JP2009151183A - Multi-channel voice sound signal coding device and method, and multi-channel voice sound signal decoding device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform high audibility quality reproduction, while suppressing a bit rate. <P>SOLUTION: A multi-channel voice sound signal coding device 1a comprises: time frequency conversion sections 11a to 11x for outputting a conversion coefficient in a form in which the voice sound signal of a time domain is converted to a frequency domain and divided into frequency bands; a most occupied channel determination section 12 for determining the most occupied channel for each frequency band, on the basis of the conversion coefficient and most occupied channel index information of each channel; a sub-occupied channel group determination section 13 for determining 0 or more sub-occupied channel groups for each frequency band; a most occupied channel conversion coefficient coding section 14 for coding the conversion coefficient of the most occupied channel, on the basis of the conversion coefficient and the most occupied channel index information of each channel; and a sub-occupied channel group conversion coefficient coding section 15 for coding the conversion coefficient of the sub-occupied channel group, on the basis of the conversion coefficient and the sub-occupied channel group index information of each channel. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a multi-channel audio / acoustic signal encoding apparatus and method for encoding a multi-channel input audio / acoustic signal, and a multi-channel audio / acoustic signal decoding apparatus for decoding encoded data into a plurality of output audio / acoustic signals. And methods.

  Multi-channel audio-acoustic signal encoding (hereinafter referred to as “multi-channel encoding”) that compresses and encodes a multi-channel input audio-acoustic signal is known. In many multi-channel encodings, more efficient compression is realized by using correlation between channels than when all input signals are encoded individually.

  On the other hand, rather than playing back multiple input signals individually, when using multi-channel playback as stereo sound using headphones or multiple speakers, it is more efficient to use human auditory characteristics. There is BCC (Binaural Cue Coding) that realizes compression (see Patent Document 1 below).

  Multi-channel coding based on these three-dimensional sound reproduction is an effective technique for realistic movie sound and music reproduction and multi-point sound communication.

  In BCC, input signals of a plurality of channels are encoded into a mixed signal obtained by mixing all input signals and auxiliary information called auditory scene parameters. The auditory scene parameter is, for example, index information of the most occupied channel for each frequency band. The decoding device decodes the three-dimensional reproduced sound from the mixed signal and the auditory scene parameter. In other words, BCC reproduces each frequency band so that it can be heard as a multi-channel audio-acoustic signal in a pseudo manner by flowing a mixed sound from the direction in which the human ear can hear the channel most easily.

  As described above, since BCC can encode a sound of a plurality of channels into a mixed signal and auxiliary information, significant information compression is possible. In other words, for example, a transmission band corresponding to a bit rate of 256 kbps is required to transmit a 64 kbps signal for four channels. However, by applying BCC mixed in one channel, the bit rate is (64 kbps + several kbps). It becomes possible to suppress to the extent.

  However, while BCC can realize encoding at a high compression rate, degradation of reproduction quality becomes a problem. That is, in the BCC that reproduces the mixed signal in the respective frequency band as the most occupied channel in a pseudo manner, if the gain of the channel other than the most occupied channel is low, distortion is hardly perceived, but the gain of the channel other than the most occupied channel is If it is high, a sense of missing or distorted signal occurs.

Therefore, in Patent Document 2, by selecting whether or not BCC is applied depending on the frequency band, reproduction with higher auditory quality than that in the case where BCC is applied to all frequency bands is realized. That is, for example, in a frequency band in which quality degradation is relatively permissible, such as a high frequency band, signals of a plurality of channels are mixed by applying BCC, and in a frequency band in which quality degradation is relatively unacceptable, such as a low frequency band, Encode the channel signal without mixing.
US Pat. No. 7,116,787 B2 JP 2004-78183 A

  However, the multi-channel encoding device according to the above prior art has the following problems. That is, for a frequency band to which BCC is not applied, there is a problem that the bit rate increases because a signal of a plurality of channels is encoded.

  Therefore, in order to solve the above-described problem, the present invention encodes the most occupied channel signal and zero or more variable number of semi-occupied channel signals together with the channel index information for each frequency band. Therefore, it is an object to perform encoding / decoding so as to realize reproduction with high auditory quality while suppressing the bit rate.

  In the present invention, the signals of a plurality of channels are not mixed as in Patent Document 1 and Patent Document 2 described above. This is because, when a mixed signal of a plurality of channels is reproduced, sound of channels other than the sound of the original most occupied channel signal is easily perceived as distortion. The present invention is characterized in that reproduction with high auditory quality is realized by selecting and encoding only the corresponding signal using the determined most occupied channel information and semi-occupied channel group information. In other words, the invention is essentially different from BCC that presupposes mixing of signals.

  In order to solve the above-described problem, the multi-channel audio-acoustic signal encoding apparatus of the present invention converts an input time-domain audio-acoustic signal into a frequency domain and divides it into at least one frequency band. And the most occupied channel index information for the determined channel, and the most occupied channel is determined for each frequency band based on the conversion coefficient of each channel output from the time frequency conversion unit. Based on the most occupied channel determination means for outputting each channel, the conversion coefficient of each channel output from the time frequency conversion means, and the most occupied channel index information output from the most occupied channel determination means. Determine zero or more quasi-occupied channels and quasi-occupy for the determined channel Based on the semi-occupied channel group determination means for outputting channel group index information, the conversion coefficient of each channel output from the time frequency conversion means and the most occupied channel index information output from the most occupied channel determination means, The most occupied channel transform coefficient encoding means for encoding the most occupied channel transform coefficient of the corresponding frequency band and outputting the most occupied channel acoustic coding data after encoding, and each channel output from the time frequency transform means Based on the transform coefficient and the quasi-occupied channel group index information output from the quasi-occupied channel group determination means, the transform coefficient of the quasi-occupied channel in the corresponding frequency band is encoded, Quasi-occupied channel group transform coefficient coding means for outputting group acoustic coded data

  According to the present invention, the most occupied channel determining means and the most occupied channel transform coefficient coding means efficiently encode the component most audible to the human ear, and the semi-occupied channel group determining means and the semi-occupied channel group By encoding other easily audible components with the transform coefficient encoding means, distortion can be suppressed and reproduction with high auditory quality can be realized. Further, the semi-occupied channel group determining means can determine the number of occupied channel groups determined to be required to be reproduced for each frequency band to be a variable number of 0 or more, thereby eliminating unnecessary channel encoding, An increase in rate can be prevented.

