JP2019113866A - Acoustic device, missing band estimation device, signal processing method, and frequency band estimation device - Google Patents

Abstract

To output high-quality sound whose high-frequency band has been interpolated with a simple structure.SOLUTION: In high-frequency interpolation, a harmonic generation unit first generates a harmonic signal of an input compressed sound signal CAD. Concurrently with generation of this harmonic signal, an HPF unit 121having a cutoff frequency Fextracts a high-frequency component of the compressed sound signal CAD. In addition, an HPF unit 121having a cutoff frequency F(>F) extracts a high-frequency component of the compressed sound signal CAD. An estimation unit 124 estimates a missing band of the compressed sound signal CAD on the basis of the ratio between the signal level of a difference signal SBD obtained by subtracting an output signal HPDof the HPF unit 121from an output signal HPDof the HPF unit 121and the signal level of the signal HPD. Then, on the basis of the estimated missing band, the estimation unit 124 controls a cutoff frequency of a variable HPF unit 130 that extracts a signal component for high-frequency interpolation from the harmonic signal.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an acoustic device, a missing band estimation device, a signal processing method and a signal processing program, a recording medium on which the signal processing program is recorded, and a frequency band estimation device.

BACKGROUND In recent years, audio devices that reproduce audio content recorded in digital format have become widespread. Such audio content data is often subjected to digital compression processing by a method such as MP3 (Moving Picture Expert Group) Audio Layer-3, in order to reduce the file size. The compressed audio signal obtained by decompressing the compressed audio data generated by performing such digital compression processing has a band limited by the sampling frequency (F _S ) adopted when obtaining the audio data before compression processing. Furthermore, the treble band is a limited audio signal. Then, the high-tone band limited by the compression processing becomes wider as the bit rate becomes lower (that is, as the compression rate becomes higher) in the case of compression processing by the same method.

  For this reason, a technique has been proposed for interpolating high-tone bands in which signal components are dropped due to compression processing according to the bit rate, that is, according to the compression rate (see Patent Document 1: below) It is called "conventional example"). In this prior art technique, the determination means reads information such as a bit rate separated from the compressed audio signal obtained by decompressing the compressed audio data. Subsequently, the determination means sets the cut-off frequency of the high-pass filter that passes the harmonic signal generated by the harmonic generation means based on the read bit rate and the like. Thus, the signal component of the high-tone band is interpolated by combining the signal that has passed through the high-pass filter with the cut-off frequency set with the compressed voice signal.

JP 2004-317622

  In the above-described prior art, the discrimination means separates from the compressed voice data in order to appropriately set the cut-off frequency of the high-pass filter that performs high-pass filtering on the harmonic signal generated by the harmonic generation means. It is designed to read information such as the bit rate being used. That is, in the prior art, the determination means can access a storage device in which information such as compressed audio data and bit rate is stored.

  By the way, in recent years, information such as compressed voice data and bit rate is often downloaded from a server on a network as acoustic content by a small portable terminal device. For this reason, compressed audio signals generated by a small portable terminal device are sent to another audio device, and high-quality audio signals are generated after interpolation of the high-tone band by the audio device, and then audio output is performed. There is.

  In such a case, when applying the technique of the conventional example, it is necessary to transmit information such as a bit rate from the portable terminal device to another audio device via a route different from the route of the compressed audio signal. Therefore, it is difficult to add a function in the portable terminal device, and it is difficult to output high-quality voice obtained by interpolating the high-tone band with a simple configuration.

  For this reason, there is a great demand for a technology that can output high-quality voice with high frequency band interpolation with a simple configuration. Responding to such a request is one of the problems to be solved by the present invention.

  According to a first aspect of the present invention, there is provided a harmonic generation unit that generates harmonics of an input audio signal; a cutoff frequency is variable, and the high frequency range of the harmonics generated by the harmonic generation unit A variable high-pass filter unit for extracting a component; a first high-pass filter unit having a first cutoff frequency and extracting a high-frequency component of the input audio signal; and a higher than the first cutoff frequency A second high-pass filter unit having a high second cutoff frequency and extracting high-frequency components of the input audio signal; a level of an output signal of the first high-pass filter unit; A control unit configured to control a cut-off frequency of the variable high-pass filter unit based on a level of an output signal of the high-pass filter unit.

