JP2002372993A - Audio band extending device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To expand the reproducing band of digital audio signals that are compressed and encoded so as to improve the reproducing tone quality to be excellent from the viewpoint of audibility. SOLUTION: Decoded signals decoded by a decoder 1 are provided to an oversampling LPF 2 to conduct a band limiting and a oversampling. Moreover, higher harmonics having a band that is equal to or more than that of the decoded signals are generated using a band-pass filter 3, a rectifying circuit 4 and a sliding band HPF 5. Then, a level control circuit 6 conducts level control of the output of the HPF 5 based on the spectrum information contained in the compressed and encoded signals and an adding circuit 7 adds the output of the HPF 5 to the decoded signals and output the result. Higher harmonics having a spectrum that is equal to or more than the band of the compressed and encoded signals are generated and the generated higher harmonics components are added to input signals in accordance with the higher frequency spectrum strength of the input signals.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio band extending apparatus capable of improving the reproduction sound of an audio signal in an audio apparatus, particularly the reproduction sound quality in a high frequency range, and outputting a comfortable audio signal to a human ear. .

[0002]

2. Description of the Related Art A signal having a spectrum whose frequency exceeds the upper limit of the high frequency range of the reproduction frequency band or the upper limit of the high frequency range of the audible frequency band is created from the audio reproduction signal, and this signal is added to the original audio reproduction signal. There was an audio band extension method to do it.

[0003] A conventional audio band extending apparatus for realizing such an audio band extending method will be described below. FIG. 7 is a block diagram showing a configuration example of a conventional audio band extending apparatus. This audio band extending device includes an oversampling type low-pass filter (oversampling LPF) 71, a band-pass filter 7
2. It includes a rectifier circuit 73, a high-pass filter 74, a spectrum analysis circuit 75, a level control circuit 76, and an addition circuit 77.

As an input signal S701, for example, a digital signal reproduced from a compact disk (hereinafter, referred to as a CD) is input. In the case of a CD, this digital signal is a signal having a sampling frequency (Fs) of 44.1 kHz and a word length of 16 bits.

[0005] The oversampling LPF 71 upsamples the sampling frequency of the input signal S701 to twice, and cuts off an unnecessary band exceeding 22.5 kHz. The bandpass filter 72 limits the band of the output S702 of the oversampling LPF 71. Assuming that the sampling frequency of the input signal S701 is Fs,
From the band-pass filter 72, for example, Fs / 4 to Fs
A signal having a band of / 2 is output.

[0006] The rectifier circuit 73 generates a harmonic of the input signal S701 by rectifying the output S703 of the bandpass filter 72 to half-wave or both-wave. High pass filter 74
Cuts off the low-frequency component of the output S704 of the rectifier circuit 73 and outputs a signal having a spectrum equal to or higher than Fs / 2. The spectrum analysis circuit 75 includes an oversampling LPF 7
The spectrum intensity of the high-frequency component of the output S702 of No. 1 is detected. For example, a spectrum intensity of Fs / 4 to Fs / 2 is detected as a high frequency component.

[0007] The level control circuit 76 controls the level of the output signal S705 of the high-pass filter 74 according to the output S706 of the spectrum analysis circuit 75. When the spectrum intensity of the input signal S702 at Fs / 4 to Fs / 2 is large, the level control circuit 76
Is increased, and when the spectrum intensity is small, the output level of the high-pass filter 74 is decreased. The addition circuit 77 adds the output S702 of the oversampling LPF 71 and the output S707 of the level control circuit 76 and outputs an output signal S708.

As described above, by generating a harmonic having a spectrum higher than the band of the input signal and adding the harmonic component generated according to the high-band spectrum intensity of the input signal to the input signal, The audio band had been extended.

[0009] The audio band extending apparatus as described above is
There has been a demand to use not only a CD but also an apparatus using a compression encoding technique such as a mini disk (MD).

FIG. 8 is an explanatory diagram showing an example of the configuration (signal arrangement) of an encoded signal used in the MD. This figure shows one frame of the encoded signal. In one frame,
For example, audio data such as 512 samples is compressed and encoded. One frame includes system information S801,
Bit allocation information S802, scale information S803,
It consists of spectrum information S804. System information S80
1 indicates system information such as the number of divisions of the frequency band. Bit allocation information S802 indicates the number of bits allocated to each band when audio data is divided into frequency bands. The scale information S803 indicates the peak level of each band. The spectrum information S804 contains spectrum data quantized to the number of bits indicated by the bit allocation information S802 in a form normalized by the scale information S803 for each band.

FIG. 9 shows a block diagram of a decoder for decoding the above coded signal. The demultiplexer 91 receives the encoded signal S901, performs demultiplexing, and outputs bit allocation information S902, scale information S903, and spectrum information S904.

[0012] The inverse quantization unit 92 generates the bit allocation information S
According to the number of bits indicated by 902, spectrum information S904 of each frequency band is inversely quantized, and inversely quantized signal S905 is output. The inverse normalization unit 93 inversely normalizes the inverse quantization signal S905 according to the scale information S903, and outputs a spectrum signal S906.

The frequency synthesizing section 94 generates a spectrum signal S9.
06 by orthogonal transform filter processing, and outputs a decoded signal S907 which is a digital audio signal.

[0014]

However, the above-mentioned conventional configuration has the following problems. That is, when a signal obtained by decoding a compression-encoded signal, such as a reproduction output of an MD, is input as an input signal, a spectrum does not always exist up to 1/2 (Fs / 2) of the sampling frequency. In some cases, there is a valley between the spectrum and the signal added to extend the band.

The above problem will be specifically described below. FIG. 10A shows bit allocation information of an encoded signal in each frequency band, and FIG. 10B shows an example of scale information in each frequency band. It is assumed that, for example, up to 22.05 kHz is divided into 16 frequency bands. Since compression encoding uses the characteristics of human hearing, bit allocation at high frequencies, for example, above 16 kHz from the low frequency band, may be reduced or, in some cases, may not be performed. In the example of FIG. 10A, no allocation is made to the fourteenth to sixteenth bands. When such an encoded signal is decoded, a decoded signal having only signals in the first to thirteenth bands is output.