  Note that the invention relating to the multi-channel audio-acoustic signal encoding apparatus can also be regarded as an invention relating to a method, and can be described as follows along with the flowchart of FIG. A multi-channel audio / acoustic signal encoding method according to the present invention is a multi-channel audio / acoustic signal encoding method executed by an apparatus for encoding an audio-acoustic signal, wherein an input audio-acoustic signal in a time domain is converted into a frequency domain. A time frequency conversion step (step S1 in FIG. 8) for converting and outputting the conversion coefficient in a format divided into at least one frequency band, and based on the conversion coefficient of each channel output by the time frequency conversion step, The most occupied channel determination step (step S2) for determining the most occupied channel for each frequency band and outputting the most occupied channel index information regarding the determined channel, and the conversion coefficient of each channel output by the time frequency conversion step And the most occupied channel output in the most occupied channel determination step. A semi-occupied channel group determining step (step S3) for determining zero or more semi-occupied channels for each frequency band based on the index information and outputting semi-occupied channel group index information on the determined channel; Based on the conversion coefficient of each channel output by the time frequency conversion step and the most occupied channel index information output by the most occupied channel determination step, encode the conversion coefficient of the most occupied channel of the corresponding frequency band, The most occupied channel transform coefficient coding step (step S4) for outputting the most occupied channel acoustic coded data after coding, the transform coefficient of each channel output by the time frequency transform step and the semi-occupied channel group determining step Quasi-occupied channel group index output by A semi-occupied channel group transform coefficient coding step (step S5) for encoding the transform coefficient of the semi-occupied channel in the corresponding frequency band based on the information and outputting the coded semi-occupied channel group acoustic coding data. ).

  Further, in the multi-channel audio-acoustic signal encoding device of the present invention, the semi-occupied channel group determining means is configured to use the semi-occupied channel based on at least one of a masking effect in human hearing and a frequency dependence of auditory sensitivity. Is preferably determined.

  According to the present invention, the semi-occupied channel group determining means determines the semi-occupied channel group based on the masking effect in which a small sound component in the vicinity thereof cannot be heard by a large sound component in the human auditory system or the level of auditory sensitivity depending on the frequency. By reflecting in the standard, it is possible to make a decision closer to human auditory characteristics, and to realize reproduction with higher auditory quality.

  Further, in the multi-channel audio-acoustic signal encoding device of the present invention, the semi-occupied channel group determining means sets a channel whose gain difference from the most occupied channel is within a predetermined value in each frequency band as the semi-occupied channel. It is preferable to determine.

  According to the present invention, the semi-occupied channel group determining means can determine the semi-occupied channel group by comparing the gain for each frequency band for each input channel, and can determine the semi-occupied channel group with a smaller amount of computation. Can do.

  In the multi-channel audio / acoustic signal encoding apparatus of the present invention, it is preferable that the semi-occupied channel group determining means determines the semi-occupied channel only for a predetermined frequency band.

  According to the present invention, the semi-occupied channel group determining means can operate only in a predetermined frequency band, and can encode with a smaller amount of calculation.

  In addition, the multi-channel audio-acoustic signal encoding apparatus according to the present invention is based on the quasi-occupied channel group index information output from the quasi-occupied channel group determining means in order to perform encoding at a fixed or variable predetermined bit rate. Preferably, the apparatus further comprises bit rate control means for controlling at least one of the semi-occupied channel group determining means, the most occupied channel transform coefficient encoding means, and the semi-occupied channel group transform coefficient encoding means.

  According to this invention, the bit rate control means is based on the quasi-occupied channel group index information output from the quasi-occupied channel group determination means so as to maintain a fixed or variable arbitrary bit rate. It is possible to control at least one of the most occupied channel transform coefficient coding means and the semi-occupied channel group transform coefficient coding means, and obtain the best coding quality at an arbitrary bit rate.

  In the multi-channel audio / acoustic signal encoding apparatus according to the present invention, the semi-occupied channel group transform coefficient encoding means includes a part or all of intermediate information and output encoded data of the most occupied channel transform coefficient encoding means. Is preferably used.

  According to the present invention, the semi-occupied channel group transform coefficient encoding means can use a part or all of the intermediate information and the output encoded data of the most occupied channel transform coefficient encoding means, and a more efficient code Can be made.

  Further, in the multi-channel audio-acoustic signal encoding apparatus according to the present invention, the most occupied channel determining means, the semi-occupied channel group determining means, the most occupied channel transform coefficient coding means, and the semi-occupied channel group transform coefficient encoding means. It is preferable that at least one of the above uses sound source localization position information indicating a position where the inputted audio-acoustic signal of each channel is localized.

  According to the present invention, at least one of the most occupied channel determining means, the semi-occupied channel group determining means, the most occupied channel transform coefficient encoding means, and the semi-occupied channel group transform coefficient encoding means is each input to the decoding device. Sound source localization position information indicating where the audio-acoustic signal of the channel is localized and reproduced can be used, and reproduction with higher audible quality can be realized by performing encoding according to the reproduction environment.

  Further, the multi-channel speech and audio signal encoding device of the present invention holds at least the most occupied channel index information output from the most occupied channel determination means, and outputs the most occupied channel index information in the subsequent processing. Further comprising: occupied channel information holding means, wherein the semi-occupied channel group determining means is the conversion coefficient of each channel output from the time frequency converting means, the most occupied channel index information output from the most occupied channel determining means, and Based on the past occupied channel information output from the past occupied channel information holding means, 0 or more quasi-occupied channels are determined for each frequency band, and quasi-occupied channel group index information on the determined channel is output. It is preferable to do.

  According to the present invention, the semi-occupied channel group determining means can use at least the past most occupied channel index information held by the past occupied channel information holding means at the time of determining the semi-occupied channel group, By preventing temporal discontinuity of the channel to be encoded, reproduction with higher auditory quality can be realized.

  In the multi-channel audio-acoustic signal encoding apparatus of the present invention, the quasi-occupied channel group determining means includes the transform coefficient of each channel output from the time frequency converting means and the most occupied channel determining means. In addition to the occupied channel index information, zero or more quasi-occupied channels for each frequency band on the basis of the encoded most occupied channel acoustic coded data output from the most occupied channel transform coefficient coding means Is preferably determined.

  According to the present invention, since the most occupied channel acoustic coded data output from the most occupied channel transform coefficient coding means can be reflected, it can be reflected under conditions close to the reproduction environment in the decoding apparatus. Channels can be determined and auditory quality can be further improved.

  The multi-channel audio / acoustic signal decoding apparatus according to the present invention determines the sound image control information for controlling the sound image of the sound / acoustic signal of the most occupied channel for each frequency band based on the input most occupied channel index information. Channel sound image control information determining means and quasi-occupied channel group for determining sound image control information for controlling the sound image of the audio sound signal of the quasi-occupied channel group for each frequency band based on the input quasi-occupied channel group index information Sound image control information determination means, most occupied channel transform coefficient decoding means for decoding the input most occupied channel acoustic coding data and outputting the most occupied channel transform coefficient after decoding, and the input semi-occupied channel group acoustic code Quasi-occupied channel group transform coefficient decoding means for decoding coded data and outputting the quasi-occupied channel group transform coefficient after decoding Then, sound image control is performed by adding the most occupied channel sound image control information output from the most occupied channel sound image control information determination means to the most occupied channel conversion coefficient output from the most occupied channel conversion coefficient decoding means, The most occupied channel sound image control means for outputting the most occupied channel deformation transform coefficient for at least two reproduction channels, and the semi-occupied channel group conversion coefficient output from the semi-occupied channel group transform coefficient decoding means with respect to the semi-occupied channel group transform coefficient Semi-occupied channel group sound image control for performing sound image control by adding the semi-occupied channel group sound image control information output from the group sound image control information determining means, and outputting quasi-occupied channel group deformation transform coefficients for at least two reproduction channels, respectively. And the most occupied channel sound image control means for each reproduction channel A combination of the most occupied channel deformation conversion coefficient output from the quasi-occupied channel group sound image control means and a quasi-occupied channel group deformation conversion coefficient output from the quasi-occupied channel group sound image control means, and outputting a combined conversion coefficient for each reproduction channel after combining Coefficient synthesis means, and frequency time conversion means for converting the combined transform coefficients output from the respective transform coefficient synthesis means from the frequency domain to the time domain and outputting the output signals of the respective reproduction channels after the conversion. And

  According to the present invention, the most occupied channel transform coefficient decoding means and the most occupied channel sound image control means efficiently decode a component most audible to the human ear as a three-dimensional sound image, and a semi-occupied channel group transform coefficient decoding means and By decoding other easily audible components as a three-dimensional sound image by the semi-occupied channel group sound image control means, distortion can be suppressed and reproduction with high auditory quality can be realized.

