JP3690591B2 - Encoder - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a coding apparatus that performs coding after converting a signal on a time axis, which has a possibility of switching the type of signal for each frame, into a signal on a frequency axis for each block composed of a plurality of frames.
[0002]
[Prior art]
As a high-efficiency encoding method for audio signals, a signal on the time axis is converted into a signal on the frequency axis for each frame (predetermined time unit), divided into a plurality of frequency bands, and encoded for each band. Transform coding method, band division coding method that divides into multiple frequency bands with the signal on the time axis and encodes each band, or a method that combines band division coding method and transform coding method It has been known.
[0003]
In the band division coding method, a signal on the time axis is divided into a plurality of frequency bands by using a filter such as a quadrature mirror filter. Also, in transform coding, spectrum transform is performed by using discrete Fourier transform, discrete cosine transform, modified discrete cosine transform, etc. (on the time axis) Is converted to a signal on the frequency axis).
[0004]
Here, for example, when performing spectrum conversion using a modified discrete cosine transform, if the number of sample data in one frame is M, M data in the latter half of 2M sample data that was spectrum-converted one before. Is converted to M data in the first half of 2M sample data to be subjected to spectrum conversion next, whereby spectrum conversion is performed for every 2M sample data (that is, sample data for two frames). . By doing in this way, the connection distortion between the unit blocks which perform encoding can be reduced.
[0005]
Further, in the publication of International Publication No. WO98 / 46045, when encoding a plurality of correlated signals such as audio stereo signals, the correlation between the plurality of signals is used to further increase the efficiency. A technique for realizing encoding is disclosed.
[0006]
For example, when an L (left) channel signal and an R (right) channel signal constituting an audio stereo signal are encoded, an average of the L channel signal and the R channel signal is used as the first signal. A first channel signal, and a difference between an L channel signal and an R channel signal divided by two, an L channel signal, or an R channel signal as a second channel signal; And the signal of the second channel are encoded separately.
[0007]
When there is a correlation between the L channel signal and the R channel signal, a signal obtained by dividing the difference between the L channel signal and the R channel signal by 2 is set as the second channel signal. On the other hand, when there is no correlation between the L channel signal and the R channel signal, the signal having the relatively low signal level is set as the second channel signal. Note that which of the three signals is set as the second channel signal is determined for each frame.
[0008]
As a result, the level of the second channel signal is lower than that of the first channel signal, so that the bit rate when encoding the second channel signal is lowered to realize higher efficiency encoding. can do.
[0009]
[Problems to be solved by the invention]
However, if the above-mentioned two techniques, that is, a technique for reducing connection distortion between unit blocks to be encoded, and a technique for encoding a plurality of correlated signals with higher efficiency are used in combination, While encoding is performed with a plurality of frames (2 frames in the above example) as one unit, the signal type may be switched for each frame, so the signal level rapidly increases within the unit block for encoding. As a result, since the bit rate assigned to the channel for switching the signal type is low, there is a problem that quantization noise increases.
[0010]
Therefore, the present invention provides an encoding apparatus capable of suppressing an increase in quantization noise due to switching of the type of a signal to be encoded within a unit block to be encoded. For the purpose.
[0011]
[Means for Solving the Problems]
In order to achieve the above object, in the present invention,
First channel conversion means for outputting a signal obtained by combining all input signals of a plurality of channels as a first channel signal;
A second channel conversion means for outputting a signal obtained by synthesizing all input signals of a plurality of channels or an input signal of a part of channels as a second channel signal;
Channel conversion control means for setting a signal output from the second channel conversion means for each frame in accordance with the relationship of the input signals of each channel;
A coding unit comprising: a coding unit configured to convert a signal output from the first channel conversion unit and the second channel conversion unit into a signal on a frequency axis for each block composed of a plurality of frames and then perform coding. In the conversion device,
The bit rate for encoding the signal output from the second channel conversion unit is not switched when the type of the signal output from the second channel conversion unit is switched in the block to be encoded. Bit rate setting means for setting each block to be encoded is higher than the case .
[0012]
With this configuration, it is possible to assign a higher bit rate to each unit block to be encoded when the type of signal to be encoded within the block is switched than when the type is not switched.
