JP3338573B2 - Sub-band division operation circuit - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sub-band division operation circuit.

[0002]

2. Description of the Related Art In recent years, compression coding techniques for image signals and audio signals have been remarkably developed. One of the compression encoding techniques is subband encoding using subband division. This is a method in which an image signal or an audio signal is frequency-analyzed and divided into a plurality of band signals having different frequencies, and encoding is performed for each band signal by changing the number of bits and the like according to the characteristics of the signal and the like. .

[0003] As an application example of sub-band coding of an audio signal, MPEG (Moving Picture Image Coding Experts) has been proposed.
Group) have been put to practical use. this is,
By applying human auditory characteristics calculated from an input audio signal to subband encoding, the original audio signal can be compressed and encoded while maintaining its quality.

FIG. 10 shows an example of the configuration of a conventional sub-band division arithmetic circuit for performing sub-band division on a digital audio signal. In FIG. 10, the original data X _{i ′} of the audio signal indicates each sample value extracted at each time from the temporally continuous audio signal.
For example, if the number of samples to be taken is 512,
The subscript i 'takes a value from 0 to 511.

The original data X _{i ′} is subjected to predetermined preprocessing by the windowing operation unit 1 and the folding operation unit 51 before the subband division operation is performed by the subband division operation unit 52. That is, the windowing operation unit 1 performs a windowing operation on the original data Xi _' based on the following (Equation 1), and obtains an audio signal _yi' for each time.

[0006]

(Equation 1)

[0007] As shown in (Equation 1), windowing operation is carried out by multiplying the _'window function C _i relative _to' the original data X _i to be inputted, when the window function C _{i For '} , those previously stored in the window function ROM 5 are used.

The windowing operation is performed in such a manner that the audio signals extracted at each time are not always continuous at the end point of each section, so that the end point of the section is set to a predetermined level. This is to reduce the occurrence of noise in the analysis result by converging and suppressing the discontinuity points.

Next, the audio signal y _{i '} for each time obtained as described above is input to the folding operation unit 51, and the folding operation is performed based on the following (Equation 2). Note that i and k in (Equation 2) are 0 to 6 respectively.
3. Takes values from 0 to 7. Thus, by folding operations per cycle 64 is executed, the pre-processing operation signal Y _i is obtained.

[0010]

(Equation 2)

The pre-processing operation signal Y _i obtained by the folding operation unit 51 is then converted into a sub-band division operation unit 52
And subjected to a sub-band division operation. The sub-band division calculator 52 is specifically configured as shown in FIG. Here, a configuration in the case of dividing into 32 subbands is shown as an example.

As apparent from FIG. 11, the sub-band division operation is performed based on the following (Equation 3). In addition,
I in (Equation 3) is a parameter in the time axis direction, and takes a value of 0 to 63 as described above. In addition, j is a parameter indicating a subband number, and takes a value of 0 to 31.

[0013]

(Equation 3)

As described above, the sub-band division calculator 52 performs the calculation based on (Equation 3) to obtain the sub-band information _Sj divided into 32 frequency bands. The value of cos ((2j + 1) (i-16) π / 64) is
What is stored in advance in the cosine function ROM 53 is used. Note that the value of 64 in this equation is M
It is unique to the PEG standard, and other sub-band coding schemes use unique values.

[0015]

As described above, in the conventional subband division operation circuit to which, for example, MPEG is applied, the windowing operation unit 1 performs a multiplication process based on (Equation 1), and the folding operation unit 51 Performs the addition processing based on (Equation 2), and performs the multiplication and addition processing based on (Equation 3) by the sub-band division calculator 52, thereby performing the sub-band division processing. Table 1 summarizes the number of times these multiplications and additions are performed.

[0016]

[Table 1]

As described above, in the conventional sub-band division operation circuit, many multiplications and additions have to be performed. For this reason, there is a problem that a calculation load for performing the sub-band division increases. In particular, since the multiplication process has a higher calculation load than the addition process, it is not preferable that the number of times of multiplication is large in performing the calculation. Further, since the number of times of performing the multiplication and the addition is large as described above, there is a problem that the time required for the processing increases accordingly.