  Note that the invention relating to the multi-channel audio-acoustic signal decoding apparatus can also be regarded as an invention relating to a method, and can be described as follows along with the flowchart of FIG. A multi-channel audio / acoustic signal decoding method according to the present invention is a multi-channel audio / acoustic signal decoding method executed by an apparatus for decoding an audio / acoustic signal, and for each frequency band based on input most occupied channel index information. Based on the most occupied channel sound image control information determining step (step S11 in FIG. 9) for determining sound image control information for controlling the sound image of the sound acoustic signal of the most occupied channel, and the input semi-occupied channel group index information, A quasi-occupied channel group sound image control information determining step (step S13) for determining sound image control information for controlling the sound image of the audio sound signal of the quasi-occupied channel group for each frequency band, and the input most occupied channel acoustic coding data And outputs the most occupied channel transform coefficient after decoding. A step (step S12), a quasi-occupied channel group transform coefficient decoding step (step S14) for decoding the input quasi-occupied channel group acoustic coding data and outputting a quasi-occupied channel group transform coefficient after decoding; Sound image control is performed by adding the most occupied channel sound image control information output in the most occupied channel sound image control information determination step to the most occupied channel conversion coefficient output in the occupied channel conversion coefficient decoding step, and at least two The most occupied channel sound image control step (step S15) for outputting the most occupied channel deformation transform coefficient for each reproduction channel, and the semi-occupied channel group transform coefficient output by the semi-occupied channel group transform coefficient decoding step The quasi output by the channel group sound image control information determination step. A quasi-occupied channel group sound image control step (step S16) for performing sound image control by adding channel group sound image control information and outputting quasi-occupied channel group deformation transform coefficients for at least two reproduction channels, and for each reproduction channel The most occupied channel deformation transform coefficient output by the most occupied channel sound image control step and the semi-occupied channel group deformation transform coefficient output by the semi-occupied channel group sound image control step are synthesized, and each reproduction channel after synthesis is synthesized. A transform coefficient synthesis step (step S17) for outputting the synthesized transform coefficients, and transforming the synthesized transform coefficients output in the respective transform coefficient synthesis steps from the frequency domain to the time domain, An output frequency time conversion step (step S18). Features.

  In the multi-channel audio-acoustic signal decoding apparatus of the present invention, the semi-occupied channel group transform coefficient decoding means uses part or all of the intermediate information and output encoded data of the most occupied channel transform coefficient decoding means. Is preferred.

  According to this invention, the semi-occupied channel group transform coefficient decoding means can use part or all of the intermediate information and output encoded data of the most occupied channel transform coefficient decoding means, and performs more efficient decoding. be able to.

  According to the present invention, it is possible to encode / decode a multi-channel audio-acoustic signal so as to realize reproduction with high auditory quality while suppressing the bit rate.

  The present invention can be readily understood by considering the following detailed description with reference to the accompanying drawings shown for the embodiments. Subsequently, embodiments of the present invention will be described with reference to the accompanying drawings. Where possible, the same parts are denoted by the same reference numerals, and redundant description is omitted.

[First Embodiment]
First, the configuration of the multi-channel speech / acoustic signal encoding apparatus according to the first embodiment will be described. FIG. 1 is a block diagram showing a configuration of a multi-channel audio-acoustic signal encoding apparatus 1a according to the first embodiment.

  As shown in FIG. 1, the multi-channel audio / acoustic signal encoding apparatus 1a according to the first embodiment converts a plurality of input audio / acoustic signals from a time domain to a frequency domain, and divides it into at least one frequency band. Based on the conversion coefficient of each channel output from the time-frequency conversion units 11a to 11x and the time-frequency conversion units 11a to 11x that output the conversion coefficients in the format, the most occupied channels (hereinafter “ The most occupied channel determination unit 12 that outputs the most occupied channel index information regarding the determined channel, and the conversion coefficient and the most occupied channel of each channel output from the time frequency conversion units 11a to 11x. Based on the most occupied channel index information output from the determination unit 12, 0 for each frequency band. The semi-occupied channel group determination unit 13 that determines the above semi-occupied channels (hereinafter referred to as “quasi-occupied channel group”) and outputs the semi-occupied channel group index information regarding the determined channel, and the time frequency conversion units 11a to 11x On the basis of the conversion coefficient of each channel output from the above and the most occupied channel index information output from the most occupied channel determination unit 12, the conversion coefficient of the most occupied channel in the corresponding frequency band is encoded, The most occupied channel transform coefficient coding unit 14 that outputs the occupied channel acoustic coding data, the transform coefficient of each channel output from the time frequency conversion units 11a to 11x, and the semi-occupancy output from the semi-occupied channel group determination unit 13 Based on the channel group index information, the conversion coefficient of the semi-occupied channel group of the corresponding frequency band is encoded, and the code It is configured to include a quasi-occupied channel group transform coefficient coding section 15, the outputting of quasi occupied channel group acoustic coded data after. Hereinafter, each configuration will be described in detail.

  The time-frequency conversion units 11a to 11x each convert the input time-domain audio-acoustic signal into the frequency domain, and output the conversion coefficient in a format divided into at least one frequency band. This process is for determining the most occupied channel group and the semi-occupied channel group for at least one frequency band. As a method for transforming to the frequency domain, orthogonal basis transforms such as discrete Fourier transform (DFT) and modified discrete cosine transform (MDCT) can be considered. However, the purpose of this processing is to make it possible to divide for each arbitrary frequency band, and not only the above-mentioned quadrature basis transformation but also a filter bank such as a quadrature mirror filter (QMF) or a polyphase filter bank. good. When a filter bank is used, the conversion coefficient is read as a filter output in the following description. Here, as a setting method of the frequency band, for example, it is possible to divide into 16 equal frequency bands or to divide into a critical bandwidth close to the human auditory mechanism, but the setting method does not matter. . Further, when the input signal has already undergone these processes, the processes of the time frequency converters 11a to 11x can be omitted.

  Moreover, the conversion method applied in the time-frequency conversion units 11a to 11x is not limited to one type, but the most occupied channel determination unit 12, the semi-occupied channel group determination unit 13, the most occupied channel transform coefficient encoding unit 14, the semi-occupancy In order to obtain an output for each unit of the channel group transform coefficient coding unit 15, it is also conceivable to apply different types of transform methods. For example, QMF output can be used for determining the most / semi-occupied channel, and MDCT output can be used for transform coefficient coding. However, when a plurality of types of conversion methods are used, the amount of calculation naturally increases.

