JP2010020333A - Scalable coder and scalable decoder - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a scalable coder and a scalable decoder capable of contracting a circuit scale and also reducing a processing operational quantity of coding. <P>SOLUTION: A frequency area conversion part 103 performs the analysis of frequency of a signal sampled at a sampling rate Fx with an analysis length 2×Na to calculate a first spectrum S1(k)(0≤k<Na). A band extension part 104 extends an effective frequency band of the first spectrum S1(k) to 0≤k<Nb to permit the impartment of new spectrum on and after the frequency k=Na of the first spectrum S1(k). An extension spectrum-imparting part 105 imparts an extension spectrum S1'(k)(Na≤k<Nb) which is inputted to an extended frequency band from an external part. A spectrum information specification part 106 outputs information required for specifying the extension spectrum S1'(k) among spectra imparted from the extension spectrum-imparting part 105 as coding codes. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a scalable encoding device that encodes an audio signal, an audio signal, and the like, and a scalable decoding device that decodes a signal encoded by the scalable encoding device.

  Today, there are many different things, such as 44.1 kHz for compact discs, 32 or 48 kHz for digital audio tape (DAT), digital VTR, or satellite television, or 48 or 96 kHz for DVD audio signals. There is a sampling rate. Therefore, when the internal sampling rate of the decoder of the playback device or recording device is different from the sampling rate of the data to be decoded, it is necessary to convert the sampling rate. An example of a conventional apparatus that performs this sampling rate conversion is disclosed in Patent Document 1.

  Also, in recent years, transmission line capacity in networks has been increasing due to the widespread use of ADSL (Asymmetric Digital Subscriber Line) and optical fiber in wired systems, or practical application of W-CDMA (Wideband-Code Division Multiple Access) and wireless LANs in wireless systems. Accordingly, there has been a demand for higher realism and higher quality by expanding the signal band in voice communication.

  Currently, G. standardized by ITU (International Telecommunication Union) is a typical method for encoding narrowband signals. 726, G.G. 729 etc. Moreover, as a typical method for encoding a wideband signal, there are G722 and G722.1 of ITU-T (International Telecommunication Union Telecommunication Standardization Sector), AMR-WB of 3GPP (The 3rd Generation Partnership Project), and the like.

  In recent years, there has been a demand for the implementation of a scalable function in a speech coding system intended to be used in various network environments such as an IP (Internet Protocol) network. The scalable function represents a function capable of decoding an audio signal even from a part of an encoded code. By having this scalable function, packet loss can be achieved by decoding high-quality audio signals using all coded codes in a good channel and transmitting only a part of the coded code in a bad channel. The frequency of occurrence can be suppressed. In addition, it is possible to obtain an effect such as efficiency of network resources at the time of communication between multiple points.

  In order to realize a high-quality encoding method having a scalable function, it is necessary to perform encoding using signals of various sampling rates. For example, a signal with a sampling rate of 8 kHz is standardized by ITU-T. 726, G.G. By performing encoding using a method such as 729 and further encoding the error signal in a region where the sampling rate is 16 kHz, quality improvement and scalability can be realized by extending the signal band.

  FIG. 22 is a block diagram showing a typical configuration of a conventional coding apparatus that performs scalable coding. In this example, the number of layers N = 3, the sampling rate of the signal handled in layer n is represented as FS (n), FS (1) = 16 [kHz], FS (2) = 24 [kHz], FS ( 3) Assume that 32 [kHz].

  The acoustic signal (sound signal, audio signal, etc.) input to the downsampling unit 12 via the input terminal 11 is downsampled from 32 kHz to 16 kHz and supplied to the first layer encoding unit 13. The first layer encoding unit 13 determines the first encoded code so that audible distortion between the input acoustic signal and the decoded signal generated after encoding is minimized. The first encoded code is sent to the multiplexing unit 26 and also sent to the first layer decoding unit 14. The first layer decoding unit 14 generates a first layer decoded signal using the first encoded code. The upsampling unit 15 upsamples the sampling frequency of the first layer decoded signal from 16 kHz to 24 kHz, and supplies this signal to the subtracter 18 and the adder 21.

  The acoustic signal input to the down-sampling unit 16 via the input terminal 11 is down-sampled from 32 kHz to 24 kHz and supplied to the delay unit 17. The delay unit 17 delays the downsampled signal by a predetermined time length. The subtractor 18 obtains a difference between the output signal of the delay unit 17 and the output signal of the upsampling unit 15, generates a second layer residual signal, and provides the second layer encoding unit 19. The second layer encoding unit 19 encodes the second layer residual signal so that the quality improvement is made audibly, determines a second encoded code, and multiplexes the second encoded code To the unit 26 and the second layer decoding unit 20. Second layer decoding section 20 performs a decoding process using the second encoded code, and generates a second layer decoded residual signal. The adder 21 calculates the sum of the first layer decoded signal and the second layer decoded residual signal, and generates a second layer decoded signal. The upsampling unit 22 upsamples the sampling frequency of the second layer decoded signal from 24 kHz to 32 kHz, and supplies this signal to the subtractor 24.

  Further, the acoustic signal input to the delay unit 23 via the input terminal 11 is delayed by a predetermined time length and is given to the subtractor 24. The subtractor 24 takes the difference between the output signal of the delay unit 23 and the output signal of the upsampling unit 22 and generates a third layer residual signal. This third layer residual signal is provided to the third layer encoding unit 25. The third layer encoding unit 25 encodes the third layer residual signal so that quality improvement is made audibly, determines a third encoded code, and sends the encoded code to the multiplexing unit 26. give. The multiplexing unit 26 multiplexes the encoded codes obtained from the first layer encoding unit 13, the second layer encoding unit 19, and the third layer encoding unit 25 and outputs the multiplexed code via the output terminal 27.

JP 2000-68948 A

  However, as described above, G.M. 726 and G.G. In the conventional coding apparatus that realizes the scalable function based on the time domain coding scheme such as 729 or AMR-WB, it is necessary to convert the sampling rate of various signals (in the above example, downsampling is performed). The sampling unit 12, the upsampling unit 15, the downsampling unit 16, and the upsampling unit 22 are required), and the configuration of the encoding device becomes complicated, and there is a problem that the amount of processing computation for encoding increases. In addition, the circuit configuration of the decoding device that decodes the signal encoded by the encoding device becomes complicated, and the amount of processing for decoding increases.

  The present invention has been made in view of this point, and an object of the present invention is to provide a scalable encoding device and a scalable decoding device that can reduce the circuit scale and the amount of processing computation.

  The sampling rate conversion device according to the first aspect of the present invention includes a conversion unit that obtains a first spectrum by performing frequency domain conversion on an input time domain signal, and an expansion unit that extends a frequency band of the obtained first spectrum. And an insertion means for inserting the second spectrum into the expanded frequency band of the expanded first spectrum.

  According to this configuration, an input time-domain signal is converted into a frequency-domain signal, and a signal equivalent to an up-sampled signal in the time domain can be obtained by extending the frequency band of the obtained spectrum. it can.

  An encoding apparatus according to a second aspect of the present invention includes an acquisition means for acquiring first and second time domain signals, and frequency domain transforms of the acquired first and second time domain signals, respectively. Conversion means for obtaining first and second spectra; expansion means for extending the frequency band of the first spectrum in accordance with the frequency band of the second spectrum; and the second spectrum in the expanded frequency band of the first spectrum. A configuration is provided that includes an inserting unit that inserts a spectrum and an encoding unit that encodes the spectrum obtained by the inserting unit.