  According to a second aspect of the present invention, there is provided a first high-pass filter unit having a first cutoff frequency and extracting a high-frequency component of an input audio signal; and a first high-pass filter section than the first cutoff frequency A second high-pass filter unit having a high second cutoff frequency and extracting high-frequency components of the input audio signal; a level of an output signal of the first high-pass filter unit; An estimation unit configured to estimate a treble band in which a signal component is missing in the input audio signal based on a level of an output signal of the high-pass filter unit.

  According to a third aspect of the present invention, there is provided a harmonic generation unit that generates harmonics of an input voice signal; a cutoff frequency is variable, and the high frequency range of the harmonics generated by the harmonic generation unit A variable high-pass filter unit for extracting a component; a first high-pass filter unit having a first cutoff frequency and extracting a high-frequency component of the input audio signal; and a higher than the first cutoff frequency A signal processing method for use in an acoustic apparatus, comprising: a second high-pass filter unit having a high second cutoff frequency and extracting a high-frequency component of the input audio signal; Acquiring the level of the output signal of the high-pass filter unit and the level of the output signal of the second high-pass filter unit; the variable high-pass filter based on the acquisition result in the acquisition step A signal processing method characterized by comprising a; and a control step of controlling the cut-off frequency of the filter unit.

  According to a fourth aspect of the present invention, there is provided a signal processing program that causes a computer of an audio device to execute the signal processing method of the present invention.

  According to a fifth aspect of the present invention, there is provided a recording medium in which the signal processing program of the present invention is recorded so as to be readable by a computer of an audio device.

  According to a sixth aspect of the present invention, there is provided a first high-pass filter unit having a first cutoff frequency and extracting a high-frequency component of an input audio signal; and a first high-pass filter section than the first cutoff frequency A second high-pass filter unit having a high second cutoff frequency and extracting high-frequency components of the input audio signal; a level of an output signal of the first high-pass filter unit; An estimation unit configured to estimate a frequency band of the input audio signal based on a level of an output signal of the high-pass filter unit.

FIG. 1 is a block diagram schematically showing a configuration of an acoustic device according to an embodiment of the present invention. It is a figure which shows the average spectrum of the compression audio | voice signal for every bit rate input into the acoustic apparatus of FIG. It is a block diagram which shows the structure of the missing band estimation apparatus of FIG. It is a figure for demonstrating the filtering characteristic of two types of high pass filter (HPF) parts of FIG. FIG. 5 is a diagram for explaining signal components corresponding to detection targets of two types of level detection units in FIG. 3 for every three bit rates. It is a block diagram for demonstrating the structure of the synthetic | combination part of FIG. It is a figure for demonstrating the high region interpolation by the apparatus of FIG.

  Hereinafter, an embodiment of the present invention will be described with reference to FIGS. 1 to 7. In the drawings, the same or equivalent elements will be denoted by the same reference symbols, without redundant description.

[Constitution]
FIG. 1 is a block diagram showing a schematic configuration of an acoustic device 100 according to an embodiment. As shown in FIG. 1, the acoustic device 100 is connected to a compressed voice decompression device (CADD) 200 and a sound output device 300.

  Here, the above-described compressed audio decompression apparatus 200 decompresses compressed audio data generated in accordance with a predetermined standard such as the MP3 standard to generate a compressed audio signal (audio signal) CAD. The compressed audio signal CAD generated in this way is sent to the acoustic device 100.

  In the present embodiment, the compressed audio signal CAD is a compressed audio signal corresponding to one of three bit rates of “BR1”, “BR2 (> BR1)” and “BR3 (> BR2)”. There is.

  Further, the sound output device 300 described above is configured to include the speaker SP. The sound output device 300 receives the signal HID after high frequency interpolation sent from the sound device 100. Then, the sound output device 300 outputs a sound according to the signal HID after high-frequency interpolation from the speaker SP.

<Configuration of Sound Device 100>
The acoustic device 100 described above includes a harmonic generation unit (HMG) 110 and a missing band estimation device (MBE) 120. The acoustic device 100 further includes a variable high pass filter (HPF) unit 130 and a combining unit 140.

The harmonic generation unit 110 receives the compressed audio signal CAD sent from the compressed audio decompression apparatus 200. Subsequently, the harmonic generation unit 110 generates first to Nth harmonics of components of a predetermined frequency band (0 to F _H ) of the compressed sound signal CAD. Then, among the generated harmonics, a component equal to or lower than the highest frequency F _MAX (= F _S / 2) of the band of the voice before compression determined by the sample frequency F _S is sent to the variable HPF unit 130 as a signal HMD.