[0013]
In addition, the bit rate for encoding the signal output from the second channel conversion means depends on the number of frames that have elapsed since the last change of the type of signal output from the second channel conversion means. May be configured.
[0014]
With this configuration, the signal type is not switched in a unit block to be encoded, but a high bit rate can be assigned to a block including a frame whose signal type has been switched from the previous frame.
[0015]
Further, the bit rate for encoding the signal output from the second channel conversion means may be set according to the type of the signal output from the second channel conversion means.
[0016]
With this configuration, it is possible to reduce the quantization noise by setting a high bit rate when encoding a signal of a type that greatly affects the degradation of the signal decoded by the quantization noise compared to other types. it can.
[0017]
The signal output from the second channel conversion unit is encoded first, and the signal output from the first channel conversion unit and the second channel conversion unit Of the bit rates given in advance to encode the output signal, the first channel conversion means at the bit rate remaining after the signal output from the second channel conversion means is encoded. The signal output from may be configured to be encoded.
[0018]
With this configuration, even if the bit rate when encoding a signal that may switch the signal type in the unit block to be encoded is set high, the bit rate that was not used in actual encoding is The bit rate for encoding other signals is assigned.
[0019]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of an encoding apparatus for encoding audio stereo signals (L (left) channel signal and R (right) channel signal) according to an embodiment of the present invention. PCM (Pulse Code Modulation) modulated L-channel signal and R-channel signal, which are input signals to the encoding device, are divided into a plurality of frequency bands by a frequency band dividing circuit 1 composed of an orthogonal mirror filter or the like. Then, the signal is supplied to the first channel conversion circuit 2-1, the second channel conversion circuit 2-2, and the control circuit 5.
[0020]
The first channel conversion circuit 2-1 takes an average of the L channel signal and the R channel signal given from the frequency band dividing circuit 1 (that is, (L channel signal + R channel signal) / 2. ) Is output. The second channel conversion circuit 2-2 is a signal obtained by dividing the difference between the L channel signal and the R channel signal given from the frequency band dividing circuit 1 by 2 (that is, (L channel signal−R channel signal). ) / 2), either an L channel signal or an R channel signal is output. Which signal the second channel conversion circuit 2-2 outputs depends on whether or not there is a correlation between the L channel signal and the R channel signal given from the frequency band dividing circuit 1 and the magnitude relationship between the two signals. Accordingly, it is set by the control circuit 5 for each frame.
[0021]
Specifically, the L channel signal and the R channel signal are as shown in FIGS. 2A and 2B, respectively, and there is a correlation between the L channel signal and the R channel signal. In this case, as shown in FIG. 3B, since the signal level of (L channel signal−R channel signal) / 2 becomes very small, (L channel signal−R channel signal) / Set to output 2.
[0022]
On the other hand, when the L channel signal and the R channel signal are as shown in FIGS. 4A and 4B, respectively, there is no correlation between the L channel signal and the R channel signal. 5B, since the signal level of (L channel signal−R channel signal) / 2 does not decrease, the signal level of the L channel signal and the R channel signal is relatively low. The lower one (R channel signal in the case of FIG. 4) is set to be output.
[0023]
As a result, a signal having a level as low as possible is output from the second channel conversion circuit 2-2. Therefore, the bit rate for encoding the signal output from the second channel conversion circuit 2-2. Even if the value is lowered, the quantization noise is not so noticeable. Therefore, the bit rate assigned to the signal output from the second channel conversion circuit 2-2 can be lowered to realize more efficient encoding.
[0024]
The first encoding circuit 3-1 outputs a signal output from the first channel conversion circuit 2-1 (hereinafter referred to as “A channel signal”) as the two adjacent frames, the x-th frame and Each encoded block consisting of the (x + 1) th frame is converted into a signal on the frequency axis by the modified discrete cosine transform, and then encoded at the bit rate set by the control circuit 5, and the code obtained thereby is converted to the A channel. It is output as the code of the (x + 1) th frame of the signal.
