JP4625709B2 - Stereo audio signal encoding device - Google Patents

Description

  The present invention involves encoding a stereo audio signal such as AAC (Advanced Audio Coding) or MP3 (MPEG-1 Audio Layer-3), for example, with processing for time-frequency conversion of left and right channel audio signals in units of variable lengths. The present invention relates to a stereo audio signal encoding apparatus that performs encoding by a method.

In encoding such as AAC and MP3, variable-length block switching is performed in order to suppress pre-echo due to a sudden amplitude change of an input signal. In this block switching, when a determination parameter such as perceptual entropy based on a psychoacoustic model or a time domain amplitude change (power ratio) exceeds a threshold value, the block is switched to a short block with high temporal resolution.
"Advanced Audio Coding (AAC)", ISO / IEC 13818-7 Information technology -Generic coding of moving pictures and associated audio information-, International Organization for Standardization / International Electrotechnical Commission, December 1, 1997

  When encoding a stereo signal with a high correlation between left and right, if the block switching determination results differ due to subtle differences in the determination parameters of the left and right channels, the encoding efficiency of the long block and short block, etc. Along with the difference, the difference between the left and right quantization noises was sometimes detected as an abnormal sound.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a stereo audio signal encoding apparatus capable of suppressing the generation of abnormal sounds due to the difference between left and right quantization noises. It is in.

A stereo audio signal encoding apparatus according to one aspect performs a stereo audio signal encoding process including a time frequency conversion of a left channel audio signal and a right channel audio signal in units of variable length blocks. Means for obtaining a cross-correlation value between the left channel and the right channel , a determination parameter for the left channel indicating characteristics of the audio signal of the left channel and necessary for block determination, and the audio of the right channel Calculating means for calculating a determination parameter for the right channel necessary for block determination and representing a characteristic of the signal; and a block length of the left channel; when the determination parameter for the left channel is less than a first threshold As the first block length and for the left channel When a constant parameter is greater than or equal to the first threshold value, the second block length is determined to be shorter than the first block length, and the block length of the right channel is determined as the determination parameter for the right channel. A determination unit configured to determine the first block length when it is less than a first threshold, and the second block length when the determination parameter for the right channel is equal to or greater than the first threshold ; Unlike block length before Symbol the determined left and right channels from each other, and for the when the cross-correlation value is not smaller than the second threshold value, the determined block length is the second block length channel When the determination parameter of the left channel and the right channel is less than a third threshold value that is greater than the first threshold value and less than a third threshold value, When the first block length is determined as the block length to be used for the time frequency, and the determination parameter for a channel whose determined block length is the second block length is equal to or greater than the third threshold Comprises determining means for determining the second block length as the block length used for the time-frequency conversion for both the left channel and the right channel .

  According to the present invention, it is possible to suppress the generation of abnormal noise due to the difference between left and right quantization noise.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram of a stereo audio signal encoding apparatus (hereinafter referred to as an encoding apparatus) according to this embodiment.

  This encoding apparatus performs an encoding process on an input stereo PCM signal and outputs an encoded bit stream. The encoding apparatus includes a frame cutout unit 1, a psychoacoustic model unit 2, a filter bank unit 3, a quantization distortion / rate control unit 4, a host processor 5, a scaling unit 6, a quantization unit 7, an encoding unit 8, and a formatter. 9 is included. Each of these units can be configured by hardware, and the function of each unit can be realized by software processing using a DSP (Digital Signal Processor) or the like.

  The frame cutout unit 1 is provided with a stereo PCM signal to be encoded. This stereo PCM signal is a time domain signal. The frame cutout unit 1 cuts out data from the stereo PCM signal for each number of samples having a prescribed frame size that is a processing unit of the codec. The frame cutout unit 1 outputs the cutout signal.

