JP2019144527A - Audio signal coding apparatus, audio signal decoding apparatus, audio signal coding method, and audio signal decoding method - Google Patents

Abstract

To provide a coding technique and a decoding technique for realizing high-quality audio signals while reducing the overall bit rate.SOLUTION: An audio signal coding apparatus (100) includes: a time-frequency transformer (101) that outputs sub-band spectra from an input signal; a sub-band energy quantizer (102); a tonality calculator (103) that analyzes tonality of the sub-band spectra; a bit allocator (104) that selects a second sub-band on which quantization is performed by a second quantizer on the basis of the analysis result of the tonality and quantized sub-band energy, and determines a first number of bits to be allocated to a first sub-band on which quantization is performed by a first quantizer; the first quantizer (106) that performs coding by using the first number of bits; the second quantizer (107) that performs coding by using a pitch filter; and a multiplexer (108).SELECTED DRAWING: Figure 1

Description

  The present disclosure relates to an encoding technique and a decoding technique that improve sound quality of an audio signal such as a voice signal or a music signal.

  An encoding technique for compressing an acoustic signal at a low bit rate is an important technique for realizing effective use of radio waves or the like in mobile communication. Furthermore, in recent years, expectations for improving the quality of telephone conversation voice have increased, and realization of a telephone service with a high sense of reality is desired. In order to realize this, an acoustic signal having a wide frequency band may be encoded at a high bit rate. However, this approach conflicts with the effective use of radio waves and frequency bands.

  Here, as an example, G.I. The acoustic signal encoding technology adopted in the 719 standard (Non-patent Document 1) will be examined.

  G. In the 719 standard, when an acoustic signal is encoded, a predetermined bit is assigned to a spectrum obtained by frequency-converting the acoustic signal. Specifically, the units (units of the required number of bits) for dividing the spectrum into subbands having a predetermined frequency bandwidth and performing quantization by lattice vector quantization in order from the subband with the largest energy are as follows: To distribute.

(1)
One unit is allocated to the subband with the maximum energy among all the subbands.

  Since 1 bit is allocated per spectrum, for example, if the number of spectral samples in a subband is 8, 1 unit is 8 bits (note that the maximum number of bits that can be allocated per spectrum is 9 bits, for example, sub (If the number of spectral samples in the frame is 8, 72 bits can be finally assigned).

(2)
The subband to which 1 unit is allocated lowers the quantization subband energy by 2 levels (6 dB). If the bit allocation to the subband to which 1 unit is allocated exceeds the maximum value (9 bits), it is excluded from the quantization target in the subsequent loop.

(3)
Returning to (1) above, the same processing is repeated.

  FIG. 6 shows the subband energy in each subband. The horizontal axis represents frequency, and the vertical axis represents logarithmic scale amplitude. In the figure, the subband energy is represented by a horizontal line instead of a point, but each width represents the frequency bandwidth of each subband.

  7 and FIG. It is a figure which shows the example of a bit allocation result to each subband at the time of using the encoding method defined by 719 standard. In each figure, the horizontal axis represents frequency, and the vertical axis represents the number of allocated bits. FIG. 7 shows a case where the bit rate is 128 kbit / s, and FIG. 8 shows a case where the bit rate is 64 kbit / s.

  In the case of 128 kbit / s, since there are abundant bit assets that can be allocated, it is possible to allocate 9 bits which is the maximum value to many subbands (spectrums), and the acoustic signal can be kept in high quality. .

  On the other hand, in the case of 64 kbit / s, there is no subband to which 9 bits, which is the maximum value, is assigned, but there is no subband to which no bit is assigned. It can be said that both effective use of the frequency band can be achieved.

Special table 2013-534328 gazette International Publication No. 2005/027095

ITU-T Standard G. 719, 2008

  However, more effective use of radio waves and frequency bands is required. Here, G. When encoding an acoustic signal having a sampling frequency of about 32 kHz at a low bit rate of about 20 kbp / s or less using the above method adopted in the 719 standard, a unit for quantizing all subbands ( There is a problem that the number of bits) cannot be secured.

  FIG. 9 shows G.M. at 20 kbit / s. It is a figure which shows the example of a bit allocation result to each subband at the time of using the encoding method defined by 719 standard. In this way, as a result of not being able to allocate bits not only to the high frequency part but also to the auditory important low frequency part, the spectrum in that subband cannot be encoded, and the quality of the acoustic signal Deterioration becomes remarkable.

  On the other hand, it is conceivable to adopt a method of dynamically changing the bit allocation method (Patent Document 1).

  However, by changing the bit allocation method with a single encoding method (quantization method) without changing the encoding method (quantization method), there is a limit in countermeasures against quality degradation of the acoustic signal.

  The present disclosure provides an encoding technique and a decoding technique for realizing a high-quality acoustic signal while reducing the overall bit rate.