  The most occupied channel determination unit 12 determines the most occupied channel for each frequency band based on the input conversion coefficients of a plurality of channels. “Occupied” means that it is the most audibly important. The channel is determined by selecting the channel with the largest absolute value of the transform coefficient in each frequency band. A method of selecting a channel having the largest gain of the conversion coefficient is exemplified, but the present invention is not limited to this. The output information may be information indicating which channel is the most occupied channel for each frequency band, and a channel index is generally considered, but is not limited thereto.

  The semi-occupied channel group determination unit 13 is configured for each frequency band based on the conversion coefficient of each channel output from the time frequency conversion units 11a to 11x and the most occupied channel index information output from the most occupied channel determination unit 12. Zero or more semi-occupied channels are determined, and semi-occupied channel group index information is output. Semi-occupied means that it is not the most occupied but important, and two examples of channel determination methods are shown below. However, the determination method is not limited to these. The output information may be information indicating which of the semi-occupied channel groups for each frequency band, and the channel index is generally considered, but is not limited thereto.

  A first example of the channel determination method is a method of selecting all channels that are not masked by the components of the most occupied channel. In this method, the audible quality of the reproduced sound is high, but a lot of information needs to be encoded. In addition, the amount of calculation may increase in order to perform strict masking determination.

  A second example of the channel determination method is a method of selecting a channel having a small gain difference from the gain of the most occupied channel. For example, a channel having a gain larger than the gain of the most occupied channel (−3 dB) is determined as the semi-occupied channel. In this method, processing can be performed with a smaller amount of computation than in the case of performing masking determination.

  Here, the information about the component and gain of the most occupied channel includes the transform coefficients output from the time-frequency transform units 11a to 11x and the output transform after being quantized by the later-described most occupied channel transform coefficient encoding unit 14. It is also possible to use a coefficient. As a result, since the change of the most occupied channel component and the gain value by the most occupied channel transform coefficient encoding unit 14 can be reflected, the quasi-occupied channel group can be determined under conditions closer to the reproduction environment in the decoding apparatus, Quality can be improved.

  In addition, the semi-occupied channel group determination method does not need to be performed independently for each frequency band, but the auditory important ranking is performed by the above-described method over all frequency bands and channels, and the important ones are in order. It is also possible to select as a semi-occupied channel group. At this time, it is possible to reproduce a higher audible quality by performing weighting determination based on the frequency dependence of sensitivity in human hearing.

  The semi-occupied channel group determination unit 13 determines zero or more variable number of channels for each frequency band. That is, depending on the frequency band, there may be no semi-occupied channel. The upper limit of the number of semi-occupied channels corresponds to the number of input channels, but can be limited to an arbitrary number in advance so as not to increase the bit rate. The limit number may naturally vary depending on the frequency band.

  Further, it can be considered that the semi-occupied channel group determination unit 13 determines the semi-occupied channel group only for a predetermined frequency band. For example, in the high frequency band, the sensitivity of human hearing is reduced, so even if a plurality of channels are faithfully reproduced, they cannot be distinguished. Therefore, by limiting the existence of the semi-occupied channel group only in a frequency band lower than a predetermined frequency, it is possible to suppress the amount of calculation for the bit rate and encoding.

  The most occupied channel transform coefficient encoding unit 14 applies the corresponding frequency based on the transform coefficient of each channel output from the time frequency transform units 11a to 11x and the most occupied channel index information output from the most occupied channel determination unit 12. The transform coefficient of the most occupied channel in the band is encoded and the most occupied channel acoustic coded data is output. As this encoding method, the transform coefficient is divided into arbitrary frequency bands, scaled and then scalar quantized, and a general acoustic encoding method such as entropy encoding of the quantized value is used. It doesn't matter at all.

  The semi-occupied channel group transform coefficient encoding unit 15 is based on the transform coefficient of each channel output from the time frequency transform units 11a to 11x and the semi-occupied channel group index information output from the semi-occupied channel group determination unit 13. The transform coefficient of the quasi-occupied channel group in the corresponding frequency band is encoded and quasi-occupied channel group acoustic encoded data is output. There is no limitation on the form of this encoding method.

  Through the above processing, the multi-channel audio / acoustic signal encoding apparatus converts the plurality of input signals into the most occupied channel index information, the semi-occupied channel group index information, the most occupied channel acoustic encoded data, and the semi-occupied channel group acoustic encoded data. Can be encoded. In actual processing, the output information may be further compressed by entropy coding or may be aggregated into one system of data by multiplexing, but is omitted because it is not an essential part of the present invention.

  According to the first embodiment, the most occupied channel determination unit 12 and the most occupied channel transform coefficient coding unit 14 efficiently encode the component most audible to the human ear and determine the semi-occupied channel group. By encoding other easily audible components by the unit 13 and the semi-occupied channel group transform coefficient encoding unit 15, distortion can be suppressed and reproduction with high auditory quality can be realized. Further, the semi-occupied channel group determination unit 13 can omit unnecessary channel encoding by determining the number of occupied channel groups determined to be required to be reproduced for each frequency band to be a variable number of zero or more. An increase in bit rate can be prevented.

  The semi-occupied channel group determination unit 13 determines the semi-occupied channel group based on at least one of the masking effect in human hearing and the frequency dependence of auditory sensitivity. In the auditory system, the masking effect that makes it impossible to hear small sound components in the vicinity due to large sound components and the level of auditory sensitivity due to frequency can be reflected in the criteria for determining the semi-occupied channel group. It is possible to make a near-term decision and to achieve playback with higher audible quality.

  In addition, the semi-occupied channel group determination unit 13 determines a channel having a gain difference from the most occupied channel in each frequency band within a predetermined value (3 dB as an example) as a semi-occupied channel group, and therefore requires a smaller amount of calculation. A semi-occupied channel group can be determined.

  In addition, when the semi-occupied channel group determination unit 13 determines the semi-occupied channel group only for a predetermined frequency band (for example, a frequency band lower than the predetermined frequency), it is possible to reduce the amount of calculation for the bit rate and encoding. Become.

[Second Embodiment]
Next, the configuration of the multi-channel speech / acoustic signal encoding apparatus according to the second embodiment will be described. FIG. 2 is a block diagram showing a configuration of a multi-channel audio and sound signal encoding device 1b according to the second embodiment. The multi-channel audio / acoustic signal encoding apparatus 1b has a configuration in which a bit rate control unit 16 is added to the multi-channel audio / acoustic signal encoding apparatus 1a of FIG. Hereinafter, the operation of the bit rate control unit 16 will be described.

  The bit rate control unit 16 is configured to perform coding at a fixed or variable predetermined bit rate based on the semi-occupied channel group index information output from the semi-occupied channel group determination unit 13. 13. Control at least one of the most occupied channel transform coefficient coding unit 14 and the semi-occupied channel group transform coefficient coding unit 15. Next, an example regarding a specific method of this control will be given. However, the control method is not limited to this.

