JP2003330497A - Method and device for encoding audio signal, encoding and decoding system, program for executing encoding, and recording medium with the program recorded thereon - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To minimize artifacts caused at the time of switching of a state mode of MS stereo encoding. <P>SOLUTION: A joint stereo encoder 14-1 encodes a digitized audio signal which includes a pair of stereo channels of a left channel and a right channel and has each frame divided into consecutive frames consisting of a plurality of time samples, to one of a plurality of state modes of MS stereo encoding. In this encoding, the joint stereo encoder 14-1 calculates respective energies of the left channel and the right channel of a current frame of the audio signal and determines a state mode of MS stereo encoding of the current frame on the basis of the value of a ratio of the energy of the right channel to that of the left channel and the state mode of MS stereo encoding of at least one just preceding frame. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a midside (M
S) In stereo coding, an audio signal coding method, apparatus, and code that enable transitions between various state modes related to MS stereo coding to be performed without introducing artifacts (pseudo signals) The present invention relates to an encoding / decoding system, a program for executing encoding, and a recording medium recording the program.

[0002]

The purpose of encoding an audio signal is to digitize it so that the encoded audio stream retains the highest possible fidelity to the original source even after being decompressed in a decoder. It is to compress the audio stream into as few bits as possible in the encoder. A commonly used compression method for stereo signals with left (L) and right (R) channels is known as joint stereo coding. Joint stereo coding is
It is a collective term for two different encoding methods known as midside stereo (generally abbreviated as MS stereo) and intensity stereo (generally abbreviated as IS stereo). FIG. 4 shows a stereo signal 31
A typical example of applying the MS stereo coding and the IS stereo coding to generate the joint stereo signal 32 is presented. Prior to the conversion, each channel frame is filtered into at least two subband frames, one subband frame being 0 to 5.5k.
Hz frequency band, the other represents a frequency band higher than 5.5 kHz.

IS stereo coding is based on high frequency (> 5).
Only the sum signal of the left channel and the right channel is coded by taking advantage of the fact that the ear cannot accurately discriminate the spatiality at (kHz). This effectively cuts in half the information sent from each channel of the original stereo. In the decoder, the combined signal is amplified or attenuated by different amounts depending on the predetermined amplification factor and attenuation factor, and the new signal is assigned to the left and right channels, so that the spatial sense is partially Remade. The amplification coefficient and the attenuation coefficient are determined by the encoder.

Unfortunately, for frequencies below 5 kHz of the audio signal's spectrum, the IS stereo coding cannot be applied because the ears are sensitive enough to identify spatiality. In this frequency band, MS stereo coding, which is the main point of the present invention, is usually used.

MS stereo coding transforms the original two-channel signal to produce a new signal pair in which one channel is more important (ie, contains more information) than the other. For example, for many common signals where the left channel (L) and the right channel (R) are very similar to each other, the original stereo channels are added to and subtracted from each other to create a new channel. Can be formed. The R'channel, which is the "difference" signal, carries a weaker signal than the L'channel, which is the "sum" signal, and thus requires fewer bits for encoding. There are different variants of MS stereo coding for various audio encoders. AT
In the case of RAC3, the L'channel is always the average signal (L + R) / 2 of the original channel, while the R'channel is half the difference signal (LR) / 2, half of the left channel. Signal L / 2 and the signal -R / 2 with the sign of the right channel half inverted. Hereinafter, these three possible "state modes" of MS stereo coding are referred to as S / D mode, S / L mode, and S / R mode, respectively.

Sub-band frames (ie 0 to 5.5 kHz) to which MS stereo coding has to be applied.
When the subband frame of the frequency of
The remaining parameters to be determined before the formation of the C3 compliant bitstream are the channel switching, which indicates whether the current state mode for MS stereo coding is S / D, S / L or S / R. Information SWI only.

[0007]

[Table 1] ―――――――――――――――――――――――――――――――――――― SWI state mode L'R 'Conditions for prior art ――――――――――――――――――――――――――――――――――――   3 S / D (L + R) / 2 (L-R) / 2 When L and R are similar   0 S / L (L + R) / 2 L / 2 R is more important than L   1 S / R (L + R) / 2 -R / 2 When L is more important than R ――――――――――――――――――――――――――――――――――――

In the prior art method, L channel and R
As a measure of the "importance" of a channel, those L
The channel and R channel energies are used. The subband frame energies E _L and E _R are usually approximated by the sum of the squares of all time samples in the subband frame of each channel. Based on this prior art condition, a simple method for determining the state mode of MS stereo coding of the current subband frame is as follows, where f is a threshold factor greater than one.

[0009]

## EQU1 ## When E _R > f × E _L , SWI = 0. Otherwise, when E _L > f × E _R , SWI = 1. When none of the above, SWI = 3.

[0010]

The simple and direct method as described above is such that the energy E _L and E _R values are

## EQU2 ## E _R ≈f × E _L or

Respect Equation 3] E _L ≒ f × E _R range around the value smoothly changes to drifting signal of, between the two-state mode of the MS stereo coding (S / D and S / L Fast toggle between modes or between S / D and S / R modes, i.e. it can cause a rapid round-trip switch between state modes of MS stereo coding. It has been discovered that such a fast toggle between the two state modes associated with MS stereo coding causes audible artifacts in real world implementations. FIG. 9 shows an example of this, and is a graph showing an example in which the state mode switching process of the MS stereo coding according to the related art is applied to the signals of the L channel and the R channel. In FIG. 9, using t as a discrete-time parameter, between the subband frame t and the subband frame t + 3, due to a slight change in the energy E _L , the S / D mode and the S / L mode are changed. A high-speed toggle that reciprocates is occurring. This high-speed toggle can cause artifacts in the signal where the signal strength is changing smoothly.

After conversion, MS stereo coding is used to allocate less bits to the less important channels, and the imbalanced signals (L 'and R').
Channel signal), while precautions must be taken to ensure that no artifacts result from the fast toggle between the two state modes associated with MS stereo coding. For example, as shown in FIG.
In the case of a signal whose signal strength changes smoothly between subband frames, when the S / L or S / R mode is entered, the mode is changed to S / L in the next subband frame.
A mechanism must be implemented to prevent returning (leaping back) to D mode.