  According to this configuration, the circuit scale of the encoding device can be reduced, and the amount of processing for encoding can be reduced.

  The encoding apparatus according to the third aspect of the present invention includes a conversion means that obtains a first spectrum at the Na point by performing frequency analysis of an input signal having a sampling frequency of Fx with an analysis length of 2 · Na, An extension means for extending the frequency band of one spectrum to the Nb point, a second spectrum to be inserted into the expanded frequency band of the first spectrum after extension, are specified, and an encoded code representing the second spectrum is output. And a coding means.

  According to this configuration, it is possible to obtain a spectrum of the sampling rate FS = Fx · Nb / Na without performing sampling conversion in the time domain.

  The encoding device according to a fourth aspect of the present invention employs a configuration in which, in the above configuration, the second spectrum is generated based on the first spectrum.

  According to this configuration, since an extended spectrum can be generated based on information obtained by the decoder, a low bit rate can be achieved.

  In the encoding device according to a fifth aspect of the present invention, in the configuration described above, the second spectrum includes Na ≦ of a spectrum obtained by frequency analysis of an input signal having a sampling frequency Fy at 2 · Nb points. A configuration that is determined to be similar to a spectrum included in the frequency band of k <Nb is adopted.

  According to this configuration, since the extended spectrum can be determined based on the spectrum of the original signal, an extended spectrum with higher accuracy can be obtained.

  In the encoding device according to a sixth aspect of the present invention, in the above configuration, the encoding unit divides a frequency band of Na ≦ k <Nb into two or more subbands, and each subband A configuration is adopted in which an encoded code representing the second spectrum is output.

  According to this configuration, it is possible to generate an encoded code having a scalable function.

  The encoding apparatus according to a seventh aspect of the present invention employs a configuration in which, in the above configuration, the signal having the sampling frequency Fx is a signal decoded in a lower layer in hierarchical encoding.

  According to this configuration, the present invention can be applied to hierarchical encoding composed of a plurality of layers of encoding units, and hierarchical encoding can be realized with minimal sampling conversion.

  The decoding apparatus according to the eighth aspect of the present invention comprises: acquisition means for acquiring a first spectrum in a frequency band of 0 ≦ k <Na by performing frequency analysis of a signal having a sampling frequency of Fx with an analysis length of 2 · Na. A decoding means for receiving the encoded code and decoding the second spectrum in the frequency band of Na ≦ k <Nb, and combining the first and second spectrums in the frequency band of 0 ≦ k <Nb A configuration is provided that includes generation means for generating a spectrum and conversion means for converting a spectrum included in a frequency band of 0 ≦ k <Nb into a signal in the time domain.

  According to this configuration, it is possible to decode the encoded code generated by any of the encoding apparatuses described above.

  The decoding device according to a ninth aspect of the present invention employs a configuration in which, in the above configuration, the second spectrum is generated based on a spectrum in a frequency band of 0 ≦ k <Na.

  According to this configuration, it is possible to decode the encoded code by the encoding method for generating the extended spectrum based on the information obtained by the decoder, so that the bit rate can be reduced.

  In the decoding device according to the tenth aspect of the present invention, in the above configuration, the combined frequency band of the combined spectrum obtained by the generating unit is combined with the predetermined width so as to match a predetermined width. And a means for inserting a specified value into the high band part of the spectrum or discarding the high band part of the combined spectrum.

  According to this configuration, even when the bandwidth of the received spectrum fluctuates due to factors such as network conditions, a desired sampling rate is generated in order to generate a decoded signal after processing is performed so that the spectrum bandwidth is constant. Can be stably generated.

  A decoding device according to an eleventh aspect of the present invention employs a configuration in which, in the above configuration, the signal having the sampling frequency Fx is a signal decoded in a lower layer in hierarchical encoding.

  According to this configuration, it is possible to decode an encoded code obtained by hierarchical encoding composed of a plurality of layers of encoding units.

  A transmission / reception device according to a twelfth aspect of the present invention employs a configuration including any of the above-described encoding device or decoding device.

  According to this configuration, it is possible to provide a transmission / reception device having the same function and effect as described above.

  The sampling rate conversion method according to the thirteenth aspect of the present invention includes a conversion step of obtaining a first spectrum by frequency domain conversion of an input time domain signal, and an expansion step of extending the frequency band of the obtained first spectrum. And an insertion step of inserting the second spectrum into the expanded frequency band of the first spectrum after expansion.

  According to this method, a signal equivalent to an upsampled signal in the time domain can be obtained by converting the input time domain signal into a frequency domain signal and extending the frequency band of the obtained spectrum. it can.

  An encoding method according to a fourteenth aspect of the present invention includes an acquisition step of acquiring first and second time domain signals, and first and second frequency domain transforms of the acquired first and second time domain signals, respectively. A conversion step of obtaining first and second spectra, an expansion step of extending a frequency band of the first spectrum in accordance with a frequency band of the second spectrum, and the second frequency band into the expanded frequency band of the first spectrum. An insertion step for inserting a spectrum and an encoding step for encoding the spectrum obtained by the insertion step are provided.

  According to this method, the circuit scale of the encoding device can be reduced, and the amount of processing for encoding can be reduced.

  The encoding method according to the fifteenth aspect of the present invention includes a conversion step of obtaining a first spectrum of Na point by performing frequency analysis of an input signal having a sampling frequency of Fx with an analysis length of 2 · Na, and the obtained first step. An expansion step for extending the frequency band of one spectrum to the Nb point, a second spectrum to be inserted into the expanded frequency band of the expanded first spectrum, are specified, and an encoded code representing the second spectrum is output. And an encoding step.

  According to this method, it is possible to obtain a spectrum of the sampling rate FS = Fx · Nb / Na without performing sampling conversion in the time domain.

  The decoding method according to the sixteenth aspect of the present invention includes an obtaining step of obtaining a first spectrum in a frequency band of 0 ≦ k <Na by performing frequency analysis of a signal having a sampling frequency of Fx with an analysis length of 2 · Na. A decoding step of receiving the encoded code and decoding a second spectrum in a frequency band of Na ≦ k <Nb, and combining the first and second spectrums in a frequency band of 0 ≦ k <Nb A generation step for generating a spectrum and a conversion step for converting a spectrum included in a frequency band of 0 ≦ k <Nb into a signal in a time domain are provided.

  According to this method, the encoded code generated by any one of the encoding methods described above can be decoded.

  According to the present invention, the circuit scale can be reduced and the amount of processing calculations can also be reduced.