The present invention has been made to solve such a problem, and it is an object of the present invention to reduce a calculation load in subband division processing.

[0019]

A sub-band division operation circuit according to the present invention is a sub-band division operation circuit for analyzing the frequency of an input signal and dividing the signal into a plurality of sub-bands having different frequencies. Windowing operation means for performing a windowing operation on the signal of the analysis source, folding operation means for performing a folding operation on the signal output from the windowing operation means, and performing butterfly operation on the signal output from the folding operation means The multi-stage connected butterfly operation means for obtaining the sub-band information, and the phase of the sub-band information output from the butterfly operation means,
A phase correction calculation means is provided for correcting so as to compensate for a deviation from the phase of the sub-band information obtained when normal frequency analysis is performed without applying the butterfly calculation.

Further, another sub-band division operation circuit of the present invention is a sub-band division operation circuit for frequency-analyzing an input signal and dividing the signal into a plurality of sub-bands having different frequencies. A windowing operation means for performing a windowing operation on the signal, a folding operation means for performing a folding operation on the signal output from the windowing operation means, and a butterfly operation on the signal output from the folding operation means. Multi-stage connected butterfly operation means for obtaining sub-band information, wherein the folding operation means corresponds to a phase shift of the sub-band information obtained when normal frequency analysis is performed without applying the butterfly operation. It is characterized by including an offset amount correcting means for correcting the offset amount.

[0021]

Another feature of the present invention is that, in the bit-reversed sequence of the result obtained after the butterfly operation, an operation for controlling a part of the butterfly operation corresponding to a regular part to be omitted. A control means is provided.

According to another feature of the present invention, a judgment means for judging whether or not an input audio signal of a frequency analysis source is a signal having only a real number part, and a real number part for the audio signal by the judgment means. And an arithmetic control means for controlling so as to omit the butterfly operation performed based on the imaginary part of the audio signal when it is determined that only the audio signal is included.

Another feature of the present invention is that a designation means for designating either the real part or the imaginary part of the subband information obtained after the butterfly operation, and a real part or an imaginary number by the designation means. When any one of the sections is designated, a computation control means for controlling so as to omit the butterfly computation for obtaining the other information not designated is further provided.

Another feature of the present invention is that when a part of the butterfly operation is omitted and the regularity of the bit reverse order obtained after the butterfly operation deteriorates, all the butterfly operations are executed. It is assumed that.

[0026]

Since the present invention comprises the above technical means, it is possible to perform subband division effectively utilizing the similarity between the basic expression of the conventional subband division operation and the basic expression of FFT (fast Fourier transform). In addition, the calculation load at the time of subband division becomes smaller than before.

According to another feature of the present invention, since the offset amount correcting means is provided, the phase shift is corrected in advance before performing the butterfly operation, and the butterfly operation is performed on the corrected signal. As a result, it is not necessary to perform the phase correction calculation after the butterfly calculation.

Further, according to another feature of the present invention, since the arithmetic control means is provided, it is not necessary to perform an unnecessary operation which is performed when the butterfly operation is simply applied. Thereby, for example, to see the regularity of the arrangement of the bit reverse order, to determine whether the signal of the analysis source includes only the real part, or to obtain one of the real part or the imaginary part of the subband information obtained as a solution By observing whether or not is specified, it is possible to omit a part of the butterfly operation corresponding to the regular part, omit the butterfly operation performed based on the imaginary part of the input signal, It is possible to omit the butterfly operation corresponding to the unspecified one of the real part and the imaginary part of the information.

According to another feature of the present invention, all butterfly operations are performed without omitting the butterfly operation for the portion that deteriorates the regularity of the bit-reversed array. Can be prevented from becoming inconvenient.

[0030]

DESCRIPTION OF THE PREFERRED EMBODIMENTS As described in the description of the prior art, in subband coding, video and audio signals are frequency-analyzed and divided into a plurality of subbands having different frequencies, and the characteristic of the band signal is assigned to each subband. And encoding by changing the assignment of the number of bits and the like according to the degree of importance.

