JP2006259291A - Audio encoder - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain compatibility with the conventional MPEG standard AAC system even when the encoded signal of Spatial Codec becomes large in size. <P>SOLUTION: An audio encoder is equipped with a 1st encoding section 101 which encodes a down-mix signal, a 2nd encoding section 102 which encodes information to put the down-mix signal back into a multichannel signal, a division section 103 which divides the encoded signal of a channel expansion section into A (A≥1) partial signals, and a multiplexing section 104 which stores them in a fill element, so the encoded signal of the channel expansion section is disregarded by an old type AA decoder. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an audio encoder that encodes a multi-channel signal. In particular, a downmix unit encoded signal obtained by encoding a signal obtained by downmixing an input multichannel signal into a stereo signal, and a channel expansion unit encoded signal obtained by encoding a signal for returning it to the original multichannel signal. And an audio encoder in which the bit stream is compatible with the MPEG standard AAC stereo encoded signal.

  Currently, Spatial Codec is being standardized in MPEG audio standardization activities. Spatial Codec is a downmix unit encoded signal obtained by encoding a signal obtained by downmixing an input multichannel signal into a stereo signal, and a channel expansion unit encoding a signal for returning the signal to the original multichannel signal. This is an encoding method in which an encoded signal is separated and stored in a bit stream.

On the other hand, an encoded signal obtained by encoding a signal obtained by down-mixing a stereo signal of two input channels into a monaural signal and an encoded signal obtained by encoding a signal for returning it to the original stereo signal are separated. As an encoding method for storing in a bitstream, the Parametric Coding technique has already been standardized in the MPEG standard (for example, see Non-Patent Document 1).
ISO / IEC 14496-3: 2001 / FDAM2 (Parametric Coding for High Quality Audio)

  However, the conventional MPEG standard AAC system is a standard in which, for example, when the input signal is 5.1 channel, it is compressed and encoded separately into the front 2 channel, the back 2 channel, the center channel, and the LFE channel. Like the Spatial Codec, a downmix unit encoded signal obtained by encoding a signal obtained by downmixing an input multichannel signal into a stereo signal, and a signal for returning it to the original multichannel signal are encoded. The method of separating the channel expansion portion encoded signal and storing it in the bit stream has a problem that compatibility with the conventional MPEG standard AAC method cannot be achieved.

  In addition, Parametric Coding technology in the MPEG standard encodes a signal for returning a monaural signal to an original stereo signal in the fill element in the MPEG standard AAC system in order to ensure compatibility with the MPEG standard AAC system. Although storing the encoded signal is disclosed, if the channel expansion portion encoded signal is stored in the fill element in the Spatial Codec by the same method as the prior art, the following problems occur. This will be described below.

  FIG. 4 is a diagram illustrating a problem that is predicted when the channel expansion unit coded signal is stored in a fill element. In FIG. 4, the horizontal axis represents the bit rate in compression encoding, and the vertical axis represents the sound quality. A curve labeled MP2 indicates the relationship between the bit rate and the sound quality in the MPEG2-Layer2 system. There is a tendency that the sound quality decreases when the bit rate is lowered. FIG. 4 does not show specific and quantitative values on the horizontal axis / vertical axis, but merely shows trends. The curve attached with MP3 shows the relationship between the bit rate and the sound quality in the MPEG2-Layer3 system, and the curve attached with AAC shows the relationship between the bit rate and the sound quality in the MPEG2-AAC system.

  As shown in FIG. 4, at any bit rate, the newly developed encoding method has the feature that the sound quality is better than the old encoding method. That is, MP3 developed after MP2 has better sound quality than MP2 at any bit rate, and AAC developed next to MP3 has better sound quality than MP3 at any bitrate. However, in the case of the Spatial Codec that attempts to maintain compatibility with the AAC scheme by storing the channel expansion portion encoded signal in a fill element, as shown by the dashed curve in FIG. 4, when the bit rate is low, Although it can be expected that the sound quality will be better than the conventional method, there is a problem that even if the bit rate is increased, the sound quality is not better than the conventional method.