FIG. 1 is a block diagram showing the main configuration of a spectrum encoding apparatus according to Embodiment 1 (a) The figure showing the first spectrum, (b) The figure showing the spectrum after the effective frequency band is expanded. Diagram for explaining in principle the effect of the process of extending the effective frequency band of the spectrum FIG. 2 is a block diagram showing a main configuration of a radio transmission apparatus according to Embodiment 1 FIG. 2 is a block diagram showing an internal configuration of the encoding apparatus according to Embodiment 1 FIG. 2 is a block diagram showing an internal configuration of a spectrum encoding unit according to Embodiment 1 FIG. 9 is a block diagram showing variations of the spectrum encoding unit according to Embodiment 1. FIG. 2 is a block diagram showing a main configuration of a radio reception apparatus according to Embodiment 1 FIG. 2 is a block diagram showing an internal configuration of a decoding apparatus according to Embodiment 1 FIG. 3 is a block diagram showing an internal configuration of a spectrum decoding unit according to Embodiment 1 The figure explaining the process performed in the band expansion part which concerns on Embodiment 1. FIG. The figure which showed how a spectrum produced | generated through the process in the coupling | bond part and time domain conversion part which concern on Embodiment 1 (a) The block diagram which showed the main structures of the transmission side when the encoding apparatus which concerns on Embodiment 1 is applied to a wired communication system, (b) The decoding apparatus which concerns on Embodiment 1 is a wired communication system Block diagram showing the main configuration of the receiving side when applied to FIG. 9 is a block diagram showing the main configuration of a decoding apparatus according to Embodiment 2. FIG. 7 is a block diagram showing an internal configuration of a spectrum decoding unit according to Embodiment 2 The figure for demonstrating in detail the process of the correction part which concerns on Embodiment 2. FIG. The figure for demonstrating in detail the process of the correction part which concerns on Embodiment 2. FIG. The figure for further explaining operation | movement of the spectrum decoding part which concerns on Embodiment 2. FIG. The figure for further explaining operation | movement of the spectrum decoding part which concerns on Embodiment 2. FIG. The figure which shows the main structures of the communication system which concerns on Embodiment 3. The figure which shows the main structures of the communication system which concerns on Embodiment 4. The block diagram which showed the typical structure of the conventional encoding apparatus which performs scalable encoding

  The essence of the present invention is that upsampling is performed on the time domain signal by extending the effective frequency band of the spectrum in the frequency domain instead of performing sampling conversion (particularly upsampling) on the input signal. It is to obtain a signal equivalent to the case.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

(Embodiment 1)
FIG. 1 is a block diagram showing the main configuration of spectrum coding apparatus 100 according to Embodiment 1 of the present invention.

  Spectrum coding apparatus 100 according to the present embodiment has sampling rate conversion section 101, input terminal 102, spectrum information specifying section 106, and output terminal 107. In addition, the sampling rate conversion unit 101 includes a frequency domain conversion unit 103, a band extension unit 104, and an extended spectrum addition unit 105.

  A signal sampled at the sampling rate Fx is input to the spectrum encoding device 100 via the input terminal 102.

  The frequency domain transform unit 103 performs frequency analysis of this signal with an analysis length of 2 · Na to transform a time domain signal into a frequency domain signal (frequency domain transformation), and the first spectrum S1 (k) (0 ≦ k). <Na) is calculated. Then, the obtained first spectrum S <b> 1 (k) is given to the band extending unit 104. Here, the frequency analysis uses a modified discrete cosine transform (MDCT). In MDCT, analysis is performed by superimposing adjacent frames in front and back and analysis frames, and the distortion between frames is canceled by using an orthogonal basis in which the first half of the analysis frame is an odd function and the second half is an even function. There is a feature that. As a frequency analysis method, discrete Fourier transform (DFT), discrete cosine transform (DCT), or the like can be used.

  The band extension unit 104 secures a new region (frequency band) so that a new spectrum can be assigned after the frequency k = Na of the input first spectrum S1 (k), and the first spectrum S1 (k) The effective frequency band is extended to 0 ≦ k <Nb. The process of extending the effective frequency band will be described in detail later.

  The extended spectrum giving unit 105 gives an extended spectrum S1 ′ (k) (Na ≦ k <Nb) input from the outside to the frequency band extended by the band extending unit 104, and outputs it to the spectrum information specifying unit 106. .

  The spectrum information specifying unit 106 outputs information necessary for specifying the extended spectrum S <b> 1 ′ (k) out of the spectrum given from the extended spectrum giving unit 105 as an encoded code via the output terminal 107. This encoded code is information representing the subband energy of the extended spectrum S1 '(k), information representing the effective frequency band, and the like. Details of this will also be described later.

  Next, details of the process in which the band extending unit 104 extends the effective frequency band of the first spectrum S1 (k) will be described with reference to FIG.

  2A shows the first spectrum S1 (k) given from the frequency domain transform unit 103, and FIG. 2B shows the spectrum S1 after the effective frequency band is extended by the band extension unit 104. (k) is shown. The band extension unit 104 secures an area where new spectrum information can be stored in a frequency band in which the frequency k of the first spectrum S1 (k) is represented by a range of Na ≦ k <Nb. The size of this new region is represented by Nb-Na.

により、Ｎｂは設定される。また、Ｎｂが決まっているときに、復号化部で復号される信号のサンプリングレートＦｙは次式

により決定される。例えば、Ｎａ＝１２８、Ｆｘ＝１６ｋＨｚの条件で符号化部を設計し、復号化部でＦｙ＝３２ｋＨｚの復号信号を生成する場合には、Ｎｂ＝１２８・３２／１６＝２５６とする必要がある。よって、この場合には、１２８≦ｋ＜２５６の領域を確保することになる。また、別の例としては、Ｎａ＝１２８、Ｎｂ＝３８４、Ｆｘ＝８ｋＨｚの条件で符号化部を設計した場合には、復号化部で生成される復号信号のサンプリングレートは、Ｆｙ＝８・３８４／１２８＝２４ｋＨｚとなる。 Here, Nb is the sampling rate Fx of the signal given from the outside via the input terminal 102, the analysis length 2 · Na of the frequency domain conversion unit 103, and the sampling rate Fy of the signal decoded by the decoding unit (not shown). Determined from the relationship. Specifically, the following formula

Thus, Nb is set. Also, when Nb is determined, the sampling rate Fy of the signal decoded by the decoding unit is

Determined by. For example, when the encoding unit is designed under the conditions of Na = 128 and Fx = 16 kHz, and the decoding unit generates a decoded signal of Fy = 32 kHz, it is necessary to set Nb = 128 · 32/16 = 256. . Therefore, in this case, an area of 128 ≦ k <256 is secured. As another example, when the encoding unit is designed under the conditions of Na = 128, Nb = 384, and Fx = 8 kHz, the sampling rate of the decoded signal generated by the decoding unit is Fy = 8 · 384/128 = 24 kHz.

  FIG. 3 is a diagram for explaining in principle the effect of the process of extending the effective frequency band of the spectrum performed in the band extending unit 104. FIG. 3A shows a spectrum Sa (k) obtained when frequency analysis is performed on the signal of the sampling rate Fx with the analysis length 2 · Na. The horizontal axis represents frequency and the vertical axis represents spectrum intensity.

  The effective frequency band of the signal is 0 to Fx / 2 from the Nyquist theorem. At this time, since the analysis length is 2 · Na, the range of the frequency index k is 0 ≦ k <Na, and the frequency resolution of the spectrum Sa (k) is Fx / (2 · Na). On the other hand, when the spectrum Sb (k) obtained by frequency analysis with the analysis length 2 · Nb after up-sampling the same signal to the sampling rate Fy is shown in FIG. The frequency index k is expanded to Fy / 2, and the range of the frequency index k is 0 ≦ k <Nb. Here, when Nb satisfies (Expression 1), the frequency resolution Fy / (2 · Nb) of the spectrum Sb (k) is equal to Fx / (2 · Na). That is, the spectrum Sa (k) and the spectrum Sb (k) in the band 0 ≦ k <Na are equal. In other words, the spectrum Sb (k) obtained when the band of the spectrum Sa (k) (0 ≦ k <Na) is expanded to Nb is obtained by upsampling the signal of the sampling Fx to the sampling Fy, It means that it matches the spectrum obtained by frequency analysis with Nb. By utilizing this principle, a spectrum equivalent to the upsampled signal can be obtained without upsampling in the time domain.