One technique for analyzing what frequency components are contained in a video or audio signal is FFT (Fast Fourier Transform). This FFT is
The amount of DFT (Discrete Fourier Transform) operation can be reduced by repeatedly performing a basic operation called a butterfly operation.

The sub-band division operation circuit of the present embodiment can reduce the load of the operation performed in the sub-band division processing by applying the butterfly operation of the FFT to the sub-band division processing. It is.

[0033] That is, the general formula of the FFT is shown by the following equation (4), W _i in the equation (4) is expressed by (Equation 5). As is clear from comparison of (Equation 4) and (Equation 5) with the above (Equation 3), the cosine function shown in (Equation 5) is a modification of the cosine function in (Equation 3). It can be thought that it did. Therefore, in the present embodiment, attention is paid to this point, and the sub-band division can be performed by applying the butterfly operation of the FFT.

[0034]

(Equation 4)

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram illustrating a functional configuration of the subband division operation circuit according to the first embodiment. The sub-band division operation circuit shown in FIG. 1 is configured by applying the compression coding technology based on the layers I and II of the MPEG audio standard, and the same block as the block shown in FIG. Is attached.

In FIG. 1, the original data X _{i ′} (i ′ = 0 to 511) of the audio signal extracted at each time is input to the windowing operation unit 1, where By performing the windowing operation on the basis of this, the audio signal y _{i ′ at} each time is obtained. In this windowing operation, the window function C _{i ′} stored in the window function ROM 5 in advance is used as described in the conventional example. By this windowing operation, the end points of the audio signal in each section extracted at each time are converged to a predetermined level, and discontinuous points of the audio signal are suppressed.

Next, the audio signal y _{i ′ at} each time obtained by the windowing operation as described above is input to the folding operation unit 2 and subjected to the folding operation based on the following (Equation 6). You. Note that i and k in (Equation 6) take values of 0 to 127 and 0 to 3, respectively. Thus, period 1
By performing the folding operation for each 28, a pre-processing operation signal Y _i is obtained.

[0038]

(Equation 5)

The pre-processing operation signal Y _i obtained by the folding operation unit 2 is next input to a butterfly operation unit 3 and subjected to butterfly operation, which is a basic operation of FFT. Generally, when performing a Fourier transform with a period N, it is necessary to perform (log ₂ N) stages of radix-2 butterfly operations as shown in FIG. That is, when the period is 128, seven stages of butterfly operations need to be performed. In the present embodiment, the radix-4 butterfly operation is performed in the second and third stages, as shown in FIG. , A total of five stages of butterfly operation are performed. Therefore, the first shown in FIG.
In the butterfly operation unit 31 to the fifth butterfly operation unit 35, the operations shown in (Expression 7) to (Expression 11) are performed, respectively.

[0040]

(Equation 6)

[0041]

(Equation 7)

[0042]

(Equation 8)

[0043]

(Equation 9)

[0044]

(Equation 10)

Here, the variable names in the polynomials shown in (Equation 7) to (Equation 11) will be described. The character "b" at the beginning of the variable name indicates that the variable is data after the completion of the butterfly operation in each stage. The first subscript indicates whether it is data of a real part or data of an imaginary part.
“I” indicates an imaginary part. The second suffix indicates what stage of the data the butterfly operation has been performed, and numbers 1 to 5 are used. The third subscript indicates the sequence number, i = 0
Any number from to 127 is used.

As described above, the first butterfly operation units 31 to 31
(Equation 7) to (Equation 11) by the fifth butterfly operation unit 35
Is performed, data b _R5i and data b _I5i are obtained as a result. This data b _I5i is shown in FIG.
, _Which substantially corresponds to the sub-band information S _j obtained by the conventional sub-band division calculating unit 52, as shown in (Equation 3) to (Equation 5), i and ( i-16)
There is a phase shift by an amount corresponding to. Therefore, the phase shift is corrected by performing a rotation calculation as shown in the following (Equation 12) in the phase correction unit 4.