  The reason why the sound quality is improved at a low bit rate is that when the 2-channel signal is downmixed to a monaural signal and returned to a stereo signal based on that as shown by the MPEG4 standard Parametric Coding method, It has been shown that high sound quality can be obtained at a low bit rate by using gain difference information and the degree of correlation. At least using such a technique, a multi-channel signal is based on a signal that is downmixed to a stereo signal. This is because high sound quality can be obtained at a low bit rate when the multi-channel signal is restored.

  On the other hand, FIG. 5 is a diagram for explaining the reason why the sound quality is not improved at a high bit rate. The upper part of FIG. 5 shows an outline of the configuration of an encoded signal of Spatial Codec at 128 kbos. The shaded portion is the downmix portion encoded signal, and the white portion is the channel expansion portion encoded signal. The downmix part encoded signal is relatively larger than the channel expansion part encoded signal. On the other hand, the lower part of FIG. 5 shows an outline of the configuration of an encoded signal of Spatial Codec at 320 kbsp. The problem here is that the downmixed portion encoded signal in the shaded portion is obtained by encoding a two-channel signal. Therefore, as shown in the lower part of FIG. Saturates. This is because when AAC is used, the improvement in sound quality is saturated at about 128 kbps in the case of a 2-channel encoded signal. If so, it can be expected that the sound quality can be improved by reducing the size of the downmixed portion encoded signal in the shaded portion and increasing the size of the channel expanded portion encoded signal in the whitened portion. This is not possible when trying to store a partial encoded signal in a fill element. This is because the size of information that can be stored in the fill element is limited to 269 bytes because of the AAC standard. This is because the fill element is defined by the syntax shown in FIG. 7 in the AAC standard, and is 269 bytes at the maximum (ISO / IEC 13818-7). FIG. 6 shows a summary of the problems when attempting to store such a channel expansion portion encoded signal in a fill element.

  That is, even if the bit rate is increased (the size is increased), the downmix part encoded signal indicated by the shaded part does not improve because the sound quality is saturated, and the channel enlarged part encoded signal indicated by the white part is The bit rate cannot be increased (the size cannot be increased) due to the restriction of the fill element. In the case of Parametric Coding, since the information for expanding the number of channels is only information for making the monaural into stereo, such a restriction on the size of the fill element can be ignored, but in the case of Spatial Codec Since the stereo signal is expanded to multi-channels such as 5 channels and 7 channels, there is a problem that the restriction on the size of the fill element hinders improvement in sound quality.

  The present invention has been made in view of such a conventional problem, and when the downmix part encoded signal and the channel expansion part encoded signal are separated and stored in a bitstream, An audio encoder is provided for allowing the downmix part encoded signal to be compatible with the conventional MPEG standard AAC system as a stereo encoded signal, and the size of the channel expansion part encoded signal is increased. Therefore, it is an object of the present invention to provide an audio encoder that can be compatible with the conventional MPEG standard AAC system even when the sound quality is improved.

  In order to solve the above problems, the invention according to claim 1 of the present application is directed to downmix means for downmixing an M channel (M> 2) multichannel signal into a stereo signal, and encoding the downmix signal to First encoding means for generating a mix part encoded signal, second encoding means for encoding information for returning the downmix signal to a multi-channel signal and generating a channel expansion part encoded signal, and the channel Dividing means for dividing the expanded portion encoded signal into A (A ≧ 1) partial signals each having N bytes or less, and multiplexing for multiplexing the downmix portion encoded signal and the A partial signals And a converting means.

  The invention described in claim 2 of the present application is characterized in that, in the invention described in claim 1, the multiplexing means further sets the A to 2 or more when the value of M is a predetermined value or more. Is.

  The invention according to claim 3 of the present application is the invention according to claim 1 or 2, wherein the second encoding means further includes gain difference information between predetermined channels of the input multi-channel signal and a degree of correlation. And when the second encoding unit encodes the gain difference information and the degree of correlation with a fineness equal to or greater than a predetermined time resolution, the A is set to 2 or more. It is characterized by this.

  The invention according to claim 4 of the present application is the invention according to claim 1 or 2, wherein the second encoding means further includes information on gain difference between predetermined channels of the input multi-channel signal and the degree of correlation. When the second encoding unit encodes the gain difference information and the degree of correlation with a fineness equal to or higher than a predetermined frequency resolution, the A is set to 2 or more. It is characterized by this.