  The missing band estimation device 120 described above receives the compressed voice signal CAD sent from the compressed voice decompression device 200. Subsequently, the missing band estimation device 120 estimates, based on the compressed speech signal CAD, a treble band (hereinafter also referred to as a “missing band”) in which a signal component is missing in the compressed speech signal CAD. Then, the missing band estimation device 120 sends, to the variable HPF unit 130, a cutoff frequency designation HPC in which the lowest frequency of the estimated missing band is designated.

  Here, estimating the missing band in the compressed speech signal CAD is nothing but estimating the frequency band of the compressed speech signal CAD. For this reason, it can be said that the missing band estimation device 120 also has a function as a frequency band estimation device that estimates the frequency band of the compressed speech signal CAD.

  The details of the configuration of the missing band estimation device 120 will be described later.

  The variable HPF unit 130 described above receives the signal HMD sent from the harmonic generation unit 110. Also, the variable HPF unit 130 receives the cutoff frequency designation HPC sent from the missing band estimation device 120. Then, the variable HPF unit 130 performs high-pass filtering processing on the signal HMD in which the frequency designated by the cutoff frequency designation HPC is the cutoff frequency. The result of the high pass filtering process is sent to the combining unit 140 as a signal HBD.

  The synthesis unit 140 described above receives the compressed speech signal CAD sent from the compressed speech decompression apparatus 200. Further, the combining unit 140 receives the signal HBD sent from the variable HPF unit 130. Then, the synthesis unit 140 synthesizes the compressed speech signal CAD and the signal HBD, and generates a signal HID after high-frequency interpolation. The signal HID after high frequency interpolation generated in this way is sent to the sound output device 300.

  The details of the configuration of the combining unit 140 will be described later.

Here, the relationship between the bit rate and the band of compressed speech will be described. In FIG. 2 (A), the average spectrum of the pre-compression audio corresponding to the digital music sound generated is sampled by the sample frequency F _S is shown schematically. As shown in FIG. 2A, the upper limit frequency of the pre-compression speech band is the highest frequency F _MAX (= F _S / 2).

The signal bands of the compressed speech obtained by decompressing the compressed speech data of the above-mentioned bit rate BR1 to BR3 obtained by compressing the data of the speech before compression are shown in FIGS. 2 (B) to (D). Here, FIG. 2 (B) shows a signal band of compressed speech at bit rate BR1. As shown in FIG. 2B, the compressed voice of bit rate BR1 has the upper limit frequency of the signal band at frequency F _BR1 , and the frequency bands (F _{BR1 to} F _MAX ) are compared with the voice before compression, It is a missing band of the signal component.

Further, FIG. 2C shows a signal band of compressed speech at a bit rate BR2 (> BR1). As shown in FIG. 2C, in the compressed voice of bit rate BR2, the upper limit frequency of the signal band is frequency F _BR2 (> F _BR1 ), and the frequency band (F _{BR2 to} F _MAX ) is not compressed. Compared to voice, it is a missing band of signal components.

Further, FIG. 2 (D) shows a signal band of compressed speech at a bit rate BR3 (> BR2). As shown in FIG. 2D, the compressed voice at bit rate BR3 has the upper limit frequency of the signal band at frequency F _BR3 (> F _BR2 ), and the frequency band (F _{BR3 to} F _MAX ) is not compressed. Compared to voice, it is a missing band of signal components.

(Configuration of Missing Band Estimation Device 120)
Next, the configuration of the above missing band estimation device 120 will be described.

Missing band estimator 120, as shown in FIG. 3, a pi-pass filter (HPF) 121 _1, 121 _2, and a subtraction unit 122. In addition, the missing band estimation device 120 includes level detection units 123 ₁ and 123 ₂ and an estimation unit 124.

HPF 121 ₁ above, performs high-pass filtering at a cut-off frequency F _C1. The HPF 121 ₁ receives the compressed audio signal CAD sent from compressed audio decompressor 200. Then, HPF 121 ₁ performs a high-pass filtering of the cut-off frequency F _C1 the compressed audio signal CAD. The high-pass filtering process result is sent as a signal HPD ₁ to subtraction unit 122.