[0047]

[Equation 11]

When the butterfly operation unit 3 executes various rotation operations in the above (Equation 7) to (Equation 11), the sine / cosine function stored in advance in the first sine / cosine function ROM 6 is used. Is used. Also, when the phase correction unit 4 executes the rotation calculation of the above (Equation 12), the second sine / cosine function R
A sine / cosine function value stored in the OM 7 in advance is used.

As described above, by performing the frequency analysis operation (subband division operation) by applying the FFT,
Data b ' _5i corresponding to the sub-band information S _j can be obtained. However, the data b ' _5i needs to be rearranged because it is given as an array in the bit reverse order.
Table 2 shows the rearranged data b ′ _5i and subband information S _j
The following shows the correspondence.

[0050]

[Table 2]

As described above, in the sub-band division operation circuit according to the first embodiment, the windowing operation unit 1 performs the multiplication processing based on (Equation 1), and the folding operation unit 2
Performs the addition process based on (Equation 6). Further, the butterfly operation unit 3 performs multiplication and addition / subtraction processing based on (Equation 7) to (Equation 11), and the phase correction unit 4 (Equation 1
The sub-band division of the audio signal is performed by performing the multiplication and addition processing based on 2). Table 3 below
Summarizes the number of times these multiplication and addition / subtraction are performed.

[0052]

[Table 3]

As can be seen by comparing Table 3 with Table 1 described above, the number of times of multiplication can be reduced according to the subband division operation circuit of this embodiment as compared with the conventional subband division operation circuit. . Although the number of additions / subtractions is increased compared to the conventional case, the computational load is much larger in the multiplication process than in the addition / subtraction process. Therefore, reducing the number of multiplications reduces the overall computational load compared to the conventional case. be able to.

Next, a second embodiment of the present invention will be described. The sub-band division operation circuit according to the second embodiment includes:
With the configuration shown in FIG. 4, the calculation for the phase correction performed in the first embodiment can be omitted.

That is, in the above-described first embodiment, the phase shift is corrected after performing the butterfly operation. On the other hand, in the second embodiment, before the butterfly operation is performed, Offset amount corresponding to (i-16) in (Equation 3) (offset amount corresponding to phase shift)
By making corrections in advance,
It is not necessary to perform the phase correction after the butterfly operation.

In FIG. 4, the original data X _{i ′} (i ′ = 0 to 511) of the audio signal extracted at each time is subjected to a windowing operation based on (Equation 1) by the windowing operation unit 1. You. The resulting audio signal y _{i ′ at} each time is
The data is input to the addition unit 9 in the folding operation unit 8 and is subjected to predetermined addition processing. The result is stored in the memory 10.

The addition process and the storage of the addition result in the memory 10 are controlled by the control unit 11.
That is, the control unit 11 controls the pre-processing operation signal Y _i calculated for the i = 0 to 127 by the addition unit 9 so as to be stored in the memory 10 with a shift by an offset amount corresponding to (i−16). Is controlled by

In this case, the corrected signal Y _i is Y ₀ to Y
_The first i
= 0 to 111, the operation shown in (Equation 13-1) is performed, and a signal obtained by this operation is a pre-processing operation signal Y ₁₆
It is stored in the memory 10 as to Y _127. Next, i = 1
The calculation shown in (Equation 13-2) is performed for 12 to 127, and the signal obtained by this is a pre-processing calculation signal Y _0.
ＹY ₁₅ are stored in the memory 10.

By correcting such an offset amount, the phase of the pre-processed operation signal Y _i , which is the basis of the frequency analysis by the butterfly operation, is previously shifted on the time axis by an amount corresponding to (i-16). The same effect as that described above can be obtained.