  According to the first aspect of the present invention, when the downmix section encoded signal and the channel expansion section encoded signal are separated and stored in a bitstream, the downmix section encoded signal is a stereo encoded signal. As a result, compatibility with the conventional MPEG standard AAC system can be obtained. In addition, compatibility with the conventional MPEG standard AAC system can be achieved even when the size of the encoded signal of the channel expansion unit is made very large to improve the sound quality.

  According to the second aspect of the present invention, even when the number of channels of the input multi-channel signal is large, compatibility with the conventional MPEG standard AAC system can be achieved.

  According to the third aspect of the present invention, compatibility with the conventional MPEG standard AAC system can be obtained even when the sound quality is improved by increasing the time resolution.

  According to the fourth aspect of the present invention, compatibility with the conventional MPEG standard AAC system can be obtained even when the sound quality is improved by increasing the frequency resolution.

(Embodiment 1)
The audio encoder according to Embodiment 1 of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing a configuration of an audio encoder according to the first embodiment.
As shown in FIG. 1, the audio encoder includes a downmix unit 100 that downmixes an M channel (M> 2) multichannel signal into a stereo signal, encodes the downmix signal, and encodes a downmix unit encoded signal. A first encoding unit 101 that generates a channel expansion unit encoded signal by encoding information for returning the downmix signal to a multi-channel signal, and the channel expansion unit code A dividing unit 103 that divides the encoded signal into A (A ≧ 1) partial signals each having N bytes or less, and a multiplexing unit that multiplexes the downmix unit encoded signal and the A partial signals 104.

The operation of the audio encoder configured as described above will be described below.
First, in the present embodiment, the downmix unit 100 receives a multi-channel signal of 4 channels (front left ch, front right ch, back left ch, back right ch) as input and downmixes it to a stereo signal. For example, the front left channel + rear left channel is newly set as the left channel, and the front right channel + rear right channel is newly set as the right channel. A method of weighting and rear channels may be used.

  Next, the first encoding unit 101 encodes the downmix signal to generate a downmix unit encoded signal. In the present application, the first encoding unit 101 encodes the downmix signal as a stereo signal in the MPEG standard AAC system.

  Next, the second encoding unit 102 encodes information for returning the downmix signal to a multi-channel signal, and generates a channel expansion unit encoded signal. The method may be, for example, a method in which a gain difference or a degree of correlation between a plurality of channel signals before downmixing is encoded, and the encoded signal is used as a channel expansion unit encoded signal. In recent years, a technique of encoding the gain difference and the degree of correlation as information for returning a signal obtained by downmixing a stereo signal of Lch and Rch to a monaural signal to the original stereo signal has been standardized as an MPEG4 standard Parametric Coding method. It has become. What is necessary is just to produce | generate the channel expansion part encoding signal in this application using such a technique.

  Next, the dividing section 103 divides the channel expansion section encoded signal into A (A ≧ 1) partial signals each having N bytes or less. In this method, the channel expansion portion encoded signal may be simply divided into A (A ≧ 1) partial signals each having N bytes or less, or the front left channel and the rear left channel. In order to return the signal obtained by down-mixing the front right channel and the rear right channel to the original signal, the encoded signal for returning the signal obtained by down-mixing the signal to the original signal is the first partial signal. The encoded signal may be divided according to which channel the encoded signal is, such as the second partial signal.

  Alternatively, of the channel expansion unit encoded signal, an encoded signal obtained by encoding gain difference information is used as the first partial signal, and an encoded signal obtained by encoding the degree of correlation is used as the second partial signal. In this way, division may be made according to the type of encoded information. Alternatively, in the case where the second encoding unit 102 generates a channel expansion unit encoded signal for each input frequency band, the encoded signal for the low frequency band is used as the first partial signal. And the coded signal for the high frequency band may be divided by the physical position of the input signal, such as the second partial signal. Of course, in that case, it may be divided not by frequency position but by temporal position.