  As the control in the semi-occupied channel group determination unit 13, for example, a restriction on the number of channels determined as the semi-occupied channel group can be considered. For example, the determination result of the semi-occupied channel group is output to the bit rate control unit 16, and the bit rate control unit 16 calculates the number of bits necessary for the entire encoding. When the required number of bits obtained by calculation exceeds a predetermined value, the quasi-occupied channel group determination unit 13 is controlled so as to reduce the number of quasi-occupied channel groups.

  As control in the most occupied channel transform coefficient coding unit 14 and the semi-occupied channel group transform coefficient coding unit 15, for example, it is conceivable to control allocation of bits required for coding. From the output result of the semi-occupied channel group determination unit 13, the bit rate control unit 16 determines the most occupied channel conversion coefficient and the semi-occupied channel group conversion coefficient for encoding with the highest perceptual quality with a predetermined bit amount. Determine whether to encode in bits.

  According to the second embodiment, the bit rate control unit 16 is based on the semi-occupied channel group index information output from the semi-occupied channel group determination unit 13 so as to maintain a fixed or variable arbitrary bit rate. , Can control at least one of the semi-occupied channel group determination unit 13, the most occupied channel transform coefficient encoding unit 14, and the semi-occupied channel group conversion coefficient encoding unit 15, and can perform the best encoding at an arbitrary bit rate. Quality can be obtained.

[Third Embodiment]
Subsequently, the configuration of the multi-channel speech and acoustic signal encoding device according to the third embodiment will be described. FIG. 3 is a block diagram showing a configuration of a multi-channel audio / acoustic signal encoding apparatus 1c according to the third embodiment. This multi-channel audio / acoustic signal encoding apparatus 1c is composed of the same components as the multi-channel audio / acoustic signal encoding apparatus 1a of FIG. 1, but the intermediate information of the most occupied channel transform coefficient encoding unit 14 and the output encoded data , A part or all of them are output to the semi-occupied channel group transform coefficient encoding unit 15. Hereinafter, operations of the most occupied channel transform coefficient coding unit 14 and the semi-occupied channel group transform coefficient coding unit 15 will be described.

  The semi-occupied channel group transform coefficient encoding unit 15 uses part or all of the intermediate information and the output encoded data of the most occupied channel transform coefficient encoding unit 14 in order to increase the encoding efficiency. That is, some or all of the intermediate information and the output encoded data may be used. An example is shown below.

  For example, the gain information of the most occupied channel is used for encoding the gain information of the semi-occupied channel. In the above example, when a channel having a gain greater than the gain of the most occupied channel (−3 dB) is determined as the semi-occupied channel group, the gain difference between the semi-occupied channel group and the most occupied channel is limited to a range of 0 to 3 dB. The Therefore, when the gain of the semi-occupied channel group is encoded, the gain difference between the semi-occupied channel group and the most occupied channel is encoded instead of encoding the absolute value of the gain of the semi-occupied channel group. In this case, since the gain difference is limited to the range of 0 to 3 dB, it can be compressed more efficiently.

  In order to further reduce the amount of information, the gain of the most occupied channel may be treated as the gain of the semi-occupied channel. In addition, when there is a correlation between the most occupied channel and the semi-occupied channel, not limited to the gain, encoding can be performed more efficiently by using the correlation.

  According to the third embodiment, the semi-occupied channel group transform coefficient encoding unit 15 uses a part or all of the intermediate information and the output encoded data of the most occupied channel transform coefficient encoding unit 14, More efficient encoding can be performed.

[Fourth Embodiment]
Subsequently, the configuration of the multi-channel speech and acoustic signal encoding device according to the fourth embodiment will be described. FIG. 4 is a block diagram showing a configuration of a multi-channel speech / acoustic signal encoding apparatus 1d according to the fourth embodiment. The multi-channel audio / acoustic signal encoding apparatus 1d includes the same components as those of the multi-channel audio / acoustic signal encoding apparatus 1a of FIG. 1, but includes the most occupied channel determination unit 12, the semi-occupied channel group determination unit 13, and the most occupied channel. At least one of the transform coefficient encoding unit 14 and the semi-occupied channel group transform coefficient encoding unit 15 uses sound source localization position information that indicates a position where the input audio-acoustic signal of each channel is localized. Hereinafter, handling of sound source localization position information will be described.

  The sound source localization position information is information that determines a position where the input signal of each channel is virtually localized at the time of reproduction, and is 1-3-dimensional position information including sound arrival direction information. As the sound source localization position information, information arbitrarily set by the user in the encoding device or the decoding device may be input, or may be set in advance according to any rule, and the generation method thereof does not matter.

  Next, as an example of the influence of sound source localization position information, a case where 2-channel input signals A and B are encoded will be described. For simplicity, only a certain frequency band will be described. The gain of A is 80 dB, and the gain of B is 76 dB. It is assumed that the most occupied channel determination unit 12 compares the gains of both channels and determines A as the most occupied channel. Furthermore, it is assumed that the semi-occupied channel group determination unit 13 determines a channel whose gain difference from the most occupied channel is within 3 dB as the semi-occupied channel. In this case, since B is not selected as a semi-occupied channel, the B signal is not encoded. Here, in the decoding apparatus, when A is positioned in the 3 o'clock direction and B is positioned in the 3 o'clock direction, since B is easily erased by the sound of A, it is difficult to perceive the loss of B, resulting in extreme quality degradation. Does not occur. However, if A was positioned in the 3 o'clock direction and B was positioned in the 9 o'clock direction, the sound would be in the right and left opposite positions, and A and B would be separated and easy to hear, making it easier to perceive B missing. It becomes a factor of quality degradation. Also, in the decoding device, if A was to be located 1 km away and B was located 1 m away, the sound actually heard by B would be louder, so B should be determined as the most occupied channel become.

  Therefore, in the multi-channel audio / acoustic signal encoding apparatus 1d, the most occupied channel determination unit 12, the semi-occupied channel group determination unit 13, the most occupied channel transform coefficient encoding unit 14, the semi-occupied channel group conversion based on the sound source localization position information. At least one of the coefficient encoding units 15 is controlled. That is, the most occupied channel determination unit 12 and the semi-occupied channel group determination unit 13 determine a channel to be encoded based on the positional relationship of the reproduced sound so that quality degradation does not occur. In addition, the most occupied channel transform coefficient encoding unit 14 and the semi-occupied channel group transform coefficient encoding unit 15 allocate the number of bits necessary for encoding so as to improve the perceived quality of the reproduced sound. The specific control method is not limited to this, and each component can operate based on the input sound source localization position information, thereby realizing reproduction in consideration of the influence on auditory quality caused by the positional relationship during reproduction. .

  According to the fourth embodiment, at least one of the most occupied channel determination unit 12, the semi-occupied channel group determination unit 13, the most occupied channel transform coefficient encoding unit 14, and the semi-occupied channel group conversion coefficient encoding unit 15 is By using sound source localization position information that indicates where the input audio-acoustic signal of each channel is localized, reproduction with higher audible quality can be realized by performing encoding according to the reproduction environment.

[Fifth Embodiment]
Subsequently, the configuration of the multi-channel speech and acoustic signal encoding device according to the fifth embodiment will be described. FIG. 5 is a block diagram showing a configuration of a multi-channel audio / acoustic signal encoding apparatus 1e according to the fifth embodiment. The multi-channel audio / acoustic signal encoding apparatus 1e has a configuration in which a past occupied channel information holding unit 17 is added to the multi-channel audio / acoustic signal encoding apparatus 1a of FIG. Hereinafter, operations of the past occupied channel information holding unit 17 and the semi-occupied channel group determining unit 13 will be described.