[0025]
The second encoding circuit 3-2 outputs the signal output from the second channel conversion circuit 2-2 (hereinafter referred to as “B channel signal”) and the two adjacent frames, the xth frame and Each encoded block consisting of the (x + 1) th frame is converted into a signal on the frequency axis by modified discrete cosine transform, and then encoded at the bit rate set by the control circuit 5, and the code obtained thereby is encoded in the B channel. It is output as the code of the (x + 1) th frame of the signal.
[0026]
The code string generation circuit 4 generates and outputs a code string as shown in FIG. The code string output from the code string generation circuit 4 is recorded on a recording medium, for example. FIG. 6 shows an example of a format for storing codes of the A channel and the B channel in a frame having a predetermined bit rate. In the x-th frame of the code string, the code of the x-th frame of the A-channel signal, the code of the x-th frame of the B-channel signal, and the channel configuration data are stored from the top. The channel configuration data indicates the type of the B channel signal in this frame. In this example, when the channel configuration data is 1, (L channel signal−R channel signal) / 2, Means an L channel signal, and 3 means an R channel signal.
[0027]
As described above, the control circuit 5 determines whether or not there is a correlation between the L channel signal and the R channel signal given from the frequency band dividing circuit 1 and the magnitude relationship between the two signals. -2 sets the type of signal output from -2 and controls the operation of each unit. The control circuit 5 sets a bit rate for each encoding block of each frame with respect to the first encoding circuit 3-1 and the second encoding circuit 3-2. This operation will be described with reference to the flowchart shown in FIG.
[0028]
First, it is determined whether or not the channel configuration data of the B channel signal has been switched from the previous frame in the current frame (S1). If the channel configuration data has been switched as a result of the determination in S1 (YES in S1), the value of the frame counter for counting the number of frames that have elapsed since the channel configuration data was last switched is set to 0 (S2 Then, a bit rate of 96 [kbps] is set for the second encoding circuit 3-2 (that is, a bit rate for encoding a B channel signal is set to 96 [kbps]). In addition, a bit rate of 32 [kbps] is provisionally set for the first encoding circuit 3-1 (that is, the bit rate for encoding the A channel signal is provisionally set to 32 [kbps]. (S3). When S3 ends, the process proceeds to S9 described later.
[0029]
On the other hand, as a result of the determination in S1, if the channel configuration data has not been switched (NO in S1), the value of the frame counter is incremented by 1 (S4). Note that the value of the frame counter is limited so as not to exceed a predetermined value. Next, it is determined whether or not the value of the frame counter is smaller than 2 (S5). As a result of the determination in S5, if the value of the frame counter is smaller than 2 (YES in S5), the process proceeds to S3 described above. On the other hand, if the value of the frame counter is not smaller than 2 (NO in S5), S6 to be described later Migrate to
[0030]
In S6, it is determined whether or not the channel configuration data of the current frame is 1. If the channel configuration data is 1 as a result of the determination in S6 (YES in S6), a bit rate of 32 [kbps] is set for the second encoding circuit 3-2 and the first encoding is performed. A bit rate of 96 [kbps] is provisionally set for the circuit 3-1 (S7). On the other hand, if the channel configuration data is not 1 (NO in S6), a bit rate of 48 [kbps] is set for the second encoding circuit 3-2 and the first encoding circuit 3-1 is set. On the other hand, a bit rate of 80 [kbps] is provisionally set (S8). When S7 and S8 are finished, the process proceeds to S9 described later.
[0031]
In S9, the second encoding circuit 3-2 is made to execute encoding of the B channel signal. As a result, the second encoding circuit 3-2 converts the block consisting of the previous frame of the B channel signal and the current frame into a signal on the frequency axis by the modified discrete cosine transform, and then the control circuit 5 Encode at the bit rate set by.
[0032]
Next, in the encoding process, not all of the allocated bit rates are used, so when the encoding of the B channel signal by the second encoding circuit 3-2 is completed, the set bits are set. A difference (that is, a surplus bit rate) from the bit rate used in actual encoding among the rates is calculated (S10). Next, the bit rate obtained by adding the bit rate calculated in S10 to the bit rate provisionally set for the first encoding circuit 3-1 is reset (S11).