  The psychoacoustic model unit 2 performs orthogonal transform such as DFT (Discrete Fourier Transform), DCT (Discrete Cosine Transform), or MDCT (Modified DCT) on the signal output from the frame cutout unit 1, thereby Convert signal to frequency domain signal. The psychoacoustic model unit 2 calculates a parameter called perceptual entropy from the transform coefficient (frequency component) obtained by the orthogonal transform. This parameter is information required to encode one frame by analyzing the auditory frequency resolution, diffusion of frequency components, unpredictability, and signal tonality from the conversion coefficient. This is for estimating the quantity, and details of the calculation method are defined in the international standard 13818-7 established by ISO / IEC.

  The psychoacoustic model unit 2 includes a block switching unit 2a and an SMR calculation unit 2b. The block switching unit 2a determines the block length used in the orthogonal transform process (MDCT process) when actually encoding based on the perceptual entropy described above or based on the power ratio with the past signal substituted for it. This block switching is performed in order to suppress annoying noise (pre-echo) caused by the influence of a quantization error when an attack signal (a signal with a sharp amplitude change) is encoded by an encoding method based on orthogonal transform. FIG. 2 is a diagram showing how frames and blocks are switched when MDCT used in AAC or the like is used as orthogonal transform. 2A shows a long block, and FIG. 2B shows a short block. Since noise that causes pre-echo spreads to the MDCT processing unit, when it is determined that an attack signal is included, the processing unit for MDCT is processed by a short block divided into eight to suppress the noise generation range. The joint between the short block and the long block is an MDCT process using a special window shape for matching the both. The block switching unit 2a outputs block length information indicating the result of this determination to the filter bank unit 3. The SMR calculation unit 2b is a permissible noise that is not perceived even if it exists as a sound masked with respect to a signal, that is, noise, for each equally-spaced band on a scale (bark, mel, etc.) in consideration of auditory frequency resolution. The SMR (Signal to Mask Ratio) indicating the ratio of the amount of is calculated. The SMR calculation unit 2 b outputs the calculated SMR to the quantization distortion / rate control unit 4.

  The filter bank unit 3 orthogonally transforms the output signal of the frame cutout unit 1 according to the block length information output from the psychoacoustic model unit 2. The filter bank unit 3 outputs the frequency transform coefficient obtained by the orthogonal transform. In addition, when AAC is employ | adopted as an encoding system, the orthogonal transformation in the filter bank part 3 becomes MDCT.

  The quantization distortion / rate control unit 4 calculates a code amount that can be allocated for each frame based on the coding rate instructed by the host processor 5 and the SMR output from the psychoacoustic model unit 2. The quantization distortion / rate control unit 4 controls the scaling unit 6, the quantization unit 7, and the encoding unit 8 using the calculated code amount as the target code amount of the encoded frame. For example, the quantization distortion / rate control unit 4 calculates a quantization distortion amount from the quantization coefficient given from the quantization unit 7, and the scaling unit 6 and the quantum according to the result and the SMR obtained from the psychoacoustic model unit 2. An output instruction is given to the conversion unit 7. The quantization distortion / rate control unit 4 repeatedly performs control until the code amount notified from the encoding unit 8 falls within the target code amount, and when the code amount falls within the target code amount, the quantization distortion / rate control unit 4 Give output instructions.

  The scaling unit 6 scales the frequency conversion coefficient (MDCT coefficient) output from the filter bank unit 3 with a predetermined scaling coefficient. Then, the scaling unit 6 outputs the scaling result to the quantization unit 7. The scaling unit 6 outputs a scaling coefficient to the formatter 9 in response to an instruction from the quantization distortion / rate control unit 4.

  The quantization unit 7 corrects the frequency conversion coefficient output from the scaling unit 6 according to a prescribed formula, and then quantizes all the frequency conversion coefficients. The quantization unit 7 outputs the quantized coefficient to the quantization distortion / rate control unit 4 as information for determining whether the quantization distortion error is an allowable error based on the SMR value. The quantization unit 7 outputs the quantized frequency transform coefficient to the encoding unit 8 in response to an output instruction from the quantization distortion / rate control unit 4.

  The encoding unit 8 compresses and encodes the output of the quantization unit 7 according to a predetermined encoding method. For example, in the case of AAC, the Huffman encoding method is applied as the above-described encoding method. The encoding unit 8 outputs the encoded code amount to the quantization distortion / rate control unit 4. The encoding unit 8 outputs the encoded frequency transform coefficient to the formatter 9 in response to an output instruction from the quantization distortion / rate control unit 4.