  Finally, the multiplexing unit 104 multiplexes the downmix unit encoded signal and the A partial signals. Here, the multiplexing unit 104 multiplexes the downmix unit encoded signal as an MPEG standard AAC stereo encoded signal. Each of the A partial signals is formatted and multiplexed in the form of a fill_element in the MPEG standard AAC system.

  It should be noted here that when A is 2 or more, each partial signal alone also multiplexes information indicating that it does not constitute the channel expansion portion encoded signal. In other words, information indicating which part of the channel expansion portion encoded signal is stored in the A fill_elements is also multiplexed.

  For example, when the dividing method in the dividing unit 103 is a method in which the channel expansion unit coded signal is simply divided into A (A ≧ 1) partial signals each having N bytes or less, 1 If the first fill element is not complete, is followed by the second fill element, and if the second fill element is not completed, it is followed by the third fill element, or if it is complete Also multiplexes information such as “end with this fill element”. Such information may be included in each fill element, or only such information may be multiplexed separately as a fill element.

  Alternatively, in the separation method in the dividing unit 103, an encoded signal for returning a signal obtained by downmixing the front left channel and the rear left channel to the original signal is used as the first partial signal, and the front right channel In the case of a method in which a coded signal for returning a signal obtained by downmixing the signal and the rear right channel to the original signal is used as the second partial signal, information on the channel included in the fill element, respectively. The information indicating is also multiplexed. Such information may be included in each fill element, or only such information may be multiplexed separately as a fill element.

  Alternatively, the separation method in the dividing unit 103 is an encoding in which the degree of correlation is encoded by using the encoded signal obtained by encoding the gain difference information as the first partial signal in the channel expansion unit encoded signal. When the signal is divided by the type of encoded information, such as a second partial signal, information indicating the type of encoded information included in each fill element is also multiplexed. Such information may be included in each fill element, or only such information may be multiplexed separately as a fill element.

  Alternatively, in the case of an encoding method in which the second encoding unit 102 generates a channel expansion unit encoded signal for each frequency band of the input signal, the dividing method in the dividing unit 103 is a low frequency band. When the method is such that the coded signal for the input signal is the first partial signal, the coded signal for the high frequency band is the second partial signal, and so on, according to the physical position of the input signal, respectively. The information indicating the content of the encoded information included in the fill element is also multiplexed. Such information may be included in each fill element, or only such information may be multiplexed separately as a fill element. Of course, in that case, it is needless to say that it may be divided not by frequency position but by temporal position.

  In the above description, it has been described that A is 2 or more, but it is needless to say that A may be 1. For example, the second encoding unit 102 decomposes an input multi-channel signal into a plurality of subband signals using a polyphase filter bank or the like and collects the subband signals for each of several frequency bands, or If the subband signals are grouped into several time slots and the gain difference information, the degree of correlation, etc. are encoded for each group, the details of the grouping are coarse. In this case, since the amount of encoded signal is small, it is not necessary to set A to 2 or more. For example, as shown in FIG. 2, when the frequency direction and the time direction are not so fine, the A may be 1, that is, the fill element may be one. On the other hand, as shown in FIG. 3, when the frequency direction summarization method and time direction summarization method are finer than those in FIG. 2, the amount of the encoded signal becomes large. There are cases where it is not possible.

  In the above description, both the time axis direction and the frequency axis direction are divided and collected, but it goes without saying that either one may be used. 2 and 3, the number of subbands is 32. Of course, this is only an example, and it is needless to say that any value such as 16, 64, or 79 may be used.

  As described above, according to the present embodiment, downmix means for downmixing an M channel (M> 2) multichannel signal to a stereo signal, the downmix signal is encoded, and a downmix unit encoded signal is encoded. First encoding means for generating a signal, second encoding means for encoding information for returning the downmix signal to a multi-channel signal, and generating a channel expansion part encoded signal, and the channel expansion part encoded signal Is divided into A (A ≧ 1) partial signals each having N bytes or less, and multiplexing means for multiplexing the downmix part encoded signal and the A partial signals. The first encoding means encodes the downmix signal as a stereo signal in the MPEG standard AAC system, and the multiplexing means converts the A partial signals. When each is formatted in the form of fill_element in the MPEG standard AAC system, the A number of fill_elements are multiplexed, and when A is 2 or more, each of the partial signals individually constitutes the channel expansion portion encoded signal. By multiplexing information indicating that there is no signal, the downmix signal can be made to conform to the MPEG standard AAC system, and the channel expansion portion encoded signal is stored as a fill element. A Spatial Codec decoder having a function of converting the downmix signal back to a multi-channel signal based on the encoded signal can decode the encoded signal into a multi-channel signal if generated by the multiplexing means. Spatial Non decoder Odec, the traditional AAC decoder, the channel expansion unit encoded signal as fill element, since the presence is ignored, so that the 2-channel signal down-mix can be generated.