[0060]

(Equation 12)

The pre-processing operation signal Y _i obtained by the folding operation unit 8 is input to the butterfly operation unit 3 and the butterfly operation is performed based on the above (Equation 7) to (Equation 11). By _performing this butterfly operation, data b _R5i and data b _I5i are obtained. As described above, in the present embodiment, the offset amount corresponding to the phase shift of (i-16) is corrected in advance. So
The phase of the data b _I5i after the frequency analysis by the butterfly operation has already been corrected. Therefore, the data b _I5i directly corresponds to the sub-band information S _j .

Therefore, according to the second embodiment, the first
In this embodiment, 64 multiplication processes and 32 addition processes (see Table 3) for phase correction performed in the embodiment can be omitted, and the calculation load as a whole can be further reduced.

Next, a third embodiment of the present invention will be described. The sub-band division operation circuit according to the third embodiment includes:
The number of times of multiplication and addition / subtraction operations can be further reduced by deleting the operation of a part having regularity from the values given as the bit reverse order after the butterfly operation.

The sub-band division operation circuit of this embodiment is configured as shown in FIG. That is, in this embodiment,
In addition to the configuration of the sub-band division operation circuit according to the second embodiment, the operation control unit 12 and the table memory 1
3 is newly provided. It should be noted that the operation control unit 12
It is constituted by a microcomputer system comprising U, RAM and the like.

As described above, when the frequency analysis operation (sub-band division operation) is performed by applying the FFT, the result obtained after this operation is given as a bit-reverse array as shown in FIG. As is clear from FIG. 6, the least significant bits in the first half of the arrangement in the bit reverse order are all "0", and the least significant bits in the second half are all "1". That is, FIG.
The upper half of the dashed line shown therein has a regularity that the least significant bits are all "0" and the least significant bits of the lower half are all "1".

Further, when the audio signal encoding technique by MPEG is applied, the value range of j is (2) in (Equation 3).
j + 1) and j = 0 to 31. Therefore, only the odd number of 0 to 63 is required as a result of the calculation by the FFT. Therefore, due to the above-described regularity, the calculation for the upper half of the dashed line shown in FIG. 2 can be omitted.

As a method for finding the regularity as described above, for example, there is the following method. That is, the arithmetic control unit 12 tabulates the bit-reversed array in the table memory 13 based on the arithmetic expression of the frequency analysis and the value range of j. Then, a regularity is found by detecting a portion where the bit does not change based on the table information. The operation control unit 12 controls the folding operation unit 8 and the butterfly operation unit 3 so as to omit a part of the operation corresponding to the detected regular part.

As a result, the folding operation unit 8 performs the operation shown in (Equation 14), and the butterfly operation unit 3
In the first butterfly computation unit 31 of the above, computation as shown in (Equation 15) is performed. When the lower half of the one-dot chain line shown in FIG. 2 is operated, the first butterfly operation unit 31 originally performs 64 sets of subtraction and rotation operations, but here, the subtraction is performed by the folding operation unit 8. It has been made like that. That is, the folding operation unit 8 performs subtraction as shown in (Equation 14), and the first butterfly operation unit 31 performs only rotation operation as shown in (Equation 15).

[0069]

(Equation 13)

In the second butterfly operation unit 32, the operation of the above (Equation 8) is performed only for i = 0 to 15, and in the third butterfly operation unit 33, the above (Equation 9)
Is i = 0-3, 16-19, 32-35, 48-
Only performed for 51. In the fourth butterfly computation unit 34, the computation of (Equation 10) described above is performed with i = 0 to 63.
Is performed only for a multiple of 4, and the fifth butterfly calculation unit 35 performs the calculation of (Equation 11) only for even numbers of i = 0 to 63.

Table 4 below summarizes the number of times of multiplication and addition / subtraction performed by the sub-band division operation circuit according to the third embodiment described above. As is apparent from Table 4, according to the sub-band division operation circuit of this embodiment, the number of times of multiplication and addition / subtraction can be greatly reduced, and the operation load can be further reduced.

