CN1722845A - Digital signal conversion method - Google Patents

Abstract

The input digital signal of the first format (DV video signal) is canceled by the deframing section (11) by framing it, then decoded by the variable length decoding (VLD) section (12) and dequantized by the inverse quantization (IQ) section (13) ) inverse quantization, inverse weighted by the inverse weighting (IW) section (14), is restored to a variable length code. Then, the inverse-weighted video signal is subjected to required resolution conversion by the resolution conversion section (16) in the orthogonal transform domain (frequency domain). Thereafter, the resolution-converted video signal is weighted by a weighting (W) section (18), then quantized by a quantization (Q) section (19), and variable-length coded by a variable-length coding (VLC) section (20) to Two-format (MPEG video signal) digital signal output.

Description

Digital Signal Converter

This invention patent application is a divisional application of the following invention patent application:

Application number: 98801684.2; Application date: November 05, 1998; Invention name: Digital signal conversion method and digital signal conversion device

technical field

The present invention relates to the conversion processing of compression-encoded digital signals by using an orthogonal transform such as discrete cosine transform (DCT), and more particularly to a digital signal conversion method and a digital signal conversion device for converting the resolution between compressed video signals of different formats Rate.

Background technique

Conventionally, discrete cosine transform (DCT), a type of orthogonal transform coding, has been used as a coding system for efficiently compression-coding still picture data and moving picture data. When dealing with such digital signals that have undergone an orthogonal transformation, it is sometimes necessary to change the resolution or transform base.

For example, if a first orthogonally transformed digital signal having a resolution of, say, 720×480 pixels in a domestic digital video format is to be converted into a second orthogonally transformed digital signal having a resolution of 360×240 pixels in a so-called MPEG1 format , then perform an inverse orthogonal transformation on the first signal to restore the signal in the space domain, and then perform transformation processing such as interpolation and thinning (thinning) to perform an orthogonal transformation again, thus converting the first signal into a second signal .

In this way, it is often the case that the orthogonally transformed digital signal is inversely transformed once in order to restore the original signal, then processed by the required transformation operation, and then the orthogonal transformation is performed again.

FIG. 28 shows an example structure of a conventional digital signal processing device for performing the above-mentioned resolution conversion on a digital signal that has been subjected to DCT.

In this conventional digital signal conversion device, a video signal (hereinafter referred to as a DV video signal) of the so-called "DV format" (a format of digital video signals for home use) is input as a digital signal of the first format conforming to the so-called MPEG ( A video signal in a format conforming to the Moving Picture Experts Group (hereinafter referred to as an MPEG video signal) is output as a digital signal in a second format.

A de-framing section 51 is used to de-framing the DV video signal. In this deframing section 51, the DV video signal framed according to the so-called DV format is restored to a variable length code.

A variable length decoding (VLD) section 52 performs variable length decoding on the video signal restored to a variable length code by the deframing section 51 . Data compressed in the DV format is compressed at a fixed rate so that its data amount is reduced to about 1/5 of an original signal, and encoded by variable length coding in order to improve data compression efficiency. The variable length decoding section 52 performs decoding corresponding to this variable length encoding.

An inverse quantization (IQ) section 53 performs inverse quantization on the video signal decoded by the variable length decoding section 52 .

The inverse weighting (IW) section 54 performs inverse weighting which is an inverse operation of the weighting performed on the video signal inversely quantized by the inverse quantization section 53 .

The weighting operation is to reduce the value of the DCT coefficient of the higher frequency component of the video signal by utilizing the property that human vision is not very sensitive to distortion on the high frequency side. In this way, the number of high-frequency coefficients having a value of 0 increases, and variable-length coding efficiency can be improved. The result is that, in some cases, the amount of arithmetic operations for the DCT transform can be reduced.

An inverse discrete cosine transform (IDCT) section 55 performs inverse DCT (inverse discrete cosine transform) on the video signal inversely weighted by the inverse weighting section 54, thus restoring DCT coefficients to data of the spatial domain, ie, pixel data.

Then, the resolution conversion section 56 performs required resolution conversion on the video signal restored as pixel data by the inverse discrete cosine transform section 55 .

A discrete cosine transform (DCT) section 57 performs discrete cosine transform (DCT) on the video signal resolution-converted by the resolution converting section 56, thus converting the video signal into orthogonal transform coefficients (DCT coefficients) again.

A weighting (W) section 58 weights the resolution-converted and converted video signal into DCT coefficients. The weighting is the same as above.

The quantization (Q) section 59 quantizes the video signal weighted by the weighting section 58 .

Then, a variable length coding (VLC) section 60 performs variable length coding on the video signal quantized by the quantization section 59, and outputs the resultant signal as an MPEG video signal.

