JP4004525B1 - Signal processing method, signal processing apparatus, and computer program - Google Patents

Abstract

Provided is a signal processing method capable of dealing with various encoding methods on the decoding side and capable of appropriately interpolating energy.
An index value calculation unit calculates a pure tone index value indicating a degree of pure tone of a dequantized acoustic signal. This pure tone index value is obtained, for example, by subtracting the average value from the maximum value related to the scale factor of each frequency band. The output unit outputs white noise from a white noise storage unit that stores white noise. The pure tone determination unit 28 compares the pure tone index value calculated by the index value calculation unit 27 with a reference value stored in advance, and determines whether or not the acoustic signal is a pure tone. Then, when the pure tone determination unit 28 does not determine that the acoustic signal is a pure tone, the correction frequency coefficient calculation unit calculates the coefficient of each frequency band of the acoustic signal based on the white noise output from the white noise storage unit. The corrected coefficient is calculated by adding values related to white noise.
[Selection] Figure 16

Description

  The present invention relates to a signal processing method for processing an acoustic signal obtained by dequantizing an encoded acoustic signal, a signal processing device, and a computer program for causing the signal processing device to function as a computer.

  MP3 (MPEG 1 Audio Player 3), AAC (Advanced Audio Coding), ATRAC (Adaptive TRansform Acoustic Coding), WMA (Windows (registered trademark) Media Audio) or AC-3 (Audio Code Number) 3) etc. are known. For example, in the MP3 system, in order to compress with high efficiency, the acoustic signal is divided into a plurality of frequency bands and is blocked in units of variable length. The blocked digital data is converted into a spectrum signal by MDCT (Modified Discrete Cosine Transform) processing, and each spectrum signal is encoded with the number of bits assigned using the psychoacoustic characteristics (for example, Patent Documents 1 to 3).

  The acoustic signal encoded in this way is decoded by a decoding device. FIG. 26 is a block diagram showing a hardware configuration of a conventional decoding device. In the figure, reference numeral 100 denotes a conventional decoding device, which includes an unpacking unit 101, an inverse quantization unit 102, a frequency time conversion unit 103, a frequency band synthesis unit 104, and an acoustic signal output unit 105. The encoded acoustic signal is input to the unpacking unit 101, and the quantization coefficient, scale factor, scale factor multiplexer, global gain, and sub-block gain are unpacked from the frame information of the acoustic signal. Then, the inverse quantization unit 102 performs inverse quantization on the IMDCT coefficient using the quantization coefficient, the scale factor, the scale factor multiplexer, the global gain, and the sub-block gain.

  An IMDCT coefficient (Inverse Modified Discrete Cosine Transform) inversely quantized by the inverse quantization unit 102 is subjected to IMDCT processing by the frequency time conversion unit 103 for each frequency band, and converted to time-axis data. Further, the inversely converted frequency band is subjected to band synthesis by an IPFB (Inverse Polyphase Filter Bank), which is a band synthesis filter, in the frequency band synthesis unit 104 and then output to the acoustic signal output unit 105 (for example, Patent Literature 1). 3).

Further, in order to compensate for the lack of power feeling associated with compression, a technique for supplementing a spectrum for power adjustment with a spectrum at the time of decoding is disclosed (for example, see Patent Document 4). In the technique described in Patent Literature 4, power adjustment information to be supplemented is generated in the power adjustment information determination unit in the encoding device based on the characteristics of the input audio signal during encoding. Next, the power adjustment information is encoded together with the encoded audio signal. Then, the power adjustment information encoded in the power adjustment information decoding unit in the decoding apparatus is decoded, and further, the power adjustment information is generated in the power correction spectrum generation / synthesis unit to supplement the decoded audio signal.
JP 2002-351500 A JP-A-2005-195983 JP 2005-26940 A JP 2003-323198 A

  However, since the acoustic signal is quantized during encoding, there is a problem that energy of the acoustic signal is lost due to rounding or truncation by quantization. For this reason, even at the time of decoding, an unsatisfactory acoustic signal is generated due to energy loss. Patent Documents 1 to 3 do not provide means for solving such problems. Moreover, although the technique described in Patent Document 4 supplements power, it is necessary to analyze the input audio signal in the encoding device at the time of encoding, generate power adjustment information, and encode it. In addition, it is necessary to decode the encoded power adjustment information by providing a power adjustment information decoding unit on the decoding device side as well, and the energy signal is completely interpolated for the acoustic signal in which such power adjustment information is not stored. There was a problem that I could not. In particular, in recent years, encoding schemes associated with various standards have become prominent, and the technique described in Patent Document 4 has a problem that encoded acoustic signals of various schemes cannot be appropriately interpolated.

  The present invention has been made in view of such circumstances, and its object is to add a value related to white noise to a coefficient of an inversely quantized acoustic signal and correct the acoustic signal when the acoustic signal is not determined to be a pure tone. The signal processing method, the signal processing apparatus, and the computer that can cope with various encoding methods on the decoding side and can appropriately interpolate energy function as the signal processing apparatus. It is in providing the computer program for making it do.

  Another object of the present invention is to determine whether or not the bit rate of the acoustic signal is smaller than a pre-stored reference bit rate, and when it is determined that the bit rate of the acoustic signal is smaller than the reference bit rate, By calculating the corrected coefficient by adding a value related to white noise to the coefficient of the signal, it is possible to cope with various encoding methods on the decoding side, and signal processing that can appropriately interpolate energy It is to provide a method and the like.

  Another object of the present invention is to read out the reference bit rate corresponding to the sampling frequency of the acoustic signal from the table, and determine whether the bit rate of the acoustic signal is smaller than the read reference bit rate. An object of the present invention is to provide a signal processing apparatus capable of executing optimum energy interpolation even when acoustic signals relating to different sampling frequencies are input by correcting the coefficients.

  Another object of the present invention is to calculate a correction coefficient based on the level of the coefficient of the dequantized acoustic signal and the white noise output from the white noise storage unit, and multiply the white noise to calculate the correction value. Then, it is to provide a signal processing apparatus capable of interpolating energy lost at the time of encoding by adding to a coefficient and eliminating deficiency due to lack of energy.

  Another object of the present invention is to provide a signal processing apparatus capable of calculating an optimum correction coefficient by dividing the coefficient level by the white noise level stored in the white noise storage unit. .

The signal processing method according to the present invention is a signal processing method for processing an acoustic signal obtained by dequantizing an encoded acoustic signal, and an index value for calculating a pure tone index value indicating a degree of pure tone of the dequantized acoustic signal. A calculation step, a pure tone determination step for determining whether or not the acoustic signal is a pure tone by comparing the pure tone index value calculated by the index value calculation step with a reference value stored in advance, and a white noise storing white noise When the acoustic signal is not determined to be pure tone by the output step of outputting white noise from the storage unit and the pure tone determination step, a value related to the white noise output by the output step is added to each coefficient of the acoustic signal. a correction frequency coefficient computing step of computing a corrected coefficient Te, a level detecting step of detecting a level of inverse quantization to coefficients of the acoustic signal, the white noise storage A correction coefficient calculating step for calculating a correction coefficient based on the white noise output from the level and the level; a correction value calculating step for calculating a correction value by multiplying the white noise stored in the white noise storage unit by the correction coefficient; The level detection step detects a minimum value excluding zero of the coefficient of the dequantized acoustic signal, and the correction frequency coefficient calculation step includes an acoustic signal obtained by dequantizing the correction value as a value related to white noise. The corrected coefficient is calculated by adding to the coefficient of the signal .

The signal processing method according to the present invention is a signal processing method for processing an acoustic signal obtained by dequantizing an encoded acoustic signal, and a bit for determining whether or not the bit rate of the acoustic signal is smaller than a reference bit rate stored in advance. When it is determined that the bit rate of the acoustic signal is smaller than the reference bit rate by the rate comparison step, the output step of outputting the white noise from the white noise storage unit storing the white noise, and the bit rate comparison step, A correction frequency coefficient calculation step of calculating a corrected coefficient by adding a value related to white noise output in the output step to each coefficient of the acoustic signal, and the bit rate comparison step corresponds to the sampling frequency of the acoustic signal The reference bit rate used as a reference is stored for each sampling frequency. A bit rate of the acoustic signal is determined based on whether the bit rate of the acoustic signal is smaller than the read reference bit rate or not. When it is determined that the coefficient is smaller than the bit rate, a coefficient after correction is calculated by adding a value related to white noise output by the output step to a coefficient of each frequency band of the acoustic signal .

The signal processing apparatus according to the present invention is an index value for calculating a pure tone index value indicating a degree of pure tone of a dequantized acoustic signal in a signal processing apparatus that processes an acoustic signal obtained by dequantizing an encoded acoustic signal. A calculation unit, a pure tone determination unit that determines whether or not the acoustic signal is a pure tone by comparing the pure tone index value calculated by the index value calculation unit with a reference value stored in advance, and a white noise storing white noise When the acoustic signal is not determined to be pure tone by the output unit that outputs white noise from the storage unit and the pure tone determination unit, a value related to the white noise output from the output unit is added to each coefficient of the acoustic signal. Correction frequency coefficient calculation unit for calculating the corrected coefficient, level detection unit for detecting the level of the coefficient of the dequantized acoustic signal, white noise output from the white noise storage unit and correction based on the level Person in charge And a correction value calculation unit that calculates a correction value by multiplying the correction coefficient by white noise stored in the white noise storage unit, and the level detection unit includes the inverse quantization The corrected frequency coefficient calculation unit corrects the correction value by adding the correction value to the coefficient of the dequantized acoustic signal as a value related to white noise. It is configured to calculate a later coefficient .

  The signal processing device according to the present invention is characterized in that the index value calculation unit is configured so that a value obtained by subtracting an average value from a maximum value related to a scale factor of each frequency band is a pure tone index value. To do.

A signal processing device according to the present invention is a signal processing device that processes an acoustic signal obtained by dequantizing an encoded acoustic signal, and determines whether or not the bit rate of the acoustic signal is smaller than a pre-stored reference bit rate. When it is determined by the rate comparison unit, the output unit that outputs white noise from the white noise storage unit that stores white noise, and the bit rate comparison unit that the bit rate of the acoustic signal is smaller than the reference bit rate, A correction frequency coefficient calculation unit that calculates a corrected coefficient by adding a value related to white noise output from the output unit to each coefficient of the acoustic signal, and a table that stores a reference bit rate that is a reference for each sampling frequency The bit rate comparison unit reads a reference bit rate corresponding to the sampling frequency of the acoustic signal from the table, It is configured to determine whether or not the bit rate of the acoustic signal is smaller than the read reference bit rate, and the correction frequency coefficient calculation unit uses the bit rate comparison unit to determine the bit rate of the acoustic signal as a reference. When it is determined that the coefficient is smaller than the bit rate, the corrected coefficient is calculated by adding a value related to white noise output from the output unit to the coefficient of each frequency band of the acoustic signal. And

  A signal processing apparatus according to the present invention includes a level detection unit that detects a level of a coefficient of a dequantized acoustic signal, a white noise output from the white noise storage unit, and a correction coefficient that calculates a correction coefficient based on the level A calculation unit; and a correction value calculation unit that calculates a correction value by multiplying the white noise stored in the white noise storage unit by the correction coefficient, and the correction frequency coefficient calculation unit converts the correction value to white noise. The corrected coefficient is calculated by adding to the coefficient of the inversely quantized acoustic signal as a value relating to.