HPF 121 ₂ above, performs high-pass filtering with a cutoff frequency F _C2 (> F _C1). The HPF 121 ₂ receives the compressed audio signal CAD sent from compressed audio decompressor 200. Then, HPF 121 ₂ performs a high pass filtering process cutoff frequency F _C2 the compressed audio signal CAD. The high-pass filtering process result is sent as a signal HPD ₂ to subtraction unit 122 and the level detector 123 _2.

It said subtraction unit 122 receives a signal HPD ₁ sent from the HPF 121 _1. Further, the subtraction unit 122 receives a signal HPD ₂ sent from the HPF 121 _2. Then, the subtraction unit 122 subtracts the signal HPD ₂ from the signal HPD ₁ . The results thus calculated are sent to a level detector 123 ₁ as the signal SBD.

Level detector 123 ₁ of the receive signal SBD sent from the subtraction unit 122. Then, level detector 123 ₁ detects the power level of the signal SBD. Detection result by the level detecting unit 123 ₁ is transmitted to the estimating unit 124 as a detection level DL _1.

Above the level detector 123 ₂ receives a signal HPD ₂ sent from the HPF 121 _2. Then, level detector 123 ₂ detects the power level of the signal HPD _2. Detection result by the level detecting unit 123 ₂ are sent to the estimating unit 124 as a detection level DL _2.

The above estimation unit 124 receives the detection level DL ₁ sent from the level detection unit 123 _1. Further, the estimating unit 124 receives the detection level DL ₂ sent from the level detection unit 123 _2. Then, the estimation unit 124 estimates a missing band in the compressed sound signal CAD based on a ratio R (= DL ₁ / DL ₂ ) of the detection level DL ₁ and the detection level DL ₂ .

  Subsequently, the estimation unit 124 generates a cutoff frequency designation HPC in which the lower limit frequency of the estimated missing band is designated. The cutoff frequency designation HPC generated in this manner is sent to the variable HPF unit 130.

The example of the filtering characteristic of the HPF 121 ₁ and HPF 121 ₂ is shown in FIG. Here, in FIG. 4 (A), examples of the filtering characteristic of the HPF 121 ₁ is shown. Further, in FIG. 4 (B), examples of the filtering characteristic of the HPF 121 ₂ is shown.

Here, the filtering characteristics of the HPF 121 _2, bit rate "BR1", "BR2", be either "BR3", when calculating the ratio R, dividing the resource in the estimation unit 124 does not cause overflow Is set as

The signal component corresponding to a detection target of the HPF 121 ₁ and HPF 121 ₂ are shown schematically in FIG. In FIG. 5, together with the signal component corresponding to a detection target of the HPF 121 ₁ is indicated by a horizontal line hatch signal component corresponding to the detection target HPF 121 ₂ are indicated by a vertical line hatching.

Here, in FIG. 5 (A), if the bit rate BR1, signal components corresponding to the detection target of the HPF 121 ₁ and HPF 121 ₂ are shown schematically. Further, in FIG. 5 (B), when the bit rate BR2, the signal component corresponding to a detection target of the HPF 121 ₁ and HPF 121 ₂ are shown schematically. Further, in FIG. 5 (C), if the bit rate BR3, signal components corresponding to the detection target of the HPF 121 ₁ and HPF 121 ₂ are shown schematically.

  As can be seen by comparing FIGS. 5A to 5C with each other, when the bit rate is different, the calculated ratio R is different. For this reason, since the missing area in the compressed speech signal CAD can be estimated based on the value of the ratio R thus calculated, the estimation unit 124 can calculate the value of the compressed speech signal CAD when the missing band for the bit rate is uniquely determined. The bit rate can be estimated.

(Configuration of the combining unit 140)
Next, the configuration of the combining unit 140 described above will be described.

The combining unit 140 includes a delay unit 141 and multiplying units 142 ₁ and 142 ₂ as shown in FIG. The combining unit 140 further includes an adding unit 143.