Nevertheless, such preventive measures are
It should not be applied excessively enough to undermine a signal that can reasonably be justified in S stereo coding to transition to a different state mode than the previous frame. This is because the artifacts are also introduced when the state mode switching for MS stereo coding fails when needed. Therefore, the mechanism described above for preventing excessive switching of state modes must be applied sparingly.

The object of the present invention is to solve the above problems,
Audio signal encoding method and apparatus, encoding and decoding system capable of preventing occurrence of undesirable artifacts in MS stereo encoding state mode switching, program for executing encoding, and recording of the program To provide the recording medium.

[0014]

SUMMARY OF THE INVENTION An audio signal encoding method according to an aspect of the present invention includes a pair of stereo channels consisting of a left channel and a right channel, each frame being a sequence of a plurality of time samples. A digitized audio signal divided into typical frames is encoded into any one of a plurality of state modes of midside stereo encoding, and the encoding method of the audio signal comprises: Calculating each of the left and right channel energies of the current frame of the signal, a value of the ratio of the right channel energy to the left channel energy, and at least one preceding side mid-side stereo code The current frame's midsize based on the Characterized in that it comprises the step of determining the status mode of de stereo coding.

In the audio signal encoding method, the mid-side stereo encoding includes a first state mode including a sum signal and a difference signal of left and right channel signals and a sum signal of left and right channel signals. It is characterized by including a second state mode including a left channel signal and a third state mode including a sum signal of left and right channel signals and a right channel signal.

In the audio signal encoding method, the step of determining the state mode of the mid-side stereo encoding of the current frame is performed when the value of the ratio is larger than a predetermined first threshold value. , Transitioning the mid-side stereo coding state mode of the current frame to the second state mode; and, when the value of the ratio is less than the reciprocal of the first threshold, mid of the current frame. Transitioning the side stereo coding state mode to the third state mode, and the mid-side stereo coding state mode of the immediately preceding preceding frame is in the third state mode, and the value of the ratio is , Less than or equal to the first threshold and greater than a predetermined second threshold less than the first threshold, the current frame A step of shifting the state mode of the mid-side stereo coding in the first state mode, the state mode of the mid-side stereo coding of the previous frame to the preceding is in the second state mode,
The value of the ratio is greater than or equal to the reciprocal of the first threshold,
And transitioning the mid-side stereo coding state mode of the current frame to the first state mode when it is smaller than the reciprocal of the second threshold value.

Further, in the audio signal encoding method, the step of determining the state mode of the midside stereo encoding of the current frame is performed by determining the state mode of the midside stereo encoding of the immediately preceding frame. Holding the mid-side stereo coding state mode of the current frame in the second state mode when in the second state mode and the value of the ratio is greater than or equal to the reciprocal of the second threshold. And the mid-side stereo coding state mode of the preceding immediately preceding frame is in the third state mode and the ratio value is less than or equal to the second threshold value, the mid-side of the current frame is Holding the stereo encoding state mode in the third state mode.

An audio signal encoding apparatus according to another aspect of the present invention includes a continuous frame including a pair of stereo channels consisting of a left channel and a right channel, each frame including a plurality of time samples. A digitized audio signal divided into a plurality of state modes of midside stereo encoding is encoded into one of the state modes, Means for calculating the respective energies of the left channel and the right channel of the frame, a value of the ratio of the energy of the right channel to the energy of the left channel, and at least one preceding mid-side stereo coding of the preceding frame. Based on the state mode and the midsides of the current frame Characterized by comprising a determining means for determining a state mode of Leo coding.

In the audio signal encoding device, the mid-side stereo encoding is performed in a first state mode including a sum signal and a difference signal of left and right channel signals, and a sum signal of left and right channel signals. It is characterized by including a second state mode including a left channel signal and a third state mode including a sum signal of left and right channel signals and a right channel signal.

Further, in the audio signal encoding apparatus, the determining means determines the mid-side stereo encoding state mode of the current frame when the value of the ratio is larger than a predetermined first threshold value. Transition to the second state mode, and when the value of the ratio is less than the reciprocal of the first threshold, transition the mid-side stereo coding state mode of the current frame to the third state mode. Then, the state mode of the mid-side stereo coding of the preceding immediately preceding frame is in the third state mode,
When the ratio value is less than or equal to the first threshold value and greater than a predetermined second threshold value that is less than the first threshold value, the mid-side stereo coding of the current frame is performed. The state mode is transited to the first state mode, and the state mode of mid-side stereo coding of the preceding immediately preceding frame is in the second state mode,
The value of the ratio is greater than or equal to the reciprocal of the first threshold,
And, when it is smaller than the reciprocal of the second threshold value, the mid-side stereo coding state mode of the current frame is transited to the first state mode.

Further, in the audio signal encoding apparatus, the determining means has the state mode of the midside stereo encoding of the immediately preceding preceding frame in the second state mode, and the value of the ratio is the above. When the reciprocal of the second threshold value or more, the state mode of the midside stereo encoding of the current frame is held in the second state mode, and the state of the midside stereo encoding of the immediately preceding frame is held. Keep the mid-side stereo coding state mode of the current frame in the third state mode when the mode is in the third state mode and the ratio value is less than or equal to the second threshold value. It is characterized by

An audio signal encoding and decoding system according to another aspect of the present invention includes the audio signal encoding device and the audio signal decoding device, and the audio signal decoding device includes: Means for decoding the encoded audio signal based on the determined state mode of the mid-side stereo encoding, and obtaining and outputting the decoded audio signal including a pair of stereo channels Is characterized by.

A program according to still another aspect of the present invention is characterized by including steps related to the audio signal encoding method.

According to another aspect of the present invention, there is provided a computer-readable recording medium in which the above program is recorded.

To achieve the above-mentioned object, the criterion for switching between the various state modes considers not only the energy of the original channel but also the state mode of the MS stereo coding of the preceding subband frame. There must be.