The aforementioned "MPEG" is an abbreviation of the Moving Picture Experts Group of ISO/IEC JTC1/SC29 (International Organization for Standardization/International Electrotechnical Commission, Joint Technical Committee 1/Subcommittee 29). The ISO11172 standard is used as the MPEG1 standard, and the ISO13818 standard is used as the MPEG2 standard. Among these international standards, ISO11172-1 and ISO13818-1 are standardized in the multimedia multiplexing part, ISO11172-2 and ISO13818-2 are standardized in the video part, and ISO11172-3 and ISO13818-3 are standardized in the audio part.

According to ISO11172-2 or ISO13818-2, which is an image compression coding standard, the image signal is compressed and coded by using the correlation of the picture in the time or space direction on the basis of the picture (frame or field), and the use of the correlation in the space direction is through It is implemented using DCT coding.

In addition, the orthogonal transform such as DCT is widely adopted by various types of picture information compression coding such as JPEG (Joint Picture Coding Experts Group).

Generally, orthogonal transform enables compression coding to have high compression efficiency and excellent reproducibility by converting an original signal in time domain or space domain to an orthogonal transform domain, such as frequency domain.

The above-mentioned "DV format" is for compressing the data amount of a digital video signal to about 1/5, and recording it on a magnetic tape as a component. The DV format is used in digital video equipment for home use and some digital video equipment for professional use. The DV format achieves efficient compression of video signals by combining discrete cosine transform (DCT) and variable length coding (VLC).

Meanwhile, for orthogonal transform and inverse orthogonal transform such as discrete cosine transform (DCT), a large amount of calculation is generally required. Therefore, there is a problem that the above-mentioned resolution conversion of the video signal cannot be performed efficiently. Also, since errors are accumulated as the amount of computation increases, there is a problem of signal degradation.

Disclosure of the invention

In view of the foregoing state of the present technology, an object of the present invention is to provide a digital signal conversion method and a digital signal conversion device capable of reducing the data volume of a signal processed by resolution conversion for conversion to a different format by arithmetic Amount of processing, making conversion processing such as resolution conversion efficient and less degraded.

In order to solve the foregoing problems, a digital signal conversion method according to the present invention includes: a data extraction step of extracting a part of the orthogonal transform coefficients from a corresponding block of the digital signal of the first format including a predetermined unit of orthogonal transform coefficient blocks, Thus forming partial blocks; the inverse orthogonal transformation step, on the basis of the partial blocks, carries out inverse orthogonal transformation to the orthogonal transformation coefficients constituting each partial block; the partial block connection step, connects the said partial blocks, thereby constituting a new block of said predetermined unit; and an orthogonal transformation step, on the basis of said block, performing orthogonal transformation on said new block, thereby generating said new block including said predetermined unit Orthogonally transform the second digital signal of the block.

Likewise, in order to solve the aforementioned problems, a digital signal conversion method according to the present invention includes: an inverse orthogonal transform step of performing an inverse orthogonal transform based on the block of the digital signal in the first format including a block of orthogonal transform coefficients of a predetermined unit. an orthogonal transform; a block dividing step of dividing each block of the digital signal of the first format processed by the inverse orthogonal transform; an orthogonal transform step of, based on the divided blocks, performing an orthogonal transform on the blocks constituting each divided block the coefficients are subjected to orthogonal transformation; and a data amplification step of interpolating each of the orthogonally transformed blocks with predetermined values of orthogonal transformation coefficients to constitute said predetermined unit, thereby generating a digital signal of a second format.

Also, in order to solve the foregoing problems, a digital signal conversion device according to the present invention includes: decoding means for decoding a digital signal in a first format including orthogonal transform coefficients of predetermined units; inverse quantization means for converting the performing inverse quantization on the decoded digital signal; the resolution conversion device is used to extract a part of the orthogonal transform coefficients from the adjacent blocks of the predetermined unit of the orthogonal transform coefficient block of the inverse quantized digital signal, thereby constituting a partial block, converting resolution; quantization means for quantizing said digital signal processed by resolution conversion; and encoding means for encoding said quantized digital signal, thereby generating a digital signal of a second format.

Also, in order to solve the foregoing problems, a digital signal conversion device according to the present invention includes: decoding means for decoding a digital signal in a first format compressed and encoded by using orthogonal transform; inverse quantization means for said Inverse quantization of the decoded digital signal; resolution conversion means for interpolating an orthogonal transform coefficient of a predetermined value to each predetermined block of said inverse quantized digital signal, thereby constituting said predetermined unit, converting resolution; quantization means , for quantizing said digital signal processed by resolution conversion; and encoding means for encoding said inversely quantized digital signal, thereby generating a digital signal of a second format.