FIG. 11 is a spectrum diagram when a conventional audio band extending device is applied to such a compression-encoded decoded signal. As shown in this figure, there is no signal energy in the fourteenth to sixteenth bands, and a band-extended signal as indicated by the hatched area in the figure is added only to the band exceeding the sixteenth band. For this reason, a valley of the spectrum is generated, which results in an unnatural band expansion in terms of audibility.

Another problem is that the level control is performed after the spectrum analysis of the input signal, so that there is a problem that a time delay occurs in the level control. Also, when a signal having a single spectrum such as a sine wave is input, there is a problem that random noise is added and the signal is deteriorated.

The present invention has been made in view of such a conventional problem. Even when a signal obtained by reproducing a compression-encoded signal does not generate a valley of a spectrum in a high-frequency portion, the level control time is not increased. It is an object of the present invention to provide an audio band extending apparatus that does not cause signal degradation even when a concentrated spectrum signal such as a sine wave is input without delay.

[0019]

According to the first aspect of the present invention, a compression coded signal obtained by dividing an audio signal into a plurality of frequency bands in frame units is input, and at least bit allocation information for each band and What is claimed is: 1. An audio band extending apparatus for reproducing said compression-encoded signal using scale information for each band, wherein said decoded signal includes a decoded signal including said bit allocation information, said scale information and an audio signal. Decoding means, a non-linear means for distorting the audio signal output from the decoding means, and referring to the bit allocation information extracted by the decoding means, higher than the highest frequency indicated by the bit allocation information A frequency-variable high-pass filter for extracting the audio component from the output signal of the non-linear means, Level control means for controlling an output level of the frequency-variable high-pass filter means in accordance with the scale information extracted by the decoding means; and an output of the frequency-variable high-pass filter means controlled by the level control means And an output signal adding means for adding the decoded signal of the decoding means and outputting an audio signal whose band has been extended.

According to a second aspect of the present invention, a compression-encoded signal obtained by dividing an audio signal into a plurality of frequency bands in frame units is input, and at least bit allocation information for each band and scale information for each band are input. An audio band extending device that reproduces the compression-encoded signal using
Decoding means for decoding the compression-encoded signal and outputting a decoded signal including the bit allocation information, the scale information, and the audio signal; dither generating means for generating high-frequency noise; and With reference to the bit allocation information, a frequency-variable high-pass filter for extracting a noise component higher than the highest frequency indicated by the bit allocation information from the output signal of the dither generating means, and the scale information extracted by the decoding means. A level control means for controlling an output level of the frequency variable high-pass filter means,
Output signal adding means for adding an output of the frequency-variable high-pass filter means controlled by the level control means and a decoded signal of the decoding means, and outputting an audio signal whose band is extended. It is characterized by the following.

According to a third aspect of the present invention, a compression-encoded signal obtained by dividing an audio signal into a plurality of frequency bands in frame units is input, and at least bit allocation information for each band, scale information for each band, and An audio band extending device that reproduces the compression-encoded signal using
Decoding means for decoding the compression-encoded signal and outputting a decoded signal including the bit allocation information, the scale information, and the audio signal; non-linear means for distorting the audio signal output from the decoding means; , A high-frequency signal adding means for adding the output of the non-linear means and the output of the dither generating means, and referring to the bit allocation information extracted by the decoding means, A frequency-variable high-pass filter for extracting an audio component higher than the indicated highest frequency from the output signal of the high-frequency signal adding means, and the frequency-variable high-pass filter according to the scale information extracted by the decoding means. Level control means for controlling the output level of the filter means;
Output signal adding means for adding an output of the frequency-variable high-pass filter means controlled by the level control means and a decoded signal of the decoding means, and outputting an audio signal whose band is extended. It is characterized by the following.

According to a fourth aspect of the present invention, a compression coded signal obtained by dividing an audio signal into a plurality of frequency bands in frame units is input, and at least bit allocation information for each band and scale information for each band are input. An audio band extending device that reproduces the compression-encoded signal using
Decoding means for decoding the compression-encoded signal and outputting a decoded signal including the bit allocation information, the scale information, and the audio signal; non-linear means for distorting the audio signal output from the decoding means; , A high-frequency signal adding means for adding the output of the non-linear means and the output of the dither generating means, and referring to the bit allocation information extracted by the decoding means, A frequency-variable high-pass filter for extracting an audio component higher than the indicated highest frequency from the output signal of the high-frequency signal adding means, and the frequency-variable high-pass filter according to the scale information extracted by the decoding means. Level control means for controlling the output level of the filter means and outputting a band extension signal Referring to the scale information extracted by the decoding means, shuts down the band extension signal of the level control means in the case of a concentrated spectrum, and outputs the band extension signal of the level control means in the case of a plurality of spectra. And an output signal adding unit that adds an output of the switching unit and a decode signal of the decoding unit and outputs an audio signal whose band has been extended.

According to a fifth aspect of the present invention, in the audio band extending apparatus according to any one of the first to fourth aspects, the audio signal output from the decoding means is at least twice the sampling frequency at the time of compression processing. And an oversampling filter for removing aliasing components due to the oversampling is provided between the decoding means and the output signal adding means.

The invention of claim 6 of the present application is directed to claims 1, 3,
5. The audio band extending apparatus according to claim 4, wherein an operation word length of the non-linear means and the frequency variable high-pass filter is larger than a word length of a decoded signal of the decoding means.

The invention of claim 7 of the present application is the invention of claims 2, 3,
5. The audio band extending apparatus according to claim 4, wherein the operation word length of the dither generation means and the frequency variable high-pass filter is larger than the word length of the decoded signal of the decoding means. .