An audio signal encoding method according to another aspect of the present invention comprises: 1) calculating energies E _L and E _R for both channels of the current subband frame; and 2) preceding subband. and recording the state mode of the MS stereo coding applied to the frame, 3) for a given range of energy values of the ratio E _{R /} E L _(or E _{L /} E _R), the status mode is preceded Ensuring that the state mode is inherited, which means that the predetermined state mode A to the predetermined state mode B
Can be achieved if the condition for switching to state mode B is not the reverse of the condition for switching from state mode B, 4) one channel rises sharply compared to the other. The step of allowing fast state mode switching for MS stereo coding in the event of an indication of energy to occur.

A method for encoding an audio signal according to another aspect of the present invention determines the state mode for MS stereo encoding of a digitized audio signal while minimizing the occurrence of artifacts or A method of encoding an audio signal that does not generate any artifacts, said audio signal comprising a pair of stereo channels divided into consecutive frames of a plurality of time samples, wherein: Energy E _L of the left and right channels of the frame
And E _R , and (b) the current frame based on the value of the ratio between the energies E _L and E _R and the state mode of MS stereo coding of at least one preceding immediately preceding frame. Determining the state mode of the MS stereo encoding of.

In the above audio signal encoding method
And the value E of the above ratio within a predetermined range_R/ E _LOr E_L/ E_R
For the current frame MS stereo coding state
The mode is the status mode of MS stereo coding of the previous frame.
It is characterized by inheriting the code.

Further, in the above audio signal encoding method, the condition for switching from the state mode A to the state mode B is not the reverse of the condition for switching from the state mode B to the state mode A.

Further, in the above audio signal encoding method, when the energy of one channel rises sharply with respect to the energy of the other channel, M
It is characterized in that fast switching to another state mode of S stereo coding is allowed.

Furthermore, in the above-mentioned audio signal encoding method, the condition is such that the value of the ratio is below or above a predetermined value.

Further, in the above-mentioned audio signal encoding method, the abruptness of the energy increase of one channel with respect to the energy of the other channel is also evaluated by the value of the ratio.

[0033]

DETAILED DESCRIPTION OF THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. In the present specification, the ATR
Although a mini disk recording / reproducing system for AC3 will be described with reference to an embodiment, the encoding method of the present invention can be applied to other encoders using joint stereo encoding.

FIG. 1 is a block diagram showing the configuration of a mini disk recording / reproducing system according to an embodiment of the present invention. This embodiment is particularly characterized by the detailed configuration inside the audio encoder 2.

In FIG. 1, an A / D converter 1 A / D-converts an audio input signal to convert it into a digitized audio sample signal, and then an audio encoder 2 uses a joint stereo coding as will be described later in detail. The converted audio sample signal is compressed and encoded using processing to generate an ATRAC3 bitstream signal. Then, the mini disk recording device 3
Modulates the bit stream signal of the ATRAC3 into a predetermined modulation recording signal, and then records the modulation recording signal on the mini disk 4. On the other hand, the mini disc reproducing apparatus 5 reproduces the recording signal from the mini disc 4 and demodulates it to reproduce the bit stream signal of the ATRAC 3.
In addition, the audio decoder 6 has a compressed ATR
The bit stream signal of AC3 is decoded into a digital audio signal by using a joint stereo decoding process which is the reverse of the above-mentioned joint stereo coding process, and finally the D / A converter 7 converts the digital audio signal into a D / A signal. A-converts and outputs an audio output signal.

By selecting an appropriate MS stereo coding state mode in the above-mentioned joint stereo coding processing, it is possible to prevent the occurrence of artifacts that occur during coding and decoding of digital audio signals. In this embodiment, the recording / reproducing system is configured to record the bit stream signal of the ATRAC3 on the mini disc 4, but the present invention is not limited to this, and recording is performed using another recording medium or transmission medium. A reproduction system and a transmission system may be configured.

FIG. 2 is a block diagram showing the detailed arrangement of the audio encoder 2 shown in FIG. In FIG. 2, the audio encoder 2 is a typical conversion encoder using the audio signal compression technology of ATRAC3. Among the detailed configurations of the audio encoder 2, in particular, the details of the joint stereo encoder 14-1. It has a characteristic in the configuration. ATRAC3 is basically a hybrid coding technique of subband coding and transform coding designed for a stream of audio sample signals sampled at 44.1 kHz.

As shown in FIG. 2, the audio encoder 2 uses a two-stage QMF filter 10 to convert a frame of a continuous audio sample signal with 1024 samples in the time domain into four sub-bands ( A subband) signal is frequency-band-divided, and then a signal (hereinafter referred to as a subband frame) divided into frames each having 256 sample signals in the time direction is output for each partial band. Here, in detail, the QMF division filter 10 further includes a low-pass filter 11 and a high-pass filter 12 that equally divide the audio sample signal into a low-pass signal and a high-pass signal, and further divides the divided low-pass and high-pass signals with respect to frequency. Filter bank 13 for equal division
-1 to 13-4. Therefore, QM
The output of the F filter is four subband frames that are downsampled, each with 256 samples. Each of these is M on the time axis.
Half the length of the transform block for the DCT transform,
0 to 5.5 kHz (subband 0), 5.5 kHz to 11 kHz (subband 1), 11 kHz to 16.
5 kHz (subband 2), and 16.5 kHz to 2
It is an audio signal obtained from each frequency band of 2 kHz (subband 3).

Next, the joint stereo encoder 14
-1 to 14-4, the joint stereo coding process is applied to these four time domain signals at the output of the QMF filter. Due to the sensitivity to spatiality mentioned above, MS stereo coding is only necessary for subband frames of subband 0, and IS stereo coding is better for subband frames of higher frequency bands. As it seems more useful and effective,
The joint stereo encoder 14-1 performs only MS stereo encoding, and the joint stereo encoder 14-
2 to 14-4 perform IS stereo coding only.