  An acoustic signal encoding device according to the present disclosure includes a time-frequency conversion unit that converts an input acoustic signal into a frequency domain to generate a spectrum, divides the spectrum into subbands for each predetermined frequency band, and outputs a subband spectrum. , Subband energy quantization unit for obtaining quantized subband energy for each subband, tonality calculation unit for analyzing the tonal property of the subband spectrum and outputting the analysis result, tonal property analysis result and quantization subband Based on the energy, the second subband to be quantized by the second quantizer is selected from the subbands, and the first number of bits allocated to the first subband to be quantized by the first quantizer is determined. The bit allocation unit, the encoding information output from the first quantization unit and the second quantization unit, the quantization subband energy, and the tonal property The information including the analysis result and multiplexes, constituting a multiplexing unit to output. The first quantizing unit pulse-codes the subband spectrum included in the first subband using the first bit number, and the second quantizing unit includes the subband included in the second subband. The spectrum is encoded using a pitch filter.

  Note that these comprehensive or specific aspects may be realized by a system, method, integrated circuit, or computer program, or realized by any combination of the system, apparatus, method, integrated circuit, and computer program. Also good.

  According to the encoding device, the decoding device, and the like of the present disclosure, it is possible to encode and decode a high-quality acoustic signal while reducing the overall bit rate.

Configuration diagram of encoding apparatus according to Embodiment 1 of the present disclosure Detailed configuration diagram of the bit distribution unit of the encoding device according to the first embodiment of the present disclosure Explanatory drawing which shows operation | movement of the encoding apparatus in Embodiment 1 of this indication. Configuration diagram of decoding apparatus according to Embodiment 2 of the present disclosure Detailed configuration diagram of the bit distribution unit of the decoding device according to the second embodiment of the present disclosure Explanatory drawing explaining the subband energy in the encoding apparatus of a prior art Explanatory drawing explaining the bit allocation result to the subband in the encoding apparatus of a prior art Explanatory drawing explaining the bit allocation result to the subband in the encoding apparatus of a prior art Explanatory drawing explaining the bit allocation result to the subband in the encoding apparatus of a prior art

  Hereinafter, the configuration and operation of the embodiment of the present disclosure will be described with reference to the drawings. Note that the input signal to the encoding device of the present disclosure and the acoustic signal that is the output signal from the decoding device are a concept that includes an audio signal, a wider-band music signal, and a signal in which these signals are mixed.

  In the present disclosure, the “input sound signal” is a concept including a music signal, a sound signal, or a signal in which both are mixed. In addition, “quantized subband energy” is obtained by quantizing the subband energy that is the sum or average of the energy of the subband spectrum in the subband, and the subband energy is, for example, the subband spectrum in the subband. The sum of squares of “Tonality” refers to the degree to which a spectrum peak stands at a specific frequency component, and the analysis result can be expressed by a numerical value or a sign. “Pulse coding” refers to coding that approximates a spectrum using pulses.

  The term “relatively low” refers to a lower value compared between subbands, for example, a case where it is lower than the average of all subbands or a case where it is lower than a predetermined value. “High frequency subband” refers to a subband located on the high frequency side among a plurality of subbands.

  In addition, the first (spectrum) quantization unit, the second (spectrum) quantization unit, the first (spectrum) decoding unit, the second (spectrum) decoding unit, the first sub, described in the embodiments and claims. The band, the second subband, the third subband, the fourth subband, the first bit number, the second bit number, the third bit number, and the fourth bit number each mean a category. , Does not mean the order.

(Embodiment 1)
FIG. 1 is a block diagram illustrating the configuration and operation of the acoustic signal encoding device 100 according to the first embodiment. 1 includes a time-frequency conversion unit 101, a subband energy quantization unit 102, a tonality calculation unit 103, a bit distribution unit 104, a normalization unit 105, a first spectrum quantization unit 106, The second spectrum quantization unit 107 and the multiplexing unit 108 are included. In addition, the antenna A is connected to the multiplexing unit 108. Then, the acoustic signal encoding apparatus 100 and the antenna A are combined to constitute a terminal apparatus or a base station apparatus.

  The time-frequency conversion unit 101 converts an input acoustic signal in the time domain into a frequency domain to generate an input acoustic signal spectrum (hereinafter referred to as “spectrum”). Examples of time-frequency conversion include MDCT (Modified Discrete Cosine Transform), but the present invention is not limited to this. For example, DCT (Discrete Cosine Transform), DFT (Discrete Fourier Transform), Fourier Transform, and the like may be used.

  In addition, the time-frequency conversion unit 101 divides the spectrum into subbands that are predetermined frequency bands. In addition to the case where the predetermined frequency bands are equally spaced, the predetermined frequency bands may be different intervals, for example, wide in the high frequency range and narrow in the low frequency range.

  Time-frequency conversion section 101 then outputs the spectrum divided for each subband to subband energy quantization section 102, tonality calculation section 103, and normalization section 105 as a subband spectrum.

  The subband energy quantization unit 102 obtains subband energy that is energy of a subband spectrum for each subband, and quantizes the subband energy to obtain quantized subband energy. Specifically, the subband energy can be obtained from the sum of squares of the subband spectra in the subband, but the present invention is not limited to this. For example, the subband energy can be obtained by integrating the amplitude of the subband spectrum for each subband. In addition, when averaging the subband energy, the sum of squares is divided by the number of spectra in the subband (subband width). Then, the subband energy thus obtained is quantized with a predetermined step size.