[0072]

[Table 4]

As described above, when the sub-band division operation circuit of this embodiment is applied to MPEG,
Since only the odd number of 0 to 63 is required as the result of the operation by T, other operations can be omitted.
That is, in the example shown in FIG. 6, the operation corresponding to the portion where the value of the most significant bit, that is, the value determined by the operation by the fifth butterfly operation unit 35 is “1” can be omitted. Therefore, in the fifth butterfly computation unit 35, the subtraction and the rotation computation become unnecessary, and eventually (Equation 1)
The calculation as shown in 6) is performed.

[0074]

[Equation 14]

As described above, when the sub-band division operation circuit of this embodiment is applied to MPEG, not only the regularity of the least significant bit of the result obtained as a reverse-ordered array after the frequency analysis operation, but also Some of the operations can be omitted depending on the regularity of the upper bits. Table 5 shows the number of operations in each block in this case.

[0076]

[Table 5]

The correspondence between the data b _I5i and the sub-band information S _j obtained by _performing the above-described operations by the folding operation unit 8 and the butterfly operation unit 3 based on the control by the operation control unit 12. Table showing
It is shown in Table 6.

[0078]

[Table 6]

Next, a fourth embodiment of the present invention will be described. As shown in FIG. 7, in the sub-band division operation circuit according to the fourth embodiment, the determination unit 14 that determines whether or not the original data X _{i ′} of the audio signal includes only the real part.
And a designating unit 15 for designating one of a real part and an imaginary part of the subband information obtained by the butterfly operation, which is required as a solution.

When the original data X _{i ′} consists of only a real part, the first butterfly operation unit 3 of the butterfly operation unit 3
Preconditioner signal Y _i which is input to the 1 is also a signal of the real part only. Therefore, in the first butterfly operation unit 31,
The operation shown in the following (Equation 17) may be performed instead of the above (Equation 15).

[0081]

(Equation 15)

Therefore, in the present embodiment, the determination unit 14 shown in FIG. 7 determines whether or not the original data X _{i ′} is only a real part, and supplies the determination result to the arithmetic control unit 12.
When the determination unit 14 determines that the original data X _{i ′} is only a real part, the arithmetic control unit 12 controls the first butterfly operation unit 31 in the butterfly operation unit 3 to calculate the above equation (17). Such calculations are performed. Table 7 summarizes the number of operations in each block in this case.

[0083]

[Table 7]

Further, when the subband information to be obtained as a final solution requires only one of the real part and the imaginary part, the fourth butterfly operation unit 34 and the fifth butterfly operation unit 35 obtain the necessary solution. It is only necessary to perform the calculation for

Therefore, for example, when only the solution of the real part is necessary, that is, when the real part is designated by the designation unit 15, the control of the arithmetic control unit 12
The fourth butterfly computation unit 34 performs the computation shown in (Equation 18), and the fifth butterfly computation unit 35 performs the computation shown in (Equation 19).

[0086]

(Equation 16)

That is, the operation for the imaginary part is omitted. Table 8 summarizes the number of operations in each block in this case.

[0088]

[Table 8]

As described above, according to the fourth embodiment, when the original data X _{i ′} of the audio signal includes only the real part, or when the data required as the solution is only one of the real part or the imaginary part, In some cases, unnecessary calculations in each case can also be reduced. Therefore, the first
-The number of operations of multiplication and addition / subtraction can be further reduced as compared with the third embodiment.

According to the fourth embodiment, although the number of operations that can be reduced based on the regularity of the result obtained after the frequency analysis operation is almost reduced, the second butterfly operation unit 32 and the second In the case where the rotation angle is 0, π / 2 in the rotation calculations in (Equation 8) and (Equation 9), the three-butterfly calculation unit 33 can further omit the rotation calculation. The number of operations in each block in this case is as shown in Table 9.

[0091]

[Table 9]

However, since the above rotation operation does not require so many multiplications and additions / subtractions, the effect is not so great even if it is reduced. On the contrary, it is preferable not to reduce the rotation calculation because the regularity of the whole is deteriorated.

Although the first to fourth embodiments have been described above, the following Table 10 summarizes the number of times of multiplication and addition / subtraction in each block in each embodiment. I will describe it.