The joint stereo encoder 14-1 is
Select the appropriate MS stereo coding state mode for each temporally consecutive subband frame of subband 0. As described above, the MS stereo coding state mode is the S / D mode (that is, the sum signal and the difference signal of the left and right channel signals for each subband frame).
First with (L + R) / 2 and (LR) / 2 signals
State mode), and an S / L mode including a sum signal of left and right channel signals and a left channel signal (that is,
An S / R mode including a second state mode composed of (L + R) / 2 and L / 2 signals and a sum signal of left and right channel signals and a right channel signal (that is, (L + R)
R) / 2 and a third state mode composed of signals of −R / 2). The joint stereo encoder 14-1 calculates the energy of the left channel and the energy of the right channel of the current frame of the audio signal, and calculates the ratio of the energy of the right channel to the energy of the left channel. The MS of the current frame based on the value and the state mode of MS stereo coding of at least one preceding immediately preceding frame.
And a determination means (not shown) for determining a stereo encoding state mode. MS stereo coded L'channel signal (ie (L + R) / 2 signal)
The distribution of the number of bits for the R'channel signal and the R'channel signal (that is, the (LR) / 2 signal, the L / 2 signal, or the -R / 2 signal) can be variable for each subband frame. Channel switching information S of the MS stereo coding state mode determined for each subband frame
The WI and the information on the allocation of the number of bits for the L ′ and R ′ channel signals are transmitted to the audio decoder 6 as side information. Joint stereo encoder 14-1
The details of the state mode switching process of the MS stereo encoding executed by will be described later.

Following the joint stereo coding process,
In gain control sections 15-1 to 15-4, gain control is executed by calculating a gain coefficient (that is, an amplification factor or an attenuation factor) for each subband frame. Then MDC
The T processing unit 16-1 has a modified discrete cosine transform (MDCT) for each signal of at least two subband frames that are gain-controlled in one frequency band and are cascade-connected to a transform block for MDCT processing. Is executed, and the resulting spectrum information signal is output to the tone component encoder 17-1. The tone component encoder 17-1 includes a tone component detection unit, a scale factor processing unit, a bit allocation unit, and the like, and separates the tone component signal and the non-tone spectrum signal from the MDCT-processed information signal. And a tone component quantizer 18-1 and a spectrum quantizer 19 respectively.
Output to -1. Here, the tone component quantizer 18-1
The spectrum quantizer 19-1 separately processes the tone component signal and the non-tone spectrum signal for encoding (quantization). Next, the bit stream multiplexer 20 is configured to include a Huffman encoder and a multiplexer, and outputs a plurality of signals (for example, 14 bits) encoded by the tone component quantizer 18-1 and the spectrum quantizer 19-1. Huffman table)
Next, tone component and spectrum information in the tone component encoders 17-1 to 17-4, each encoded tone component, each encoded non-tone component, and the joint stereo encoder 14-1. To 14-4 output the joint stereo coding processing data (for example, MS stereo coding channel switching information SW).
Information on the allocation of the number of bits for the I, L'and R'channel signals, and the amplification factor and / or of IS stereo coding
Or attenuation coefficient), gain coefficient data output from the gain control units 15-1 to 15-4 (for example, data of a position for gain control in a subband frame, data on a level for gain control, and gain control). Data of the number of points to be included.) And the side information in accordance with the ATRAC3 standard are multiplexed to obtain and output the bitstream signal of ATRAC3.

FIG. 3 is a block diagram showing the detailed structure of the audio decoder 6 shown in FIG. First, the bit stream demultiplexer 21 reads out from a recording medium such as a mini disk 4 or a transmission medium and then demodulates the AT.
The bitstream signal of the RAC3 is separated into a tone component code string and a spectrum code string (non-tone component code string) for each subband, and side information including joint stereo coding processing and gain coefficient data. . Next, the tone component dequantizer 22-1 and the spectrum dequantizer 23-1 respectively decode (dequantize) each code string into spectrum coefficients, and the tone component decoding unit 24-1 is separated. The decoded tone component signal and the non-tone spectrum signal are combined. The same applies to the tone component code strings and spectrum code strings in other subbands. Next, the inverse MDCT processing units 25-1 to 25-4 generate time-domain subband frames by inverse MDCT for each frequency band, and then the inverse gain control units 26-1 to 26-4 perform bitstream processing. Based on the gain control data in the side information input from the demultiplexer 21, inverse gain control is performed with a gain coefficient opposite to that when gain control is performed for each subband frame, that is, amplified by gain control. The part is suppressed, the suppressed part is amplified, and the inverse-gain-controlled sub-band frame of each band is joint stereo decoder 27-
1 to 27-4.

Next, the joint stereo decoder 27-
1 to 27-4 perform MS stereo decoding and IS stereo decoding on the inverse gain-controlled subband frame based on the side information transmitted from the audio encoder 2 to output the audio encoder. The original stereo signal input to 2 is restored and output to the QMF synthesis filter 28. Here, the joint stereo decoder 27-1 performs MS stereo decoding on the input subband frame to restore and output an audio signal including a stereo channel before being MS stereo encoded. And joint stereo decoder 27-2
27 to 27-4 each perform IS stereo decoding on the input subband frame, and output an audio signal in which the L channel and the R channel are amplified and / or attenuated by a predetermined amount. . MS stereo decoding and IS stereo decoding are collectively referred to as joint stereo decoding. Finally, the QMF synthesis filter 28 synthesizes the sub-band frames for each band (frequency band) and outputs the synthesized digital audio signal.

As described above, the mini-disc recording / reproducing system of FIG. 1 has an audio signal code including the audio encoder 2 which is an audio signal encoding device and the audio decoder 6 which is an audio signal decoding device. It is an encryption and decryption system. The audio encoder 2 includes a joint stereo encoder 14-1 that encodes a subband frame of an input audio signal into any one of a plurality of state modes of midside stereo encoding. ,Also,
The audio decoder 6 decodes the encoded audio signal based on the determined state mode of the midside stereo encoding, obtains and outputs the decoded audio signal including a pair of stereo channels. The joint stereo decoder 27-1 is provided.