  The formatter 9 multiplexes the output of the encoding unit 8 and the scaling coefficient output from the scaling unit 6 according to a predetermined format. The formatter 9 outputs the result of the above multiplexing as an encoded audio signal.

  Next, the operation of the encoding apparatus configured as described above will be described. The point of the present invention lies in the block length determination process in the block switching unit 2a. Therefore, here, this process will be described in detail. For other processing, processing similar to that of an existing encoding device that conforms to AAC can be applied.

FIG. 3 is a diagram showing an operation flow of block switching processing in the block switching unit 2a in FIG.
When a stereo PCM signal having a prescribed frame size cut out by the frame cutout unit 1 is input, the processing shown in FIG. 3 is executed with this frame as a processing target.

  In step Sa1, the block switching unit 2a calculates a determination parameter necessary for block determination for each of the left and right channels. The determination parameters include perceptual entropy based on the psychoacoustic model, power ratio for each frame or sub-block (unit smaller than the frame) in the time domain in order to detect an attack signal. In step Sa2, the block switching unit 2a determines the block shape (block length) for each channel based on the determination parameter. The block shape is determined by comparing the determination parameter with the threshold value α. When the determination parameter is less than the threshold value α, it can be determined that the signal is a stationary signal, so the block switching unit 2a determines that the block shape is a long block. On the other hand, when the determination parameter is greater than or equal to the threshold value α, it is determined that the signal is a steep signal including an attack, so the block switching unit 2a determines the block shape to be a short block.

  In step Sa3, the block switching unit 2a checks whether or not the block shapes determined for the left and right channels are the same. If the result of this confirmation is YES, the block switching unit 2a proceeds from step Sa3 to step Sa9. However, if the result of the above confirmation is NO, the block switching unit 2a proceeds from step Sa3 to step Sa4.

  In step Sa4, the block switching unit 2a calculates a cross-correlation value between the left and right channels. The cross-correlation value Corr is obtained by the following equation when the L channel side signal L (k), the R channel side signal R (k), and the signal range to be determined are N. The cross-correlation value calculated here may be a time domain or a frequency domain.

  In step Sa5, the block switching unit 2a checks whether or not the cross-correlation value is equal to or greater than the threshold value β. If the correlation between the left and right channels is small and the result of the above confirmation is NO, the block switching unit 2a proceeds from step Sa5 to step Sa9 and uses the block shape determined in step Sa2 as it is. However, if the correlation between the left and right channels is large and the result of the above confirmation is YES, the block switching unit 2a proceeds from step Sa5 to step Sa6.

  In step Sa6, the block switching unit 2a confirms whether or not the determination parameter calculated in step Sa1 is greater than or equal to the threshold value α ′ for the channel determined to be a short block. The threshold value α ′ is set to a value slightly larger than the threshold value α. If the result of the above confirmation is YES, the block switching unit 2a proceeds from step Sa6 to step Sa7. In step Sa7, the block switching unit 2a changes a channel determined to be a long block (long side channel) to a short block. After this, the block switching unit 2a proceeds to Step Sa9. On the other hand, if the result of the above confirmation is NO, the block switching unit 2a proceeds from step Sa6 to step Sa8. In step Sa8, the block switching unit 2a changes a channel determined to be a short block (short side channel) to a long block. After this, the block switching unit 2a proceeds to Step Sa9.

  In step Sa9, the block switching unit 2a determines the block shape of each channel at the time of proceeding here as the block shape related to the block to be processed this time, and notifies this to the filter bank unit 3.

  Thus, according to the present embodiment, if the block shape determined based on the comparison between the determination parameter and the threshold value α is the same for the left and right channels, the determined block shape is determined for both the left and right channels.

  If the block shape determined based on the comparison between the determination parameter and the threshold α is different between the left and right channels and the correlation between the left and right channels is relatively large, the block shape of the left and right channels is forcibly shared. Is done. The difference between the left and right quantization noises is kept small, and the generation of abnormal noise is suppressed.