  The past occupied channel information holding unit 17 holds at least the most occupied channel index information output from the most occupied channel determination unit 12 and outputs it at the next and subsequent processing. In addition to this, the past occupied channel information holding unit 17 may hold the semi-occupied channel index information, the most occupied channel conversion coefficient, and the semi-occupied channel group conversion coefficient.

  The semi-occupied channel group determination unit 13 converts the conversion coefficients of each channel output from the time frequency conversion units 11a to 11x, the most occupied channel index information output from the most occupied channel determination unit 12, and the past occupied channel information holding unit 17 A semi-occupied channel is determined so as to prevent temporal discontinuity of the channel to be encoded based on the past dedicated channel information output. Whether or not it is selected as the most occupied or semi-occupied means whether or not the component of the channel is present in the reproduced sound, and when these are suddenly switched between the frames, Prone to distortion. For this reason, in order to make it difficult to cause discontinuity of selection, a channel selected in the past in each frequency band is selected or easily selected in subsequent processing. As a specific method, it is conceivable to add a weight obtained by multiplying a gain of a channel selected in the past when determining the semi-occupied channel by a time attenuation constant, but the present invention is not limited to this. By these processes, for example, it is possible to prevent the sound from being suddenly interrupted in the falling edge of the sound, and to improve the audibility quality by reproducing the sound until the volume is sufficiently low.

  According to the fifth embodiment, the semi-occupied channel group determination unit 13 performs at least the past most occupied channel index information held by the past occupied channel information holding unit 17 in the determination criteria of the semi-occupied channel group. Can be used, and reproduction with higher auditory quality can be realized by preventing temporal discontinuity of the channel to be encoded.

[Sixth Embodiment]
Next, the configuration of the multi-channel speech / acoustic signal encoding device according to the sixth embodiment will be described. FIG. 6 is a block diagram showing a configuration of a multi-channel audio / acoustic signal decoding apparatus 2a according to the sixth embodiment.

  The multi-channel audio / acoustic signal decoding apparatus 2a according to the sixth embodiment controls sound image control information for controlling the sound image of the audio / acoustic signal of the most occupied channel for each frequency band based on the input most occupied channel index information. Sound image control information for controlling the sound image of the audio signal of the semi-occupied channel group for each frequency band on the basis of the input semi-occupied channel group index information. A semi-occupied channel group sound image control information determining unit 22 to determine, a most occupied channel transform coefficient decoding unit 23 for decoding the input most occupied channel acoustic coding data and outputting the decoded most occupied channel transform coefficient; That decodes the quasi-occupied channel group acoustic encoded data and outputs the quasi-occupied channel group transform coefficient after decoding The most occupied channel sound image control information output from the most occupied channel sound image control information determination unit 21 is added to the most occupied channel conversion coefficient output from the group group conversion coefficient decoding unit 24 and the most occupied channel conversion coefficient decoding unit 23. The most occupied channel sound image control unit 25 that performs sound image control and outputs the most occupied channel deformation transform coefficient for each of at least two reproduction channels, and the semi-occupied channel group output from the semi-occupied channel group transform coefficient decoding unit 24 Sound image control is performed by adding the quasi-occupied channel group sound image control information output from the quasi-occupied channel group sound image control information determination unit 22 to the conversion coefficient, and quasi-occupied channel group deformation transform coefficients for at least two reproduction channels, respectively. The semi-occupied channel group sound image control unit 26 for outputting The most occupied channel deformation conversion coefficient output from the sound image control unit 25 and the semi-occupied channel group deformation conversion coefficient output from the semi-occupied channel group sound image control unit 26 are combined and combined for each reproduction channel after combining. The transform coefficient synthesizers 27a to 27y that output the coefficients and the synthesized transform coefficients output from the transform coefficient synthesizers 27a to 27y are converted from the frequency domain to the time domain, and the output signals of the respective reproduction channels after conversion are output. Frequency time conversion units 28a to 28y. Hereinafter, each configuration will be described in detail.

  The most occupied channel sound image control information determination unit 21 determines sound image control information for controlling the sound image of the most occupied channel in each frequency band based on the input most occupied channel index information. The sound image control information is, for example, a time difference or intensity difference between reproduction channels. When reproducing with a plurality of speakers, the sound image can be controlled in a pseudo manner by reproducing with a difference in time and intensity. Also, when reproducing with stereo headphones, reproduction with a pseudo three-dimensional effect can be realized by reproducing with different time and intensity between the right and left ears. As sound image control information in headphone playback, a head related transfer function (HRTF) can be used for playback with more stereoscopic effect. The sound image control information may be determined by calculation, or may be selected from a database held in advance, and the generation method is not limited.

  The semi-occupied channel group sound image control information determining unit 22 operates in the same manner as the most occupied channel sound image control information determining unit 21 and outputs semi-occupied channel group sound image control information.

  The most occupied channel transform coefficient decoding unit 23 decodes the input most occupied channel acoustic coded data and outputs the most occupied channel transform coefficient. This decoding process is not limited to any form as long as it corresponds to the encoding process.

  The semi-occupied channel group transform coefficient decoding unit 24 outputs quasi-occupied channel group transform coefficients by a decoding process corresponding to the encoding process, similarly to the most occupied channel transform coefficient decoding unit 23.

  The most occupied channel sound image control unit 25 adds the most occupied channel sound image control information output from the most occupied channel sound image control information determination unit 21 to the most occupied channel conversion coefficient output from the most occupied channel conversion coefficient decoding unit 23. By doing so, sound image control is performed. Specifically, for example, when the conversion coefficient is a DFT coefficient and the sound image control information is a DFT coefficient of a transfer function such as HRTF, both may be complex-multiplied. Even when the sound image control information is a time difference or an intensity difference, it can be easily added to the conversion coefficient by multiplication or addition. In the case of stereo headphones, since there are transfer functions for the right ear and the left ear, two channels of a modified transformation coefficient for right ear reproduction and a modified transformation coefficient for left ear reproduction are output. Similarly, when the playback device is M speakers, an output of M channels can be obtained.

  The semi-occupied channel group sound image control unit 26 performs sound image control in the same manner as the most occupied channel sound image control unit 25, and outputs a semi-occupied channel group deformation conversion coefficient for the number of reproduction channels. However, when a plurality of semi-occupied channels exist in a certain frequency band, sound image control is performed for each channel. At this time, a plurality of modified transform coefficients may be combined into one for each frequency band, or may be individually output and combined in transform coefficient combining units 27a to 27y described later.

  Each of the transform coefficient synthesis units 27a to 27y uses the most occupied channel deformation conversion coefficient output from the most occupied channel sound image control unit 25 and the semi-occupied channel group deformation conversion coefficient output from the semi-occupied channel group sound image control unit 26. Synthesize. As a combining method, simple addition or weighted addition that emphasizes the most occupied channel is conceivable, but is not limited thereto.