  The signal processing apparatus according to the present invention is characterized in that the level detection unit is configured to detect a minimum value excluding zero of a coefficient of the dequantized acoustic signal.

  The signal processing apparatus according to the present invention is configured such that the correction coefficient calculation unit calculates a correction coefficient by dividing a level detected by the level detection unit by a level of white noise stored in the white noise storage unit. It is characterized by being.

  The signal processing device according to the present invention is characterized in that the level of the white noise is a maximum value or an average value of the white noise stored in the white noise storage unit.

The computer program according to the present invention is a computer program for processing an acoustic signal obtained by dequantizing an encoded acoustic signal by a computer. The computer program is a pure tone index value indicating the degree of pure tone of the inverse quantized acoustic signal. An index value calculating step for calculating the tone value, a pure tone index value calculated by the index value calculating step, and a reference value stored in advance to determine whether or not the acoustic signal is a pure tone; and a white noise White noise output from the output step to each coefficient of the acoustic signal when the acoustic signal is not determined to be pure tone by the output step of outputting white noise from the white noise storage unit storing a correction frequency coefficient computing step of computing a corrected coefficient by adding a value for, the coefficients of the acoustic signal dequantized A level detection step for detecting a bell; a white noise output from the white noise storage unit; a correction coefficient calculation step for calculating a correction coefficient based on the level; and a white color stored in the white noise storage unit. A correction value calculation step of calculating a correction value by multiplying noise, the level detection step detects a minimum value excluding zero of the coefficient of the dequantized acoustic signal, and the correction frequency coefficient calculation step includes: The correction value is added to the coefficient of the acoustic signal that is inversely quantized as a value related to white noise, and the corrected coefficient is calculated .

The computer program according to the present invention is a computer program for processing an acoustic signal obtained by dequantizing an encoded acoustic signal by a computer, and whether or not the bit rate of the acoustic signal is smaller than a reference bit rate stored in advance in the computer. A bit rate comparison step for determining whether or not, a step of outputting white noise from a white noise storage unit storing white noise, and a bit rate comparison step determine that the bit rate of the acoustic signal is smaller than a reference bit rate. A correction frequency coefficient calculation step of calculating a corrected coefficient by adding a value related to white noise output by the output step to each coefficient of the acoustic signal, and the bit rate comparison step includes: Set the reference bit rate corresponding to the sampling frequency of the signal. Reading from a table storing a reference bit rate serving as a reference for each sampling frequency, determining whether the bit rate of the acoustic signal is smaller than the read reference bit rate, and the correction frequency coefficient calculating step includes When the bit rate comparison step determines that the bit rate of the acoustic signal is smaller than the reference bit rate, the corrected value is obtained by adding the value related to the white noise output by the output step to the coefficient of each frequency band of the acoustic signal. The coefficient is calculated .

  In the present invention, the index value calculation unit calculates a pure tone index value indicating the degree of pure tone of the dequantized acoustic signal. This pure tone index value is obtained, for example, by subtracting the average value from the maximum value related to the scale factor of each frequency band. The output unit outputs white noise from a white noise storage unit that stores white noise. The pure tone determination unit compares the pure tone index value calculated by the index value calculation unit with a reference value stored in advance, and determines whether or not the acoustic signal is a pure tone. Then, when the acoustic signal is not determined to be a pure tone by the pure tone determination unit, the positive frequency coefficient calculation unit calculates a value related to the white noise calculated based on the white noise output from the white noise storage unit as the coefficient of the acoustic signal. Is added to calculate the corrected coefficient, so that it is possible to determine whether or not to interpolate energy independently on the decoding side, and it is possible to appropriately interpolate energy for various standards.

  In the present invention, the bit rate comparison unit determines whether or not the bit rate of the acoustic signal is smaller than a reference bit rate stored in advance. When the bit rate comparison unit determines that the bit rate of the acoustic signal is smaller than the reference bit rate, the correction frequency coefficient calculation unit adds a value related to white noise output from the white noise storage unit to the coefficient and performs correction. Therefore, it is possible to determine whether to interpolate energy independently on the decoding side, and to appropriately interpolate energy for various standards.

  In the present invention, the table stores a reference bit rate serving as a reference for each sampling frequency. The bit rate comparison unit reads a reference bit rate corresponding to the sampling frequency of the acoustic signal from the table, and determines whether the bit rate of the acoustic signal is smaller than the read reference bit rate. When the bit rate comparison unit determines that the bit rate of the acoustic signal is smaller than the reference bit rate, the correction frequency coefficient calculation unit adds the value related to the white noise output from the white noise storage unit to the coefficient of the acoustic signal. Thus, the corrected coefficient is calculated. Thus, a reference bit rate is provided for each sampling frequency, white noise can be added according to the sampling frequency and the reference bit rate, and various types of different acoustic signals can be appropriately interpolated.

  In the present invention, in the present invention, the level detection unit detects the level of the coefficient of the acoustic signal obtained by dequantizing the encoded acoustic signal. The correction coefficient calculation unit calculates the correction coefficient based on the white noise and level output from the white noise storage unit that stores the white noise. In this process, for example, the correction coefficient is calculated by dividing the minimum value excluding zero of the coefficient by the level (for example, maximum value or average value) of the white noise stored in the white noise storage unit. The correction value calculation unit calculates a correction value by multiplying the correction coefficient calculated by the correction coefficient calculation unit by the white noise stored in the white noise storage unit. Since the correction value obtained in this manner is added to the coefficient of the inversely quantized acoustic signal to calculate the corrected coefficient, the optimum value for each band is determined according to the level of the coefficient existing in the band. A correction value related to the amount of white noise can be added. As a result, the energy lost during quantization is interpolated by an optimal amount.

  In the present invention, the index value calculation unit calculates a pure tone index value indicating the degree of pure tone of the dequantized acoustic signal, and the pure tone determination unit calculates the pure tone index calculated by the index value calculation unit. The value is compared with a reference value stored in advance, and it is determined whether or not the acoustic signal is a pure tone. Then, when it is not determined to be a pure tone, the corrected coefficient is calculated by adding the value related to the white noise output from the white noise storage unit to the coefficient of the acoustic signal, so that energy is interpolated independently on the decoding side. Whether or not the energy can be appropriately interpolated with respect to various standards. Further, it is not necessary to give information on whether or not to interpolate energy at the time of encoding. In addition, when it is determined to be a pure tone, white noise is not added, so that dynamic range, distortion, frequency characteristics, etc. are not deteriorated, and reproduction is possible without impairing the sound quality of an instrument such as a piano that is close to a sine wave. is there.

  In the present invention, the bit rate comparison unit determines whether or not the bit rate of the acoustic signal is lower than the reference bit rate stored in advance, and determines that the bit rate of the acoustic signal is lower than the reference bit rate. In this case, since the corrected coefficient is calculated by adding the value related to the white noise output from the white noise storage unit to the coefficient, it is possible to determine whether to interpolate energy independently on the decoding side. It becomes possible to appropriately interpolate energy for various standards. Also, when the bit rate of the acoustic signal is larger than the reference bit rate, white noise is not added, and reproduction is possible without adding excessive energy.

  In the present invention, the reference bit rate corresponding to the sampling frequency of the acoustic signal is read from the table, and it is determined whether or not the bit rate of the acoustic signal is smaller than the read reference bit rate. Then, when it is determined that the bit rate of the acoustic signal is smaller than the reference bit rate, the corrected coefficient is calculated by adding a value related to the white noise to the coefficient of the acoustic signal. Therefore, according to the sampling frequency and the reference bit rate Thus, white noise can be added, and various types of acoustic signals can be appropriately interpolated.

  In the present invention, the correction coefficient is calculated based on the coefficient level and white noise, the correction value is calculated by multiplying the correction coefficient by white noise, and the corrected coefficient is calculated by adding to the coefficient. The correction value related to the optimum amount of white noise can be added according to the level of the coefficient existing in the band. As a result, the energy lost at the time of quantization is interpolated, and it is possible to prevent an unsatisfactory sense of energy and to reproduce an acoustic signal closer to the original sound. .

Embodiment 1
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a hardware configuration of a decoding device as a signal processing device. In the figure, reference numeral 20 denotes a decoding device that decodes an encoded acoustic signal, and includes an acoustic signal input unit 21, an unpacking unit 22, an inverse quantization unit 23, an interpolation processing unit 1, a frequency time conversion unit 24, and a frequency band synthesis. The unit 25 and the acoustic signal output unit 26 are included. In the present embodiment, an example in which MP3 is applied as a compression encoding method will be described. However, other methods may be applied in the same manner.

  The encoded acoustic signal read from the recording medium or the encoded acoustic signal received by the digital tuner is input to the acoustic signal input unit 21, and the input encoded acoustic signal is input to the unpacking unit (demultiplexer) 22. Is output. The unpacking unit 22 unpacks the quantization coefficient, scale factor, scale factor multiplexer, global gain, and sub-block gain from the frame information of the acoustic signal. Using the unpacked quantization coefficient, the number of quantization bits, the scale factor, the scale factor multiplexer, the global gain, and the sub-block gain, the inverse quantization unit 23 performs inverse quantization on the IMDCT coefficient. From the inverse quantization unit 23, an IMDCT coefficient represented by the following equation (1) is output for each frequency band according to the block length (long block or short block).

  In Expression (1), the variable m is an IMDCT coefficient index, MK (m) is a quantized coefficient (Huffman decoded value), sgn (MK (m)) is a sign of the quantized coefficient, scalefac_multiplier is 1 or 0.5, gr Is the index of the granule, wnd is the index of the window shape, sfb is the index of the scale factor band, preflag [gr] is the pre-emphasis flag, 0 or 1, and pretab [sfb] is obtained by a predetermined pre-emphasis table Represents a value. Note that the scale factor in ATRAC (for example, expressed by 6 bits each and can be specified in units of about 2 dB) is the same as the value related to the scale factor in MP3, and the value related to the scale factor in MP3 is expressed by Equation (1). As shown in FIG. 4, the scale factor, the scale factor multiplexer, the global gain, and the sub-block gain (the part after the 2nd multiplier in the equation (1)), the pre-emphasis presence / absence flag, and the value obtained from the pre-emphasis table are used. . Hereinafter, the scale factor in ATRAC and the value related to the scale factor in MP3 are collectively described as a scale factor.