In FIG. 5, the state mode switching process of MS stereo coding according to the present embodiment is presented in the form of a state transition diagram so as to emphasize the property of the switching characteristic depending on the state mode. The nodes 103, 100, 101 are
3 illustrates three possible state modes (S / D, S / L, S / R) for MS stereo encoding of ATRAC3.
Branches indicate conditions under which transitions between state modes may occur. If a transition occurs, it occurs once for each temporally consecutive subband frame.

First, the energies E _L and E _R of the channel are calculated. The energy E _{L of the} left channel is
It is defined as the sum of the squares of all 256 time domain samples of the subband frame in subband 0 of the left channel. The right channel energy E _R is likewise the sum of the squares of all 256 time domain samples of the subband frame in subband 0 of the right channel.

When the state mode of MS stereo coding of the immediately preceding subband frame is the S / D mode (SWI = 3 as in the case of node 103), the threshold coefficient f
₁ >> as ₁

## EQU00004 ## If E _R > f ₁ × E _L holds and the energy E _R indicates an imbalanced state larger than E _L , the state mode of MS stereo coding of the current subband frame is the S / L mode. (Node 10
SWI = 0) as in the case of 0. However, in order for the S / L mode to be switched back to the S / D mode, the condition imposed is the reverse of the above condition, ie

[Equation 5] Not E _L ≧ E _R / f ₁ ,

## EQU6 ## where f ₁ × E _R ≧ E _L > f ₂ × E _R. Here, the threshold coefficient f ₂ is

Meets Equation 7] _{f 1 »f 2 »(1 /} f 1), the above equation, the energy _{E L} is substantially _E R / f
Greater than ₁ (since f ₂ >> ₁ / f ₁ ), but f ₁ × E
Below _R , it is in a state not too large (because f ₁ >> f ₂ ). By doing so, the state mode switching process of the MS stereo coding of the present embodiment, the energy E _L of the left channel E _R
A signal that rises slowly and changes smoothly with respect to / f _{1 is} protected from a high-speed toggle between the S / L mode (SWI = 0) and the S / D mode (SWI = 3). That is, the condition of the following equation

As long as E _L ≦ f ₂ × E _R is satisfied, the state mode of MS stereo coding of the current subband frame is kept in S / L mode. FIG. 8 shows an example in which the state mode switching process of the MS stereo coding according to the present embodiment is applied to the same audio signal as in FIG. 9, and the high-speed toggle generated in the conventional example of FIG. It can be seen that the embodiment shown in FIG. The condition (Equation 6) for switching the S / L mode back to the S / D mode is not the reverse (Equation 5) of the condition for switching the S / D mode to the S / L mode.
This prevents high speed toggle between the S / D and S / L modes.

On the other hand, the energy E _L of the left channel is E _R / f ₁ between subband frames that are temporally close to each other.
A sharp rise with respect to, it is allowed to switch from the S / L mode to the S / D mode at a high speed in the next subband frame. This means that in the next subband frame

## EQU9 ## If f ₁ × E _R ≧ E _L > f ₂ × E _R occurs. In the case of signals with rapidly changing signal strength, the fast switching of the state modes of MS stereo coding means that any artifacts that may be introduced by the fast state mode switching are usually psychoacoustic due to a sudden surge of signal strength. It is legally accepted because it is masked. Furthermore, due to the extension of this method, the situation expressed by the following equation in the next subband frame

If E _L > f ₁ × E _R occurs, fast switching from S / L mode to S / R mode (SWI = 1 as in node 101) is also allowed.

Similarly, the MS of the immediately preceding subband frame
The stereo encoding state mode is the S / D mode (node 1
When SWI = 3) as in the case of 03, the energy
-E _LIs energy E_RIndicates a greater imbalance

If E _L > f ₁ × E _R holds, the state mode of MS stereo coding of the current subband frame is switched to S / R mode (SWI = 1 as in the case of node 101). However, in order for the S / R mode to be switched back to the S / D mode again, the condition imposed is the opposite of the above condition, ie

[Equation 12] Not E _R ≧ E _L / f ₁ ,

Equation 13 is _{f 1 × E L ≧ E R} > f 2 × E L, this condition, the energy _{E R} of the right channel is substantially larger than _{E L / f 1 (f 2} »1 / Since f ₁ ), the state is not too large below f ₁ × E _L (because f ₁ >> f ₂ ).
By doing so, the state mode switching process of the MS stereo coding according to the present embodiment, the signal energy E _R of the right channel changes smoothly and gradually increases with respect to E _{L /} f _1, S / R mode (SWI =
Protect from fast toggle between 1) and S / D mode (SWI = 3). That is, the condition of the following equation

As long as E _R ≦ f ₂ × E _L is satisfied, the state mode of MS stereo coding of the current subband frame is kept in S / R mode. If the condition mode of the MS stereo coding of the immediately preceding subband frame is the S / D mode (SWI = 3 as in the case of the node 103), if the conditions of Eqs. The sub-band frame state mode is maintained as it is in the S / D mode.

In a similar manner, the energy E _R of the right channel is E _L between subband frames that are close in time.
When it rapidly rises with respect to / f ₁ , it is allowed to switch from the S / R mode to the S / D mode at a high speed in the next subband frame. This means that in the next subband frame

## EQU15 ## If f ₁ × E _L ≧ E _R > f ₂ × E _L occurs. Furthermore, in the next subband frame, the situation expressed by

If E _R > f ₁ × E _L occurs, fast switching from S / R mode to S / L mode (SWI = 0 as in node 100) is also allowed.