[0094]

[Table 10]

In the above description, the case where MPEG is used as the frequency analysis method has been described, but the present invention is not limited to this. For example, the present invention can be applied to a case where a signal filtering process is performed.

In general, when only a signal in a certain frequency band of an input signal is useful, a filtering process is performed to extract only the useful signal. The following filters are used for the filtering process. That is, there are an LPF (low-pass filter) that extracts only a signal in a low-frequency region, an HPF (high-pass filter) that extracts only a signal in a high-frequency region, and a BPF (band-pass filter) that extracts only a signal in a frequency region having a certain characteristic. . In addition, the filtering processing by the digital method includes FIR (finite impulse response).
e) Filters and IIR (infinite impulse response) filters are widely used, but filtering can also be performed by a method as shown in FIG.

In FIG. 8, first, a frequency analysis is performed by the frequency analyzer 21 on an input analysis source signal. next,
The gain multiplier 22 multiplies the signal of the analysis result output from the frequency analyzer 21 by a gain specific to the filter. Then, the signal combiner 23 combines the signals output from the gain multiplier 22 to perform a filtering process.

For example, when filtering is performed by the BPF, the frequency component in the region where the gain value is small is multiplied by “0” by the gain multiplier 22 in FIG. Therefore, when the BPF is used, only the frequency components in a specific region determined by the filter characteristics need to be obtained. So B
If the technical means of the present invention is applied to the frequency analyzer 21 to which the PF is applied, it is possible to reduce operations other than operations for obtaining useful frequency components.

For example, the sample length for frequency analysis is set to N
= 512, and 64 to 1 in the sub-band after frequency analysis.
Consider a case where only signals in the 27 region are used as useful signals. In this case, if the original signal is sampled at 44.1 KHz, 5.5125 KHz
Only the signal in the frequency band of ~ 11.025 KHz needs to be analyzed.

FIG. 9 shows the arrangement of the binary numbers 64 to 127 in the bit reverse order. As is clear from FIG. 9, the lower three bits of the arrangement in the bit reverse order are all "100", and there is a regularity that this portion does not change. Therefore, by detecting this regularity, it is possible to reduce unnecessary butterfly computations outside the useful frequency domain.

[0101]

As described above, according to the present invention, the signal sub-band division is performed by applying the fast Fourier transform butterfly operation which requires a relatively small amount of computation. The number of multiplication operations with a large load can be reduced as compared with the conventional case. Therefore, the calculation load can be reduced, and the calculation processing speed as a whole can be improved.

By providing the phase correction operation means, it is possible to correct the phase shift of the sub-band information output from the butterfly operation means. Alternatively, before performing the butterfly operation, by correcting the offset amount corresponding to the phase shift of the subband information obtained when performing a normal frequency analysis without applying the butterfly operation, the phase shift can be corrected in advance. The butterfly operation can be performed on the corrected signal, and the operation of the phase correction after the butterfly operation can be omitted. Therefore, the number of operations for performing the subband division is further reduced, and the operation load is reduced. Therefore, the calculation processing speed can be further improved.

According to another feature of the present invention, since the arithmetic control means for controlling so as to omit a part of the butterfly operation is provided, the unnecessary operation which is performed when the butterfly operation is simply applied is provided. Is not performed, and only the butterfly operation for obtaining the originally required subband information need be performed, and the number of multiplication and addition / subtraction operations can be further reduced, thereby reducing the operation load.

Further, according to another feature of the present invention, all butterfly operations are performed on a portion that degrades the regularity of the bit-reversed array. Can be prevented, and the calculation load can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a sub-band division operation circuit according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram for explaining processing content of a butterfly operation.

FIG. 3 is a block diagram illustrating a configuration of a butterfly operation unit.

FIG. 4 is a block diagram illustrating a configuration of a sub-band division operation circuit according to a second embodiment of the present invention.

FIG. 5 is a block diagram illustrating a configuration of a sub-band division operation circuit according to a third embodiment of the present invention.