  In the present embodiment, as shown in the figure, 32 frequency band block (0) to block (31) include IMDCT coefficients I (0), I (1),..., I (m),. ) Is output. When the sampling frequency is 44.1 kHz, the frequency of block (0) is 0 Hz to 689.0625 Hz, block (1) is 689.0625 Hz to 1378.125 Hz, and block (31) is 21360.9375 Hz to 22050 Hz. In the following, a block of an arbitrary frequency band is assumed to be block (k). Here, k is an integer and satisfies 0 ≦ k ≦ 31. The IMDCT coefficients I (0) to I (575) of each frequency band are input to the interpolation processing unit 1.

  The IMDCT coefficient of each frequency band is composed of a plurality of coefficients (spectrums) according to the block length. The long block consists of 18 coefficients, and the short block consists of 6 coefficients. In the present embodiment, description will be made assuming that the block length is a long block. In addition, 576 samples of IMDCT coefficients constitute one granule, and a total of 1152 coefficient samples of granule 0 and granule 1 are processed as one frame. Hereinafter, one granule is referred to as one frame. explain.

  FIG. 2 is a graph showing changes in the IMDCT coefficient with respect to frequency. The horizontal axis represents frequency, and the vertical axis represents the coefficient. In the case of a long block, an IMDCT coefficient (hereinafter, represented by a coefficient I (m)) has 18 coefficients I (18 × k) to I (18 × k + 17) in one frequency band. In the graph of FIG. 2, changes in coefficients I (18 × k), I (18 × k + 1),..., I (18 × k + 17) corresponding to frequencies 18 × k, 18 × k + 1,. It is shown. This coefficient takes a positive, negative or zero value.

  In FIG. 1, the coefficient I (m) is input to the interpolation processing unit 1, and the corrected coefficient I (m) to which the correction value related to white noise is added is output from the interpolation processing unit 1. The frequency time conversion unit 24 performs an IMDCT process and converts it into a time axis acoustic signal. Further, the inversely converted acoustic signal is subjected to band synthesis by an IPFB (Inverse Polyphase Filter Bank) which is a band synthesis filter in the frequency band synthesis unit 25 and then output to the acoustic signal output unit 26.

  FIG. 3 is a block diagram showing a hardware configuration of the interpolation processing unit 1. The interpolation processing unit 1 includes a minimum value detection unit 12 as a level detection unit, a correction coefficient calculation unit 18, a correction value calculation unit 13, a white noise storage unit 16, an output unit 14, a maximum spectrum detection unit 15, and a correction frequency coefficient calculation unit 17. It is comprised including. The coefficient I (m) of each frequency band output from the inverse quantization unit 23 is input to the minimum value detection unit 12. The minimum value detection unit 12 detects the minimum value of the absolute value of the coefficient (spectrum) excluding zero from the coefficients I (0) to I (575) of one frame. In the example of FIG. 2, I (18 × k + 1) is detected instead of zero I (18 × k + 2). The detected coefficient I (m) having the minimum absolute value is output to the correction coefficient calculation unit 18 as the minimum value. In the present embodiment, the minimum value detection unit 12 is used as the level detection unit and the minimum value of the absolute value of the coefficient excluding zero of the coefficient I (m) is detected as the level. It is not a thing. For example, the level of the coefficient I (m) may be the second smallest value of the coefficient I (m), three average values or variance values having the smallest absolute value, and the like.

  The white noise storage unit 16 stores a plurality of white noises (frequency component sources of white noise) in units of blocks. FIG. 4 is an explanatory diagram showing a record layout of the white noise storage unit 16. The white noise storage unit 16 is configured by orthogonal transform coefficients (spectrums) obtained by dividing white noise of a time corresponding to one frame into a plurality of bands (blocks) and performing orthogonal transform for each band. The white noise storage unit 16 includes a block field, a number m field, and a spectrum Iwn (m) field. The block field stores the number of blocks constituting white noise corresponding to the long block, and a total of 32 blocks from block 0 to block 31 are prepared. The number m field stores a number for identifying the white noise spectrum for each block.

  Further, the spectrum of white noise is stored in the spectrum Iwn (m) field in association with the number m. In the present embodiment, a total of 576 spectra are stored, with 18 spectra per block for a total of 32 blocks. For example, the white noise spectrum “0.003125” is stored in the number “1” of the block “0”. The white noise spectrum may be set to have a predetermined value when the entire spectrum is averaged, for example, about -20 dB. Note that Iwn (m) is a table for creating white noise at a level of about 1/2 of the minimum value of I (m), and need not be limited to a numerical value of about −20 dB. In the white noise storage unit 16, conversion coefficient (spectrum) data obtained by performing time frequency conversion of a white noise source of a time signal at the time of compression is stored in a memory or the like according to the block length. That is, the white noise per frame is divided into a plurality of bands, and is configured by orthogonal transform coefficients obtained by orthogonal transform for each band. In the case of a long block, it is divided into 18 time data by frequency division at the time of compression, and the conversion coefficient converted into 18 frequency component data is recorded in a memory or the like, and in the case of a short block, by frequency division at the time of compression. A conversion coefficient that is divided into six pieces of time data and converted into six pieces of frequency component data is recorded in a memory or the like. In the present embodiment, an embodiment using white noise prepared in advance for ease of explanation will be described. However, the average spectrum of the whole is set to a predetermined value by a random number generator (not shown) in the apparatus. The following white noise may be sequentially generated and temporarily stored in the white noise storage unit 16 for output.

  The output unit 14 connected to the white noise storage unit 16 randomly or regularly selects a plurality of blocks of the white noise storage unit 16, and the spectrum related to the white noise of the selected block is the maximum spectrum detection unit 15 and Output to the correction value calculation unit 13. The maximum spectrum detector 15 detects the maximum value related to the absolute value of the spectrum in the block output from the output unit 14. In the present embodiment, the maximum spectrum detector 15 detects the maximum value of the spectrum, but an average spectrum detector (not shown) is provided in place of the maximum spectrum detector 15, and the average of the spectra in the block is The value may be detected. The spectrum (or spectrum average value) of the white noise whose absolute value detected by the maximum spectrum detection unit 15 is the maximum value is output to the correction coefficient calculation unit 18. In the present embodiment, the maximum spectrum detection unit 15 or the average spectrum detection unit is provided to output the maximum value or average value of white noise. It is not limited to form. For example, an average value or a variance value of a plurality of spectra having a large white noise value of a selected block may be detected.

  The correction coefficient calculation unit 18 calculates the correction coefficient by dividing the minimum value of the coefficient I (m) output from the minimum value detection unit 12 by twice the maximum value (or average value) of the white noise spectrum. To do. The correction coefficient calculation unit 18 outputs the calculated correction coefficient to the correction value calculation unit 13. In the present embodiment, the minimum value of the coefficient I (m) is divided by twice the maximum value (or average value) of the spectrum of white noise. You may divide by. The correction value calculation unit 13 multiplies the spectrum relating to one block of white noise output from the output unit 14 and the correction coefficient output from the correction coefficient calculation unit 18 as a value related to the white noise to be corrected, to obtain a coefficient I ( Calculate a correction value to be corrected to m). For example, when the white noise spectrums Iwm (1) to Iwm (18) of the block “0” are output, each spectrum is multiplied by a correction coefficient, and a correction value as a value related to the white noise of “18” is calculated. The The correction frequency coefficient calculation unit 17 calculates the corrected coefficient I (m) by adding the correction value output from the correction value calculation unit 13 and the coefficient I (m). The corrected coefficient I (m) is output to the frequency time conversion unit 24. In the above example, 18 correction values are added to the coefficients I (18 × k) to I (18 × k + 17), respectively, and corrected coefficients I (18 × k) to I (18 × k + 17) are calculated. Here, the correction coefficient calculated by the correction coefficient calculation unit 18 is 0 ≦ k ≦ 31 between frequency bands, and the same value is used. Further, the white noise related to the block selected randomly or regularly from the output unit 14 and output to the correction value calculation unit 13 is 0 ≦ k ≦ 31 between frequency bands, and the white noise spectrum related to the same block is used. Alternatively, the spectrum of white noise related to different blocks that are selected randomly or regularly in a frequency band between 0 ≦ k ≦ 31 may be used. Note that the correction value may not be added to the frequency band where the quantization value is non-zero.

  The processing of each hardware described above will be described using a flowchart. FIG. 5 is a flowchart showing the procedure for calculating the corrected coefficient. First, the minimum value detection unit 12 detects the minimum value of the absolute value of the coefficient excluding zero of the coefficient I (m) from the coefficients in one frame (step S51). The output unit 14 connected to the white noise storage unit 16 selects one block of the white noise storage unit 16 randomly or regularly, and the white noise (spectrum) of the selected block is selected as the maximum spectrum detection unit 15. And it outputs to the correction value calculation part 13 (step S52). The maximum spectrum detector 15 detects the maximum value related to the absolute value of the spectrum in the block output from the output unit 14 (step S53). The spectrum (or spectrum average value) of the white noise whose absolute value is maximum detected by the maximum spectrum detector 15 is output to the correction coefficient calculator 18.

  The correction coefficient calculation unit 18 calculates the correction coefficient by dividing the minimum value of the coefficient I (m) output from the minimum value detection unit 12 by twice the maximum value (or average value) of the white noise spectrum. (Step S54). The correction coefficient calculation unit 18 outputs the calculated correction coefficient to the correction value calculation unit 13. The correction value calculation unit 13 multiplies the correction coefficient output from the correction coefficient calculation unit 18 by each spectrum related to the white noise of the block randomly or regularly output from the output unit 14 and corrects the coefficient to I (m). A correction value to be calculated is calculated (step S55). The correction frequency coefficient calculation unit 17 adds each correction value output from the correction value calculation unit 13 to the coefficient I (m), and calculates the corrected coefficient I (m) (step S56). For example, when the correction coefficient calculated in step S54 is J, the spectra Iwn (1) to Iwn (18) in the block 0 shown in FIG. 4 are multiplied by J and corrected values Iwn ′ (1) to Iwn ′ ( 18) is calculated. Then, correction values Iwn ′ (1) to Iwn ′ (18) are added to the coefficients I (0) to I (17) in block (0), and the corrected coefficients I ′ (0) to I '(17) can be obtained. The same applies to the coefficient I (18) to the coefficient I (35) of the next block (1). That is, correction values Iwn ′ (19) to Iwn ′ (36) are calculated by multiplying the spectrums Iwn (19) to Iwn (36) in the block 1 shown in FIG. 4 selected randomly or regularly by J. To do. Then, the correction values Iwn ′ (19) to Iwn ′ (36) are added to the coefficients I (18) to I (35) in block (1), and the corrected coefficients I ′ (18) to I '(35) can be obtained. Such processing is repeatedly applied to different blocks. As described above, since the correction value calculated based on the minimum value of the IMDCT coefficient and the white noise is added to the IMDCT coefficient, the optimum amount of white noise that matches the sound pressure can be added, and the energy lost by the quantization can be appropriately set. It becomes possible to make up for.