FIG. 6 is another state transition diagram for explaining the principle of the state mode switching process of the MS stereo coding described with reference to FIG. In FIG. 6A, when the state mode of the MS stereo coding of the immediately preceding subband frame is the S / D mode, in order to transit to the S / L mode, the energy ratio value E _R / E _L Must exceed the threshold factor f ₁ . Further, in order to transit from the S / D mode to the S / R mode, the ratio value E _R / E _L
Must be smaller than the threshold coefficient 1 / f ₁ (that is, the ratio value E _L / E _R must be larger than the threshold coefficient f ₁ ). In FIG. 6B, when the state mode of MS stereo coding of the immediately preceding subband frame is S / L,
In order to transit to the S / D mode, the ratio value E _R / E _L is smaller than the threshold coefficient 1 / f ₂ and the threshold value 1 / f _2.
Must be greater than or equal to f ₁ . Further, in order to transit from the S / L mode to the S / R mode, the ratio value E _R / E _L needs to be smaller than the threshold coefficient 1 / f ₁ . On the other hand, when the ratio value E _R / E _L is equal to or greater than the threshold coefficient 1 / f ₂ , the current MS stereo coding state mode of the subband frame is held in the S / L mode. Similarly, in FIG. 6C, when the MS stereo coding state mode of the immediately preceding subband frame is S / R, S / D
In order to enter the mode, the ratio value E _R / E _L must be greater than the threshold coefficient f ₂ and less than or equal to the threshold coefficient f ₁ . Further, in order to transit from the S / R mode to the S / L mode, the ratio value E _R / E _L is equal to the threshold coefficient f.
Must be greater than ₁ . On the other hand, the ratio value E _R / E _L
Is a threshold coefficient f ₂ or less, the state mode of MS stereo coding of the current subband frame is S / S.
Hold in R mode.

As a result of the adjustment, for ATRAC3, it has been found that the threshold coefficients f ₁ = 100 and f ₂ = 1 are reasonable values, but other encoders have other values. It may be more appropriate. In general, it is desirable that the difference between the two threshold factors is several tens of dB.

In accordance with the above principle, joint stereo
The encoder 14-1 uses the MS stereo as shown in FIG.
The encoding state mode switching process is executed. Of FIG.
In step S1, the immediately preceding subband is used as initialization processing.
The state mode of MS stereo coding of the frame is set to S / D mode.
The current status is set in step S2.
Band channel L channel and R channel
Energy E_LAnd E_RTo calculate. Then, step S
In 3, the MS stereo of the immediately preceding subband frame
Determine what the encoding state mode is. The last service
If the band frame is in S / L mode,
At S4, the energy of the L and R channels
The ratio value E_R/ E_LIs the reciprocal of the threshold coefficient 1 / f_TwoYo
Is less than (ie E_L> F_Two× E_RIs
Or not) is determined. When step S4 is NO,
At step S6, the MS stereo of the current subband frame
Set the encoding status mode to S / L mode and
Go to step S13. When step S4 is YES
Is the energy ratio value E in step S5._R/ E_LBut
Reciprocal of threshold coefficient 1 / f₁Is less than (
Wow, E_L> F₁× E _ROr not) is determined.
When step S5 is YES, step S12
In the current MS stereo coding of the subband frame
Set the state mode to S / R mode, and step S13
Proceed to. When step S5 is NO, step S9
In the current MS stereo coding of the subband frame
Set the status mode to S / D mode, and step S13
Proceed to.

In step S3, the immediately preceding subband frame
If the system is in S / D mode, then in step S7, L
Value E of the energy ratio of the channel and the R channel_R
/ E _LIs the threshold coefficient f₁Whether it is greater than
Chi, E_R> F₁× E_LOr not) is determined. Su
If step S7 is YES, the process proceeds to step S6.
If step S7 is NO, step S8
Energy ratio value E_R/ E_LIs the reciprocal of the threshold coefficient 1 /
f₁Less than (ie E_L> F ₁× E_R
Or not) is determined. If step S8 is YES
If there is, go to step S12, step S8 NO
When it is, it progresses to step S9. Also, step S3
Then, of the MS stereo coding of the immediately preceding subband frame
If the state mode is S / R, step S10
And the energy ratio value E_R/ E_LIs the threshold coefficient f_Two
Greater than or equal to (ie E_R> F_Two× E_LAnd
Or not) is determined. Step S10 is NO
If so, the process proceeds to step S12. YES in step S10
If it is, in step S11, the value of the energy ratio
E_R/ E_LIs the threshold coefficient f₁Is greater than (
N, E_R> F₁× E_LOr not) is determined
It When step S11 is YES, step S6
Proceed to step S11 if NO in step S11.
Proceed to 9.

Then, in step S13, it is determined whether or not the current subband frame is the last subband frame of the input signal. Step S13 is NO
If so, the sub-band frame next to sub-band 0 is acquired in step S14, and steps S6, S9 or S
The MS stereo coding state mode of the current subband frame determined in 12 is set as the MS stereo coding state mode of the immediately preceding subband frame, and then the process returns to step S2. If step S13 is YES, the process ends.

The state mode switching process of the MS stereo coding of the present embodiment described above is well known to those skilled in the computer art, and a conventional general purpose computer programmed according to the teachings of the present application. Alternatively, it can be implemented using a microprocessor, or a computer-embedded encoding device. Appropriate software coding can be readily prepared by skilled programmers based on the teachings of the present disclosure, as will be apparent to those skilled in the software art. The processing of this embodiment may also be implemented by the provision of application specific integrated circuits, or by interconnecting appropriate networks of conventional component circuits, as will be readily apparent to those skilled in the art. It is possible. It should be noted that the computer not only includes the audio encoder 2 disclosed herein, but also the corresponding audio decoder 6, which is the MS stereo code generated by the audio encoder 2. It may be implemented to decode the encoded bitstream signal to operate to obtain and output an audio signal containing the decoded stereo channels.

The present invention may be implemented as a computer program for operating the above-mentioned computer which is the audio encoder 2 or the audio decoder 6. The computer program is M
A plurality of instructions for performing an S stereo coding state mode switching process, the computer executing the computer program to determine a MS stereo coding state mode for an audio signal, and / or Alternatively, the MS stereo encoded audio signal is decoded. The computer program may be in the form of a computer program recording medium or a computer program product provided with a computer program recording medium, that is, in the form of a software program provided on the computer program recording medium. The recording medium includes a computer-readable recording medium in which a computer program is recorded. The computer-readable recording medium is, for example, a magnetic recording medium such as a magnetic disk or a magnetic tape (such as a floppy (registered trademark) disk), an optical disk, an optical tape, or a machine-readable memory (RAM). Or, including but not limited to, an optical recording medium such as read only memory (ROM), or any other physical device or medium. The computer program may be transmitted from a computer such as a server to a computer such as a client device via an electric communication line and / or a wireless line such as the Internet or a wireless LAN.