[0026] The invention of claim 8 of the present application provides claims 1, 3,
4. The audio band extending apparatus according to any one of 4, wherein the non-linear means and the frequency variable high-pass filter are:
The operation word length is made larger than the word length of the decode signal of the decoding means, and the word length LS of the decode signal is
The processing is performed by adding a signal component having substantially the same amplitude to the B width.

[0027] The invention of claim 9 of the present application is the invention of claims 2, 3,
5. The audio band extending apparatus according to claim 4, wherein the dither generation means and the frequency-variable high-pass filter add a signal component having substantially the same amplitude to the calculated word length to the LSB width of the word length of the decoded signal. It is characterized by processing.

The invention of claim 10 of the present application relates to claim 1,
3. The audio band extending apparatus according to claim 1, wherein the non-linear means is a half clipping means for extracting only one side of a signal having positive and negative amplitudes, or an absolute value converting means for returning the amplitude of one side to the opposite side. It is a feature.

The invention of claim 11 of the present application relates to claim 1,
5. The audio band extending apparatus according to claim 3, wherein the non-linear means distorts a signal limited to a predetermined frequency band from a decoded signal of the decoding means.

The twelfth invention of the present application is the second invention.
5. The audio band extending apparatus according to claim 3, wherein the dither generating means is a diamond dither generating means having a probability density of a triangular distribution within a predetermined amplitude.

According to a thirteenth aspect of the present invention, in the audio band extending apparatus according to the twelfth aspect, the diamond dither generating means includes an output of two independent PN sequence generating means PA and an output of the PN sequence generating means PB. Are added.

According to a fourteenth aspect of the present invention, in the audio band extending apparatus according to the twelfth aspect, the diamond dither generating means includes a partial element PPA and a partial element PP which are different from each other and do not overlap each other in one PN sequence generating means PP.
B is added.

The invention of claim 15 of the present application is the invention of claim 2
5. The audio band extending apparatus according to any one of items 3 and 4, wherein the dither generating means is a bell-shaped dither generating means having a bell-shaped distribution of probability density within a predetermined amplitude.

According to a sixteenth aspect of the present invention, in the audio band extending apparatus of the fifteenth aspect, the bell-shaped dither generating means includes an output of three independent PN sequence generating means PA and an output of the PN sequence generating means PB. , PN sequence generating means PC.

According to a seventeenth aspect of the present invention, in the audio band extending apparatus according to the fifteenth aspect, the bell-shaped dither generating means includes partial elements PPA and PPB which are different from each other and do not overlap each other in one PN sequence generating means PP. , And a partial element PPC are added.

The invention of claim 18 of the present application is directed to claim 2
3. The audio band extending apparatus according to claim 1, wherein the dither generating means has an output spectrum distribution of 1 /
It is characterized by f-characteristics.

The invention of claim 19 of the present application is the invention of claim 2
In the audio band extending apparatus according to any one of the items (3) and (4), when it is determined that only the concentrated spectrum exists according to the scale information extracted by the decoding means, When the level is set to substantially zero and it is determined that there are two or more spectra, the level is set so that the dither spectrum height is substantially continuous with the highest spectrum level indicated by the bit allocation information extracted by the decoding means. Is controlled.

The invention of claim 20 of the present application is directed to claim 2,
In the audio band extending apparatus according to any one of (3) and (4), when it is determined that two or more spectra exist according to the scale information extracted by the decoding means, It is characterized in that a frequency characteristic slope is obtained and a dither level is controlled so that a spectral distribution is smoothly continuous.

The invention of claim 21 of the present application is directed to claims 1 to 4
The audio band extending apparatus according to any one of the above, wherein the level control means mutes the output level to zero when the signal is determined to be zero according to the scale information extracted by the decoding means. Things.

The invention of claim 22 of the present application is the invention of claims 1-4
In the audio band extending apparatus according to any one of the above, the level control means smoothes the scale information taken out by the decoding means using a predetermined smoothing filter, and makes the fluctuation of the output level slow. Things.

[0041]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An audio band extending apparatus according to each embodiment of the present invention will be described below with reference to the drawings. (Embodiment 1) FIG. 1 is a block diagram showing a configuration of an audio band extending apparatus according to Embodiment 1 of the present invention. This audio band extending apparatus includes a decoder 1, an oversampling type low-pass filter (oversampling LPF) 2, a band-pass filter 3, a rectifying circuit 4, a sliding band type high-pass filter (sliding band HPF) 5, a level control circuit 6, and an adding circuit. 7.

The decoder 1 decodes the compression-encoded signal to generate bit allocation information S102 and scale information S10.
3 and a decoding means for outputting a decode signal S104 including an audio signal. The bandpass filter 3 and the rectifier circuit 4 output the oversampling LP from the decoder 1.
Non-linear means for distorting the audio signal output via F2.

The sliding band HPF 5 refers to the bit allocation information S102 extracted by the decoder 1 and extracts a frequency component higher than the highest frequency indicated by the bit allocation information from the output signal of the non-linear means. It is. Level control circuit 6
Is a level control unit that controls the output level of the sliding band HPF5 according to the scale information S103 extracted by the decoder 1.

The adding circuit 7 is an output signal adding means for adding the output of the sliding band HPF 5 controlled by the level control circuit 6 and the decoded signal of the oversampling LPF 2 to output an audio signal whose band has been extended.

The operation of the audio band extending apparatus having such a configuration will be described. As an input signal, the coded signal S101 described with reference to FIG. If this signal is reproduced from, for example, a mini disk (MD), the sampling frequency (Fs) is 4
This is a 4.1 kHz signal.

The decoder 1 performs a decoding process in the same manner as the decoding process described with reference to FIG. The decoder 1 demultiplexes the coded signal S101, extracts system information, bit allocation information, scale information, and spectrum information, outputs bit allocation information S102 and scale information S103, and dequantizes these signals. , Inverse normalization, and frequency synthesis to obtain the decoded signal S
104 is output.