Embodiment 2
The second embodiment relates to a mode in which the minimum value detected by the minimum value detection unit 12 is corrected. FIG. 6 is a block diagram illustrating a hardware configuration of the interpolation processing unit 1 according to the second embodiment. In addition to the configuration of the first embodiment, the interpolation processing unit 1 according to the second embodiment includes an energy calculation unit 121, a previous frame energy storage unit 122, an energy change rate calculation unit 123, a determination unit 124, a minimum value correction unit 126, a minimum A value change rate calculation unit 125, a corrected minimum value setting unit 127, a minimum value setting unit 128, and a previous frame minimum value storage unit 129 are configured. In the following, a frame that means a time interval in units of encoding (decoding) processing is referred to as a previous frame.

  The frequency band coefficient I (m) is input to the energy calculation unit 121, the minimum value detection unit 12, and the correction frequency coefficient calculation unit 17. The energy calculation unit 121 calculates energy by a value obtained by squaring the coefficient I (m), that is, the sum of power spectra. This energy is calculated by taking the square sum of the coefficients of the frequency bands I (0) to I (575) of the frame, as shown in equation (2). Here, energy is used, but an average power spectrum calculated by dividing by the number of spectra, for example, 576, may be used as shown in Equation (3). Below, the energy calculation part 121 demonstrates as what calculates energy.

  Here, E indicates the energy of one frame, m is a natural number, and the maximum value is M. In the example of FIG. 2, energy is calculated by taking the square sum of I (0) to I (575). The previous frame energy storage unit 122 stores the energy of the previous frame calculated by the energy calculation unit 121. The energy change rate calculation unit 123 receives the energy of the previous frame stored by the previous frame energy storage unit 122 and the energy calculated by the energy calculation unit 121, respectively. The energy change rate calculation unit 123 divides the energy calculated by the energy calculation unit 121 by the energy of the previous frame output from the previous frame energy storage unit 122 as shown by the equation (4), and takes the square root to obtain the energy. Calculate the rate of change. That is, the energy of the decoding processing unit (frame) at the predetermined time calculated by the energy calculation unit 121 and the energy of the decoding processing unit (frame) past the predetermined time stored by the previous frame energy storage unit 122. Calculate the rate of change.

  Here, Ep indicates the energy of the previous frame, and ER indicates the energy change rate. In the second embodiment, the energy change rate is calculated by dividing the energy by the previous energy and taking the square root. However, the present invention is not limited to this as long as the energy fluctuation between frames can be calculated. Alternatively, for example, the previous energy is subtracted from the calculated energy and squared to obtain the rate of change. Conversely, the previous energy is divided by the energy calculated by the energy calculating unit 121 to obtain a square root.

  The calculated energy change rate is output to the determination unit 124. The coefficient I (m) input to the minimum value detection unit 12 is the coefficient I (m) related to the minimum value of the absolute value excluding zero in the frequency band as in the first embodiment. The previous frame minimum value storage unit 129 stores the minimum value of the previous frame detected by the minimum value detection unit 12. The minimum value change rate calculation unit 125 receives the minimum value of the previous frame stored by the previous frame minimum value storage unit 129 and the minimum value detected by the minimum value detection unit 12. The minimum value change rate calculation unit 125 calculates the minimum value change rate by dividing the minimum value detected by the minimum value detection unit 12 by the minimum value of the previous frame output from the previous frame minimum value storage unit 129. That is, the minimum value of the decoding processing unit (frame) at the predetermined time detected by the minimum value detection unit 12 and the decoding processing unit (frame) of the past from the predetermined time stored in the previous frame minimum value storage unit 129. The rate of change from the minimum value is calculated. In addition, if the change rate of the minimum value can also be calculated from the minimum value of the previous frame, for example, a value obtained by subtracting the minimum value of the previous frame from the detected minimum value and squared is calculated as the change rate. Conversely, the minimum value of the previous frame may be divided by the detected minimum value.

  The determination unit 124 compares the energy change rate output from the energy change rate calculation unit 123 with the minimum value change rate output from the minimum value change rate calculation unit 125, and the energy change rate is greater than the minimum value change rate. Determine whether or not. This corrects the minimum value described in the first embodiment when the fluctuation of the minimum value is large between frames, and uses the energy change rate between frames as a comparison target as a threshold for the determination. . If the determination unit 124 determines that the energy change rate is larger than the minimum value change rate, it is not necessary to correct the energy change rate. Output to 128. The minimum value setting unit 128 outputs the minimum value to the previous frame minimum value storage unit 129 so that the minimum value is stored in the previous frame minimum value storage unit 129 as the minimum value of the previous frame. That is, the minimum value output from the minimum value setting unit 128 is used as the minimum value of the previous frame output from the previous frame minimum value storage unit 129 when calculating the minimum value change rate in the next frame.

  On the other hand, when the determination unit 124 determines that the energy change rate is smaller than the minimum value change rate, the determination unit 124 outputs the energy change rate and the minimum value of the previous frame to the minimum value correction unit 126. The minimum value correction unit 126 corrects the minimum value so that the minimum value change rate and the energy change rate are equal. Specifically, the minimum value is corrected so as to satisfy the conditional expression represented by Expression (5).

  Here, MSR indicates a minimum value change rate obtained by dividing the minimum value of the coefficient I (m) by the minimum value of the previous frame. As described above, the minimum value change rate on the right side of Expression (5) excludes the minimum value of | I (m) | that excludes zero of | I (m) | in the previous frame, as shown in Expression (6). It can be transformed as the one divided by the minimum value. The minimum value after correction to be obtained is a value obtained by multiplying the energy change rate ER by the minimum value of the previous frame. The minimum value correcting unit 126 outputs the minimum value corrected in this way to the corrected minimum value setting unit 127 and the correction coefficient calculating unit 18. Similar to the minimum value setting unit 128, the corrected minimum value setting unit 127 outputs the corrected minimum value to the previous frame minimum value storage unit 129 so as to store the corrected minimum value as the minimum value of the previous frame.

  The previous frame minimum value storage unit 129 stores the minimum value output from the corrected minimum value setting unit 127 or the minimum value setting unit 128. The previous frame minimum value storage unit 129 outputs the minimum value output from the corrected minimum value setting unit 127 or the minimum value setting unit 128 to the minimum value change rate calculation unit 125. In the second embodiment, the minimum value is corrected so that the minimum value change rate becomes equal to the energy change rate. However, the minimum value needs to be corrected strictly until the minimum value change rate becomes equal to the energy change rate. However, it is sufficient to correct until substantially equal, and some error may be given. Further, the corrected minimum value output from the minimum value correction unit 126 is output to the correction coefficient calculation unit 18.

  In the above hardware configuration, the corrected minimum value calculation processing procedure will be described with reference to a flowchart. FIG. 7 is a flowchart showing the procedure for calculating the corrected minimum value. First, the storage contents of the previous frame energy storage unit 122 and the previous frame minimum value storage unit 129 are initialized (step S71). Specifically, the energy of the previous frame and the minimum value of the previous frame are set to 1. In this process, when calculating the corrected coefficient for the first frame, the energy change rate calculation unit 123 and the minimum value change rate calculation unit 125 substitute 0 for the denominator to generate an infinite solution. In the process after the next frame, the energy and minimum value of the calculated previous frame are stored and used. If there is a silent frame in the middle of the music, it starts from the start.

  The frequency band coefficient I (m) is output to the energy calculation unit 121, the minimum value detection unit 12, and the correction frequency coefficient calculation unit 17 (step S72). The energy calculation unit 121 calculates the energy of the coefficient I (m) (step S73). The energy change rate calculation unit 123 receives the energy of the previous frame stored by the previous frame energy storage unit 122 and the energy calculated by the energy calculation unit 121, respectively. The energy change rate calculation unit 123 divides the energy calculated by the energy calculation unit 121 by the energy of the previous frame output from the previous frame energy storage unit 122, and calculates the energy change rate by taking the square root (step S74). ). The previous frame energy storage unit 122 outputs the stored previous frame energy to the energy change rate calculation unit 123, and then stores the energy output from the energy calculation unit 121 as the previous frame energy used in the calculation of the next frame. .

  The calculated energy change rate is output to the determination unit 124. Based on the input coefficient I (m), the minimum value detection unit 12 detects the coefficient I (m) related to the absolute minimum value (step S75). The minimum value change rate calculation unit 125 receives the minimum value of the previous frame stored by the previous frame minimum value storage unit 129 and the minimum value detected by the minimum value detection unit 12. The minimum value change rate calculation unit 125 calculates the minimum value change rate by dividing the minimum value detected by the minimum value detection unit 12 by the minimum value of the previous frame output from the previous frame minimum value storage unit 129 ( Step S76).

  The determination unit 124 compares the energy change rate output from the energy change rate calculation unit 123 with the minimum value change rate output from the minimum value change rate calculation unit 125, and the energy change rate is greater than the minimum value change rate. Whether or not (step S77). If the determination unit 124 determines that the energy change rate is greater than the minimum value change rate (YES in step S77), the determination unit 124 outputs the minimum value detected by the minimum value detection unit 12 to the correction coefficient calculation unit 18 (step S78). . The minimum value is output to the minimum value setting unit 128. The minimum value setting unit 128 outputs the minimum value to the previous frame minimum value storage unit 129 in order to store the minimum value as the minimum value of the previous frame, and the previous frame minimum value storage unit 129 receives the minimum value from the minimum value setting unit 128. The output minimum value is stored (step S79).

  On the other hand, when the determination unit 124 determines that the energy change rate is smaller than the minimum value change rate (NO in step S77), the energy change rate and the minimum value of the previous frame are output to the minimum value correction unit 126 to correct the minimum value. The unit 126 corrects the minimum value so that the minimum value change rate and the energy change rate are equal (step S710). Specifically, as described above, the minimum value correction unit 126 calculates the corrected minimum value by multiplying the energy change rate by the minimum value of the previous frame. The minimum value correction unit 126 outputs the calculated minimum value after correction to the correction minimum value setting unit 127 and the correction coefficient calculation unit 18 (step S711). Similar to the minimum value setting unit 128, the corrected minimum value setting unit 127 outputs the corrected minimum value to the previous frame minimum value storage unit 129 so as to store the corrected minimum value as the minimum value of the previous frame.

  The previous frame minimum value storage unit 129 stores the minimum value output from the corrected minimum value setting unit 127 (step S712). After the processes of step S79 and step S712, the energy calculation unit 121 stores the energy calculated in step S73 in the previous frame energy storage unit 122 (step S713). The minimum value output in step S78 or the minimum value output in step S711 is output to the correction coefficient calculation unit 18. Subsequent processing is the same as step S52 to step S56 in FIG. In this way, paying attention to the energy change rate and the minimum value change rate between frames, if the minimum value change rate is significant, the minimum value is corrected by reducing this and added to the coefficient I (m). Variations in correction values related to white noise can be reduced, and as a result, unnatural interpolation processing between frames is prevented.