  Then, the obtained quantization subband energy is output to normalization section 105 and bit distribution section 104, and the encoded quantization subband energy obtained by encoding the quantization subband energy is output to multiplexing section 108. .

  The tonality calculation unit 103 analyzes the subband spectrum included in each subband and determines tonalness. Tonal property refers to the degree to which a spectrum peak stands at a specific frequency component, and is a concept that includes a peak property that means that a distinct peak exists. Quantitatively, for example, it can be obtained by the ratio of the amplitude of the average spectrum in the target subband and the amplitude of the maximum spectrum existing in the subband, and when this value exceeds a predetermined threshold, The spectrum of the subband is defined as having a tonal property (peak property). In the present embodiment, when a predetermined threshold is exceeded, 1 is generated as the peak / tonal flag, and when it is equal to or lower than the predetermined threshold, 0 is generated as the peak / tonal flag. The data is output to the multiplexing unit 108. Of course, the above ratio may be directly output as an analysis result.

  The significance of the tonality calculator is as follows.

  Under low bit rate conditions, a pitch filter-based method (ie low frequency spectrum is used for efficient quantization of a spectrum where the spectrum energy is distributed across the entire subband, such as a noisy spectrum. It is effective to use a method of expressing a high-frequency spectrum using this method. Therefore, the degree of energy dispersion in the subband is determined from the peak / tonal scale of the spectrum in the subband (such as the ratio of peak power to average power), and the spectrum sub- Bands are subject to quantization based on pitch filters.

  The bit allocation unit 104 refers to the quantized subband energy for each subband and the peak / tonal flag, and means the total number of bits that can be used for encoding for the subband spectrum in each subband. Allocate bits from bit assets. Specifically, a first number of bits, which is the number of bits assigned to the first subband that is a subband quantized by the first spectrum quantization unit, is calculated and determined, and this is calculated by the first spectrum quantization unit 106. Output as distribution bit information. Further, the second subband, which is a subband to be quantized by the second spectrum quantization unit 107, is selected and specified, and this is output to the second spectrum quantization unit 107 as a quantization mode.

  Details of the configuration and operation of the bit distribution unit 104 will be described later.

  In this embodiment, the bit allocation unit 104 refers to the peak / tonal flag and the quantized subband energy for each subband, but the reference order is arbitrary.

  In addition, the second subband to be quantized by the second spectrum quantization unit 107 may be the entire band, but generally, a band with a low quantization subband energy and a band with a low tonal property are Since it is mainly in the high frequency range, only a subband existing in a specific high frequency range may be targeted. For example, only four or five subbands in the high frequency range can be targeted.

  Alternatively, the acoustic signal usually has a high tonal property on the low frequency region side and a low tonal property on the high frequency region side, so that the subband on the high frequency region side is substantially subject to quantization based on the pitch filter. For this reason, a method may be used in which the high frequency region side from the subband selected by tonal property is all subject to quantization by the pitch filter, and only the subband number is transmitted as the quantization mode.

  The normalization unit 105 generates a normalized subband spectrum by normalizing (dividing) each subband spectrum with the input quantized subband energy. Thereby, the difference in amplitude between subbands is normalized. Then, normalization section 105 outputs the normalized subband spectrum to first spectrum quantization section 106 and second spectrum quantization section 107.

  Note that the normalization unit 105 has an arbitrary configuration.

  In addition, the normalization unit 105 has one configuration in the present embodiment, but two normalization units 105 may be arranged in front of each of the first spectrum quantization unit 106 and the second spectrum quantization unit 107.

  The first spectrum quantizing unit 106 is an example of a first quantizing unit, and uses the bit composed of the first number of bits allocated by the bit distributing unit 104 to input the normalized subband spectrum. The first spectrum quantization unit 106 quantizes the subband spectrum belonging to the first subband to be quantized. Then, the quantization result is output to the second spectrum quantization unit 107 as a quantized spectrum, and the first encoded information generated by encoding the quantized spectrum is output to the multiplexing unit 108.

  The first spectrum quantization unit 106 uses a pulse encoding unit. As an example of the pulse encoding unit, a lattice vector quantization unit that performs lattice vector quantization, and a pulse that performs pulse encoding that approximates a subband spectrum with a small number of pulses. An encoding unit may be mentioned. In other words, any quantization unit can be used as long as it is a quantization method suitable for quantization of a spectrum with high tonal characteristics or a method of quantization with a small number of pulses.

  Note that at a very low bit rate, the effect of maintaining the sound quality can be expected by quantization by pulse coding that approximates the subband spectrum with a small number of pulses rather than lattice vector quantization.

  The second spectrum quantizing unit 107 is an example of a second quantizing unit, and can employ a quantization method based on the following extension band (prediction model using a pitch filter), for example.

  Here, the pitch filter is a processing block that performs processing represented by the following Expression 1.

  In general, the pitch filter refers to a filter that emphasizes the pitch period (T) with respect to the time-axis signal (emphasizes the pitch component on the frequency axis). When the number of taps is 1, the discrete signal x For example, a digital filter represented by Formula 1 with respect to [i]. However, the pitch filter in the present embodiment is defined as a processing block that performs the processing represented by Expression 1, and does not necessarily perform pitch emphasis on a time-axis signal.