  In this embodiment, if the determination parameter of the short side channel is equal to or larger than the threshold value α ′ slightly larger than the threshold value α, the block shape of the left and right channels is made common to the short block, and otherwise, common to the long block. It becomes. Therefore, if the short side channel is determined to be a short block because the determination parameter is slightly larger than the threshold value α, even if this is changed to a long block, the quality will not be greatly reduced. Coding efficiency can be improved by sharing. However, if the determination parameter of the short-side channel is sufficiently larger than the threshold value α, the quality deterioration due to changing this to a long block will increase. Therefore, the quality can be improved by sharing the short block.

  Even if the block shape determined based on the comparison between the determination parameter and the threshold α is different between the left and right channels, if the correlation between the left and right channels is relatively small, the determined block shape for both the left and right channels To be confirmed. As a stereo PCM signal having a relatively small correlation between the left and right channels, a signal related to a dual mono sound source is assumed. In the case of this dual mono sound source, it is assumed that each channel is listened to independently, so it is desirable to encode them independently so as not to be affected by the channel. Deterioration may be caused. However, in the present embodiment, when the correlation between the left and right channels is relatively small as described above, the block shape of the left and right channels is allowed to be different. Can be

This embodiment can be variously modified as follows.
When the block shape is forcibly shared, it may be unconditionally shared with either the long block or the short block. For example, it is conceivable to share a long block if priority is given to coding efficiency, and to a short block if priority is given to quality.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment.

1 is a block diagram of a stereo audio signal encoding device according to an embodiment of the present invention. The figure which shows the mode of the block switching in MDCT processing. The figure which shows the operation | movement flow of the block switching process in the block switching part 2a in FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Frame cutout part, 2 ... Psychological auditory model part, 2a ... Block switching part, 2b ... SMR calculation part, 3 ... Filter bank part, 4 ... Quantization distortion / rate control part, 5 ... Host processor, 6 ... Scaling part , 7 ... quantization unit, 8 ... encoding unit, 9 ... formatter.

Claims (4)

In stereo audio signal coding apparatus which performs a stereo audio signal coding with Hidarichi Yaneru audio signal and the right channel audio signals respectively processing for time-frequency transform in blocks of variable length,
It means for obtaining a cross-correlation value between the right channel and the left channel,
The left channel determination parameter necessary for block determination and the right channel determination signal necessary for block determination and the right channel determination parameter necessary for block determination are calculated. A calculation means;
The block length of the left channel is set as a first block length when the determination parameter for the left channel is less than a first threshold, and the determination parameter for the left channel is greater than or equal to the first threshold Are determined as second block lengths shorter than the first block length, and the block length of the right channel is determined when the determination parameter for the right channel is less than the first threshold. Determining means for determining the second block length when the determination parameter for the right channel is equal to or greater than the first threshold ;
A channel in which the determined block length is the second block length when the determined block lengths of the left channel and the right channel are different from each other and the cross-correlation value is equal to or greater than a second threshold. When the determination parameter for use is less than a third threshold value that is greater than the first threshold value, the first block length is determined as the block length to be used for the time frequency for both the left channel and the right channel. When the determination parameter for a channel whose determined block length is the second block length is equal to or greater than the third threshold, the second block length is set for both the left channel and the right channel. Stereo audio signal, comprising: a determining means for determining a block length to be used for the time frequency conversion Goka apparatus. The determining means determines each of the determined block lengths as a block length to be used for the time-frequency conversion when the determined block lengths of the left channel and the right channel are the same. 1. The stereo audio signal encoding device according to 1. The determining means determines each of the determined block lengths when the block lengths of the determined left channel and right channel are different from each other and the cross-correlation value is less than a second threshold. The stereo audio signal encoding device according to claim 1, wherein the stereo audio signal encoding device is determined as a block length used for frequency conversion . The stereo audio signal encoding apparatus according to claim 1 , wherein the calculation means calculates perceptual entropy based on a psychoacoustic model or a power ratio for each frame or sub-block in the time domain as the determination parameter .

Images