  FIG. 8 is a graph showing a comparison of correction coefficients when the determination unit 124 is not used and when it is used. FIG. 8A is a graph showing the change of the correction coefficient when the process in the determination unit 124, that is, the comparison process between the energy change rate and the minimum value change rate is not performed. The horizontal axis of the graph in FIG. 8A indicates the frame, and the vertical axis indicates the correction coefficient calculated by the correction coefficient calculation unit 18 (the minimum value of the coefficient I (m) is divided by the maximum value of the spectrum of white noise. Stuff). The correction coefficient “1” is −20 dB. The graph of FIG. 8A shows the change of the correction coefficient for each frame, and there are some places where the correction coefficient protrudes at several places.

  FIG. 8B is a graph showing the change in the correction coefficient when the process in the determination unit 124, that is, the comparison process between the energy change rate and the minimum value change rate is performed. Compared with the graph of FIG. 8A, it can be understood that the protruding correction coefficient is reduced. As shown in FIG. 8, in consideration of the energy change rate and the minimum value change rate between frames, the correction value with less variation is added to the coefficient I (m) by appropriately correcting the minimum value, and more natural. The original sound can be reproduced.

  FIG. 9 is a graph showing a comparison result with the prior art. In FIG. 9, the horizontal axis represents the encoding bit rate, and the unit is kbps (bits per second). The vertical axis represents ODG (Objective Difference Grade). A solid line indicates a change in ODG with respect to each encoding bit rate of the MP3 decoder to which the present technology is applied, and a dotted line indicates a change in ODG with respect to each encoding bit rate of the conventional MP3 decoder. In the evaluation test, a measurement method defined by CPX-2601 of JEITA (Japan Electronics and Information Technology Industries Association) was used, and a LAME core encoder was used.

  The evaluation standard is “ITU-R Rec. BS. 1387-1 (2001) Basic Version”, and the test source is the track of Cat No. 42204-2 (1988) EBU / Sound Quality Assessment Material recordings for subjective tests (SQAM). No.27 (castanets), 32 (triangles), 35 (iron koto), 40 (harpsichord), 65 (orchestra), 66 (xylophone ensemble), 69 (hops / ABBA), and 70 (hops / E.RABBITT) The average of 8 songs above was used.

  ODG is an objective evaluation value obtained as a final result of measurement, and takes a negative value from 0 to −4. The larger the value, the smaller the deterioration from the original sound. "0.0" means "I don't know the difference from the original sound", "-1.0" means "I don't care about the difference from the original sound", and "-2.0" "Do not", "-3.0" is defined as "get in the way" and "-4.0" is defined as "very disturbing", respectively. Comparing the solid line graph and the dotted line graph in FIG. 9, it can be understood that the ODG is improved over the entire region, and particularly good results are obtained in the range of 64 kbps to 160 kbps.

  The second embodiment is configured as described above, and the other configurations and operations are the same as those of the first embodiment. Therefore, corresponding parts are denoted by the same reference numerals, and detailed description thereof is omitted.

Embodiment 3
FIG. 10 is a block diagram showing the configuration of the signal processing device 20 according to the third embodiment. Each process of the signal processing device 20 according to the third embodiment may be realized by software executed on a personal computer. Hereinafter, the signal processing device 20 will be described as being a personal computer 20. The personal computer 20 is a publicly known computer, and a CPU (Central Processing Unit) 61 via a bus 67, a RAM (Random Access Memory) 62, a storage unit 65 such as a hard disk, an input unit 63, an output unit 64 such as a speaker, the Internet The communication unit 66 is connectable to a communication network such as the above.

  A computer program for operating the personal computer 20 can also be provided by a portable recording medium 1A such as a CD-ROM, MO, or DVD-ROM as in the third embodiment. Further, the computer program can be downloaded from a server computer (not shown) via the communication unit 66. The contents will be described below.

  A portable recording medium on which a computer program for causing a reader / writer (not shown) of the personal computer 20 shown in FIG. 10 to detect a minimum value, calculate a correction coefficient, calculate a correction value, and calculate a corrected coefficient is recorded. 1A (CD-ROM, MO, DVD-ROM or the like) is inserted, and this program is installed in the control program of the storage unit 65. Alternatively, such a program may be downloaded from an external server computer (not shown) via the communication unit 66 and installed in the storage unit 65. Such a program is loaded into the RAM 62 and executed. Thereby, it functions as the signal processing apparatus 20 of the present invention as described above.

  The third embodiment is configured as described above, and the other configurations and operations are the same as those of the first and second embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 4
The fourth embodiment relates to a form in which a plurality of white noise storage units 16 are provided. FIG. 11 is a block diagram illustrating a hardware configuration of the interpolation processing unit 1 according to the fourth embodiment. In addition to the configuration of the first or second embodiment, the interpolation processing unit 1 according to the fourth embodiment adds a block length determination unit 161, a selection unit 162, a first white noise storage unit 16L, and a second white noise storage. The unit 16S is configured. The block length determination unit 161 determines whether the block length of the input acoustic signal is a long block or a short block. The block length determination unit 161 refers to the blocksplit_flag [ch] [gr] flag in the frame side information area in the bit stream to determine whether each frame is a long block or a short block. One frame is composed of three sub-blocks of 18 samples × 32 blocks and 576 samples, and in the case of a short block, three sub-blocks of 6 samples × 32 blocks and 192 samples.

  The block length determination unit 161 outputs the determined block length to the selection unit 162 and the energy calculation unit 121. When calculating the average power spectrum as energy, the energy calculation unit 121 calculates the energy of the coefficient I (m) by dividing the square of the sum of the coefficients I (m) by the number of samples. In order to provide the number of samples, the block length determination unit 161 outputs the block length to the energy calculation unit 121. The energy calculation unit 121 appropriately sets the number of samples for division according to the block length.

  Based on the block length determined by the block length determination unit 161, the selection unit 162 selects either the first white noise storage unit 16L or the second white noise storage unit 16S according to the determination block length, and the selected Information is output to the output unit 14. In the fourth embodiment, two types of white noise storage units are provided. The first white noise storage unit 16L is the same as that shown in the first embodiment (see FIG. 4), and is selected when the selection unit 162 determines that the block length is a long block. On the other hand, the second white noise storage unit 16S is selected when the selection unit 162 determines that the block length is a short block. In the fourth embodiment, two first white noise storage units 16L and a second white noise storage unit 16S corresponding to the long block and the short block are provided for ease of explanation. When there are more than one type of different block lengths, three or more types of white noise storage units may be prepared according to the block length.

  FIG. 12 is an explanatory diagram showing a record layout of the second white noise storage unit 16S. The second white noise storage unit 16S includes a block field, a number m field, and a spectrum Iwn (m) field. The block field stores the number of blocks constituting white noise corresponding to the short block, and a total of 32 blocks from block 0 to block 31 are prepared. The number m field stores a number for identifying the white noise spectrum for each block.

  Further, the spectrum of white noise is stored in the spectrum Iwn (m) field in association with the number m. In the fourth embodiment, six spectra per block are stored in a total of 32 blocks, for a total of 192 spectra. For example, the white noise spectrum “0.007248” is stored in the number “1” of the block “0”. Similar to the long block, the white noise spectrum may be set to be about -20 dB when the entire spectrum is averaged. This Iwn (m) is a table for creating white noise at a level of about 1/2 of the minimum value of I (m), and need not be limited to a numerical value of about −20 dB.

  When the selection unit 162 selects the first white noise storage unit 16L, that is, when the block length is a long block, the output unit 14 corresponds to a block in which a spectrum with zero I (m) exists, as shown in FIG. The spectrum of a plurality of blocks selected randomly or regularly is output to the maximum spectrum detector 15 and the correction value calculator 13 from the first white noise storage unit 16L shown in FIG. On the other hand, when the selection unit 162 selects the second white noise storage unit 16S, that is, when the block length is a short block, the second white noise storage unit 16S illustrated in FIG. 12 randomly or regularly. The spectrums of the selected blocks are output to the maximum spectrum detector 15 and the correction value calculator 13.

  In the white noise of the plurality of blocks selected in this way, the maximum value of the spectrum belonging to the plurality of blocks is detected by the maximum spectrum detection unit 15, and the minimum value of the coefficient I (m) is calculated by the correction coefficient calculation unit 18. A correction coefficient is calculated by dividing by the maximum value of the spectrum of white noise, and the correction value output to the correction frequency coefficient calculation unit 17 is obtained by multiplying each spectrum of the block by the correction value calculation unit 13. Is calculated.

  As for the output of the plurality of blocks at the output unit 14, the output unit 14 randomly selects an arbitrary block from the first white noise storage unit 16L or the second white noise storage unit 16S by a random number generator (not shown), for example. Or you may make it select a block periodically so that all the blocks may be selected equally. For example, all the blocks “0” to “31” of the first white noise storage unit 16L or the second white noise storage unit 16S may be output so as to be selected once. In this way, by randomly or regularly selecting a block of white noise, the correction value added to the coefficient I (m) is flattened without being biased to specific white noise, and more natural interpolation processing is performed. Can be realized.

  FIG. 13 is a flowchart showing the procedure of the selection process of the first white noise storage unit 16L or the second white noise storage unit 16S. The block length determination unit 161 refers to the frame side information of the input bit stream and determines whether the block length is a long block or a short block (step S141). The block length determination unit 161 outputs the determined block length to the energy calculation unit 121 and the selection unit 162 (step S142). The selection unit 162 determines whether the block length determined by the block length determination unit 161 is a long block (step S143). If the selection unit 162 determines that the block is a long block (YES in step S143), the first white noise storage unit 16L is selected (step S144), and a signal indicating that the block is a long block is output to the output unit 14.

  On the other hand, if the selection unit 162 determines that the block is not a long block (NO in step S143), the second white noise storage unit 16S is selected (step S145), and a signal indicating a short block is output to the output unit 14. . When the selection unit 162 selects the first white noise storage unit 16L, that is, when the block length is a long block, the output unit 14 randomly from the first white noise storage unit 16L illustrated in FIG. A plurality of regularly selected blocks are selected (step S146). Similarly, when the selection unit 162 selects the second white noise storage unit 16S, that is, when the block length is a short block, the output unit 14 randomly selects from the second white noise storage unit 16S illustrated in FIG. Alternatively, a plurality of regularly selected blocks are selected (step S146). This block selection processing procedure will be described later.

  The output unit 14 reads the white noise of a plurality of blocks selected randomly or regularly (step S147). Then, the output unit 14 outputs the read white noise (spectrum) to the maximum spectrum detection unit 15 (step S148). The maximum spectrum detector 15 detects the maximum absolute value of the spectrum of the output white noise. Further, the output unit 14 outputs the white noise of the block selected randomly or regularly to the correction value calculation unit 13 (step S149). Subsequent calculation of the correction coefficient and calculation of the correction value have been described in detail in the first and second embodiments, and a description thereof will be omitted.