The oversampling LPF 2 samples the sampling frequency of the input decoded signal S104 at a positive integer multiple, for example, twice, and attenuates an unnecessary band. In the case of twice, the bandwidth of half or more of the sampling frequency of the input signal, that is, 22.05 kHz
z or more is attenuated by, for example, 60 dB or more.

The bandpass filter 3 limits the band of the output S105 of the oversampling LPF2. Assuming that the sampling frequency of the input signal S104 is Fs, the bandpass filter 3 outputs a signal S106 having a band of, for example, Fs / 4 to Fs / 2. The rectifier circuit 4 rectifies the output S106 of the bandpass filter 3 by half-wave (half clipping means) or double-wave rectification (absolute value converting means).
Generates harmonics of the input signal.

The sliding band HPF 5 outputs the bit allocation information S 10 to the output S 107 of the rectifier circuit 4.
The frequency below the highest frequency where the assignment of 2 exists, that is, the frequency below the highest frequency where the assignment is not 0 is cut off, and a signal S108 having a frequency spectrum exceeding that frequency is output.

The level control circuit 6 has the scale information S10
3, the spectrum intensity of the high frequency component, for example, F
The signal level of s / 4 to Fs / 2 is obtained, and the output level of the sliding band HPF5 is controlled according to the level. The level control circuit 6 calculates Fs / 4 to Fs of the encoded signal.
When the signal level at s / 2 is high, the output level of sliding band HPF5 is increased, and when the signal level is low, the output level of sliding band HPF5 is reduced.

The adder circuit 7 includes an oversampling LP
The output S105 of the F2 and the output S10 of the level control circuit 6
9 is added, and an output signal S110 having an expanded band is output.

As described above, according to the present embodiment, the bandpass filter 3, the rectifier circuit 4, and the sliding band HPF 5 are used to generate harmonics higher than the spectrum band included in the decoded signal. Utilizing the level information, level control is performed in accordance with the high-band spectrum intensity of the decoded signal, and is added to the decoded signal and output. In this way, it is possible to realize an audio band extending apparatus that does not cause a valley in a spectrum even if a band is extended for a signal obtained by decoding a compression-encoded signal in which bits are not necessarily allocated to all bands.

According to the present embodiment, the decoder 1
Is a demultiplexing process, which uses the scale information and bit allocation information extracted as preprocessing for decoding the frame, and performs level control and sliding band HP.
By specifying the cutoff frequency of F, control can be performed at the same timing as the decode signal. For this reason, it is possible to realize an audio band extending device in which the control of the signal to be added for the band extension is delayed and the audibility becomes unnatural.

(Embodiment 2) Next, an audio band extending apparatus according to Embodiment 2 of the present invention will be described.
FIG. 2 is a block diagram showing the configuration of the audio band extending apparatus according to the present embodiment, and the same parts as those in FIG. This audio band extending device includes a decoder 1, an oversampling LPF 2, a sliding band HPF 5, a level control circuit 6, an adding circuit 7, and a dither generating circuit 8.

The dither generating circuit 8 is a dither generating means for generating white noise or high frequency noise in an audio band. The sliding band HPF 5 refers to the bit allocation information S102 extracted by the decoder 1, and extracts a white noise component higher than the highest frequency indicated by the bit allocation information from the output signal of the dither generation circuit 8 as a frequency-variable high-pass filter. It is.

The operation of the audio band extending apparatus having such a configuration will be described. As an input signal, the coded signal S101 described with reference to FIG. This signal is a signal having a sampling frequency Fs of 44.1 kHz if reproduced from, for example, an MD.

The decoder 1 performs the decoding process in the same manner as the decoding process described with reference to FIG. The decoder 1
The coded signal S101 is demultiplexed to extract system information, bit allocation information, scale information, and spectrum information, respectively, and output bit allocation information S102 and scale information S103. , Frequency synthesis, and decode signal S
104 is output.

The oversampling LPF 2 performs sampling by increasing the sampling frequency of the input decoded signal S104 by a positive integer, for example, twice, and attenuates unnecessary bands. In the case of twice, the band of half or more of the sampling frequency of the input signal, that is, 22.05k
Hz or more is attenuated, for example, by 60 dB or more.

The dither generating circuit 8 generates white noise whose spectrum intensity is uniformly distributed in the band from 0 to Fs. The sliding band HPF 5 outputs the bit allocation information S 10 to the output S 111 of the dither generation circuit 8.
The frequency at which the assignment of 2 exists, that is, the frequency below the highest non-zero assignment frequency is cut off, and a signal S108 having a frequency spectrum exceeding that frequency is output.

The level control circuit 6 has the scale information S10
3, the spectrum intensity of the high frequency component, for example, F
The signal level of s / 4 to Fs / 2 is obtained, and the output level of the sliding band HPF5 is controlled according to the level. The level control circuit 6 calculates Fs / 4 to Fs of the encoded signal.
When the signal level at s / 2 is high, the output level of the high-pass filter 5 is increased, and when the signal level is low, the output level is reduced. The addition circuit 7 adds the output S105 of the oversampling LPF2 and the output S109 of the level control circuit 6, and outputs this addition signal as an output signal S110.

As described above, according to the present embodiment, the dither generating circuit 8 and the sliding band HPF 5 are used to generate harmonics equal to or higher than the spectral band included in the decoded signal. Is used to perform level control according to the high-frequency spectrum intensity of the decoded signal, add the decoded signal to the decoded signal, and output the resultant signal. In this way, it is possible to realize an audio band extending apparatus that does not cause a valley of a spectrum even if a band is extended to a signal obtained by decoding a compression-encoded signal in which bits are not allocated to all bands.

According to the present embodiment, the decoder 1
Uses the scale information and bit allocation information extracted as preprocessing for decoding the frame in the demultiplexing processing, and performs level control and sliding band HP.
By specifying the cutoff frequency of F, control can be performed at the same timing as the decode signal. For this reason, it is possible to realize an audio band extending device in which the control of the signal to be added for the band extension is not delayed and the audibility becomes unnatural.