  The frequency time conversion units 28a to 28y convert the combined conversion coefficients output from the conversion coefficient combining units 27a to 27y from the frequency domain to the time domain, and output the output signals of the respective reproduction channels. When a filter bank such as QMF is used in the encoding device, an output signal is obtained by a synthesis filter.

  Through the above processing, the multi-channel audio / acoustic signal decoding apparatus 2a converts the most occupied channel index information, the semi-occupied channel group index information, the most occupied channel acoustic encoded data, and the semi-occupied channel group acoustic encoded data into a plurality of output signals. Can be decrypted. In actual processing, window processing or overlap processing may be further applied to the output signal, but it is omitted because it is not an essential part of the present invention.

According to the sixth embodiment, the most occupied channel transform coefficient decoding unit 23 and the most occupied channel sound image control unit 25 efficiently decode a component that is most audible to the human ear as a three-dimensional sound image, and a semi-occupied channel. The group conversion coefficient decoding unit 24 and the semi-occupied channel group sound image control unit 26 decode other easily audible components as a three-dimensional sound image, so that distortion can be suppressed and reproduction with high auditory quality can be realized.
[Seventh Embodiment]
Subsequently, the configuration of the multi-channel audio-acoustic signal decoding apparatus according to the seventh embodiment will be described. FIG. 7 is a block diagram showing a configuration of a multi-channel audio / acoustic signal decoding apparatus 2b according to the seventh embodiment. The multi-channel audio / acoustic signal decoding apparatus 2b includes the same components as the multi-channel audio / acoustic signal decoding apparatus 2a of FIG. 6, but a part of the intermediate information and the output encoded data of the most occupied channel transform coefficient decoding unit 23. Alternatively, the entire configuration is output to the semi-occupied channel group transform coefficient decoding unit 24. Hereinafter, operations of the most occupied channel transform coefficient decoding unit 23 and the semi-occupied channel group transform coefficient decoding unit 24 will be described.

  The semi-occupied channel group transform coefficient decoding unit 24 uses part or all of the intermediate information and the output encoded data of the most occupied channel transform coefficient decoding unit 23 in order to increase the coding efficiency. That is, some or all of the intermediate information and the output encoded data may be used. An example of the operation is a decoding process corresponding to the operations of the most occupied channel transform coefficient coding unit 14 and the semi-occupied channel group transform coefficient coding unit 15 in the above-described multi-channel audio-acoustic signal coding apparatus 1c of FIG.

  According to the seventh embodiment, the semi-occupied channel group transform coefficient decoding unit 24 uses a part or all of the intermediate information and the output encoded data of the most occupied channel transform coefficient decoding unit 23, thereby making it more efficient. Decryption can be performed.

  Note that the multi-channel audio-acoustic signal encoding apparatus and decoding apparatus in the first to seventh embodiments described above are various information processing apparatuses (for example, a computer, a mobile phone, a communication apparatus) composed of a CPU, RAM, ROM, And the like, which are provided in the audio equipment and the like, and are executed when the CPU executes a program stored in a memory such as a ROM.

It is a block diagram which shows the structure of the multichannel audio | voice acoustic signal encoding apparatus which concerns on 1st Embodiment. It is a block diagram which shows the structure of the multichannel audio | voice acoustic signal encoding apparatus which concerns on 2nd Embodiment. It is a block diagram which shows the structure of the multichannel audio | voice acoustic signal encoding apparatus which concerns on 3rd Embodiment. It is a block diagram which shows the structure of the multichannel audio | voice sound signal encoding apparatus which concerns on 4th Embodiment. It is a block diagram which shows the structure of the multichannel audio | voice sound signal encoding apparatus which concerns on 5th Embodiment. It is a block diagram which shows the structure of the multichannel audio | voice sound signal decoding apparatus which concerns on 6th Embodiment. It is a block diagram which shows the structure of the multichannel audio | voice sound signal decoding apparatus which concerns on 7th Embodiment. 3 is a flowchart showing a basic procedure of a multi-channel audio and sound signal encoding method according to the present invention. It is a flowchart which shows the basic procedure of the multichannel audio | voice sound signal decoding method which concerns on this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1a, 1b, 1c, 1d, 1e ... Multi-channel audio | voice audio signal encoding apparatus, 11a-11x ... Time frequency conversion part, 12 ... Most occupied channel determination part, 13 ... Semi-occupied channel group determination part, 14 ... Most occupied channel Transform coefficient coding unit, 15 ... Semi-occupied channel group transform coefficient coding unit, 16 ... Bit rate control unit, 17 ... Past occupied channel information holding unit, 2a, 2b ... Multi-channel audio-acoustic signal decoding device, 21 ... Most occupied Channel sound image control information determination unit, 22 ... Semi-occupied channel group sound image control information determination unit, 23 ... Most occupied channel transform coefficient decoding unit, 24 ... Semi-occupied channel group conversion coefficient decoding unit, 25 ... Most occupied channel sound image control unit, 26 ... quasi-occupied channel group sound image control unit, 26 ... quasi-occupied channel group sound image control unit, 27a-27y ... conversion coefficient synthesis unit, 28a-28y ... frequency time conversion unit.

Claims (13)