  FIG. 14 is a flowchart showing the procedure of block selection processing. A procedure in which the output unit 14 selects one block from the first white noise storage unit 16L or the second white noise storage unit 16S will be described. In the following, it is assumed that the block length is a long block, and a mode in which white noise related to one block is output from the first white noise storage unit 16L will be described. First, the output unit 14 selects the block 0 of the first white noise storage unit 16L shown in FIG. 4 (step S151). Then, the output unit 14 outputs 18 white noises belonging to the block 0 of the first white noise storage unit 16L to the maximum spectrum detection unit 15 and the correction value calculation unit 13.

  Subsequently, the output unit 14 selects the next block (step S152). For example, after selecting block 0, block 1 is selected.

  The output unit 14 reads the white noise in the next block from the first white noise storage unit 16 </ b> L and outputs it to the maximum spectrum detection unit 15 and the correction value calculation unit 13. The output unit 14 determines whether or not the block selected in step S152 is the last block (block 31 in the example of FIG. 4) (step S153). If the output unit 14 determines that the selected block is not the final block (NO in step S153), the output unit 14 proceeds to step S152 and further selects the next block.

  On the other hand, if the output unit 14 determines that the block selected in step S152 is the final block (YES in step S153), the output unit 14 proceeds to step S151 if it is determined that the block 31 is selected in the above example. The next block to be selected is block 0. In this way, the white noise block is regularly selected, and the white noise of all 576 samples in one frame is selected and output equally, so that it is flattened without biasing to specific white noise, and more natural interpolation is performed. Processing can be realized. When randomly selecting a block, a block corresponding to a numerical value generated by a random number generator (not shown) may be selected.

  The fourth embodiment is configured as described above, and the other configurations and operations are the same as those of the first to third embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 5
The processing according to the fourth embodiment may be realized as software processing using the personal computer shown in FIG. FIG. 15 is a block diagram showing a configuration of the signal processing device 20 according to the fifth embodiment. A computer program for operating the personal computer 20 as a signal processing device can be provided by a portable recording medium 1A such as a CD-ROM, MO, or DVD-ROM as in the fifth embodiment. . Further, the computer program can be downloaded from a server computer (not shown) via the communication unit 66. The contents will be described below.

  A portable computer recorded with a computer program for causing a reader / writer (not shown) of the personal computer 20 shown in FIG. 15 to determine a block length, select a white noise storage unit, output white noise, and calculate a corrected coefficient. The recording medium 1A (CD-ROM, MO, DVD-ROM or the like) is inserted, and this program is installed in the control program of the storage unit 65. Alternatively, such a program may be downloaded from an external server computer (not shown) via the communication unit 66 and installed in the storage unit 65. Such a program is loaded into the RAM 62 and executed. Thereby, it functions as the signal processing apparatus 20 of the present invention as described above.

  The fifth embodiment has the above-described configuration, and the other configurations and operations are the same as those of the first to fourth embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 6
The sixth embodiment relates to a mode for determining whether or not to correct a coefficient according to the pure tone of an acoustic signal. FIG. 16 is a block diagram showing a hardware configuration of decoding apparatus 20 according to the sixth embodiment. As shown in FIG. 16, an index value calculation unit 27 and a pure tone determination unit 28 are newly provided. The unpacking unit 22 extracts the scale factor of each frequency band from the frame side information of the bit stream.

  The index value calculation unit 27 calculates a pure tone index value indicating the degree of pure tone by subtracting the average value from the maximum value of the scale factor of each frequency band. The calculated pure tone index value is output to the pure tone determination unit 28. A reference value is stored in a memory (not shown) inside the pure tone determination unit 28, and the pure tone determination unit 28 compares the input pure tone index value with the reference value to determine whether the tone is pure. For example, when the maximum value of the scale factor is 120 dB, the reference value may be 70 dB.

  When the pure tone index value is smaller than the reference value, the pure tone determination unit 28 determines that the pure tone is low, and outputs the coefficient I (m) of the entire frequency band to the interpolation processing unit 1 to perform the above-described interpolation processing. Is called. On the other hand, the pure tone determination unit 28 determines that the pure tone property is high when the pure tone index value is larger than the reference value, and directly outputs the frequency frequency coefficient I (m) to the interpolation processing unit 1 without outputting the coefficient I (m). The data is output to the conversion unit 24. As described above, by performing or not performing the interpolation process according to the characteristics of the acoustic signal, it is possible to perform an appropriate interpolation process, and it is possible to increase the processing speed and reduce the power consumption.

  FIG. 17 is a flowchart showing the procedure of pure tone determination processing. First, the unpacking unit 22 extracts the scale factor of each frequency band from the frame side information of the bit stream (step S171). The extracted scale factor is output to the index value calculation unit 27. The index value calculation unit 27 extracts the maximum value from the scale factor of each frequency band (step S172). The index value calculation unit 27 calculates the average value of scale factors (step S173). The index value calculation unit 27 calculates the pure tone index value by subtracting the average value from the maximum value of the scale factor (step S174). The index value calculation unit 27 outputs the calculated pure tone index value to the pure tone determination unit 28 (step S175).

  The pure tone determination unit 28 reads a reference value from an internal memory (not shown) (step S176). Then, the pure tone determination unit 28 compares the input pure tone index value with the reference value, and determines whether or not the pure tone index value is smaller than the read reference value (step S177). If the pure tone determination unit 28 determines that the pure tone index value is smaller than the reference value (YES in step S177), the pure tone determination unit 28 determines that the pure tone property is low, and sends the coefficient I (m) of the entire frequency band to the interpolation processing unit 1. Output (step S178).

  On the other hand, when the pure tone determination unit 28 determines that the pure tone index value is larger than the reference value (NO in step S177), the pure tone determination unit 28 determines that the pure tone property is high, and calculates the coefficient I (m) of the entire frequency band as an interpolation processing unit. The data is directly output to the frequency time conversion unit 24 without being output to 1 (step S179). In the sixth embodiment, whether or not the tone is pure is determined based on the scale factor. However, whether or not the tone is pure may be determined based on the power of each frequency band. In this case, the index value calculation unit 27 subtracts the average value from the maximum power value of the coefficient I (m) in each frequency band, and outputs this to the pure tone determination unit 28 as a pure tone index value. For example, 40 dB is stored as a reference value in the pure tone determination unit 28. When the pure tone index value is smaller than the reference value, the pure tone determination unit 28 determines that the pure tone property is low, and outputs the coefficient I (m) of the entire frequency band to the interpolation processing unit 1. On the other hand, when the pure tone index value is larger than the reference value, it is determined that the pure tone is high, and the coefficient I (m) of all frequency bands is output to the frequency time conversion unit 24 without going through the interpolation processing unit 1. . Note that the technique disclosed in Japanese Patent Application Laid-Open No. 2002-351500 or Japanese Patent Application Laid-Open No. 2005-195983 may be applied to the determination of the pure tone.

  The sixth embodiment has the above-described configuration, and the other configurations and operations are the same as those of the first to fifth embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 7
The processing according to the sixth embodiment may be realized as software processing using the personal computer shown in FIG. FIG. 18 is a block diagram showing the configuration of the signal processing device 20 according to the seventh embodiment. A computer program for operating the personal computer 20 as a signal processing device can be provided by a portable recording medium 1A such as a CD-ROM, MO, or DVD-ROM as in the seventh embodiment. . Further, the computer program can be downloaded from a server computer (not shown) via the communication unit 66. The contents will be described below.

  A reader / writer (not shown) of the personal computer 20 shown in FIG. 18 calculates a pure tone index value, determines whether or not the tone is pure, outputs white noise, and corrects the coefficient according to whether or not the tone is pure. A portable recording medium 1A (CD-ROM, MO, DVD-ROM, or the like) on which a computer program for calculating is recorded is inserted, and this program is installed in the control program of the storage unit 65. Alternatively, such a program may be downloaded from an external server computer (not shown) via the communication unit 66 and installed in the storage unit 65. Such a program is loaded into the RAM 62 and executed. Thereby, it functions as the signal processing apparatus 20 of the present invention as described above.

  The seventh embodiment has the above-described configuration, and the other configurations and operations are the same as those of the first to sixth embodiments. Therefore, corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 8
The eighth embodiment relates to a mode for determining whether or not to execute an interpolation process according to a bit rate. FIG. 19 is a block diagram showing a hardware configuration of decoding apparatus 20 according to the eighth embodiment. As shown in FIG. 19, a bit rate acquisition unit 210, a sampling frequency acquisition unit 211, a bit rate comparison unit 212, and a table 213 are newly provided. The bit rate acquisition unit 210 acquires the bit rate of the acoustic signal from the bit rate index described in the header attached to the acoustic signal. The acquired bit rate is output to the bit rate comparison unit 212 via the sampling frequency acquisition unit 211.

  The sampling frequency acquisition unit 211 acquires the sampling frequency described in the header attached to the acoustic signal. As the sampling frequency, any of 32 kHz, 44.1 kHz, and 48 kHz is acquired in the MP3 system. The sampling frequency acquisition unit 211 outputs the acquired sampling frequency to the bit rate comparison unit 212.

  FIG. 20 is an explanatory diagram showing a record layout of the table 213. The table 213 stores a reference bit rate serving as a reference for each sampling frequency. The table 213 stores bit rates in association with the sampling frequencies of 32 kHz, 44.1 kHz, and 48 kHz. In the case of 32 kHz, the reference bit rate is stored as 160 kbps, and in the case of a bit rate smaller than 160 kbps surrounded by hatching in FIG. 20, the above-described determination of pure tone and interpolation processing are performed.

  In the case of 44.1 kHz, the reference bit rate is stored as 192 kbps, and in the case of a bit rate smaller than 192 kbps surrounded by the hatching in FIG. 20, the above-described determination of pure tone and interpolation processing are performed. Further, in the case of 48 kHz, the reference bit rate is stored as 224 kbps, and in the case of a bit rate smaller than 224 kbps surrounded by hatching in FIG. 20, the above-described determination of pure tone and interpolation processing are performed. In the case of the ATRAC3 standard mini-disc, the sampling frequency is only 44.1 kHz. In this case, the reference bit rate is 292 kbps. Are determined and interpolated.

  The bit rate comparison unit 212 reads the reference bit rate from the table 213 based on the sampling frequency output from the sampling frequency acquisition unit 211. The bit rate comparison unit 212 determines whether the bit rate output from the bit rate acquisition unit 210 is smaller than the reference bit rate. When the bit rate comparison unit 212 determines that the bit rate output from the bit rate acquisition unit 210 is smaller than the reference bit rate, the bit rate comparison unit 212 outputs the coefficient I (m) of the entire frequency band to the interpolation processing unit 1. For example, when the acquired sampling frequency is 32 kHz and the acquired bit rate is 32, 64, 96, or 128 kbps, the coefficient I (m) of the entire frequency band is a target of interpolation processing.