  In this way, the sampling rate conversion unit 101 converts the input time domain signal into a frequency domain signal and expands the effective frequency band of the obtained spectrum, thereby frequency-sampling the signal up-sampled in the time domain. A spectrum equivalent to the spectrum obtained by conversion can be obtained.

  Since the signal output from the sampling rate conversion unit 101 is a frequency domain signal, if a time domain signal is required, a time domain conversion unit may be provided to perform reconversion to the time domain. . In the above example, since the sampling rate conversion unit 101 is installed in the spectrum encoding device 100, the signal is input to the spectrum information specifying unit 106 as the frequency domain signal without returning to the time domain signal, and the encoded code Is generated.

  Here, by adjusting the selection of the extended spectrum input to the extended spectrum providing unit 105 and the method of specifying the spectrum information in the spectrum information specifying unit 106, the encoded code output from the spectrum information specifying unit 106 is adjusted. The coding rate is different. That is, a part of the processing in the sampling rate conversion unit 101 has a great influence on the encoding. This means that the spectrum encoding apparatus 100 simultaneously realizes conversion and encoding of the sampling rate of the input signal.

  In addition, here, in order to simplify the explanation, the case where the extended spectrum is added to the original spectrum in the extended spectrum adding unit 105 has been described as an example. However, the processing performed in the spectrum information specifying unit 106 is performed using the extended spectrum. Since it is to output information necessary for identification as an encoded code, it is sufficient if the extended spectrum to be assigned is specified. Therefore, the extended spectrum does not necessarily have to be actually assigned. .

  In addition, although up sampling has been described as an example of sampling rate conversion here, the above principle can be applied to down sampling.

  FIG. 4 is a block diagram showing the main configuration of radio transmission apparatus 130 when encoding apparatus 120 according to the present embodiment is installed on the transmission side of the radio communication system.

  The wireless transmission device 130 includes an encoding device 120, an input device 131, an A / D conversion device 132, an RF modulation device 133, and an antenna 134.

  The input device 131 converts the sound wave W11 that can be heard by the human ear into an analog signal that is an electrical signal, and outputs the analog signal to the A / D conversion device 132. The A / D converter 132 converts this analog signal into a digital signal and outputs it to the encoder 120 (signal S1). The encoding device 120 encodes the input digital signal S1 to generate an encoded signal, and outputs the encoded signal to the RF modulation device 133 (signal S2). The RF modulation device 133 modulates the encoded signal S2 to generate a modulated encoded signal, and outputs the modulated encoded signal to the antenna 134. The antenna 134 transmits the modulated encoded signal as a radio wave W12.

  FIG. 5 is a block diagram showing an internal configuration of the encoding apparatus 120 described above. Here, a case where hierarchical coding (scalable coding) is performed will be described as an example.

  The encoding apparatus 120 includes an input terminal 121, a downsampling unit 122, a first layer encoding unit 123, a first layer decoding unit 124, a delay unit 126, a spectrum encoding unit 100a, a multiplexing unit 127, and an output terminal 128. Have

  An acoustic signal S1 having a sampling rate Fy is input to the input terminal 121. The downsampling unit 122 performs downsampling on the signal S1 input via the input terminal 121 to generate and output a signal at the sampling rate Fx. The first layer encoding unit 123 encodes the signal after downsampling, and outputs the obtained encoded code to the multiplexing unit (multiplexer) 127 and also to the first layer decoding unit 124. First layer decoding section 124 generates a first layer decoded signal based on the encoded code.

  On the other hand, the delay unit 126 gives a delay of a predetermined length to the signal S <b> 1 input via the input terminal 121. The magnitude of this delay is the same as the time delay that occurs when the signal passes through the downsampling unit 122, the first layer encoding unit 123, and the first layer decoding unit 124. The spectrum encoding unit 100a performs spectrum encoding using the signal S3 of the sampling rate Fx output from the first layer decoding unit 124 and the signal S4 of the sampling rate Fy output from the delay unit 126 to generate The encoded code S5 is output to the multiplexing unit 127. The multiplexing unit 127 multiplexes the encoded code obtained by the first layer encoding unit 123 and the encoded code S5 obtained by the spectrum encoding unit 100a, and outputs the result as an output code S2 via the output terminal 128. The output code S2 is given to the RF modulation device 133.

  FIG. 6 is a block diagram showing an internal configuration of the spectrum encoding unit 100a. The spectrum encoding unit 100a has the same basic configuration as that of the spectrum encoding device 100 shown in FIG. 1, and the same components are denoted by the same reference numerals and the description thereof is omitted. .

  A feature of the spectrum encoding unit 100a is that an extended spectrum S1 ′ (k) (Na ≦ k <Nb) is given using the spectrum of the input signal S3 having the sampling rate Fy. According to this, since the target signal for determining the extended spectrum S1 '(k) is given, the accuracy of the extended spectrum S1' (k) is improved, and as a result, the quality can be improved.

  The frequency domain converter 112 performs frequency analysis on the signal S4 of the sampling rate Fy input via the input terminal 111 with an analysis length of 2 · Nb, and obtains a second spectrum S2 (k) (0 ≦ k <Nb). Here, it is assumed that the relationship represented by (Expression 1) is established between the sampling frequencies Fx and Fy and the analysis lengths Na and Nb.

  The spectrum information specifying unit 106 determines an encoded code representing the extended spectrum S1 '(k). Here, the extended spectrum S1 ′ (k) is determined using the second spectrum S2 (k) obtained by the frequency domain transform unit 112. The spectrum information specifying unit 106 determines the encoding code through two steps: a step of determining the shape of the extended spectrum S1 '(k) and a step of determining the gain of the extended spectrum S1' (k).

  First, the step of determining the shape of the extended spectrum S1 '(k) will be described below.

に示すように、周波数軸上である固定値Ｃだけ離れた第１スペクトルＳ１(ｋ)を拡張スペクトルＳ１’(ｋ)にコピーする。ここでＣは、あらかじめ定められた固定値であり、Ｃ≦Ｎａの条件を満たす必要がある。この方法では、拡張スペクトルＳ１’(ｋ)の形状を表すための情報は符号化コードとして出力されない。 In this step, the extended spectrum S1 ′ (k) is determined using the band 0 ≦ k <Na of the first spectrum S1 (k). The specific method is as follows:

As shown, the first spectrum S1 (k) separated by a fixed value C on the frequency axis is copied to the extended spectrum S1 ′ (k). Here, C is a predetermined fixed value and needs to satisfy the condition of C ≦ Na. In this method, information for representing the shape of the extended spectrum S1 ′ (k) is not output as an encoded code.

で表される。 As another method, the extended spectrum S1 ′ (k) and the second spectrum S2 (k) are used instead of the fixed value C as described above, using a variable T that takes a value in a predetermined range T _{MIN to} T _MAX. The value T ′ of the variable T when the shapes are the most similar may be output as part of the encoded code. At this time, the extended spectrum S1 ′ (k) is given by

It is represented by

  Next, the step of determining the gain of the extended spectrum S1 ′ (k) performed by the spectrum information specifying unit 106 will be described below.