(Embodiment 3) Next, an audio band extending apparatus according to Embodiment 3 of the present invention will be described.
FIG. 3 is a block diagram showing the configuration of the audio band extending apparatus according to the present embodiment, and the same parts as those in FIG. This audio band extending device includes a decoder 1, an oversampling LPF 2, a band pass filter 3, a rectifier circuit 4, a sliding band HPF 5, a level control circuit 6, an adder circuit 7, a dither generator circuit 8, and an adder circuit 9. .

The adder 9 is a high-frequency signal adder for adding the output of the rectifier 4 and the output of the dither generator 8. The sliding band HPF 5 is a frequency-variable high-pass filter that refers to the bit allocation information S102 extracted by the decoder 1 and extracts an audio component higher than the highest frequency indicated by the bit allocation information from the output signal of the adding circuit 9. .

The level control circuit 6 is a level control means for controlling the output level of the sliding band HPF 5 according to the scale information S102 extracted by the decoder 1. The adding circuit 7 is an output signal adding unit that adds the output of the sliding band HPF 5 controlled by the level control circuit 6 and the decoded signal of the decoder 1 and outputs an audio signal whose band has been extended.

The operation of the audio band extending apparatus having such a configuration will be described. The coded signal S101 described with reference to FIG. 8 is input to the decoder 1 as an input signal. For example, if this signal is reproduced from an MD, the sampling frequency (Fs) is 44.1 kHz.
Signal.

The decoder 1 performs a decoding process in the same manner as the decoding process described with reference to FIG. The decoder 1 demultiplexes the coded signal S101 to extract system information, bit allocation information, scale information, and spectrum information, respectively, and outputs bit allocation information S102 and scale information S103. Denormalization and frequency synthesis are performed, and a decoded signal S104 is output.

The oversampling LPF 2 performs sampling by increasing the sampling frequency of the input decode signal S104 by a positive integer, for example, twice, and attenuates unnecessary bands. In the case of twice, the band of half or more of the sampling frequency of the input signal, that is, 22.05k
Hz or more is attenuated, for example, by 60 dB or more.

The band-pass filter 3 limits the band of the output S105 of the oversampling LPF2. Assuming that the sampling frequency of the input signal S104 is Fs, the bandpass filter 3 outputs a signal S106 having a band of, for example, Fs / 4 to Fs / 2. The rectifier circuit 4 generates a harmonic of the input signal by rectifying the output S106 of the bandpass filter 3 by half-wave or double-wave.

FIG. 4 shows the details of the dither generation circuit 8. The dither generation circuit 8 includes PN sequence noise generators 41, 42, 4
3. It includes an adder circuit 44 and a 1 / f characteristic filter 45. The PN sequence noise generators 41, 42, and 43 operate independently of each other, and the spectrum intensity is uniformly distributed in the band of 0 to Fs, and the probability distribution of which frequency component has a spectrum is uniform. Generates white noise.

The addition circuit 44 includes a PN sequence noise generator 4
1, 42 and 43 are added. By performing the addition in this manner, the noise after the addition becomes a bell-like shape in which the probability distribution of the amplitude is close to a Gaussian distribution, and approaches the natural sound. The 1 / f characteristic filter 45 attenuates the output signal of the adder circuit 44 as the frequency increases, that is, outputs the signal so as to have a 1 / f frequency characteristic.

The dither generating circuit 8 may be a diamond dither generating means whose probability density has a triangular distribution within a predetermined amplitude. The diamond dither generation means
The outputs of the two independent PN sequence generating means PA and PN
This is to add the output of the sequence generation means PB. Further, the diamond dither generation means may add the partial elements PPA and PPB which are different from each other and do not overlap with each other in one PN sequence generation means PP.

The dither generating circuit 8 may be bell-shaped dither generating means having a bell-shaped distribution of the probability density within a predetermined amplitude. This bell-shaped dither generation means adds the outputs of three independent PN sequence generation means PA, the output of the PN sequence generation means PB, and the output of the PN sequence generation means PC. Also, the bell-shaped dither generating means is provided with one PN.
The partial element PPA, the partial element PPB, and the partial element PPC which are different from each other and do not overlap with each other may be added.

The addition circuit 9 shown in FIG.
07 and the output S111 of the dither generation circuit 8 are added. The sliding band HPF 5 cuts off, from the output S112 of the adder circuit 9, the frequency at which the bit allocation information S102 is allocated, that is, the frequency below the highest frequency where the allocation is not 0, and a signal having a frequency spectrum exceeding that frequency. S108 is output.

The level control circuit 6 generates the scale information S10
3, the spectrum intensity of the high frequency component, for example, F
The signal level of s / 4 to Fs / 2 is obtained, and the output level of the sliding band HPF5 is controlled according to the level. The level control circuit 6 calculates Fs / 4 to Fs of the encoded signal.
When the signal level at s / 2 is high, the output level of the high-pass filter 5 is increased, and when the signal level is low, the output level of the high-pass filter 5 is reduced.
The adder circuit 7 outputs the output S of the oversampling LPF2.
105 and the output S109 of the level control circuit 6 are added,
The addition signal is output as an output signal S110.

As described above, according to the present embodiment, the bandpass filter 3, the rectifier circuit 4, the dither generation circuit 8,
Using the sliding band HPF5, a harmonic higher than the spectrum band included in the decoded signal is generated, and the level control circuit 6 performs level control in accordance with the high frequency spectrum intensity of the decoded signal using the level information, thereby obtaining the decoded signal. Is added to According to such a configuration, it is possible to realize an audio band extending apparatus that does not generate a valley of a spectrum even if the band is extended for a signal obtained by decoding a compression-encoded signal in which bits are not necessarily allocated to all bands. be able to.

According to the present embodiment, the decoder 1
Uses the scale information and bit allocation information extracted as preprocessing for decoding the frame in the demultiplexing processing, and performs level control and sliding band HPF.
By specifying the cutoff frequency of 5, control can be performed at the same timing as the decode signal. For this reason, it is possible to realize an audio band extending device in which the control of the signal to be added for the band extension is not delayed and the audibility becomes unnatural.