  In the present embodiment, the pitch filter (processing block represented by Expression 1) is applied to the quantized MDCT coefficient sequence Mq [i]. Specifically, in Equation 1, x [i] = 0 (i ≧ K, K is the lower frequency limit of the MDCT coefficient to be encoded), y [i] = Mq [i] (i <K), and y [ i] (K ≦ i ≦ K ′, where K ′ is the upper frequency limit of the MDCT coefficient to be encoded). T that minimizes the error between the MDCT coefficient Mt [i] to be encoded and the calculated y [i] is encoded as lag information. Such spectral coding based on the pitch filter is disclosed in Patent Document 2 and the like.

  The second spectrum quantization unit 107 specifies a second subband (normalized subband spectrum) to be quantized by the second spectrum quantization unit 107 with reference to the quantization mode. Thereby, the K and K ′ are specified. A normalized subband spectrum (corresponding to Mt [i], K ≦ i ≦ K ′) relating to the specified second subband (frequency K to K ′) is converted into a quantized spectrum (Mq [i], The subband or band of the quantized spectrum having the maximum correlation in relation to i <K) is searched, and the position is generated as lag information (corresponding to T). Examples of the lag information include the absolute position and relative position of subbands and bands, or subband numbers. Then, the second spectrum quantization unit 107 encodes the lag information and outputs the encoded lag information to the multiplexing unit 108 as second encoded information.

  In the present embodiment, the encoded quantization subband energy is multiplexed and transmitted by the multiplexing unit 108, and the gain can be generated on the decoding unit side, so the gain is not encoded. However, the gain may be encoded and sent. In that case, the gain between the second subband to be quantized and the subband of the quantized spectrum having the maximum correlation is calculated, and the second spectrum quantization unit 107 encodes the lag information and the gain, It outputs to the multiplexing part 108 as 2nd encoding information.

  In general, the bandwidth of the high frequency subband is set wider than that of the low frequency subband, but the lattice vector quantum is low because the energy of the subband of the low frequency region to be copied is small. In some cases, it may not be a target of conversion. In such a case, such a subband may be regarded as a zero spectrum, or noise may be added to avoid a sudden change in spectrum between subbands.

  Multiplexing section 108 multiplexes the quantized subband energy, the first encoded information, the second encoded information, and the peak / tonal flag, and outputs the multiplexed information to antenna A.

  Then, the antenna A transmits the encoded information to the acoustic signal decoding device. The encoded information reaches the acoustic signal decoding apparatus via various nodes and base stations.

  Next, details of the bit distribution unit 104 will be described.

  FIG. 2 is a block diagram illustrating a detailed configuration and operation of the bit distribution unit 104 of the audio signal encoding device 100 according to the first embodiment. The bit distribution unit 104 illustrated in FIG. 2 includes a bit reservoir 111, a bit reservoir 112, a bit distribution calculation unit 113, and a quantization mode determination unit 114.

  The bit reservoir 111 refers to the peak / tonal flag output from the tonality calculation unit 103. When the peak / tonal flag is 0, the bit reservoir 111 is a bit required for the second spectrum quantization performed by the second spectrum quantization unit 107. Secure the number.

  In this embodiment, the number of bits necessary for encoding the lag information is secured based on the pitch filter. Then, the reserved number of bits is removed from the bit assets that are the total number of bits that can be used for quantization, and the remaining bit assets are output to the bit reservoir 112. The bit assets are supplied from the subband energy quantization unit 102. This is because the bits excluding the number of bits necessary for variable-length coding of the quantized subband energy are the first spectrum quantization unit. 106, the second spectrum quantization unit 107, and the peak / tonal flag can be used for quantization (encoding). The subband energy quantization unit 102 does not always generate bit asset information.

  The bit reservoir 112 secures the number of bits used for the peak / tonal flag. For example, in the present embodiment, since the peak / tonal flag is sent in 5 subbands in the high frequency range, the bit reservoir 112 reserves 5 bits.

  Then, the bit reservoir 112 outputs the number of bits obtained by subtracting the number of bits secured in the bit reservoir 112 from the bit asset input from the bit reservoir 111 to the bit allocation calculation unit 113 in the adaptive bit allocation unit. Note that the total number of bits secured in the bit reservoir 111 and the bit reservoir 112 is the third bit number. A subband having a peak / tonal flag of zero corresponds to the third subband.

  The order of the bit reservoir 111 and the bit reservoir 112 may be changed. Further, in this embodiment, the bit reservoir 111 and the bit reservoir 112 block are separated, but this may be performed simultaneously in one block. Alternatively, these operations may be performed in the bit allocation calculation unit 113.

  The bit allocation calculation unit 113 calculates the bit allocation to the subbands quantized by the first spectrum quantization unit 106. Specifically, first, the number of bits output from the bit reservoir 112 is allocated to each subband with reference to the quantized subband energy. As described in the section of the prior art, the allocation method determines whether the quantization subband energy is large or small and whether it is important auditoryly or not, and assigns bits to the subbands that are considered important. As a result, no bits are allocated to subbands whose quantization subband energy is zero, or zero and lower than a predetermined value.