に従い、パワの偏差Ｖを算出し、この値を量子化して得られるインデックスを符号化コードとして出力端子１０７を介し出力する。 The gain of the extended spectrum S1 ′ (k) is determined so as to coincide with the power of the band Na ≦ k <Nb of the second spectrum S2 (k). Specifically, the following formula

Accordingly, the power deviation V is calculated, and an index obtained by quantizing this value is output as an encoded code via the output terminal 107.

で表される。ここで、ｊはサブバンドの番号を表し、ＢＬ(ｊ)は第ｊサブバンドの最小周波数に相当する周波数インデックス、ＢＨ(ｊ)は第ｊサブバンドの最大周波数に相当する周波数インデックスを表す。このようにサブバンド毎に符号化コードを出力する構成にすることで、スケーラブル機能を実現することができる。 Alternatively, the extended spectrum S1 ′ (k) may be divided into a plurality of subbands, and the encoding code may be determined independently for each subband. In such a case, in the step of determining the shape of the extended spectrum S1 ′ (k), T ′ represented in (Equation 4) may be determined for each subband and output as an encoded code. Only one T ′ may be determined and output as an encoded code. Then, in the step of determining the gain of the extended spectrum S1 ′ (k), a power deviation V (j) is calculated for each subband, and an index obtained by quantizing this value is used as an encoded code. Output via. The amount of power fluctuation for each subband is

It is represented by Here, j represents a subband number, BL (j) represents a frequency index corresponding to the minimum frequency of the jth subband, and BH (j) represents a frequency index corresponding to the maximum frequency of the jth subband. Thus, a scalable function is realizable by setting it as the structure which outputs an encoding code for every subband.

  In addition to the mode for calculating the second spectrum S2 (k) as shown in FIG. 6, as shown in FIG. 7, the mode (spectrum encoding unit 100b) for LPC analysis of the signal at the sampling rate Fy is also possible. good. That is, the LPC coefficient is obtained by LPC analysis of the signal of the sampling rate Fy, and the extended spectrum S1 '(k) can be determined using the LPC coefficient. In this configuration, the LPC coefficient is DFT converted into spectral information, and the extended spectrum S1 '(k) can be determined using this spectrum.

  Thus, according to the encoding apparatus of the present embodiment, the circuit scale of the encoding apparatus can be reduced, and the processing amount of encoding processing can also be reduced.

  In addition to the above effects, when the coding apparatus according to the present embodiment is applied to scalable coding, the following further effects can be obtained.

  When sampling rate conversion is performed in the time domain as in the prior art, it is necessary to pass the input signal through a low-pass filter (hereinafter referred to as LPF) in order to avoid aliasing. In general, when filtering processing is performed in the time domain, a time delay (delay) occurs in the output signal with respect to the input signal. When an FIR type filter is applied to an LPF, it is necessary to increase the filter order in order to make the cut-off characteristic steep, and a time delay corresponding to a sample value that is half the filter order increases with a large increase in the amount of calculation. It will occur.

  For example, when a 256th-order filter is applied to a signal having a sampling frequency Fs = 24 kHz, a delay of 5 ms or more occurs only by sampling rate conversion. The occurrence of such a delay is a problem when it is applied to a two-way voice call and it feels that the reaction of the other party has been delayed.

  Further, when an IIR filter is used for the LPF, the cut-off characteristic can be made steep even if the order is relatively reduced, and the delay is not increased as much as the FIR filter. However, an IIR filter cannot design a filter in which the amount of delay generated at all frequencies is constant, unlike an FIR filter. In scalable coding, when subtracting the signal after sampling rate conversion from the input signal, it is necessary to give a certain delay amount to the input signal in accordance with the time delay of the signal after sampling rate conversion. When the type LPF is used, the amount of delay with respect to the frequency is not constant, so that there is a problem that the subtraction process cannot be performed accurately.

  The encoding apparatus according to the present embodiment can solve these problems that occur in scalable encoding.

  FIG. 8 is a block diagram showing the main configuration of radio receiving apparatus 180 that receives a signal transmitted from radio transmitting apparatus 130.

  The wireless reception device 180 includes an antenna 181, an RF demodulation device 182, a decoding device 170, a D / A conversion device 183, and an output device 184.

  The antenna 181 receives a digital encoded acoustic signal as the radio wave W12, generates a digital received encoded acoustic signal of an electrical signal, and provides the RF demodulator 182 with it. The RF demodulator 182 demodulates the received encoded acoustic signal from the antenna 181 to generate a demodulated encoded acoustic signal S11 and provides it to the decoding device 170.

  The decoding device 170 receives the digital demodulated encoded acoustic signal S11 from the RF demodulation device 182 and performs a decoding process to generate a digital decoded acoustic signal S12 and supplies it to the D / A conversion device 183. The D / A conversion device 183 converts the digital decoded acoustic signal S12 from the decoding device 170 to generate an analog decoded speech signal, and provides it to the output device 184. The output device 184 converts an analog decoded audio signal, which is an electrical signal, into air vibration and outputs the sound wave W13 so that it can be heard by the human ear.

  FIG. 9 is a block diagram showing an internal configuration of the decoding apparatus 170 described above. Here, a case where a hierarchically encoded signal is decoded will be described as an example.

  The decoding apparatus 170 includes an input terminal 171, a separation unit 172, a first layer decoding unit 173, a spectrum decoding unit 150, and an output terminal 176.

  The input terminal 171 receives the code S11 that is hierarchically encoded from the RF demodulator 182. Separating section 172 separates demodulated encoded acoustic signal S11 input via input terminal 171 and generates an encoded code for first layer decoding section 173 and an encoded code for spectrum decoding section 150. . First layer decoding section 173 decodes the decoded signal of sampling rate Fx using the encoded code obtained by separating section 172, and provides this decoded signal S13 to spectrum decoding section 150. The spectrum decoding unit 150 performs spectrum decoding, which will be described later, on the encoded code S14 separated by the separation unit 172 and the signal S13 of the sampling rate Fx generated by the first layer decoding unit 173, and performs the sampling rate Fy. The decoded signal S12 is generated and output through the output terminal 176.

  FIG. 10 is a block diagram showing an internal configuration of the spectrum decoding unit 150 described above.

  The spectrum decoding unit 150 includes input terminals 152 and 153, a frequency domain conversion unit 154, a band extension unit 155, a decoding unit 156, a combining unit 157, a time domain conversion unit 158, and an output terminal 159.

  A signal S13 sampled at the sampling rate Fx is input to the input terminal 152. In addition, the input terminal 153 receives the encoded code S14 related to the extended spectrum S1 '(k).

  The frequency domain transform unit 154 performs frequency analysis on the time domain signal S13 input from the input terminal 152 with the analysis length 2 · Na, and calculates the first spectrum S1 (k). The frequency analysis method uses a modified discrete cosine transform (MDCT). In MDCT, analysis is performed by superimposing adjacent frames in front and back and analysis frames, and the distortion between frames is canceled by using an orthogonal basis in which the first half of the analysis frame is an odd function and the second half is an even function. There is a feature that. The first spectrum S1 (k) obtained in this way is given to the band extending unit 155. As a frequency analysis method, it is possible to use discrete Fourier transform (DFT), discrete cosine transform (DCT), or the like.