  In particular, even when the size of the channel expansion portion encoded signal is very large, a plurality of fill elements are used, so that compatibility with the conventional MPEG standard AAC system can be achieved. It should be noted here that when divided into a plurality of fill elements, each of them is not established as a channel expansion portion encoded signal alone, but information relating them is also multiplexed as a fill element, so Spatial The Codec decoder can accurately interpret the encoded signal if it is generated by the multiplexing means, and as a result can be decoded into a multi-channel signal.

  In the present embodiment, the number of channels of the multi-channel signal is set to 4 for simplification of explanation, but it may not be 4, and 5.1 channels or 7. Needless to say, there may be one channel. Rather, the larger the number of input multi-channel channels, the larger the size of the channel expansion portion encoded signal, so there are many cases where A must be 2 or more, which is consistent with the spirit of the present invention. It becomes a situation.

  Since the present invention can be compatible with the conventional MPEG standard AAC stereo encoded signal even when the Spatial Codec encoded signal has a large size, it has been put into practical use by using the AAC stereo system. Devices, such as digital broadcast receivers (so-called 1Seg receivers) and portable audio devices, can further enhance the functionality of the devices, while preventing inconvenience to users of legacy devices. Can be.

The figure which shows the structure of the audio encoder in this Embodiment 1. The figure which shows the time / frequency signal of an input in case the division | segmentation number A in a division | segmentation part should be one. The figure which shows the time / frequency signal of an input in case the division | segmentation number A in a division part becomes 2 or more. The figure which shows the problem at the time of storing a channel expansion part encoding signal in a fill element. The figure for demonstrating the reason why a sound quality does not improve at a high bit rate when a channel expansion part coding signal is stored in a fill element. The figure which arranged the subject at the time of trying to store a channel expansion part coding signal in a fill element. The figure which shows the syntax of the fill element in MPEG standard AAC system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Downmix part 101 1st encoding part 102 2nd encoding part 103 Dividing part 104 Multiplexing part

Claims (4)

Downmix means for downmixing an M channel (M> 2) multichannel signal to a stereo signal;
A first encoding means for encoding the downmix signal and generating a downmix section encoded signal;
A second encoding means for encoding information for returning the downmix signal to a multi-channel signal and generating a channel extension encoded signal;
Dividing means for dividing the channel extension portion encoded signal into A (A ≧ 1) partial signals each having N bytes or less;
A multiplexing means for multiplexing the downmix part encoded signal and the A partial signals;
The first encoding means encodes the downmix signal as a stereo signal in the MPEG standard AAC system, and the multiplexing means formats the A partial signals in the form of a fill_element in the MPEG standard AAC system, A number of fill_elements is multiplexed, and when A is 2 or more, information indicating which part of the channel expansion portion encoded signal is stored in the A fill_elements is also multiplexed. An audio encoder. The audio encoder according to claim 1, wherein the multiplexing means sets the A to 2 or more when the value of M is a predetermined value or more. The second encoding means encodes gain difference information and a degree of correlation between predetermined channels of the input multi-channel signal,
The multiplexing means sets the A to 2 or more when the second encoding means encodes the gain difference information and the degree of correlation with a fineness of a predetermined time resolution or more. The audio encoder according to claim 1 or 2. The second encoding means encodes gain difference information and a degree of correlation between predetermined channels of the input multi-channel signal,
The multiplexing means sets the A to 2 or more when the second encoding means encodes the gain difference information and the degree of correlation at a fineness equal to or higher than a predetermined frequency resolution. The audio encoder according to claim 1 or 2.

Images