method. A sound signal transmission method for transmitting a prediction residual of a phase component obtained by performing prediction on a phase component in each Fourier transform frame belonging to a Fourier transform frame group and the above-described tolerance data. 5. A Fourier transform is performed discretely using the same window function on signals of successive Fourier transform frames cut out from an audio signal so as to have a predetermined fixed time length. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a method of transmitting an acoustic signal transmitted after highly efficient encoding, N (N) of phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame. (Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, Also, the arithmetic mean value of the slope values individually determined by the phase components of the first Fourier transform frame in the N Fourier transform frames and the individual phase components of the other N-1 Fourier transform frames, respectively. Is transmitted as tolerance data. 6. A Fourier transform is performed discretely using the same window function on a signal of each successive Fourier transform frame cut out from an audio signal so as to have a predetermined fixed time length, and each Fourier transform is performed. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a method of transmitting an acoustic signal transmitted after highly efficient encoding, N (N) of phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame. (Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, Also, a sound signal transmission method in which a value of a gradient determined by applying the least squares method using the phase components in the N Fourier transform frames is transmitted as tolerance data. 7. A Fourier transform is performed discretely using the same window function on signals of successive Fourier transform frames cut out from an audio signal so as to have a predetermined fixed time length, and the Fourier transform is performed. Using the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each transform frame, the amplitude component and phase component for each discrete frequency obtained for each Fourier transform frame, In the method of transmitting an acoustic signal transmitted after highly efficient encoding, N (N) of phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame (Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, Also, the arithmetic mean value of the slope values individually determined by the phase components of the first Fourier transform frame in the N Fourier transform frames and the individual phase components of the other N-1 Fourier transform frames, respectively. Are represented by an arithmetic progression as the tolerance data of the respective Fourier transform frame groups, each Fourier transform frame existing between the first Fourier transform frame and the Nth Fourier transform frame in the N Fourier transform frames described above. A method for transmitting an acoustic signal, comprising transmitting a prediction residual of a phase component obtained by performing a prediction of a phase component and the above-mentioned tolerance data. 8. A Fourier transform is performed discretely using the same window function on the signals of each successive Fourier transform frame cut out from the audio signal so as to have a predetermined fixed time length, and the Fourier transform is performed. Using the same predetermined number of data for each discrete frequency obtained as a result of Fourier transform for each transform frame, the amplitude component and the phase component for each discrete frequency obtained for each Fourier transform frame described above, In a method of transmitting an acoustic signal transmitted after highly efficient encoding, N (N) of phase components obtained from the same predetermined number of discrete frequency data obtained as a result of Fourier transform for each successive Fourier transform frame. (Where N is a natural number of 3 or more) The phase component of the first Fourier transform frame in the Fourier transform frames is transmitted as initial setting data, Further, the first Fourier transform in the N Fourier transform frames is performed by an arithmetic progression in which the slope value determined by applying the least squares method using the phase components in the N Fourier transform frames as the tolerance data. A prediction residual of a phase component obtained by performing prediction of a phase component in each Fourier transform frame existing between the transform frame and the Nth Fourier transform frame;
An acoustic signal transmission method for transmitting the above-mentioned tolerance data.