The delay unit 141 receives the compressed audio signal CAD (= D ₀ (T) (T: time)) sent from the compressed audio decompression apparatus 200. Then, the delay unit 141 generates a signal DLD (= D (T)) obtained by delaying the compressed audio signal CAD by a time T _DL corresponding to the phase delay in the harmonic generation unit 110 and the variable HPF unit 130. Here, the relationship between the signal D (T) and the compressed audio signal D ₀ (T) is expressed by the following equation (1).
D (T) = D ₀ (T−T _DL ) (1)

As a result, synchronization of the signal DLD and the signal HBD output from the variable HPF unit 130 described above is achieved. Signal DLD thus generated is sent to the multiplying unit 142 _1.

Multiplying unit 142 ₁ of the receive signal DLD transmitted from the delay unit 141. Then, multiplying unit 142 ₁ generates a signal MLD signal DLD and _1x K. The signal MLD generated in this way is sent to the adding unit 143.

Multiplying portion 142 ₂ of the receive signal HBD sent from the variable HPF 130. Then, the multiplication unit 142 ₂ generates a signal MHD signals HBD is _doubled K. The signal MHD generated in this manner is sent to the addition unit 143.

The ratio between the value K ₁ and the value K _2, from the viewpoint of suitable high-frequency interpolation, experiments, simulations, are predetermined based on experience and the like.

The above adder 143 receives the signals MLD sent from the multiplying unit 142 _1. The addition unit 143 receives a signal MHD sent from the multiplication unit 142 _2. Then, the adding unit 143 adds the signal MLD and the signal MHD to generate a signal HID after high-frequency interpolation. The signal HID after high frequency interpolation generated in this way is sent to the sound output device 300.

The spectrum of the signal MHD generated as described above is shown in FIG. Here, in FIG. 7A, the spectrum of the signal MHD generated corresponding to the compressed audio signal CAD at the bit rate BR1 is shown by a broken line. Further, in FIG. 7B, the spectrum of the signal MHD generated corresponding to the compressed audio signal CAD at the bit rate BR2 is shown by a broken line. Further, in FIG. 7C, the spectrum of the signal MHD generated corresponding to the compressed audio signal CAD at the bit rate BR3 is shown by a broken line. In FIG. 7 (A) ~ (C) , the spectrum of the signal DLD the K ₁ × signal MLD (thus, the compressed audio signal CAD to ₁ × K signals) are shown by a solid line.

  As shown in FIGS. 7A to 7C, the signal HMD is a signal that appropriately interpolates the missing band of the signal component in the compressed audio signal CAD.

[Operation]
Next, the operation of the acoustic device 100 configured as described above will be described, focusing mainly on the process of generating the signal HBD (see FIG. 1) based on the compressed audio signal CAD.

  When the compression speech decompression apparatus 200 starts supplying the compression speech signal CAD, in the acoustic device 100, the harmonic generation unit 110 and the missing band estimation apparatus 120 receive the compression speech signal CAD. In the acoustic device 100, the synthesizing unit 140 receives the compressed sound signal CAD (see FIG. 1).

When receiving the compressed audio signal CAD, the harmonic generation unit 110 generates harmonics of components of a predetermined frequency band of the compressed audio signal CAD. Then, among the generated harmonics, the harmonic generation unit 110 sends the component below the highest frequency F _MAX of the pre-compression speech band determined by the sample frequency F _S to the variable HPF unit 130 as a signal HMD (see FIG. 1).

On the other hand, when receiving the compressed voice signal CAD, the missing band estimation device 120 estimates the missing band in the compressed voice signal CAD based on the compressed voice signal CAD in parallel with the high frequency generation by the above-mentioned harmonic generation unit 110. . In the estimation of such missing band, the missing band estimator 120, compression HPF 121 ₁ receives an audio signal CAD is subjected to high-pass filtering of the cut-off frequency F _C1 the compressed audio signal CAD. Then, HPF 121 ₁ sends a result of the high pass filtering process, as a signal HPD ₁ to subtraction unit 122 (see FIG. 3).

Further, when receiving the compressed audio signal CAD, HPF 121 _2, in parallel with the high pass filtering process by HPF 121 ₁ mentioned above is subjected to high-pass filtering of the cut-off frequency F _C2 the compressed audio signal CAD. Then, HPF 121 ₂ sends a result of the high pass filtering process, as a signal HPD ₂ to subtraction unit 122 and the level detector 123 ₂ (see FIG. 3).

Signal HPD ₁ sent from HPF 121 _1, and receives a signal HPD ₂ sent from the HPF 121 _2, the subtraction unit 122 calculates a difference between the signal HPD ₁ and the signal HPD _2. The subtraction unit 122 sends the calculated difference, the level detecting unit 123 ₁ as the signal SBD (see FIG. 3).