According to the present embodiment, the rectifier circuit 4
By synthesizing the harmonic component generated by the above and a dither signal having a 1 / f characteristic frequency distribution with a bell-shaped probability distribution, it is possible to realize a more audible natural audio band extending apparatus.

(Embodiment 4) Next, an audio band extending apparatus according to Embodiment 4 of the present invention will be described.
FIG. 5 is a block diagram showing the configuration of the audio band extending apparatus according to the present embodiment, and the same parts as those in FIG. 3 are described with the same reference numerals. This audio band extending device includes a decoder 1, an oversampling LPF 2, a band pass filter 3, a rectifier circuit 4, a sliding band HPF 5, a level control circuit 6, an adder circuit 7, a dither generator circuit 8, an adder circuit 9, and a switch 10. Be composed.

The switch 10 refers to the scale information S103 taken out by the decoder 1, and cuts off the band extension signal output from the level control circuit 6 in the case of a concentrated spectrum, and shuts down the level control circuit 6 in the case of a plurality of spectra. Switch means for outputting the band extension signal output from the. The adding circuit 7 is an output signal adding unit that adds the output of the switch 10 and the decoded signal output from the oversampling LPF 2 and outputs an audio signal whose band has been extended.

The operation of the audio band extending apparatus having such a configuration will be described. The coded signal S101 described with reference to FIG. 8 is input as an input signal. This signal is a signal having a sampling frequency Fs of 44.1 kHz if reproduced from, for example, an MD.

The decoder 1 performs a decoding process using the same method as the decoding process described with reference to FIG. The decoder 1 demultiplexes the coded signal S101, extracts system information, bit allocation information, scale information, and spectrum information, outputs bit allocation information S102 and scale information S103, and dequantizes these signals. , Inverse normalization, and frequency synthesis, and outputs a decoded signal S104.

The oversampling LPF 2 performs sampling by increasing the sampling frequency of the input decode signal S104 to a positive integer, for example, twice, and attenuates unnecessary bands. In the case of twice, the band of half or more of the sampling frequency of the input signal, that is, 22.05k
Hz or more is attenuated, for example, by 60 dB or more.

The bandpass filter 3 limits the band of the output S105 of the oversampling LPF2. Assuming that the sampling frequency of the input signal S104 is Fs, the bandpass filter 3 outputs a signal S106 having a band of, for example, Fs / 4 to Fs / 2.

The rectifier circuit 4 generates a harmonic of the input signal by rectifying the output S106 of the bandpass filter 3 between half waves or both waves. The dither generating circuit 8 generates white noise in which the spectral intensity is uniformly distributed in the band from 0 to Fs. The adder circuit 9 outputs the output S107 of the rectifier circuit 4.
And the output S111 of the dither generation circuit 8 are added.

The sliding band HPF 5 outputs the bit allocation information S 10 to the output S 112 of the adder 9.
The frequency at which the assignment of 2 exists, that is, the frequency below the highest non-zero assignment frequency is cut off, and a signal S108 having a frequency spectrum exceeding that frequency is output.

The level control circuit 6 generates the scale information S10
3, the spectrum intensity of the high frequency component, for example, F
The signal level of s / 4 to Fs / 2 is obtained, and the output level of the sliding band HPF5 is controlled according to the level. The level control circuit 6 calculates Fs / 4 to Fs of the encoded signal.
When the signal level at s / 2 is high, the output level of sliding band HPF5 is increased, and when the signal level is low, the output level of sliding band HPF5 is reduced.

The switch 10 turns on or off according to the output of the scale information S103. For example, when the scale information other than the frequency information adjacent thereto, that is, the frequency band adjacent to the upper and lower sides of the scale information is lower than the maximum by, for example, 20 dB or more, the switch 10 is turned off, and otherwise, the switch 10 is turned on. The addition circuit 7 is an oversampling LPF 2
Is added to the output S105 of the switch 10 and the output S113 of the switch 10, and the added signal is output as the output signal S110.

FIG. 6 is an explanatory diagram showing the operation when a concentrated spectrum signal is input. In the case of a concentrated spectrum signal such as a single sine wave, as shown in FIG. 6A, there is no scale information indicating a large value only in a specific band or a band in the vicinity thereof. Therefore, the switch 10 is turned off,
The band extension signal is not added to the decoded signal.
As shown in (b), only the original concentrated spectrum signal is output.

As described above, according to the present embodiment, the higher harmonics than the spectrum band included in the decoded signal are generated using the bandpass filter 3, the rectifier circuit 4, the dither generation circuit 8, and the sliding band HPF5. Then, the level control circuit 6 performs level control according to the high frequency spectrum intensity of the decoded signal using the level information,
It is added to the decoded signal. Therefore, it is possible to realize an audio band extending apparatus that does not cause a valley of a spectrum even if the band is extended for a signal obtained by decoding a compression-encoded signal in which bits are not necessarily allocated to all bands.

According to the present embodiment, the decoder 1
Is a demultiplexing process, which uses the scale information and bit allocation information extracted as preprocessing for decoding the frame, and performs level control and sliding band HP.
By specifying the cutoff frequency of F, control can be performed at the same timing as the decode signal. Therefore, it is possible to realize an audio band extending device that does not delay control of a signal to be added for band extension and causes unnatural sound.

Further, according to the present embodiment, when it is determined from the spectrum information that the signal is a concentrated spectrum signal,
The switch is turned off, and no band extension signal is added.
Therefore, when a signal such as a sine wave signal is output from the decoder, the signal does not deteriorate.

In the above embodiment, a smoothing filter may be provided at the input of the scale information S103 of the level control circuit 6 to smooth the scale information taken out by the decoder 1. This makes it possible to make the fluctuation of the output level of the level control circuit 6 slow. Further, the above-mentioned smoothing filter may be one which can control an attack characteristic and a release characteristic respectively.