  Further, in the allocation, the input peak / tonal flag is referred to, and the subband (third subband) in which the peak / tonal flag is 0 is excluded from the bit allocation target. That is, bits are allocated using only the subbands with high peak characteristics (subbands where the peak / tonal flag is set to 1 here) as the target subbands for bit allocation. Then, the subband (first subband) to which the bits are to be allocated is specified, and the number of bits allocated to each subband is combined to be allocated bit information, which is first output to the quantization mode determining unit 114.

  The quantization mode determination unit 114 receives the allocation bit information and the peak / tonal flag output from the bit allocation calculation unit 113. If there is a high frequency band subband that is not to be bit-distributed even though the tonal property is high (which is the quantization target of the first spectrum quantization unit 106), the subband is converted by the second spectrum quantization unit 107. Redefine the subband to be quantized (fourth subband), and calculate bit allocation to subtract the number of bits (fourth bit) required for quantization in the second spectrum quantizer from the allocated bit information Output to the unit 113. That is, the number of bits necessary for quantization by the second spectrum quantization unit 107 is assigned to the band, and the assigned number of bits (fourth bit number) is output. Alternatively, the number of allocated bits may be subtracted from the bit assets that can be used by the first spectrum quantization unit 106 and output to the bit allocation calculation unit 113.

  Also, the quantization mode determination unit 114 identifies the subband to be quantized by the second spectrum quantization unit 107 and outputs this to the second spectrum quantization unit 107 as a quantization mode. Specifically, a high frequency region subband (third subband) with low tonality (peak / tonal flag is 0) and a high frequency region subband (fourth subband) to which no bits are allocated, A subband (second subband) to be quantized by the second spectrum quantization unit 107 is determined and output as a quantization mode.

  The bit allocation calculation unit 113 again updates the bit asset by subtracting the bit number (fourth bit number) received from the quantization mode determination unit 114 from the bit number (bit asset) input from the bit reservoir 112, The bit allocation to the subbands to be quantized by the first spectrum quantization unit 106 is recalculated. When the updated bit asset is received from the quantization mode determination unit, the bit allocation to the subbands to be quantized by the first spectrum quantization unit 106 is recalculated using the updated bit asset. Finally, the first bit number is a value obtained by subtracting the third bit number and the fourth bit number from the total bit number (bit asset).

  Then, the number of bits after recalculation (first bit number) and the information of the subband (first subband) quantized by the first spectrum quantization unit 106 are used as distribution bit information, and this time, the first spectrum quantum. To the conversion unit 106.

  When the bit allocation is calculated by the bit allocation calculation unit 113 for the first time, if no sub-calculation is necessary, for example, any subband is allocated, direct allocation bit information is converted to the first spectrum quantization unit. 106 may be output.

  FIG. 3 is a flowchart illustrating the operation of the audio signal encoding device 100 according to the first embodiment, specifically, the operation of the bit distribution unit 104.

  First, the bit allocation unit 104 acquires the quantized subband energy from the subband energy quantization unit 102 (S1).

  Next, the bit allocation unit 104 acquires the peak / tonal flag in the high frequency range from the tonality calculation unit 103 (S2).

  Then, the bit distribution unit 104 specifies the subband (third subband) to be quantized by the second spectrum quantization unit 107 based on the peak / tonal flag, and in the bit reservoir 111 and the bit reservoir 112, A bit (third bit number) to be quantized by the spectrum quantization unit 107 is secured (S3).

  In the bit allocation calculation unit 113, the bit allocation unit 104 determines the number of bits to be allocated to the subbands to be quantized by the first spectrum quantization unit 106 based on the quantization subband energy (S4).

  The bit allocation unit 104 checks the allocation bit to the high frequency band subband determined by the bit allocation calculation unit 113 in the quantization mode determination unit 114 and quantizes the second spectrum quantization unit 107 as necessary. The power subband (second subband) is re-specified, and the bit asset for the first subband quantization unit 106 is updated (S5).

  Finally, the bit allocation unit 104 recalculates the bit allocation (first bit number) to the first spectrum quantization unit 106 by using the updated bit asset again in the bit allocation calculation unit 113 ( S6).

  As described above, according to the audio signal encoding device of the present embodiment, it is possible to realize high-quality audio signal encoding while reducing the overall bit rate.

  In particular, according to the configuration and operation of FIGS. 2 and 3, the first frequency can be generated without generating a subband without quantization (bit allocation is 0) in a high frequency region where the subband width is particularly wide. It is possible to realize bit allocation that maximizes the number of subbands quantized by the quantization unit. Therefore, it is possible to realize adaptive bit allocation that can bring out the best performance at a limited bit rate.

(Embodiment 2)
FIG. 4 is a block diagram illustrating a configuration and an operation of the acoustic signal decoding device 200 according to the second embodiment. The acoustic signal decoding apparatus 200 illustrated in FIG. 4 includes a separation unit 201, a subband energy decoding unit 202, a bit distribution unit 203, a first spectrum decoding unit 204, a second spectrum decoding unit 205, a denormalization unit 206, and a frequency-time. The conversion unit 207 is configured. An antenna A is connected to the separation unit 201. Then, the acoustic signal decoding device 200 and the antenna A are combined to constitute a terminal device or a base station device.