Receiving signal SBD sent from the subtraction unit 122, level detector 123 ₁ detects the power level of the signal SBD. Then, the level detecting unit 123 ₁ transmits the detection result, the estimating unit 124 as a detection level DL ₁ (see FIG. 3).

Receiving signal HPD ₂ sent from the HPF 121 _2, level detector 123 ₂ detects the power level of the signal HPD _2. Then, level detector 123 ₂ sends the detection result, the estimating unit 124 as a detection level DL ₂ (see FIG. 3).

Detection level DL ₁ sent from the level detection unit 123 _1, and receives a detection level DL ₂ sent from the level detection unit 123 _2, estimating unit 124, based on the detection level DL ₁ and the detection level DL ₂ , Generate cutoff frequency designated HPC. When generating the cutoff frequency designation HPC, the estimation unit 124 first calculates a ratio R (= DL ₁ / DL ₂ ) between the detection level DL ₁ and the detection level DL ₂ .

  Subsequently, the estimation unit 124 estimates the missing band of the compressed speech signal based on the calculated ratio R.

  Next, the estimation unit 124 generates a cutoff frequency designation HPC in which the lower limit frequency of the estimated missing band is designated. Then, the estimation unit 124 sends the generated cutoff frequency designation HPC to the variable HPF unit 130 (see FIG. 3).

  When receiving the cutoff frequency designation HPC sent from the missing band estimation device 120 (more specifically, the estimation unit 124), the variable HPF unit 130 performs a high pass setting the frequency designated by the cutoff frequency designation HPC as the cutoff frequency A filtering process is performed on the signal HMD sent from the harmonic generation unit 110 to generate a signal HBD. Then, the variable HPF unit 130 sends the generated signal HBD to the combining unit 140 (see FIG. 1).

When receiving the signal HBD sent from the variable HPF unit 130, the combining unit 140 combines the signal HBD with the compressed voice signal CAD sent from the compressed voice decompression apparatus 200. At the time of such synthesis, in the synthesis unit 140, the delay unit 141 delays the compressed voice signal CAD by the time T _DL corresponding to the phase delay in the harmonic generation unit 110 and the variable HPF unit 130 to synchronize with the signal HBD. The signal DLD is generated. The delay unit 141 sends the generated signal DLD to the multiplying unit 142 ₁ (see FIG. 6).

Receiving signal DLD transmitted from the delay unit 141, multiplying section ₁₄₂₁ generates a signal MLD signal DLD and _1x K. Then, multiplying unit 142 ₁ transmits the generated signal MLD to the adder 143 (see FIG. 6).

On the other hand, the multiplication unit 142 ₂ generates a signal MHD signals HBD is _doubled K. Then, the multiplication unit 142 ₂ sends the generated signal MHD to the adder 143 (see FIG. 6).

Multiplying unit 142 ₁ from the signal sent MLD, and receives signal MHD sent from the multiplication unit 142 _2, an adder 143 adds the signal MLD and signal MHD, signal high-frequency interpolation has been made Generate HID. Then, the adding unit 143 sends the generated signal HID after high-frequency interpolation to the sound output device 300 (see FIG. 6).

  That is, after achieving synchronization between the signal HBD and the compressed audio signal CAD, the combining unit 140 performs weighted addition with a mixing ratio that allows appropriate high-frequency interpolation, and combines the signal HBD with the compressed audio signal CAD. . The signal HID after high-frequency interpolation generated as a result of such synthesis is sent to the sound output device 300.

  Upon receiving the signal HID after high-frequency interpolation sent from the acoustic device 100 (more specifically, the combining unit 140), the sound output device 300 outputs a sound according to the signal HID after high-frequency interpolation from the speaker SP . As a result, high-quality speech in which high-frequency interpolation is appropriately performed corresponding to the bit rate of the compressed speech signal CAD is output from the sound output device 300.

  As described above, in the present embodiment, at the time of high-frequency interpolation, the harmonic generation unit 110 first generates harmonics of the input compressed audio signal CAD. In parallel with the generation of such harmonics, the missing band estimation device 120 estimates the missing band in the compressed speech signal CAD.