A time-frequency conversion means for converting an input audio-acoustic signal in the time domain into a frequency domain and outputting a conversion coefficient in a format divided into at least one frequency band;
A most occupied channel determining means for determining the most occupied channel for each frequency band based on the conversion coefficient of each channel output from the time frequency converting means and outputting the most occupied channel index information for the determined channel; ,
Based on the conversion factor of each channel output from the time frequency conversion means and the most occupied channel index information output from the most occupied channel determination means, zero or more quasi-occupied channels for each frequency band are obtained. Quasi-occupied channel group determining means for determining and outputting quasi-occupied channel group index information for the determined channel;
Based on the conversion coefficient of each channel output from the time frequency conversion means and the most occupied channel index information output from the most occupied channel determination means, encode the conversion coefficient of the most occupied channel of the corresponding frequency band, The most occupied channel transform coefficient coding means for outputting the most occupied channel acoustic coded data after coding;
Based on the conversion coefficient of each channel output from the time frequency conversion means and the semi-occupied channel group index information output from the semi-occupied channel group determination means, the conversion coefficient of the semi-occupied channel in the corresponding frequency band And a semi-occupied channel group transform coefficient encoding means for encoding the encoded semi-occupied channel group acoustic encoded data. The multi-occupied channel group determining means determines the semi-occupied channel based on at least one of a masking effect in human hearing and a frequency dependence of auditory sensitivity. Channel speech acoustic signal encoding apparatus. 2. The multi-channel according to claim 1, wherein the semi-occupied channel group determining unit determines a channel having a gain difference within a predetermined value within a predetermined value in each frequency band as the semi-occupied channel. Speech acoustic signal encoding apparatus. The multi-channel audio-acoustic signal encoding apparatus according to any one of claims 1 to 3, wherein the semi-occupied channel group determining unit determines the semi-occupied channel only for a predetermined frequency band. . The multi-channel speech and acoustic signal encoding device is
Based on the quasi-occupied channel group index information output from the quasi-occupied channel group determining means to perform encoding at a fixed or variable predetermined bit rate, the quasi-occupied channel group determining means and the most occupied channel conversion 5. The multi code according to claim 1, further comprising: a coefficient encoding unit; and a bit rate control unit that controls at least one process of the semi-occupied channel group transform coefficient encoding unit. Channel speech acoustic signal encoding apparatus. 6. The semi-occupied channel group transform coefficient encoding means uses part or all of intermediate information and output encoded data of the most occupied channel transform coefficient encoding means. The multi-channel speech acoustic signal encoding device according to one item. At least one of the most occupied channel determining means, the semi-occupied channel group determining means, the most occupied channel transform coefficient encoding means, and the semi-occupied channel group transform coefficient encoding means is an input audio-acoustic signal of each channel The multi-channel audio / acoustic signal encoding apparatus according to any one of claims 1 to 6, wherein sound source localization position information indicating a position where the sound is localized is used. The multi-channel speech and acoustic signal encoding device is
Holding at least the most occupied channel index information output from the most occupied channel determining means, and further comprising past occupied channel information holding means for outputting the most occupied channel index information in the subsequent processing,
The semi-occupied channel group determining means outputs the conversion coefficient of each channel output from the time frequency converting means, the most occupied channel index information output from the most occupied channel determining means, and the past occupied channel information holding means. 2. The method according to claim 1, wherein zero or more semi-occupied channels are determined for each frequency band based on the past past occupied channel information, and semi-occupied channel group index information relating to the determined channel is output. The multichannel audio | voice sound signal encoding apparatus as described in any one of -7. The quasi-occupied channel group determination means includes the most occupied channel conversion coefficient encoding in addition to the conversion coefficient of each channel output from the time frequency conversion means and the most occupied channel index information output from the most occupied channel determination means. 9. The method according to any one of claims 1 to 8, wherein zero or more semi-occupied channels are determined for each frequency band based on the encoded most occupied channel acoustic encoded data output from the means. The multi-channel speech acoustic signal encoding device according to claim 1. The most occupied channel sound image control information determining means for determining sound image control information for controlling the sound image of the audio signal of the most occupied channel for each frequency band, based on the input most occupied channel index information;
Quasi-occupied channel group sound image control information determining means for determining sound image control information for controlling the sound image of the audio sound signal of the quasi-occupied channel group for each frequency band based on the input quasi-occupied channel group index information;
Decoding the input most occupied channel acoustic coding data and outputting the most occupied channel transform coefficient after decoding;
Quasi-occupied channel group transform coefficient decoding means for decoding the input quasi-occupied channel group acoustic encoded data and outputting the quasi-occupied channel group transform coefficient after decoding;
Sound image control is performed by adding the most occupied channel sound image control information output from the most occupied channel sound image control information determining means to the most occupied channel conversion coefficient output from the most occupied channel conversion coefficient decoding means, and at least The most occupied channel sound image control means for outputting the most occupied channel deformation transform coefficient for each of the two reproduction channels;
Sound image control by adding the semi-occupied channel group sound image control information output from the semi-occupied channel group sound image control information determination means to the semi-occupied channel group conversion coefficient output from the semi-occupied channel group transform coefficient decoding unit Quasi-occupied channel group sound image control means for outputting quasi-occupied channel group deformation transform coefficients for at least two reproduction channels, respectively,
For each reproduction channel, the most occupied channel deformation conversion coefficient output from the most occupied channel sound image control means and the semi-occupied channel group deformation conversion coefficient output from the semi-occupied channel group sound image control means are combined, and after synthesis Transform coefficient synthesizing means for outputting a synthesized transform coefficient for each reproduction channel;
Multi-channel audio comprising: frequency-time conversion means for converting the combined transform coefficient output from each of the transform coefficient synthesis means from the frequency domain to the time domain and outputting an output signal of each reproduction channel after conversion. Acoustic signal decoding apparatus. The multi-channel speech acoustic signal according to claim 10, wherein the semi-occupied channel group transform coefficient decoding means uses a part or all of intermediate information and output encoded data of the most occupied channel transform coefficient decoding means. Signal decoding device. A multi-channel audio-acoustic signal encoding method executed by an apparatus for encoding an audio-acoustic signal, comprising:
A time-frequency conversion step of converting an input audio-acoustic signal in the time domain into a frequency domain and outputting a conversion coefficient in a format divided into at least one frequency band;
A most occupied channel determination step for determining a most occupied channel for each frequency band based on the conversion coefficient of each channel output by the time frequency conversion step, and outputting most occupied channel index information regarding the determined channel; ,
Based on the conversion coefficient of each channel output in the time frequency conversion step and the most occupied channel index information output in the most occupied channel determination step, zero or more quasi-occupied channels for each frequency band are obtained. A semi-occupied channel group determination step for determining and outputting semi-occupied channel group index information for the determined channel;
Based on the conversion coefficient of each channel output by the time frequency conversion step and the most occupied channel index information output by the most occupied channel determination step, encode the conversion coefficient of the most occupied channel of the corresponding frequency band, A most occupied channel transform coefficient coding step for outputting the most occupied channel acoustic coded data after coding;
Based on the conversion coefficient of each channel output in the time frequency conversion step and the semi-occupied channel group index information output in the semi-occupied channel group determination step, the conversion coefficient of the semi-occupied channel in the corresponding frequency band And a semi-occupied channel group transform coefficient encoding step for outputting the encoded semi-occupied channel group acoustic encoded data. A multi-channel audio-acoustic signal decoding method executed by an apparatus for decoding audio-acoustic signals,
A most occupied channel sound image control information determining step for determining sound image control information for controlling the sound image of the sound and sound signal of the most occupied channel for each frequency band based on the input most occupied channel index information;
Based on the input semi-occupied channel group index information, a semi-occupied channel group sound image control information determining step for determining sound image control information for controlling the sound image of the audio sound signal of the semi-occupied channel group for each frequency band;
A most occupied channel transform coefficient decoding step for decoding input most occupied channel acoustic coding data and outputting a decoded most occupied channel transform coefficient;
A semi-occupied channel group transform coefficient decoding step for decoding the input semi-occupied channel group acoustic coding data and outputting the decoded semi-occupied channel group transform coefficient;
Sound image control is performed by adding the most occupied channel sound image control information output by the most occupied channel sound image control information determination step to the most occupied channel conversion coefficient output by the most occupied channel conversion coefficient decoding step, and at least A most occupied channel sound image control step for outputting a most occupied channel deformation transform coefficient for each of the two reproduction channels;
Sound image control by adding the quasi-occupied channel group sound image control information output by the quasi-occupied channel group sound image control information determination step to the quasi-occupied channel group sound image control information determination step output to the quasi-occupied channel group conversion coefficient decoding step Performing a semi-occupied channel group sound image control step for outputting a semi-occupied channel group deformation transform coefficient for each of at least two reproduction channels;
For each reproduction channel, the most occupied channel deformation conversion coefficient output by the most occupied channel sound image control step and the semi-occupied channel group deformation conversion coefficient output by the semi-occupied channel group sound image control step are synthesized, and after synthesis A transform coefficient synthesis step for outputting a synthesized transform coefficient for each reproduction channel;
A multi-channel audio comprising: a frequency time conversion step of converting the combined conversion coefficient output by each of the conversion coefficient combining steps from a frequency domain to a time domain and outputting an output signal of each reproduction channel after conversion. Acoustic signal decoding method.

Images