  On the other hand, when it is determined that the bit rate output from the bit rate acquisition unit 210 is not smaller than the reference bit rate, the frequency time conversion unit converts the coefficient I (m) of each frequency band without passing through the interpolation processing unit 1. Output directly to 24. For example, when the acquired sampling frequency is 32 kHz and the acquired bit rate is 160, 192, 224, 256, 288, 320, 352, 384, 416 or 448 kbps, the coefficient I (m) of each frequency band is calculated by the interpolation process. Not eligible. As described above, since the interpolation process is configured to be executed or not executed according to the sampling frequency and the bit rate, it is possible to perform an optimal interpolation process suitable for the state of the acoustic signal, and to increase the processing speed or power consumption. Reduction can be achieved.

  FIG. 21 is a flowchart showing the comparison process. The bit rate acquisition unit 210 acquires the bit rate of the acoustic signal from the bit rate index described in the header attached to the acoustic signal (step S211). The bit rate acquisition unit 210 outputs the acquired bit rate to the bit rate comparison unit 212 via the sampling frequency acquisition unit 211 (step S212). The sampling frequency acquisition unit 211 acquires the sampling frequency described in the header attached to the acoustic signal (step S213). The sampling frequency acquisition unit 211 outputs the acquired sampling frequency to the bit rate comparison unit 212 (step S214).

  The bit rate comparison unit 212 reads the reference bit rate corresponding to the sampling frequency output from the sampling frequency acquisition unit 211 from the table 213 (step S215). Then, the bit rate comparison unit 212 determines whether or not the bit rate acquired by the bit rate acquisition unit 210 is smaller than the read reference bit rate (step S216). When the bit rate acquisition unit 210 determines that the acquired bit rate is smaller than the reference bit rate (YES in step S216), the bit rate acquisition unit 210 outputs the coefficient I (m) of the entire frequency band to the interpolation processing unit 1 (step S217). .

  On the other hand, when it is determined that the acquired bit rate is not smaller than the reference bit rate (NO in step S216), the frequency time conversion unit 24 converts the coefficient I (m) of the entire frequency band without passing through the interpolation processing unit 1. (Step S218).

  The eighth embodiment is configured as described above, and the other configurations and operations are the same as those of the first to seventh embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 9
The processing according to the eighth embodiment may be realized as software processing using the personal computer shown in FIG. FIG. 22 is a block diagram showing a configuration of the signal processing device 20 according to the ninth embodiment. A computer program for operating the personal computer 20 as a signal processing device can be provided by a portable recording medium 1A such as a CD-ROM, MO, or DVD-ROM as in the ninth embodiment. . Further, the computer program can be downloaded from a server computer (not shown) via the communication unit 66. The contents will be described below.

  A portable recording medium 1A on which a computer program for causing a reader / writer (not shown) of the personal computer 20 shown in FIG. 22 to compare bit rates, output white noise, and calculate a corrected coefficient according to the bit rate is recorded. (CD-ROM, MO, DVD-ROM or the like) is inserted and this program is installed in the control program of the storage unit 65. Alternatively, such a program may be downloaded from an external server computer (not shown) via the communication unit 66 and installed in the storage unit 65. Such a program is loaded into the RAM 62 and executed. Thereby, it functions as the signal processing apparatus 20 of the present invention as described above.

  The ninth embodiment is configured as described above, and the other configurations and operations are the same as those of the first to eighth embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 10
In the tenth embodiment, it is determined whether or not a zero orthogonal transform coefficient is present in the frequency band of the encoded audio signal or the encoded video signal, and if present, the orthogonal transform related to the white noise source is present in the frequency band. The present invention relates to a form of adding a coefficient. FIG. 23 is a block diagram illustrating a hardware configuration of the interpolation processing unit 1 according to the tenth embodiment. The decoding device 20 including the interpolation processing unit 1 in the figure performs inverse quantization on the encoded audio signal or encoded video signal obtained by encoding the audio signal or video signal orthogonally transformed for each frame unit for a predetermined time and processes it. In the present embodiment, a mode for processing an encoded audio signal will be described, and an example in which MDCT is used as a method of orthogonal transform will be described. However, the present invention is not limited to this, and an encoded video signal is used and orthogonal transform is performed. DCT or FFT (Fast Fourier Transform) may be used as the method.

  The interpolation processing unit 1 according to the tenth embodiment includes a coefficient determination unit 30, an addition unit 31, a correction coefficient calculation unit 18, a correction value calculation unit 13, an output unit 14, and a white noise storage unit 16. Since the correction coefficient calculation unit 18, the correction value calculation unit 13, the output unit 14, and the white noise storage unit 16 are the same as those described in the first to ninth embodiments, detailed description thereof is omitted. The inversely quantized coefficient I (m) is input to the coefficient determining unit 30. The coefficient determination unit 30 determines whether or not there is a coefficient I (m) having a zero spectrum (orthogonal transform coefficient) for each frequency band.

  When the coefficient determination unit 30 determines that the spectrum zero coefficient I (m) exists in the frequency band, the coefficient determination unit 30 sets the flag in the frequency band and outputs the coefficient I (m) to the addition unit 31. On the other hand, when the coefficient determination unit 30 determines that the spectrum zero coefficient I (m) does not exist in the frequency band, the coefficient I (m) is sent to the addition unit 31 without setting a flag in the frequency band. Output. The white noise storage unit 16 as the white noise source divides the white noise of the time corresponding to the above-described frame into a plurality of bands (blocks), and performs orthogonal transform coefficients (white noise spectrum) obtained by orthogonal transform for each band. It is configured. In the present embodiment, an example in which white noise according to the long block shown in FIG.

  The output unit 14 reads the white noise spectrum (orthogonal transform coefficient) associated with a predetermined band, that is, a randomly or regularly selected block, from the white noise storage unit 16 as described in the fourth embodiment. The spectrum of white noise is output to the correction value calculation unit 13. On the other hand, from the correction coefficient calculation unit 18 described in the first and second embodiments, the calculated correction coefficient is output to the correction value calculation unit 13. The correction value calculation unit 13 multiplies the output correction coefficient by the white noise spectrum related to the block output from the output unit 14 to calculate a correction value. The correction value calculation unit 13 outputs the calculated correction value to the addition unit 31. The adding unit 31 outputs the coefficient I (m) related to the frequency band for which the flag is set out of the coefficient I (m) related to the frequency band output from the coefficient determining unit 30 to the coefficient I (m) output from the correction value calculating unit 13. A correction value, that is, a value related to an orthogonal transform coefficient related to a predetermined band read from the white noise source is added. The white noise spectrum may be added only to the coefficient I (m) having a spectrum of zero. In the present embodiment, the correction value calculation unit 13 performs an operation of multiplying the correction value calculated by the correction coefficient calculation unit 18 by the white noise spectrum of the predetermined band stored in the white noise storage unit 16. However, the white noise spectrum of the predetermined band may be output to the adding unit 31 as it is without performing this multiplication processing.

  FIG. 24 is a flowchart showing a procedure of white noise addition processing. The coefficient determination unit 30 determines whether there is a coefficient of spectrum 0 in the frequency band (step S251). If it is determined that there is at least one spectrum 0 coefficient (YES in step S251), the coefficient determination unit 30 sets a flag in the frequency band (step S252). In step S251, the flag is set when at least one coefficient of spectrum 0 exists. However, the flag may be set when a predetermined number of coefficients of spectrum 0 exist. On the other hand, when the coefficient determination unit 30 determines that there is no spectrum 0 coefficient in the frequency band (NO in step S251), the process of step S252 is skipped. Then, the coefficient determining unit 30 determines whether or not the process of step S251 has been performed for all frequency bands (step S253).

  If the coefficient determination unit 30 determines that processing for all frequency bands has not been performed (NO in step S253), the process proceeds to step S251, and the above processing is repeated. On the other hand, when the coefficient determination unit 30 determines that the processing has been completed for all frequency bands (YES in step S253), the coefficient determination unit 30 outputs information indicating the coefficient of each frequency band and the presence / absence of flag setting to the addition unit 31. As described in the fourth embodiment, the output unit 14 reads the spectrum of the white noise related to the block selected randomly or regularly as described in the fourth embodiment, and the correction value calculation unit 13 reads the spectrum of the read white noise. Output to. From the correction coefficient calculation unit 18, the calculated correction coefficient is output to the correction value calculation unit 13. The correction value calculation unit 13 multiplies the output correction coefficient by the white noise spectrum related to the block output from the output unit 14 to calculate a correction value. The correction value calculation unit 13 outputs the calculated correction value to the addition unit 31. The adding unit 31 adds the correction value read from the white noise storage unit 16 to the frequency band for which the flag is set (step S254). Similarly, the adding unit 31 adds a correction value based on the spectrum of white noise related to a block selected randomly or regularly to the coefficient of another frequency band in which the flag is set. The addition unit 31 outputs the added frequency band coefficient to the frequency time conversion unit 24, while adding the correction value based on the spectrum of white noise to the coefficient related to the frequency band for which the flag is not set. It outputs to the frequency time conversion part 24.

  The tenth embodiment has the above-described configuration, and the other configurations and operations are the same as those of the first to ninth embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

Embodiment 11
The processing according to the tenth embodiment may be realized as software processing using the personal computer shown in FIG. FIG. 25 is a block diagram showing a configuration of the signal processing device 20 according to the eleventh embodiment. A computer program for operating the personal computer 20 as a signal processing device can be provided on a portable recording medium 1A such as a CD-ROM, MO, or DVD-ROM as in the eleventh embodiment. . Further, the computer program can be downloaded from a server computer (not shown) via the communication unit 66. The contents will be described below.

  A portable recording medium 1A (CD-ROM, MO, DVD-ROM or the like) on which a computer program for causing a reader / writer (not shown) of the personal computer 20 shown in FIG. And insert this program into the control program of the storage unit 65. Alternatively, such a program may be downloaded from an external server computer (not shown) via the communication unit 66 and installed in the storage unit 65. Such a program is loaded into the RAM 62 and executed. Thereby, it functions as the signal processing apparatus 20 of the present invention as described above.

  The eleventh embodiment is configured as described above, and the other configurations and operations are the same as those of the first to tenth embodiments. Accordingly, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted. To do.