  The separation unit 201 receives the encoded information received by the antenna A, and separates the encoded quantization subband energy, the first encoded information, the second encoded information, and the peak / tonal flag. The encoded quantization subband energy is a subband energy decoding unit 202, the first encoded information is a first spectrum decoding unit 204, the second encoded information is a second spectrum decoding unit 205, and the peak / tonal flag is a bit. To the distribution unit 203.

  The subband energy decoding unit 202 decodes the encoded quantized subband energy, generates decoded quantized subband energy, and outputs the decoded quantized subband energy to the bit distribution unit 203 and the inverse normalization unit 206.

  The bit allocation unit 203 refers to the decoded quantization subband energy for each subband and the peak / tonal flag, and determines the allocation of bits to be allocated by the first spectrum decoding unit 204 and the second spectrum decoding unit 205. Specifically, the number of bits (first bit number) to be assigned when the first spectrum decoding unit 204 decodes the first encoded information and the subband (first subband) to which the bits are assigned are determined and distributed. The sub-band (second sub-band) to be decoded is output to the second spectrum decoding unit 205 as the bit information, and the second encoded information decoded by the second spectrum decoding unit 205 is to be decoded. Output as quantization mode.

  Since the bit distribution unit 203 is the same as the configuration and operation of the bit distribution unit 104 described on the encoding device side as shown in FIG. 5, details of the operation are described in the description of the bit distribution unit 104 on the encoding device side. Quote.

  The first spectrum decoding unit 204 decodes the first encoded information using the first number of bits indicated in the allocated bit information, generates a first decoded spectrum, and outputs the first decoded spectrum to the second spectrum decoding unit 205.

  The second spectrum decoding unit 205 generates the second decoded spectrum by decoding the second encoded information using the first decoded spectrum in the subband specified in the quantization mode, and the second decoded spectrum and the first Combined with the decoded spectrum, a reproduction spectrum is generated and output.

  The inverse normalization unit 206 adjusts the amplitude (gain) of the reproduction spectrum with reference to the decoded quantized subband energy, and outputs this to the frequency-time conversion unit 207.

  The frequency-time conversion unit 207 converts the reproduction spectrum in the frequency domain into an output acoustic signal in the time domain and outputs it. As an example of the frequency-time conversion, there is an inverse conversion of the conversion given in frequency-time.

  As described above, according to the acoustic signal decoding device of the present embodiment, it is possible to realize high-quality acoustic signal decoding while reducing the overall bit rate.

(Summary)
As described above, the acoustic signal encoding device and the acoustic signal decoding device of the present disclosure have been described in the first and second embodiments. The encoding device and the decoding device of the present disclosure may be in a semi-finished product or component level form as represented by a system board or a semiconductor element, and also include a finished product level form such as a terminal device or a base station device. It is a concept. When the encoding device and the decoding device according to the present disclosure are in a semi-finished product or a component level form, they are combined with an antenna, a DA / AD converter, an amplifying unit, a speaker, a microphone, and the like to obtain a finished product level form.

  The block diagrams of FIGS. 1, 2, 4, and 5 represent the configuration and operation (method) of hardware designed exclusively, and execute the operation (method) of the present disclosure on general-purpose hardware. Including a case where the program is realized by installing a program for executing the program and executing the program on the processor. Examples of general-purpose hardware electronic computers include personal computers, various portable information terminals such as smartphones, and mobile phones.

  Moreover, the hardware designed exclusively is not limited to a finished product level (consumer electronics) such as a mobile phone and a fixed phone, but includes a semi-finished product and a component level such as a system board and a semiconductor element.

  The acoustic signal encoding device and the acoustic signal decoding device according to the present disclosure can be applied to an apparatus related to recording, transmission, and reproduction of an acoustic signal.

DESCRIPTION OF SYMBOLS 100 Acoustic signal encoding apparatus 101 Time-frequency conversion part 102 Subband energy quantization part 103 Tonality calculation part 104 Bit allocation part 105 Normalization part 106 1st spectrum quantization part 107 2nd spectrum quantization part 108 Multiplexing part 111 Bit reservoir 112 Bit reservoir 113 Bit allocation calculation unit 114 Quantization mode determination unit 200 Acoustic signal decoding device 201 Separation unit 202 Subband energy decoding unit 203 Bit allocation unit 204 First spectrum decoding unit 205 Second spectrum decoding unit 206 Denormalization Section 207 Frequency-time conversion section 211 Bit reservoir 212 Bit reservoir 213 Bit allocation calculation section 214 Quantization mode determination section

Claims (17)