<Modification> The audio signal encoding / decoding system according to the embodiment of the present invention has been described as the mini-disc recording / reproducing system, but may be any other recording medium or transmission medium. . Although the description of the present embodiment is based on the ATRAC3 encoder, its applicability can be extended to other encoders that utilize the same MS stereo encoding method. Also,
In the present embodiment, the MS stereo coding is performed only by the joint stereo encoder 14-1, but it is possible to perform the MS stereo coding by the other joint stereo encoders 14-2 to 14-4. There is no limit. Furthermore, in this embodiment, the MS stereo coding state mode of the current subband frame is determined based on the energy of each channel of the stereo signal and the MS stereo coding state mode of the immediately preceding subband frame. However, the state mode of MS stereo coding of the current subband frame may be determined based on two or more preceding subband frames. Further, in the present embodiment, in order to compare the signals of the left and right channels, the sum total of squares (energy) of all time samples in the subband frame of each channel is calculated. The signals of the left and right channels may be compared using the sum of the absolute values of.

As described above, according to the audio signal encoding method of the embodiment of the present invention, the following two points regarding the artifacts which may occur when the MS stereo encoding is not carefully implemented. You can avoid the potential danger of. One is to toggle rapidly between two state modes when MS stereo coding state mode switching is not envisaged, and the other is when state mode switching is needed. It means that switching fails. In this embodiment,
He has been successful in addressing the two causes of these artifacts. In particular, if the energy of one channel rises sharply with respect to the energy of the other, a fast switch to the other MS stereo coding state mode is allowed. Finally, this embodiment is very easy to implement on an LSI due to its simplicity and very few memory requirements.

[0060]

As described above in detail, according to the audio signal coding method of the present invention, each frame includes a pair of stereo channels consisting of a left channel and a right channel, and each frame has a plurality of time samples. A method for encoding an audio signal for encoding a digitized audio signal divided into successive frames into any one of a plurality of state modes of MS stereo encoding Calculating the respective energies of the left and right channels of the current frame of the audio signal, the value of the ratio of the energy of the right channel to the energy of the left channel, and the MS of at least one preceding frame.
Determining the MS stereo coding state mode of the current frame based on the stereo coding state mode.

Further, according to the audio signal coding apparatus of the present invention, a continuous frame including a pair of stereo channels consisting of a left channel and a right channel and each frame consisting of a plurality of time samples is formed. A device for encoding an audio signal for encoding a divided digitized audio signal into any one of a plurality of state modes of MS stereo encoding, wherein: Calculating means for calculating the respective energies of the left and right channels of the frame, a value of the ratio of the energy of the right channel to the energy of the left channel, and a state mode of MS stereo coding of at least one preceding frame Based on and, the current frame MS stereo coding state And a determining means for determining the mode.

According to the above audio signal encoding method and apparatus, the state mode related to the MS stereo encoding of the digitized audio signal is determined, while the generation of the artifact is minimized or the artifact is generated at all. The effect of not letting it occur is brought about.
Further, according to the present invention, it is possible to provide an audio signal encoding / decoding system including the audio signal encoding device and the audio signal decoding device. Further, according to the present invention, it is possible to provide a program including the steps related to the audio signal encoding method and a computer-readable recording medium in which the program is recorded.

MS stereo coding is a common method of reducing the bit rate for many audio coders, but it is prone to artifacts if not implemented carefully. The present invention provides an effective and efficient audio signal encoding method and apparatus for preventing high speed toggle between state modes of MS stereo encoding and delay of mode switching to minimize the occurrence of artifacts. To do.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a mini disk recording / reproducing system according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a detailed configuration of the audio encoder 2 of FIG.

3 is a block diagram showing a detailed configuration of the audio decoder 6 of FIG.

FIG. 4 shows MS stereo coding and I for stereo signal 31.
FIG. 7 is a diagram illustrating an example in which S stereo encoding is applied to generate a joint stereo signal 32.

FIG. 5 is a state transition diagram between three possible state modes according to MS stereo coding.

FIG. 6 is a diagram showing conditions for determining the MS stereo coding state mode of the current subband frame based on the MS stereo coding state mode of the immediately preceding subband frame, FIG. The condition when the MS stereo coding state mode of the immediately preceding subband frame is the S / D mode, (b) is the MS of the immediately preceding subband frame
FIG. 6 is a diagram showing conditions when the stereo encoding state mode is the S / L mode, and (c) is a diagram showing conditions when the MS stereo encoding state mode of the immediately preceding subband frame is the S / R mode.

7 is a joint stereo encoder 14-1 of FIG.
6 is a flowchart showing a state mode switching process of MS stereo encoding executed by the.

FIG. 8 is a graph showing an example in which the state mode switching process of MS stereo coding according to the embodiment of the present invention is applied to the signals of the L channel and the R channel.

FIG. 9 is a graph showing an example in which a state mode switching process of MS stereo coding according to the related art is applied to signals of L channel and R channel.