The band pass filter 3 and the rectifier circuit 4
, The dither generation circuit 8 and the sliding band HPF 5 have an operation word length longer than the word length of the decode signal of the decoder 1 and the decode signal S104.
May be obtained by adding a signal component having substantially the same amplitude to the LSB width of the word length. For example, assuming that the word length of the decoded signal S104 output from the decoder 1 is 16 bits, the LSB of this signal, that is, the 4-bit noise (dither) lower than the LSB of this signal so as to have almost the same amplitude as the 16th bit of the least significant bit Signal).

[0095]

As described above, according to the present invention, a bandpass filter, a rectifier circuit, a dither generation circuit, and a sliding band type highpass filter generate higher harmonics than the spectrum band included in the decoded signal, and the level control circuit. The level information is used to control the level in accordance with the high-band spectrum intensity of the decoded signal, and is added to the decoded signal and output. For this reason, when a compression-encoded signal in which bits are not assigned to all bands is decoded, an audio signal that does not have a valley in the spectrum is obtained even if the band of the decoded signal is expanded.

Further, since frequency information such as scale information and bit allocation information can be directly obtained from the medium on which the audio signal is recorded, the reproduced audio band can be expanded as compared with the medium on which the PCM audio signal is recorded. This has the effect of reducing the load on hardware or software for performing the operations.

In particular, according to the first to third aspects of the present invention, the scale information and the bit allocation information extracted as a pre-process by the decoder for decoding the frame in the demultiplexing process are used for the level control and the sliding band type. By using it to specify the cutoff frequency of the high-pass filter, control can be performed at the same timing as the decode signal. For this reason, an audio signal is obtained in which the control of the signal to be added for band expansion is delayed and the audibility does not become unnatural.

In particular, according to the fourth aspect of the present invention, when it is determined from the spectrum information that the signal is a concentrated spectrum signal, a signal such as a sine wave signal may be deteriorated because no band extension signal is added. Absent.

In particular, according to the second to fourth aspects of the present invention, a harmonic component generated by a rectifier circuit is combined with a dither signal having a 1 / f characteristic frequency distribution with a bell-shaped probability distribution. Thus, an audio signal that is more audible and natural can be obtained without emphasizing a specific band.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an audio band extending device according to a first embodiment of the present invention.

FIG. 2 is a configuration diagram of an audio band extending device according to a second embodiment of the present invention.

FIG. 3 is a configuration diagram of an audio band extending device according to a third embodiment of the present invention.

FIG. 4 is a block diagram of a dither generation circuit used in an audio band extending device according to the second, third, and fourth embodiments.

FIG. 5 is a configuration diagram of an audio band extending device according to a fourth embodiment of the present invention.

FIG. 6 is an explanatory diagram illustrating an operation of the audio band extending apparatus according to the fourth embodiment when a concentrated spectrum signal is input.

FIG. 7 is a configuration diagram of a conventional audio band extending apparatus.

FIG. 8 is a signal arrangement diagram showing a configuration of an encoded signal.

FIG. 9 is a block diagram of a decoder that decodes an encoded signal.

FIG. 10A shows bit allocation information of an encoded signal,
(B) is a frequency characteristic diagram showing an example of scale information.

FIG. 11 is a spectrum diagram when a conventional audio band extending device is applied to a decoding signal of compression encoding.

[Explanation of symbols]

Reference Signs List 1 decoder 2 oversampling low-pass filter (oversampling LPF) 3 band-pass filter 4 rectifier circuit 5 sliding-band high-pass filter (sliding band HPF) 6 level control circuit 7, 9, 44 adder circuit 8 dither generation circuit 10 switch 41, 42, 43 PN sequence noise generator 45 1 / f characteristic filter 91 demultiplexer 92 dequantizer 93 denormalizer 94 frequency synthesizer

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G10L 7/04 G F-term (Reference) 5D045 DA11 DB00 5J064 AA01 BA06 BB07 BC02 BC08 BC11 BC18 BD01

Claims (23)