A first encoding unit that generates a first encoded signal by encoding a low frequency signal having a predetermined frequency or less of a voice or audio input signal, and generates a low frequency decoded signal by decoding the first encoded signal;
A second encoding unit that generates a high frequency encoded signal by encoding a high frequency signal from the low frequency signal based on the low frequency decoded signal;
A first multiplexing unit that multiplexes the first encoded signal and the high-frequency encoded signal and outputs an encoded signal;
The second encoding unit includes:
Calculate the energy ratio between the high frequency noise component, which is the noise component of the high frequency signal, and the high frequency non-tonal component of the high frequency decoded signal generated from the low frequency decoded signal, and output it as a high frequency encoded signal To
Encoding device. An energy calculator that calculates energy of the voice or audio input signal and outputs the energy as quantized band energy; and
The first multiplexing unit multiplexes the quantized band energy, the first encoded signal, and the high frequency encoded signal, and outputs the result.
The encoding device according to claim 1. The second encoding unit includes:
A separation unit that separates the low frequency decoded signal into a low frequency non-tonal signal that is a non-tonal component of the low frequency decoded signal and a low frequency tonal signal that is a tonal component of the low frequency decoded signal;
A first band extension unit that outputs position information of a specific band that maximizes a correlation between the high-frequency signal and the low-frequency tonal signal as lag information;
A second band extending unit that outputs the low frequency non-tonal signal corresponding to the lag information as a high frequency non-tonal signal;
A first calculator that calculates energy of a high-frequency noise component that is a noise component from the high-frequency signal corresponding to the lag information;
Calculating the ratio from the energy ratio of the high-frequency noise component and the high-frequency non-tonal signal, and outputting as a scale factor;
A second multiplexing unit that multiplexes the lag information and the scale factor and outputs the result as a high frequency encoded signal,
The encoding device according to claim 2. The second encoding unit includes:
A noise adding unit that adds a noise signal to the low-frequency decoded signal;
The encoding device according to claim 3. The second encoding unit includes:
A noise addition unit that adds a noise signal to the low-frequency non-tonal signal output from the separation unit;
The encoding device according to claim 3. In the encoding device, a first encoded signal obtained by encoding a low frequency signal having a frequency equal to or lower than a predetermined frequency of a voice or audio input signal, and a high frequency encoded signal obtained by encoding a higher frequency signal than the low frequency signal are input. A decoding device comprising:
A separation unit that separates the first encoded signal and the high-frequency encoded signal;
A first decoding unit for decoding the first encoded signal to generate a low-frequency decoded signal;
A second decoding unit that decodes the highband encoded signal and generates a wideband decoded signal using the lowband decoded signal,
The high frequency encoded signal includes an energy ratio between a high frequency noise component which is a noise component and a high frequency non-tonal component of a high frequency decoded signal generated from the low frequency decoded signal,
The second decoding unit
Adjusting the amplitude of the low frequency non-tonal signal, which is a non-tonal component of the low frequency decoded signal, with reference to the decoded ratio;
Decoding device. In the encoding device, a first encoded signal obtained by encoding a low frequency signal having a predetermined frequency or lower of a voice or audio input signal, a high frequency encoded signal obtained by encoding a signal of a higher frequency than the low frequency signal, and a band A decoding device to which an energy encoded signal is input,
A first decoding unit for decoding the first encoded signal to generate a low-frequency decoded signal;
A second decoding unit that decodes the high frequency encoded signal and generates a wideband decoded signal using the low frequency decoded signal;
A third decoding unit that decodes the band energy encoded signal to generate quantized band energy,
The second decoding unit
A separation unit that separates the low frequency decoded signal into a low frequency non-tonal signal that is a non-tonal component of the low frequency decoded signal and a low frequency tonal signal that is a tonal component of the low frequency decoded signal;
A first band extension unit that generates a high frequency non-tonal signal by copying the low frequency non-tonal signal to a high frequency using lag information obtained by decoding the high frequency encoded signal;
A first scaling unit that adjusts an amplitude of the high-frequency non-tonal signal using a scale factor obtained by decoding the high-frequency encoded signal;
From the energy of the high frequency non-tonal signal and the quantization band energy, a tonal signal energy estimation unit that estimates the energy of the high frequency tonal signal;
A first combining unit that combines the low frequency non-tonal signal and the high frequency non-tonal signal to generate a wideband non-tonal signal;
A second band extending unit that generates a high frequency tonal signal by copying the low frequency tonal signal to a high frequency using the lag information;
A second scaling unit that adjusts the amplitude of the high frequency tonal signal based on the energy of the high frequency tonal signal;
A second combining unit that combines the low frequency tonal signal and the high frequency tonal signal whose amplitude is adjusted to generate a wideband tonal signal;
An adder that adds the wideband non-tonal signal and the wideband tonal signal to generate a wideband decoded signal;
The lag information is position information of a specific band that maximizes the correlation between the high frequency signal and the low frequency tonal signal,
The scale factor is an energy ratio between a high-frequency noise component that is a noise component of a high-frequency signal corresponding to the lag information and a high-frequency non-tonal signal.
Decoding device. The second decoding unit
A noise adding unit that adds a noise signal to the low-frequency decoded signal;
The decoding device according to claim 6. The second decoding unit
A noise addition unit that adds a noise signal to the low-frequency non-tonal signal output from the separation unit;