[0033]
Then, the first encoding circuit 3-1 is caused to execute encoding of the A channel signal (S12). As a result, the first encoding circuit 3-1 converts the block consisting of the previous frame of the A channel signal and the current frame into a signal on the frequency axis by modified discrete cosine transform, and then the control circuit 5 Encode at the bit rate set by. The above processes S1 to S12 are performed for each frame.
[0034]
With the above configuration, for example, an L channel signal and an R channel signal as shown in FIGS. 8A and 8B are input. As a result, an A channel signal and a B channel signal are respectively shown in FIG. As shown in (c) and (d) of FIG. 8, the channel configuration data of the B channel signal is 1 in both the N-1 frame and the N frame forming the N block encoding block, Since the type of the B channel signal is not switched in the encoding block, the bit rate for encoding the B channel signal is set to 32 [kbps] in the encoding block of the Nth frame.
[0035]
On the other hand, in the Nth frame and the N + 1th frame forming the N + 1th frame coding block, the channel configuration data of the B channel signal is 1 and 3, respectively, and the type of the B channel signal is within the coding block. Since the switching is performed, the bit rate for encoding the B channel signal is set to 96 [kbps] in the encoding block of the (N + 1) th frame.
[0036]
In this way, in this embodiment, in the coding block in which the type of the B channel signal is switched in the block, the bit rate for encoding the B channel signal is set higher than in the block in which the switching is not performed. It is possible to suppress an increase in quantization noise due to switching of the B channel signal type.
[0037]
Here, in the second channel conversion circuit 3-2, when the channel configuration data is switched, in order to prevent the signal level from changing suddenly at the joint between the previous frame and the switched frame, 1 In some cases, the last sample data of the previous frame and the first few sample data of the switched frame are corrected so that both frames are smoothly connected. In this case, the channel configuration data is the same in the frame where the channel configuration data is switched and the next frame, but the sample data is corrected in the former frame, and the quantization noise increases at a low bit rate. Is concerned.
[0038]
However, in the present embodiment, when the number of frames that have elapsed since the channel configuration data was last switched is 1 by the process of S5 in FIG. 7, in other words, the frame in which the channel configuration data has switched and the next In a coding block composed of frames, a bit rate for coding a B channel signal is set high. Therefore, even if the sample data is corrected as described above, it is possible to suppress an increase in quantization noise due to the switching of the signal type. If the sample data is not corrected as described above, the determination threshold value in S5 may be set to 1, or S5 may be deleted.
[0039]
Further, when the channel configuration data is 1 (that is, when the B channel signal is (L channel signal−R channel signal) / 2), the degradation of the decoded L channel signal and the R channel signal is deteriorated. The degree is determined by the quantization noise of both the A channel signal and the B channel signal, whereas when the channel configuration data is other than 1 (that is, the B channel signal is the L channel signal or the R channel signal). In the case of a signal), the degree of degradation of the decoded L-channel signal or R-channel signal depends only on the quantization noise of the B-channel signal. Therefore, when the B channel signal is encoded at the same bit rate when the channel configuration data is 1 and when it is other than 1, when the L channel signal and the R channel signal are decoded, Deterioration is conspicuous in the portion where the channel configuration data of the signal is other than 1.
[0040]
On the other hand, in the present embodiment, the processing at S6 to S8 in FIG. 7 causes the bit rate for encoding the B channel signal to be higher when the channel configuration data is other than 1 than when it is 1. Therefore, it is possible to prevent the above-described problem from occurring.
[0041]
Further, in the present embodiment, even if the bit rate for encoding the B channel signal is set high by the processing of S10 and S11 in FIG. 7, the bit rate that is not used in the actual encoding is set as follows. Since the first channel signal is assigned to the bit rate for encoding, the encoding can be performed more efficiently.
[0042]
As an embodiment of the present invention, only the case where the input signal is 2 channels and the unit block to be encoded is 2 frames is shown, but the case where the input signal is 3 channels or more, or encoding is performed. The present invention can also be applied when the unit block to be performed is three frames or more.