  The band extending unit 155 secures a region where a spectrum can be newly added after the frequency k = Na of the input first spectrum S1 (k), and the band of the first spectrum S1 (k) is 0 ≦ k <. Nb is set. The first spectrum S1 (k) whose band is extended is output to the combining unit 157.

  On the other hand, the decoding unit 156 decodes the encoded code S14 related to the extended spectrum S1 ′ (k) input through the input terminal 153, obtains the extended spectrum S1 ′ (k), and outputs the extended spectrum S1 ′ (k) to the combining unit 157.

  The combining unit 157 combines the first spectrum S1 (k) and the extended spectrum S1 '(k) given from the band extending unit 155. This coupling is realized by inserting the extended spectrum S1 '(k) in the band Na ≦ k <Nb of the first spectrum S1 (k). The first spectrum S1 (k) obtained by this processing is output to the time domain conversion unit 158.

  The time domain transform unit 158 performs a time domain transform process corresponding to the inverse transform of the frequency domain transform performed by the spectrum encoding unit 100a, undergoes appropriate window function multiplication and superposition addition, and performs time domain signal S12. Is generated. The time-domain signal S12 generated in this way is output as a decoded signal via the output terminal 159.

  Next, processing performed by the bandwidth extension unit 155 will be described with reference to FIG.

に従い、Ｎｂを設定することができる。また、Ｎｂが決まっているときには、スペクトル復号化部１５０で復号される信号のサンプリングレートＦｙは、次式

により決定される。例えば、入力信号のサンプリングレートがＦｘ＝１６ｋＨｚ、周波数領域変換部１５４の分析長がＮａ＝１２８の条件のときに、スペクトル復号化部１５０でサンプリングレートがＦｙ＝３２ｋＨｚの復号信号を生成する場合には、帯域拡張部１５５でＮｂ＝１２８・３２／１６＝２５６とする必要がある。よって、この場合には、帯域拡張部１５５にて１２８≦ｋ＜２５６の領域を確保することになる。また、別の例として、入力信号のサンプリングレートがＦｘ＝８ｋＨｚ、周波数領域変換部１５４の分析長がＮａ＝１２８、帯域拡張部１５５の拡張量がＮｂ＝３８４のときに、スペクトル復号化部１５０で生成される復号信号のサンプリングレートはＦｙ＝８・３８４／１２８＝２４ｋＨｚとなる。 FIG. 11A shows the first spectrum S1 (k) given from the frequency domain transform unit 154. FIG. FIG. 11B shows a spectrum obtained as a result of the processing of the band extending unit 155, and an area where new spectrum information can be stored in a band where the frequency k is expressed in the range of Na ≦ k <Nb is secured. The size of this new area is represented by Nb-Na. Nb is a relationship between the sampling rate Fx of the signal given from the input terminal 152, the analysis length 2 · Na of the frequency domain transform unit 154, and the sampling rate Fy of the signal decoded by the spectrum decoding unit 150. Depending on

Nb can be set according to When Nb is determined, the sampling rate Fy of the signal decoded by the spectrum decoding unit 150 is given by

Determined by. For example, when the sampling rate of the input signal is Fx = 16 kHz and the analysis length of the frequency domain transform unit 154 is Na = 128, the spectrum decoding unit 150 generates a decoded signal with the sampling rate Fy = 32 kHz. Needs to be Nb = 128 · 32/16 = 256 in the bandwidth extension unit 155. Therefore, in this case, the bandwidth extension unit 155 secures an area of 128 ≦ k <256. As another example, when the sampling rate of the input signal is Fx = 8 kHz, the analysis length of the frequency domain transform unit 154 is Na = 128, and the extension amount of the band extension unit 155 is Nb = 384, the spectrum decoding unit 150 The sampling rate of the decoded signal generated in the above is Fy = 8 · 384/128 = 24 kHz.

  FIG. 12 is a diagram showing how a spectrum is subjected to processing in the combining unit 157 and the time domain conversion unit 158 to generate a decoded signal.

  The combining unit 157 inserts the extended spectrum S1 ′ (k) (Na ≦ k <Nb) into the band of Na ≦ k <Nb of the first spectrum S1 (k) whose band is extended, and the combined spectrum obtained thereby The first spectrum S1 (k) (0 ≦ k <Nb) is sent to the time domain transform unit 158. The time domain transform unit 158 generates a decoded signal in the time domain, and thereby can obtain a decoded signal having a sampling rate FS (= Fx · Nb / Na).

  Thus, according to the decoding apparatus of the present embodiment, it is possible to decode the signal encoded by the encoding apparatus according to the present embodiment.

  Here, the case where the encoding apparatus or decoding apparatus according to the present embodiment is applied to a wireless communication system has been described as an example, but the encoding apparatus or decoding apparatus according to the present embodiment is described below. As shown in FIG. 6, the present invention can also be applied to a wired communication system.

  FIG. 13A is a block diagram showing a main configuration on the transmission side when the coding apparatus according to the present embodiment is applied to a wired communication system. The same components as those shown in FIG. 4 are denoted by the same reference numerals, and the description thereof is omitted.

  The wired transmission device 140 includes an encoding device 120, an input device 131, and an A / D conversion device 132, and an output is connected to the network N1.

  The input terminal of the A / D conversion device 132 is connected to the output terminal of the input device 131. The input terminal of the encoding device 120 is connected to the output terminal of the A / D conversion device 132. The output terminal of the encoding device 120 is connected to the network N1.

  The input device 131 converts the sound wave W11 that can be heard by the human ear into an analog signal, which is an electrical signal, and provides the analog signal to the A / D converter 132. The A / D conversion device 132 converts the analog signal into a digital signal and gives it to the encoding device 120. The encoding device 120 generates a code by encoding the input digital signal and outputs the code to the network N1.

  FIG. 13B is a block diagram showing a main configuration on the receiving side when the decoding apparatus according to the present embodiment is applied to a wired communication system. In addition, the same code | symbol is attached | subjected to the same thing as the component shown in FIG. 8, and the description is abbreviate | omitted.

  The wired receiving device 190 includes a receiving device 191 connected to the network N1, a decoding device 170, a D / A conversion device 183, and an output device 184.

  The input terminal of the receiving device 191 is connected to the network N1. An input terminal of the decryption device 170 is connected to an output terminal of the reception device 191. The input terminal of the D / A conversion device 183 is connected to the output terminal of the decoding device 170. The input terminal of the output device 184 is connected to the output terminal of the D / A converter 183.

  The receiving device 191 receives the digital encoded acoustic signal from the network N1, generates a digital received acoustic signal, and provides it to the decoding device 170. The decoding device 170 receives the received acoustic signal from the receiving device 191, performs a decoding process on the received acoustic signal, generates a digital decoded acoustic signal, and supplies the digital decoded acoustic signal to the D / A conversion device 183. The D / A conversion device 183 converts the digital decoded speech signal from the decoding device 170 to generate an analog decoded speech signal, and provides it to the output device 184. The output device 184 converts an analog decoded acoustic signal, which is an electrical signal, into vibration of the air and outputs the sound wave W13 so that it can be heard by the human ear.

  Thus, according to said structure, the wired transmission / reception apparatus which has the same effect as said wireless transmission / reception apparatus can be provided.