In the estimation of such missing band, the missing band estimator 120, together with the high pass filter 121 ₁ extracts a high frequency component of compressed speech signal CAD with a cutoff frequency F _C1, the cutoff frequency F _C2 (> F _C1 ) is high-pass filter portion 121 ₂ having, for extracting a high frequency component of compressed speech signal CAD. Subsequently, in the missing band estimation device 120, the estimation unit 124 outputs the signal HPD ₁ output from the high pass filter unit 121 ₁ (first high pass filter unit) from the high pass filter unit 121 ₂ (second high pass filter unit). the level of the difference signal SBD obtained by subtracting the signal HPD ₂ that is, for calculating the ratio R between the level of the signal HPD _2. When the bit rate of the compressed audio signal CAD is different, the filtering characteristics of the high pass filter sections 121 ₁ and 121 ₂ are set so that the ratio R is different.

  Next, the estimation unit 124 estimates the missing band of the compressed audio signal CAD based on the calculated ratio R. Then, the estimation unit 124 sends, to the variable HPF unit 130, a cutoff frequency designation HPC in which the estimated lower limit frequency of the missing band is designated, to control the high-pass filtering process by the variable HPF unit 130.

  Under such control, the variable HPF unit 130 performs high-pass filtering processing on the signal HMD sent from the harmonic generation unit 110, using the frequency designated by the cutoff frequency designation HPC as the cutoff frequency, Generate signal HBD. Then, the synthesis unit 140 synthesizes the compressed speech signal CAD and the signal HBD.

  Therefore, according to the present embodiment, it is possible to output high-quality voice in which the treble band interpolation is appropriately performed with a simple configuration.

[Modification of the embodiment]
The present invention is not limited to the above embodiments, and various modifications are possible.

  For example, in the above embodiment, the level of the difference signal obtained by subtracting the signal output from the second high pass filter unit from the signal output from the first high pass filter unit, and the level output from the second high pass filter unit The missing band of the input compressed speech signal is estimated based on the ratio to the level of the signal. On the other hand, the frequency band of the input compressed audio signal is estimated based on the ratio between the level of the signal output from the first high pass filter unit and the level of the signal output from the second high pass filter unit. You may do it. Then, cutoff frequency specification in which the upper limit frequency of the estimated frequency band is specified may be performed on the variable HPF section.

  In addition, if the bit rate of the compressed audio signal is different, if the ratio of the level of the signal output from the first high pass filter unit to the level of the signal output from the second high pass filter unit is different, A high pass filter unit having filtering characteristics other than the characteristics of the high pass filter unit exemplified in the above embodiments may be employed.

  Further, although the present invention is applied to the high-frequency interpolation of the compressed audio signal in the above embodiment, the present invention may be applied to the high-frequency interpolation of audio signals other than the compressed audio signal.

  Further, in the above embodiment, the compressed voice decompression device and the sound output device are arranged as devices different from the sound device. On the other hand, the audio device may have the function of a compressed audio decompression device, and the audio device may have the function of a sound output device.

  Note that the sound device of the above embodiment is configured as a computer as an arithmetic unit provided with a DSP (Digital Signal Processor) or the like, and the computer prepares the processing in the above embodiment by executing a program prepared in advance. Some or all of the above may be executed. This program may be acquired in the form of being recorded on a portable recording medium such as a CD-ROM or a DVD, or may be acquired in the form of delivery via a network such as the Internet. .

100 ... acoustic device 110 ... harmonic generation unit 120 ... missing band estimation device (frequency band estimation device)
121 ₁ ... High pass filter (first high pass filter)
121 ₂ ... High pass filter (second high pass filter)
124 ... estimation unit (control unit)
130 ... variable high-pass filter unit 140 ... combining unit

Claims (1)

A harmonic generation unit that generates harmonics of the input audio signal;
A variable high-pass filter unit having a variable cutoff frequency and extracting a high-frequency component of the harmonic generated by the harmonic generation unit;
A first high pass filter unit having a first cutoff frequency and extracting a high frequency component of the input audio signal;
A second high-pass filter unit having a second cutoff frequency higher than the first cutoff frequency and extracting a high-frequency component of the input audio signal;
A control unit that controls the cutoff frequency of the variable high pass filter unit based on the level of the output signal of the first high pass filter unit and the level of the output signal of the second high pass filter unit;
An acoustic device comprising:

Images