FIG. 6 is a diagram illustrating an example of an arrangement in a bit reverse order of a result obtained after a butterfly operation.

FIG. 7 is a block diagram illustrating a configuration of a sub-band division operation circuit according to a fourth embodiment of the present invention.

FIG. 8 is a block diagram illustrating a configuration for performing a filtering process on a subband-divided audio signal.

FIG. 9 is a diagram showing another example of an arrangement of a result obtained after the butterfly operation in a bit reverse order.

FIG. 10 is a block diagram illustrating a configuration of a conventional subband division operation circuit.

FIG. 11 is a diagram showing a configuration of a sub-band division operation unit in a conventional sub-band division operation circuit.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Windowing calculation part 2 Folding calculation part 3 Butterfly calculation part 4 Phase correction part 5 Window function ROM 6 First sine / cosine function ROM 7 Second sine / cosine function ROM 8 Folding calculation part 9 Addition part 10 Memory 11 Control Unit 12 Operation control unit 13 Table memory 14 Judgment unit 15 Designation unit 21 Frequency analyzer 22 Gain multiplier 23 Signal synthesizer 31 First butterfly operation unit 32 Second butterfly operation unit 33 Third butterfly operation unit 34 Fourth butterfly operation unit 35 5th butterfly operation unit

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Yasaki Sato 2-6-3 Otemachi, Chiyoda-ku, Tokyo Inside Nippon Steel Corporation (56) References JP-A-5-216496 (JP, A) JP-A-59-205670 (JP, A) JP-A-60-3789 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H03M 7/30 G06T 9/00 G10L 19/02 H03H 17/02 H04N 7/24

Claims (6)

(57) [Claims] 1. A sub-band division operation circuit for frequency-analyzing an input signal and dividing the signal into a plurality of sub-bands having different frequencies, wherein a windowing operation means for performing a windowing operation on the input frequency analysis source signal. A folding operation means for performing a folding operation on a signal output from the windowing operation means, and a multi-stage connected butterfly operation means for obtaining sub-band information by performing a butterfly operation on a signal output from the folding operation means And a phase that corrects the phase of the sub-band information output from the butterfly operation means so as to compensate for the deviation from the phase of the sub-band information obtained when normal frequency analysis is performed without applying the butterfly operation. A sub-band division operation circuit comprising a correction operation means. 2. A sub-band division operation circuit for frequency-analyzing an input signal and dividing the signal into a plurality of sub-bands having different frequencies, wherein a windowing operation means for applying a window operation to the input frequency analysis source signal. A folding operation means for performing a folding operation on a signal output from the windowing operation means, and a multi-stage connected butterfly operation means for obtaining sub-band information by performing a butterfly operation on a signal output from the folding operation means The folding operation means includes an offset amount correction means for correcting an offset amount corresponding to a phase shift of subband information obtained when normal frequency analysis is performed without applying the butterfly operation. And a sub-band division operation circuit. 3. An operation control means for controlling so as to omit a part of a butterfly operation corresponding to a part having regularity in an arrangement of a bit reverse order of a result obtained after the butterfly operation. The sub-band division operation circuit according to claim 1. 4. A determining means for determining whether an input signal of a frequency analysis source is a signal having only a real part, and when the determining means determines that the input signal includes only a real part. An arithmetic control means for controlling so as to omit a butterfly operation performed based on an imaginary part of the input signal.
4. The sub-band division operation circuit according to any one of items 3 to 3. 5. A designating means for designating one of a real part and an imaginary part of subband information obtained after said butterfly operation, and said designating means designating one of a real part and an imaginary part. 5. The control device according to claim 1, further comprising: an arithmetic control unit configured to perform a control so as to omit the butterfly operation for obtaining the other information not specified when the information is present. 6. Subband division operation circuit. 6. The method according to claim 3, wherein when the regularity of the bit-reversed array obtained after the butterfly operation deteriorates if one of the butterfly operations is omitted, all of the butterfly operations are executed. 2. The sub-band division operation circuit according to claim 1.