The decoding device according to claim 6.   A terminal device comprising the encoding device according to claim 1.   A terminal device comprising the decoding device according to claim 6. Encoding a low frequency signal of a voice or audio input signal below a predetermined frequency to generate a first encoded signal;
Decoding the first encoded signal to generate a low-frequency decoded signal;
Based on the low frequency decoded signal, a high frequency signal is encoded by generating a high frequency signal than the low frequency signal,
Calculating an energy ratio between a high frequency noise component which is a noise component of the high frequency signal and a high frequency non-tonal component of the high frequency decoded signal generated from the low frequency decoded signal;
Multiplexing the first encoded signal and the high frequency encoded signal including the ratio to output an encoded signal;
Encoding method. The encoding method according to claim 12 includes:
Calculate the energy of the voice or audio input signal and output it as quantized band energy,
Separating the low frequency decoded signal into a low frequency non-tonal signal that is a non-tornal component of the low frequency decoded signal and a low frequency tonal signal that is a tonal component of the low frequency decoded signal;
The position information of a specific band that maximizes the correlation between the high frequency signal and the low frequency tonal signal is output as lag information,
The low frequency non-tonal signal corresponding to the lag information is output as a high frequency non-tonal signal,
From the high frequency signal corresponding to the lag information, calculate the energy of the high frequency noise component that is a noise component,
Calculating an energy ratio between the high-frequency noise component and the high-frequency non-tonal signal and outputting it as a scale factor;
Encoding method. About a first encoded signal obtained by encoding a low frequency signal of a predetermined frequency or lower of a voice or audio input signal in an encoding device, and a high frequency encoded signal obtained by encoding a higher frequency signal than the low frequency signal,
Separating the first encoded signal and the high frequency encoded signal;
Decoding the first encoded signal to generate a low-frequency decoded signal;
Decoding the highband encoded signal, generating a wideband decoded signal using the lowband decoded signal;
The high frequency encoded signal includes an energy ratio between a high frequency noise component which is a noise component and a high frequency non-tonal component of a high frequency decoded signal generated from the low frequency decoded signal,
Generating the decoded ratio, and adjusting the amplitude of the low frequency non-tonal signal, which is a non-tonal component of the low frequency decoded signal, with reference to the ratio;
Decryption method. A first encoded signal obtained by encoding a low-frequency signal having a frequency equal to or lower than a predetermined frequency of an audio or audio input signal in the encoding device; a high-frequency encoded signal obtained by encoding a signal higher in frequency than the low-frequency signal; About the signal
Decoding the first encoded signal to generate a low-frequency decoded signal;
Decoding the highband encoded signal, generating a wideband decoded signal using the lowband decoded signal;
Decoding the band energy encoded signal to generate quantized band energy;
The low-frequency decoded signal is separated into a low-frequency non-tonal signal that is a non-tonal component of the low-frequency decoded signal and a low-frequency tonal signal that is a tonal component of the low-frequency decoded signal, and the high frequency encoding Using the lag information obtained by decoding the signal, the low frequency non-tonal signal is copied to the high frequency to generate a high frequency non-tonal signal,
Adjust the amplitude of the high-frequency non-tonal signal using a scale factor obtained by decoding the high-frequency encoded signal,
From the energy of the high frequency non-tonal signal and the quantization band energy, the energy of the high frequency tonal signal is estimated,
Combining the low frequency non-tonal signal and the high frequency non-tonal signal to generate a wideband non-tonal signal;
Using the lag information, the low frequency tonal signal is copied to a high frequency to generate a high frequency tonal signal,
Based on the energy of the high frequency tonal signal, adjust the amplitude of the high frequency tonal signal,
A broadband tonal signal is generated by combining the low frequency tonal signal and the high frequency tonal signal whose amplitude is adjusted,
Adding the broadband non-tonal signal and the broadband tonal signal to generate a wideband decoded signal;
The lag information is position information of a specific band that maximizes the correlation between the high frequency signal and the low frequency tonal signal,
The scale factor is an energy ratio between a high-frequency noise component that is a noise component of a high-frequency signal corresponding to the lag information and a high-frequency non-tonal signal.
Decryption method. A process of generating a first encoded signal by encoding a low frequency signal of a predetermined frequency or less of a voice or audio input signal;
A process of decoding the first encoded signal to generate a low-frequency decoded signal;
Based on the low frequency decoded signal, a process of generating a high frequency encoded signal by encoding a higher frequency signal than the low frequency signal;
A process of calculating an energy ratio between a high frequency noise component that is a noise component of the high frequency signal and a high frequency non-tonal component of the high frequency decoded signal generated from the low frequency decoded signal;
The program which makes a processor perform the process which multiplexes the said 1st encoding signal and the high frequency encoding signal containing the said ratio, and outputs an encoding signal. About a first encoded signal obtained by encoding a low frequency signal having a frequency equal to or lower than a predetermined frequency of a voice or audio input signal in the encoding device, and a high frequency encoded signal obtained by encoding a higher frequency signal than the low frequency signal,
Processing to separate the first encoded signal and the high frequency encoded signal;
A process of decoding the first encoded signal to generate a low-frequency decoded signal;
Decoding the highband encoded signal and generating a wideband decoded signal using the lowband decoded signal;
The high frequency encoded signal includes an energy ratio between a high frequency noise component which is a noise component and a high frequency non-tonal component of a high frequency decoded signal generated from the low frequency decoded signal,
A program that generates the decoded ratio and adjusts the amplitude of a low-frequency non-tonal signal that is a non-tonal component of the low-frequency decoded signal with reference to the ratio.

Images