(Embodiment 2)
FIG. 14 is a block diagram showing the main configuration of decoding apparatus 270 according to Embodiment 2 of the present invention. Note that this decoding device 270 has the same basic configuration as that of the decoding device 170 shown in FIG. 9, and the same components are denoted by the same reference numerals and description thereof is omitted.

  The feature of this embodiment is that the decoded signal is converted at a desired sampling rate by correcting the maximum frequency index Nb of the combined first spectrum S1 (k) (0 ≦ k <Nb) to a desired value Nc. Is to generate.

  The spectrum decoding unit 250 includes the encoded code S14 separated by the separation unit 172, the signal S13 of the sampling rate Fx generated by the first layer decoding unit 173, and the coefficient Nc (signal received via the input terminal 271) Spectrum decoding is performed using S21). Then, the decoded signal of the obtained sampling rate Fy is output via the output terminal 176. When the analysis length of the frequency domain transform in the spectrum decoding unit 250 is 2 · Na, the sampling rate Fy of the decoded signal is represented by Fy = Fx · Nc / Na.

  FIG. 15 is a block diagram showing the internal configuration of the spectrum decoding unit 250 described above.

  The coefficient Nc input via the input terminal 271 is given to the correction unit 251 and the time domain conversion unit 158a.

  The correcting unit 251 determines the effective band of the first spectrum S1 (k) (0 ≦ k <Nb) given from the combining unit 157 based on the coefficient Nc (signal S21) given via the input terminal 271. Correct to k <Nc. Then, the first spectrum S1 (k) (0 ≦ k <Nc) after the band correction is given to the time domain conversion unit 158a.

  The time domain conversion unit 158a converts the first spectrum S1 (k) (0 ≦ k <Nc) given from the correction unit 251 under the analysis length 2 · Nc according to the coefficient Nc given via the input terminal 271. Processing is performed, appropriate window function multiplication and overlay addition are performed, and a time domain signal is generated and output through the output terminal 159. The sampling rate of this decoded signal is FS = Fx · Nc / Na.

  16 and 17 are diagrams for explaining the processing of the correction unit 251 in more detail.

  FIG. 16 shows the processing of the correction unit 251 when Nc <Nb. The band of the first spectrum S1 (k) (signal S21) given from the combining unit 157 is 0 ≦ k <Nb. Therefore, the correcting unit 251 deletes the spectrum in the range of Nc ≦ k <Nb so that the band of the first spectrum S1 (k) satisfies 0 ≦ k <Nc. The first spectrum S1 (k) (0 ≦ k <Nc) (signal S22) obtained as a result is given to the time domain converter 158a, and a decoded signal S23 in the time domain is generated. The sampling rate of the decoded signal S23 is FS = Fx · Nc / Na.

  FIG. 17 similarly shows the processing of the correction unit 251, but shows processing when Nc> Nb. The band of the first spectrum S1 (k) (signal S25) given from the combining unit 251 is 0 ≦ k <Nb as in FIG. The correcting unit 251 expands the band of Nb ≦ k <Nc so that the band of the first spectrum S1 (k) becomes 0 ≦ k <Nc, and gives a specific value (for example, zero value) to the region. To do. The first spectrum S1 (k) (0 ≦ k <Nc) (signal S26) obtained as a result is given to the time domain conversion unit 158a, and a time domain decoded signal S27 is generated. The sampling rate of the decoded signal S27 is FS = Fx · Nc / Na.

  The operation of the spectrum decoding unit 250 will be further described with reference to FIGS. 18 and 19.

  First, it is assumed that the encoded code input via the input terminal 153 varies from frame to frame. That is, the band of the first spectrum S1 (k) output from the combining unit 157 includes 0 ≦ k <Na (band R1), 0 ≦ k <Nb1 (band R2), 0 ≦ as shown in FIG. Assume that there are three bands k <Nb2 (band R3) (where Na <Nb1 <Nb2), and one of these bands is selected for each frame.

  FIG. 19A illustrates the operation of the spectrum decoding unit 250 when the coefficient Nc is equal to Nb2, and FIG. 19B illustrates the operation of the spectrum decoding unit 250 when the coefficient Nc is equal to Nb1. FIG.

  In these figures, the spectrum band obtained in the i-th frame is any one of R1, R2, and R3. Process 1 is a process of inserting a zero value into a band of Nb1 ≦ k <Nb2, Process 2 is a process of inserting a zero value into a band of Na ≦ k <Nb2, and Process 3 is a process of deleting a band of Nb1 ≦ k <Nb2. Processing 4 and processing 4 represent processing for inserting a zero value in the band of Na ≦ k <Nb1.

  First, the case of FIG. 19A will be described.

  In this figure, since the spectrum band is R3 in the 0th frame to the 1st frame and the 7th frame to the 8th frame, that is, the band of the first spectrum S1 (k) is 0 ≦ k <Nb2, the correction unit 251 The first spectrum S1 (k) (0 ≦ k <Nb2) is output to the time domain conversion unit 158a without performing any processing.

  Further, in the second to fourth frames and the ninth frame, the spectrum band is R2, that is, the band of the first spectrum S1 (k) is 0 ≦ k <Nb1, and therefore the correcting unit 251 has the first spectrum S1 (k ) Is extended to Nb2 and a zero value is inserted into the band of Nb1 ≦ k <Nb2, and then the first spectrum S1 (k) (0 ≦ k <Nb2) is output to the time domain transform unit 158a.

  On the other hand, in the fifth to sixth frames, the spectrum band is R1, that is, the band of the first spectrum S1 (k) is 0 ≦ k <Na. Therefore, the correction unit 251 sets the band of the first spectrum S1 (k). After extending to Nb2 and inserting a zero value in the range of Na ≦ k <Nb2, the first spectrum S1 (k) (0 ≦ k <Nb2) is output to the time domain conversion unit 158a.

  Next, the case of FIG. 19B will be described.

  In this figure, in the second to fourth frames and the ninth frame, since the spectrum band is R2, that is, the band of the first spectrum S1 (k) is 0 ≦ k <Nb1, the correction unit 251 does not perform any processing. Without applying, the first spectrum S1 (k) (0 ≦ k <Nb1) is output to the time domain conversion unit 158a.

  Further, in the 0th frame to the 1st frame and the 7th frame to the 8th frame, the spectrum band is R3, that is, the band of the first spectrum S1 (k) is 0 ≦ k <Nb2, and therefore the correcting unit 251 has Nb1 ≦ After deleting the band of k <Nb2, the first spectrum S1 (k) (0 ≦ k <Nb1) is output to the time domain conversion unit 158a.

  On the other hand, in the fifth to sixth frames, the spectrum band is R1, that is, the band of the first spectrum S1 (k) is 0 ≦ k <Na. Therefore, the correction unit 251 sets the band of the first spectrum S1 (k). After extending to Nb1 and inserting a zero value in the band of Na ≦ k <Nb1, the first spectrum S1 (k) (0 ≦ k <Nb1) is output to the time domain transform unit 158a.

  As described above, according to the present embodiment, even when the effective frequency band of the received first spectrum S1 (k) fluctuates with time, by giving an appropriate coefficient Nc, a desired sampling rate can be obtained. The decoded signal can be obtained stably.