[0043]
【The invention's effect】
As described above, according to the encoding apparatus of the present invention, for each unit block to be encoded, the bit rate is set according to whether or not the type of signal to be encoded within the block is switched. Therefore, when switching, a higher bit rate is assigned than when switching, so that an increase in quantization noise due to switching of the signal type can be suppressed.
[0044]
Also, since the bit rate is set according to the number of frames that have passed since the signal type was last switched, the signal type is not switched within the unit block to be encoded, but the signal type from the previous frame Even if the block containing a frame that has been switched is assigned a high bit rate and the data is corrected so that it is smoothly connected to the previous frame when the signal type is switched, the signal type It is possible to prevent an increase in quantization noise due to switching of.
[0045]
In addition, since the bit rate is set for each unit block to be encoded according to the type of signal to be encoded in the block, the quantization noise is greatly deteriorated in the degradation of the decoded signal compared to other types. When encoding the affected signal types, a high bit rate can be set to reduce the quantization noise, thereby preventing the problem that the degradation of the decoded signal becomes partially noticeable. be able to.
[0046]
In addition, the bit rate that is not used in the actual encoding even if the bit rate for encoding a signal that may switch the signal type in the unit block to be encoded is set high. Therefore, it is possible to perform the encoding more efficiently.
[Brief description of the drawings]
FIG. 1 is a block diagram of an encoding apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating an example of waveforms of an L channel signal and an R channel signal;
3 corresponds to (L channel signal + R channel signal) / 2 and (L channel signal−R channel signal) when the waveforms of the L channel signal and the R channel signal are those shown in FIG. It is a figure which shows the waveform of) / 2.
FIG. 4 is a diagram illustrating another example of waveforms of an L channel signal and an R channel signal.
5 corresponds to (L channel signal + R channel signal) / 2 and (L channel signal−R channel signal) when the waveforms of the L channel signal and the R channel signal are those shown in FIG. It is a figure which shows the waveform of) / 2.
FIG. 6 is a diagram for explaining a code string generated by a code string generation circuit.
FIG. 7 is a flowchart of processing for setting a bit rate for each coding block;
FIG. 8 is a diagram illustrating an example of waveforms of an L channel signal and an R channel signal, and waveforms of an A channel signal and a B channel signal corresponding thereto.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Frequency band dividing circuit 2-1 1st channel conversion circuit 2-2 2nd channel conversion circuit 3-1 1st encoding circuit 3-2 2nd encoding circuit 4 Code stream generation circuit 5 Control circuit

Claims (5)

First channel conversion means for outputting a signal obtained by combining all input signals of a plurality of channels as a first channel signal;
A second channel conversion means for outputting a signal obtained by synthesizing all input signals of a plurality of channels or an input signal of a part of channels as a second channel signal;
Channel conversion control means for setting a signal output from the second channel conversion means for each frame in accordance with the relationship of the input signals of each channel;
A coding unit comprising: a coding unit configured to convert a signal output from the first channel conversion unit and the second channel conversion unit into a signal on a frequency axis for each block composed of a plurality of frames and then perform coding. In the conversion device,
The bit rate for encoding the signal output from the second channel conversion unit is not switched when the type of the signal output from the second channel conversion unit is switched in the block to be encoded. A coding apparatus comprising bit rate setting means that is higher than the case and is set for each block to be encoded. The bit rate setting means encodes the bit rate for encoding the signal output from the second channel conversion means after the last switching of the type of signal output from the second channel conversion means. The encoding apparatus according to claim 1, wherein the encoding apparatus is set according to the number of frames. The bit rate setting means sets the bit rate for encoding the signal output from the second channel conversion means according to the type of the signal output from the second channel conversion means. The encoding apparatus according to claim 1 or 2, wherein the encoding apparatus is characterized. The encoding means encodes the signal output from the second channel conversion means first, and outputs the signal output from the first channel conversion means and the signal output from the second channel conversion means. Of the bit rates set by the bit rate setting means for encoding, the signal output from the second channel conversion means is output from the first channel conversion means at the bit rate remaining after encoding. 4. The encoding apparatus according to claim 1, wherein the encoded signal is encoded. 5. The encoding apparatus according to claim 1, wherein the input signals of the plurality of channels are audio stereo signals.

Images