[Claims] 1. A compression-encoding signal in which an audio signal is divided into a plurality of frequency bands in frame units, and the compression encoding is performed using at least bit allocation information for each band and scale information for each band. An audio band extending device that reproduces a signal, wherein the decoding unit decodes the compression-encoded signal and outputs a decoded signal including the bit allocation information, the scale information, and the audio signal. A non-linear means for distorting the audio signal, and a frequency for referring to the bit allocation information extracted by the decoding means, and extracting an audio component higher than the highest frequency indicated by the bit allocation information from the output signal of the non-linear means. A variable high-pass filter; and the scale extracted by the decoding means. Level control means for controlling an output level of the frequency-variable high-pass filter means in accordance with the level information, an output of the frequency-variable high-pass filter means controlled by the level control means, and a decode signal of the decoding means. And an output signal adding means for outputting an audio signal whose band has been expanded. 2. A compression-encoding signal in which an audio signal is divided into a plurality of frequency bands in frame units, and the compression encoding is performed using at least bit allocation information for each band and scale information for each band. An audio band extending apparatus for reproducing a signal, comprising: decoding means for decoding the compression-encoded signal and outputting a decoded signal including the bit allocation information, the scale information, and an audio signal; and a dither generating high-frequency noise. Generating means; and a frequency-variable high-pass filter that refers to the bit allocation information extracted by the decoding means and extracts a noise component higher than a highest frequency indicated by the bit allocation information from an output signal of the dither generating means. And according to the scale information extracted by the decoding means, the frequency A level control means for controlling an output level of the variable high-pass filter means; an output of the frequency-variable high-pass filter means controlled by the level control means and a decode signal of the decoding means, whereby a band is extended. And an output signal adding means for outputting the output audio signal. 3. A compression coded signal obtained by dividing an audio signal into a plurality of frequency bands in frame units, and said compression coding is performed using at least bit allocation information for each band and scale information for each band. An audio band extending device that reproduces a signal, wherein the decoding unit decodes the compression-encoded signal and outputs a decoded signal including the bit allocation information, the scale information, and the audio signal. Non-linear means for distorting the audio signal, high-frequency noise generating dither generating means, high-frequency signal adding means for adding the output of the non-linear means and the output of the dither generating means, and Referring to the bit allocation information, an audio signal having a frequency higher than the highest frequency indicated by the bit allocation information A frequency-variable high-pass filter for extracting an O component from an output signal of the high-frequency signal adding means; and an output level of the frequency-variable high-pass filter means in accordance with the scale information extracted by the decoding means. Level control means, and an output signal addition means for adding an output of the frequency variable high-pass filter means controlled by the level control means and a decode signal of the decoding means, and outputting an audio signal having an extended band. , An audio band extending device. 4. A compression coded signal in which an audio signal is divided into a plurality of frequency bands in frame units, and said compression coding is performed using at least bit allocation information for each band and scale information for each band. An audio band extending device that reproduces a signal, wherein the decoding unit decodes the compression-encoded signal and outputs a decoded signal including the bit allocation information, the scale information, and the audio signal. Non-linear means for distorting the audio signal, high-frequency noise generating dither generating means, high-frequency signal adding means for adding the output of the non-linear means and the output of the dither generating means, and Referring to the bit allocation information, an audio signal having a frequency higher than the highest frequency indicated by the bit allocation information. A frequency-variable high-pass filter for extracting an E component from the output signal of the high-frequency signal adding means, and controlling an output level of the frequency-variable high-pass filter means in accordance with the scale information extracted by the decoding means. Level control means for outputting a band extension signal, and referring to the scale information extracted by the decoding means, in the case of a concentrated spectrum, the band extension signal of the level control means is cut off, and in the case of a plurality of spectra, Switch means for outputting the band extension signal of the level control means, and output signal addition means for adding an output of the switch means and a decode signal of the decoding means and outputting an audio signal whose band has been extended, An audio band extending device, comprising: 5. An oversampling filter for performing oversampling on an audio signal output from the decoding means at a sampling frequency twice or more as high as a sampling frequency at the time of compression processing, and removing an aliasing noise component due to the oversampling. 5. The audio band extending device according to claim 1, wherein the audio band extending device is provided between the decoding unit and the output signal adding unit. 6. The method according to claim 1, wherein an operation word length of the nonlinear means and the frequency variable high-pass filter is longer than a word length of a decoded signal of the decoding means.
The audio band extending apparatus according to any one of claims 3 and 4. 7. An arithmetic unit according to claim 2, wherein an operation word length of said dither generation means and said frequency variable high-pass filter is larger than a word length of a decoded signal of said decoding means. The audio band extending device according to the item. 8. The non-linear means and the frequency-variable high-pass filter have an operation word length greater than a word length of a decode signal of the decoding means and an amplitude substantially equal to an LSB width of the word length of the decode signal. The audio band extending apparatus according to any one of claims 1, 3, and 4, wherein the processing is performed by adding the above signal components. 9. The dither generating means and the frequency-variable high-pass filter process the calculated word length by adding a signal component having substantially the same amplitude to the LSB width of the word length of the decoded signal. The audio band extending apparatus according to any one of claims 2, 3, and 4. 10. The method according to claim 1, wherein the non-linear means is a half clipping means for extracting only one side of the signal having positive and negative amplitudes, or an absolute value converting means for turning back the amplitude of one side to the opposite side. The audio band extending device according to any one of claims 1 to 7. 11. The apparatus according to claim 1, wherein said non-linear means distorts a signal limited to a predetermined frequency band from a decoded signal of said decoding means. An audio band extending device according to claim 1. 12. The dither generating means according to claim 2, wherein the probability density within a predetermined amplitude is a triangular distribution.
5. The audio band extending device according to any one of 4. 13. The output of two independent PN sequence generating units PA and PN
13. The audio band extending device according to claim 12, wherein the output from the sequence generating means PB is added. 14. The audio band according to claim 12, wherein said diamond dither generating means adds the different and non-overlapping partial elements PPA and PPB of one PN sequence generating means PP. Expansion device. 15. The dither generating means according to claim 2, wherein said dither generating means is bell-shaped dither generating means having a bell-shaped distribution of probability density within a predetermined amplitude. Audio band extender. 16. The bell-shaped dither generating means comprises: an output of three independent PN sequence generating means PA;
16. The audio band extending apparatus according to claim 15, wherein an output of the sequence generation means PB and an output of the PN sequence generation means PC are added. 17. The bell-shaped dither generating means comprises: a partial element PPA, a partial element PPB, and a partial element PPC which are different from each other and do not overlap with each other;
16. The audio band extending apparatus according to claim 15, wherein 18. The audio band extending apparatus according to claim 2, wherein said dither generating means has an output spectrum distribution having a 1 / f characteristic. 19. The level control means, when it is determined that only a concentrated spectrum exists according to the scale information extracted by the decoding means, sets the level of the dither generation means to substantially zero, and When it is determined that a spectrum exists, the level is controlled such that the dither spectrum height is substantially continuous with the highest spectrum level indicated by the bit allocation information extracted by the decoding means. The audio band extending device according to any one of 2, 3, and 4. 20. When it is determined that two or more spectra exist according to the scale information extracted by the decoding unit, the level control unit obtains a frequency characteristic slope of the spectrum, and smoothes the spectrum distribution. The audio band extending apparatus according to claim 2, wherein the dither level is controlled so as to be continuous. 21. The apparatus according to claim 1, wherein the level control means mutes the output level to zero when the signal is determined to be zero according to the scale information extracted by the decoding means. The audio band extending device according to claim 1. 22. The apparatus according to claim 1, wherein said level control means smoothes the scale information taken out by said decoding means by using a predetermined smoothing filter, and makes output level fluctuation slow. The audio band extending device according to claim 1. 23. The audio band extending apparatus according to claim 22, wherein said smoothing filter controls an attack characteristic and a release characteristic, respectively.