(Embodiment 3)
FIG. 20 is a diagram showing a main configuration of a communication system according to Embodiment 3 of the present invention.

  The feature of this embodiment is to cope with a case where the effective frequency band of the first spectrum S1 (k) received on the receiving side varies with time depending on the situation (communication environment) of the communication network.

  Hierarchical encoding section 301 performs the hierarchical encoding processing shown in Embodiment 1 on the input signal of sampling rate Fy, and generates a scalable encoded code. Here, the generated encoded code is configured by information on band 0 ≦ k <Ne (R31), information on band Ne ≦ k <Nf (R32), and information on band Nf ≦ k <Ng (R33). Shall be. The hierarchical encoding unit 301 gives this encoded code to the network control unit 302.

  The network control unit 302 transfers the encoded code given from the hierarchical encoding unit 301 to the hierarchical decoding unit 303. Here, the network control unit 302 discards part of the encoded code to be transferred to the hierarchical decoding unit 303 according to the network status. Therefore, the encoded code input to the hierarchical decoding unit 303 is the case where the encoded code composed of the information R31 to R33 and the encoded code of the information R33 are discarded when no encoded code is discarded. Is either an encoded code configured by information R31 and R32, or an encoded code configured by information R31 when the encoded code of information R32 and R33 is discarded.

  Hierarchical decoding section 303 applies the hierarchical decoding method shown in Embodiment 1 or Embodiment 2 to a given encoded code to generate a decoded signal. When Embodiment 1 is applied to hierarchical decoding section 303, sampling rate Fz of the output decoded signal is Fy (because Fz = Fy · Ng / Ng). When Embodiment 2 is applied to hierarchical decoding section 303, the sampling rate of the decoded signal can be set by a desired coefficient Nc, and the sampling rate Fz of the decoded signal is Fy · Nc / Ng. It becomes.

  As described above, according to the present embodiment, even when the effective frequency band of the first spectrum S1 (k) received on the receiving side varies with time depending on the state of the communication network, the receiving side has a desired sampling rate. Can be obtained stably.

(Embodiment 4)
FIG. 21 is a diagram showing a main configuration of a communication system according to Embodiment 4 of the present invention.

  A feature of this embodiment is that one encoded code generated by one hierarchical encoder is simultaneously transmitted to a plurality of hierarchical decoders having different decoding rates (different decoding capabilities). However, the receiving side responds to this and obtains decoded signals having different sampling rates.

  Hierarchical encoding section 401 performs the encoding process shown in Embodiment 1 on the input signal of sampling rate Fy to generate a scalable encoded code. Here, the generated encoded code includes information on the band 0 ≦ k <Nh (R41), information on the band Nh ≦ k <Ni (R42), and information on the band Ni ≦ k <Nj (R43). Shall. Hierarchical encoding section 401 gives this encoded code to first hierarchical decoding section 402-1, second hierarchical decoding section 402-2, and third hierarchical decoding section 402-3, respectively.

  First layer decoding section 402-1, second layer decoding section 402-2, and third layer decoding section 402-3 perform the first or second embodiment on the given encoded code. The decoded signal is generated by applying the hierarchical decoding method shown in FIG. First hierarchy decoding section 402-1 is a decoding process when coefficient Nc = Nj, second hierarchy decoding section 402-2 is a decoding process when coefficient Nc = Ni, and third hierarchy decoding section 402-3 performs a decoding process when the coefficient Nc = Nh.

  First layer decoding section 402-1 performs a decoding process when coefficient Nc = Nj, and generates a decoded signal. The sampling rate F1 of this decoded signal is Fy (because F1 = Fy · Nj / Nj).

  Second layer decoding section 402-2 performs a decoding process when coefficient Nc = Ni, and generates a decoded signal. The sampling rate F2 of this decoded signal is Fy · Ni / Nj.

  Third layer decoding section 402-3 performs a decoding process when coefficient Nc = Nh, and generates a decoded signal. The sampling rate F3 of this decoded signal is Fy · Nh / Nj.

  As described above, according to the present embodiment, the transmitting side can transmit the encoded code without considering the decoding capability of the receiving side, so that the load on the communication network can be suppressed. Also, the decoded signals of these plural types of sampling rates can be generated with a simple configuration and a small amount of calculation.

  The present invention can also be applied to a communication terminal apparatus and a base station apparatus in a mobile communication system, thereby providing a communication terminal apparatus and a base station apparatus having the same effects as described above.

  Here, the case where the present invention is configured by hardware has been described as an example, but it can also be realized by software.

  The present invention has the effect of realizing scalable coding with a simple configuration and a small amount of computation, and can be applied to the use of a communication system such as an IP network.

103, 112, 154 Frequency domain conversion unit 104, 155 Band extension unit 105 Extended spectrum giving unit 106 Spectrum information specifying unit 113 LPC analysis unit 156 Decoding unit 157 Combining unit 158 Time domain conversion unit 251 Correction unit

Claims (4)

First encoding means for encoding a first band of an audio signal or an audio signal;
A second encoding means for encoding a second band of the audio signal or the audio signal,
The second encoding means includes
Conversion means for obtaining a spectrum by frequency domain conversion from a time domain signal at a first sampling rate obtained by the first encoding means;
Generating means for generating, based on the spectrum, an extended spectrum of a bandwidth corresponding to a ratio of the first sampling rate and a second sampling rate corresponding to the second band;
Encoding means for encoding the spectrum and the extended spectrum;
Scalable encoding device. First decoding means for decoding encoded information generated by the scalable encoding device to generate a time-domain signal having a first sampling rate corresponding to the first band of the audio signal or audio signal;
Decoding the encoded information to generate a time domain signal having a second sampling rate corresponding to the second band of the audio signal or the audio signal; and
The second decoding means includes
First transforming means for obtaining a spectrum by frequency domain transform from a time domain signal at a first sampling rate obtained by the first decoding means;
Generating means for decoding encoded information generated by the scalable encoding device and generating an extended spectrum of a bandwidth corresponding to a ratio of the first sampling rate and the second sampling rate;
Second conversion means for obtaining a time domain signal by time domain conversion from the spectrum and the extended spectrum,
Scalable decoding device. A first encoding step of encoding a first band of the audio signal or audio signal;
A second encoding step for encoding a second band of the audio signal or the audio signal,
The second encoding step includes
A conversion step of obtaining a spectrum by frequency domain conversion from a time-domain signal at a first sampling rate obtained in the first encoding step;
Generating an extended spectrum of a bandwidth corresponding to a ratio of the first sampling rate and a second sampling rate corresponding to the second band based on the spectrum;
An encoding step for encoding the spectrum and the extended spectrum;
Scalable encoding method. A first decoding step of decoding time-domain signals having a first sampling rate corresponding to a first band of an audio signal or an audio signal by decoding encoded information generated by the scalable encoding device;
A second decoding step of decoding the encoded information to generate a time domain signal having a second sampling rate corresponding to the second band of the audio signal or the audio signal,
The second decoding step includes
A first transform step for obtaining a spectrum by frequency domain transform from a time domain signal at a first sampling rate obtained in the first decoding step;
A generation step of decoding encoded information generated by the scalable encoding device to generate an extended spectrum of a bandwidth corresponding to a ratio of the first sampling rate and the second sampling rate;
A second transforming step of obtaining a time domain signal from the spectrum and the extended spectrum by time domain transform,
Scalable decoding method.

Images