It is a block diagram which shows the hardware constitutions of the decoding apparatus which is a signal processing apparatus. It is a graph which shows the change of the IMDCT coefficient with respect to a frequency. It is a block diagram which shows the hardware constitutions of an interpolation process part. It is explanatory drawing which shows the record layout of a white noise storage part. It is a flowchart which shows the calculation procedure of the coefficient after correction | amendment. 6 is a block diagram illustrating a hardware configuration of an interpolation processing unit according to Embodiment 2. FIG. It is a flowchart which shows the calculation process procedure of the minimum value after correction | amendment. It is a graph which shows the comparison of the correction coefficient with the case where a judgment part is not used, and the case where it uses. It is a graph which shows the comparison result with a prior art. FIG. 10 is a block diagram illustrating a configuration of a signal processing device according to a third embodiment. FIG. 10 is a block diagram illustrating a hardware configuration of an interpolation processing unit according to a fourth embodiment. It is explanatory drawing which shows the record layout of a 2nd white noise storage part. It is a flowchart which shows the procedure of the selection process of a 1st white noise storage part or a 2nd white noise storage part. It is a flowchart which shows the procedure of the selection process of a block. FIG. 10 is a block diagram illustrating a configuration of a signal processing device according to a fifth embodiment. FIG. 20 is a block diagram showing a hardware configuration of a decoding apparatus according to a sixth embodiment. It is a flowchart which shows the procedure of a pure tone determination process. FIG. 10 is a block diagram illustrating a configuration of a signal processing device according to a seventh embodiment. FIG. 20 is a block diagram illustrating a hardware configuration of a decoding device according to an eighth embodiment. It is explanatory drawing which shows the record layout of a table. It is a flowchart which shows the procedure of a comparison process. FIG. 10 is a block diagram illustrating a configuration of a signal processing device according to a ninth embodiment. FIG. 22 is a block diagram illustrating a hardware configuration of an interpolation processing unit according to the tenth embodiment. It is a flowchart which shows the procedure of the white noise addition process. FIG. 20 is a block diagram showing a configuration of a signal processing device according to Embodiment 11. It is a block diagram which shows the hardware constitutions of the conventional decoder.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Interpolation processing part 12 Minimum value detection part 13 Correction value calculation part 14 Output part 15 Maximum spectrum detection part 16 White noise storage part 17 Correction frequency coefficient calculation part 18 Correction coefficient calculation part 20 Decoding apparatus (signal processing apparatus, personal computer)
DESCRIPTION OF SYMBOLS 21 Acoustic signal input part 22 Unpacking part 23 Inverse quantization part 24 Frequency time conversion part 25 Frequency band synthetic | combination part 26 Acoustic signal output part 27 Index value calculation part 28 Pure tone determination part 30 Coefficient judgment part 31 Addition part 121 Energy calculation part 122 previous frame energy storage unit 123 energy change rate calculation unit 124 determination unit 125 minimum value change rate calculation unit 126 minimum value correction unit 127 corrected minimum value setting unit 128 minimum value setting unit 129 previous frame minimum value storage unit 161 block length determination unit 162 selection unit 210 bit rate acquisition unit 211 sampling frequency acquisition unit 212 bit rate comparison unit 213 table 1A portable recording medium 16L first white noise storage unit 16S second white noise storage unit

Claims (11)

In a signal processing method for processing an acoustic signal obtained by dequantizing an encoded acoustic signal,
An index value calculating step for calculating a pure tone index value indicating a degree of pure tone of the dequantized acoustic signal;
A pure tone determination step of determining whether the acoustic signal is a pure tone by comparing the pure tone index value calculated by the index value calculation step with a reference value stored in advance;
An output step of outputting white noise from a white noise storage unit storing white noise;
Correction frequency coefficient calculation for calculating a corrected coefficient by adding a value related to white noise output by the output step to each coefficient of the acoustic signal when the sound signal is not determined to be pure sound by the pure tone determination step Steps ,
A level detection step for detecting the level of the coefficient of the dequantized acoustic signal;
A correction coefficient calculation step of calculating a correction coefficient based on the white noise output from the white noise storage unit and the level;
A correction value calculating step of calculating a correction value by multiplying the white noise stored in the white noise storage unit by the correction coefficient,
The level detection step detects a minimum value excluding zero of the coefficient of the dequantized acoustic signal,
The correction frequency coefficient calculation step includes calculating a corrected coefficient by adding the correction value to a coefficient of an acoustic signal obtained by inverse quantization as a value related to white noise . In a signal processing method for processing an acoustic signal obtained by dequantizing an encoded acoustic signal,
A bit rate comparison step for determining whether the bit rate of the acoustic signal is smaller than a pre-stored reference bit rate;
An output step of outputting white noise from a white noise storage unit storing white noise;
When the bit rate comparison step determines that the bit rate of the acoustic signal is smaller than a reference bit rate, the coefficient after correction is performed by adding a value relating to the white noise output by the output step to each coefficient of the acoustic signal. a correction frequency coefficient computing step of calculating,
The bit rate comparison step reads a reference bit rate corresponding to a sampling frequency of the acoustic signal from a table storing a reference bit rate serving as a reference for each sampling frequency, and the bit rate of the acoustic signal is the read reference bit. Determine if it is less than the rate,
In the correction frequency coefficient calculation step, when it is determined by the bit rate comparison step that the bit rate of the acoustic signal is smaller than a reference bit rate, the white color output by the output step to the coefficient of each frequency band of the acoustic signal A signal processing method, wherein a coefficient after correction is calculated by adding a value related to noise . In a signal processing apparatus that processes an acoustic signal obtained by dequantizing an encoded acoustic signal,
An index value calculation unit that calculates a pure tone index value indicating a degree of pure tone of the dequantized acoustic signal;
A pure tone determination unit that determines whether the acoustic signal is a pure tone by comparing the pure tone index value calculated by the index value calculation unit with a reference value stored in advance;
An output unit for outputting white noise from a white noise storage unit storing white noise;
Corrected frequency coefficient calculation that calculates a corrected coefficient by adding a value related to white noise output from the output unit to each coefficient of the acoustic signal when the pure tone determination unit does not determine that the acoustic signal is pure tone and parts,
A level detector for detecting the level of the coefficient of the dequantized acoustic signal;
A correction coefficient calculation unit that calculates a correction coefficient based on the white noise output from the white noise storage unit and the level; and
A correction value calculation unit that calculates a correction value by multiplying the white noise stored in the white noise storage unit by the correction coefficient,
The level detection unit is configured to detect a minimum value excluding zero of the coefficient of the dequantized acoustic signal,
The signal processing apparatus, wherein the correction frequency coefficient calculation unit is configured to add the correction value to a coefficient of an acoustic signal that has been dequantized as a value related to white noise to calculate a corrected coefficient .   4. The signal processing according to claim 3, wherein the index value calculation unit is configured to use a value obtained by subtracting an average value from a maximum value related to a scale factor of each frequency band as a pure tone index value. apparatus. In a signal processing apparatus that processes an acoustic signal obtained by dequantizing an encoded acoustic signal,
A bit rate comparison unit for determining whether the bit rate of the acoustic signal is smaller than a reference bit rate stored in advance;
An output unit for outputting white noise from a white noise storage unit storing white noise;
When the bit rate comparison unit determines that the bit rate of the acoustic signal is smaller than the reference bit rate, the coefficient after correction by adding a value related to white noise output from the output unit to each coefficient of the acoustic signal a correction frequency coefficient calculation unit for calculating a,
A table storing a reference bit rate that is a reference for each sampling frequency, and
The bit rate comparison unit is configured to read a reference bit rate corresponding to a sampling frequency of an acoustic signal from the table and determine whether the bit rate of the acoustic signal is smaller than the read reference bit rate. And
When the bit rate comparison unit determines that the bit rate of the acoustic signal is smaller than a reference bit rate by the bit rate comparison unit, the corrected frequency coefficient calculation unit outputs the white color output from the output unit to the coefficient of each frequency band of the acoustic signal A signal processing apparatus configured to calculate a corrected coefficient by adding a value related to noise . A level detector for detecting the level of the coefficient of the dequantized acoustic signal;
A correction coefficient calculation unit that calculates a correction coefficient based on the white noise output from the white noise storage unit and the level; and
A correction value calculation unit that calculates a correction value by multiplying the white noise stored in the white noise storage unit by the correction coefficient;
It said correction frequency coefficient computing unit, to claim 5, characterized in that the correction value are configured to calculate a corrected coefficient by adding the coefficients of the acoustic signal dequantized as a value relating to white noise The signal processing apparatus as described. The signal processing apparatus according to claim 6 , wherein the level detection unit is configured to detect a minimum value excluding zero of a coefficient of the dequantized acoustic signal. The correction coefficient calculation unit is configured to calculate a correction coefficient by dividing a level detected by the level detection unit by a level of white noise stored in the white noise storage unit. Item 7. The signal processing device according to Item 3 or 6 . The signal processing apparatus according to claim 8 , wherein the level of the white noise is a maximum value or an average value of the white noise stored in the white noise storage unit. In a computer program for processing an acoustic signal obtained by dequantizing an encoded acoustic signal by a computer,
On the computer,
An index value calculating step for calculating a pure tone index value indicating a degree of pure tone of the dequantized acoustic signal;
A pure tone determination step of determining whether the acoustic signal is a pure tone by comparing the pure tone index value calculated by the index value calculation step with a reference value stored in advance;
An output step of outputting white noise from a white noise storage unit storing white noise;
Correction frequency coefficient calculation for calculating a corrected coefficient by adding a value related to white noise output by the output step to each coefficient of the acoustic signal when the sound signal is not determined to be pure sound by the pure tone determination step Steps ,
A level detection step for detecting the level of the coefficient of the dequantized acoustic signal;
A correction coefficient calculation step of calculating a correction coefficient based on the white noise output from the white noise storage unit and the level;
A correction value calculation step of calculating a correction value by multiplying the white noise stored in the white noise storage unit by the correction coefficient;
The level detection step detects a minimum value excluding zero of the coefficient of the dequantized acoustic signal,
The corrected frequency coefficient calculating step calculates the corrected coefficient by adding the correction value to the coefficient of the acoustic signal obtained by inverse quantization as a value related to white noise.
A computer program characterized by the above . In a computer program for processing an acoustic signal obtained by dequantizing an encoded acoustic signal by a computer,
On the computer,
A bit rate comparison step for determining whether the bit rate of the acoustic signal is smaller than a pre-stored reference bit rate;
An output step of outputting white noise from a white noise storage unit storing white noise;
When the bit rate comparison step determines that the bit rate of the acoustic signal is smaller than a reference bit rate, the coefficient after correction is performed by adding a value relating to the white noise output by the output step to each coefficient of the acoustic signal. to execute the correction frequency coefficient computing step of calculating,
The bit rate comparison step reads a reference bit rate corresponding to a sampling frequency of the acoustic signal from a table storing a reference bit rate serving as a reference for each sampling frequency, and the bit rate of the acoustic signal is the read reference bit. Determine if it is less than the rate,
In the correction frequency coefficient calculation step, when it is determined by the bit rate comparison step that the bit rate of the acoustic signal is smaller than a reference bit rate, the white color output by the output step to the coefficient of each frequency band of the acoustic signal Calculate the corrected coefficient by adding noise values
A computer program characterized by the above .

Images