[Explanation of symbols]

1 ... A / D converter, 2 ... audio encoder, 3 ... Mini disk recording device, 4 ... Mini disc, 5 ... Mini disc player, 6 ... Audio decoder, 7 ... D / A converter, 10 ... QMF division filter, 11 ... Low-pass filter, 12 ... High-pass filter, 13-1 to 13-4 ... Filter bank, 14-1 to 14-4 ... Joint stereo encoder, 15-1 to 15-4 ... Gain control section, 16-1 to 16-4 ... MDCT processing unit, 17-1 to 17-4 ... Tone component encoder, 18-1 to 18-4 ... Tone component quantizer, 19-1 to 19-4 ... Spectral quantizer, 20 ... Bitstream multiplexer, 21 ... Bitstream demultiplexer, 22-1 to 22-4 ... Tone component dequantizer, 23-1 to 23-4 ... Spectral inverse quantizer, 24-1 to 24-4 ... Tone component decoder, 25-1 to 25-4 ... Inverse MDCT processing unit, 26-1 to 26-4 ... Inverse gain control section, 27-1 to 27-4 ... Joint stereo decoder, 28 ... QMF synthesis filter.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Suahone Neo             Singapore 534415 Singapore, Thailand             Sen Avenue, Block 1022, 04-             No. 3530, Thai Sen Industrial             Estate, Panasonic Singapore             Research Institute Co., Ltd. F-term (reference) 5J064 AA00 BA16 BB07 BC02 BC06                       BC07 BC11 BC16 BC25 BC29                       BD02 BD03 BD04

Claims (11)

[Claims] 1. A digitized audio signal comprising a pair of stereo channels consisting of a left channel and a right channel, each frame being divided into consecutive frames of a plurality of time samples. A method of encoding an audio signal for encoding one of a plurality of state modes of stereo encoding, comprising calculating respective energies of a left channel and a right channel of a current frame of the audio signal. The mid-side stereo code of the current frame based on the step, the value of the ratio of the energy of the right channel to the energy of the left channel, and the state mode of mid-side stereo coding of at least one preceding frame. Determining the state mode of activation Method of encoding an audio signal, characterized in that. 2. The mid-side stereo coding includes a first state mode including a sum signal and a difference signal of left and right channel signals, and a sum signal of left and right channel signals and a left channel signal. 2. The audio signal encoding method according to claim 1, further comprising a second state mode and a third state mode including a sum signal of left and right channel signals and a right channel signal. 3. The step of determining the state mode of mid-side stereo coding of the current frame comprises the steps of: when the ratio value is greater than a predetermined first threshold value;
Transitioning the mid-side stereo coding state mode of the current frame to the second state mode; and when the value of the ratio is less than the reciprocal of the first threshold, the mid-side of the current frame. The step of transitioning the stereo coding state mode to the third state mode, and the mid-side stereo coding state mode of the immediately preceding preceding frame is in the third state mode, and the value of the ratio is When being less than or equal to the first threshold and greater than a second predetermined threshold that is less than the first threshold, the state mode of midside stereo coding of the current frame is changed to the first mode. And the state mode of the mid-side stereo coding of the preceding immediately preceding frame is in the second state mode. Is greater than or equal to the reciprocal of the first threshold and less than the reciprocal of the second threshold, the mid-side stereo coding state mode of the current frame is changed to the first state mode. The method for encoding an audio signal according to claim 2, further comprising: 4. The step of determining a mid-side stereo coding state mode of the current frame, wherein the mid-side stereo coding state mode of the immediately preceding frame is in the second state mode, When the ratio value is greater than or equal to the reciprocal of the second threshold, holding the mid-side stereo coding state mode of the current frame in the second state mode; When the state mode of midside stereo coding is in the third state mode and the value of the ratio is less than or equal to the second threshold value, the state mode of midside stereo coding of the current frame is set to the third state mode. 4. The method of encoding an audio signal according to claim 3, further comprising the step of holding in three state modes. 5. A digitized audio signal comprising a pair of stereo channels consisting of a left channel and a right channel, each frame being divided into consecutive frames of a plurality of time samples. An apparatus for encoding an audio signal for encoding into any one of a plurality of state modes of stereo encoding, calculating respective energies of a left channel and a right channel of a current frame of the audio signal. The mid-side stereo of the current frame based on the calculation means, the value of the ratio of the energy of the right channel to the energy of the left channel, and the state mode of the mid-side stereo coding of at least one preceding immediately preceding frame. And a determining means for determining the state mode of encoding. Encoding apparatus of an audio signal, characterized in that the. 6. The mid-side stereo coding includes a first state mode including a sum signal and a difference signal of left and right channel signals, and a sum signal of left and right channel signals and a left channel signal. 6. The audio signal encoding apparatus according to claim 5, further comprising a second state mode and a third state mode including a sum signal of left and right channel signals and a right channel signal. 7. The determining means, when the value of the ratio is larger than a predetermined first threshold value,
Transitioning the state mode of mid-side stereo coding of the current frame to the second state mode, and when the value of the ratio is less than the reciprocal of the first threshold, the mid-side stereo code of the current frame. The state mode of encoding is transited to the third state mode, the state mode of the mid-side stereo coding of the preceding immediately preceding frame is in the third state mode, and the value of the ratio is the first When it is less than or equal to the threshold value and is greater than a predetermined second threshold value that is smaller than the first threshold value, the mid-side stereo coding state mode of the current frame is changed to the first state mode. The state mode of mid-side stereo coding of the preceding preceding frame is transited to the second state mode, and the value of the ratio is less than or equal to the reciprocal of the first threshold. 7. And the audio signal according to claim 6, wherein the state mode of midside stereo coding of the current frame is transited to the first state mode when it is less than the reciprocal of the second threshold. Signal coding device. 8. The determination means is characterized in that the state mode of mid-side stereo coding of the immediately preceding frame is in the second state mode, and the value of the ratio is not less than the reciprocal of the second threshold value. , The mid-side stereo coding state mode of the current frame is held in the second state mode, and the mid-side stereo coding state mode of the immediately preceding preceding frame is set to the third state mode. 8. If the ratio value is less than or equal to the second threshold value, the mid-side stereo coding state mode of the current frame is held in the third state mode. Audio signal encoding device. 9. An audio signal encoding and decoding system comprising the audio signal encoding device according to claim 5 and an audio signal decoding device, wherein: An audio signal decoding apparatus decodes the encoded audio signal based on the determined state mode of the midside stereo encoding to obtain an audio signal including a pair of decoded stereo channels. A system for encoding and decoding an audio signal, characterized by comprising means for outputting. 10. A program comprising steps relating to the audio signal encoding method according to claim 1. Description: 11. A computer-readable recording medium in which the program according to claim 10 is recorded.