Also, in order to solve the foregoing problems, a digital signal conversion method according to the present invention is used for converting a digital signal in a first format including an orthogonal transform coefficient block including a predetermined unit into a new orthogonal transform including another predetermined unit A digital signal of the second format of the block of coefficients. In this method, the data amount of the digital signal of the second format is controlled by using data amount information contained in the digital signal of the first format.

Also, in order to solve the foregoing problems, a digital signal conversion device according to the present invention is used for converting a digital signal in a first format including an orthogonal transform coefficient block comprising a predetermined unit into a new orthogonal transform comprising another predetermined unit The digital signal of the second format of the coefficient block, the device includes: decoding means for decoding the digital signal of the first format; inverse quantization means for inverse quantization of the decoded digital signal; signal conversion means, For performing signal processing along with the format conversion of the inversely quantized digital signal; quantization means for quantizing the digital signal processed by the signal processing; data volume control means for controlling the data volume of the quantization means; and encoding means for encoding the digital signal whose data amount is controlled and quantized by the data amount control means, thereby generating the digital signal of the second format.

Likewise, in order to solve the foregoing problems, a digital signal conversion method according to the present invention is used to convert a digital signal in a first format into a digital signal in a second format. The method includes: a decoding step of decoding the digital signal in the first format; a signal conversion step of converting the decoded digital signal of the first format into a digital signal of the second format; an encoding step of encoding the second format a digital signal in two formats; and a weighting processing step of integrally performing inverse weighting on the digital signal in the first format and weighting on the digital signal in the second format.

Likewise, in order to solve the foregoing problems, a digital signal conversion device according to the present invention is used for converting a digital signal in a first format into a digital signal in a second format. The device includes: decoding means for decoding the digital signal in the first format; signal converting means for converting the decoded digital signal in the first format into a digital signal in the second format; encoding means, for encoding the digital signal in the second format; and weighting processing means for performing the inverse weighting of the digital signal in the first format and the weighting of the digital signal in the second format as a whole.

Also, in order to solve the foregoing problems, according to the present invention, decoding with motion compensation is performed on an input information signal compressed and encoded with motion detection, and signal conversion processing is performed on the decoded signal. Then, the converted signal with motion detection is subjected to compression coding processing based on the motion vector information of the input information signal.

Also, in order to solve the aforementioned problems, according to the present invention, a partial decoding process is performed on an input information signal processed by compression coding including predictive coding with motion detection and orthogonal transform coding to obtain an orthogonal Decoded signal in transform domain. Then, signal conversion processing is performed on the decoded signal in the orthogonal transform domain, and compression coding processing with motion compensation prediction is performed on the converted signal by using motion detection based on motion vector information of the input information signal.

Also, in order to solve the aforementioned problems, according to the present invention, a partial decoding process is performed on an input information signal processed by compression coding including predictive coding with motion detection and orthogonal transform coding, thereby obtaining an orthogonal transform domain signal of. Then, signal conversion processing is performed on the signal, and compression coding processing is performed on the converted signal by adding motion vector information converted based on the motion vector information of the input information signal.

Likewise, in order to solve the aforementioned problems, according to the present invention, the digital signal in the first format is decoded, and the digital signal in the first format has the dynamic mode/static mode information added in advance, and the signal conversion process is performed on the decoded signal. Then, for each predetermined block of the converted signal, it is distinguished whether to perform inter-frame differential coding according to the dynamic mode/static mode information. Based on the result of the distinction, the converted signal is encoded to output a digital signal in a second format processed by encoding using an interframe difference.

Likewise, in order to solve the aforementioned problems, according to the present invention, a partial decoding process is performed on the digital signal in the first format, so as to obtain a signal in the orthogonal transform domain. Performing signal conversion processing on the signal in the orthogonal transform domain, and for each predetermined block of the converted signal, according to the maximum value of the absolute value of the inter-frame difference of the converted signal, distinguishing whether to perform inter-frame differential coding . Based on the result of the distinction, the converted signal is encoded to output a digital signal in a second format.

In addition, in order to solve the aforementioned problems, according to the present invention, for signals including intra-frame coding processed by intra-frame coding and forward predictively coded signals and bidirectionally predictively coded signals processed by forward and bidirectional interframe predictive coding with motion detection The digital signal in the first format of the signal performs inverse orthogonal transformation on the intra-coded signal and the forward predictive coded signal. Based on the inverse orthogonal transform output, a motion compensated output to be added to the partially decoded forward predictive coded signal and bidirectional predictive coded signal is generated. An orthogonal transform is performed on the motion compensated output, and the orthogonal transform output is added to the partially decoded forward predictive coded signal and bidirectional predictive coded signal. Compression coding is performed on the signal output based on the addition, and a digital signal in the second format is output.

Brief description of the drawings

FIG. 1 is a block diagram showing an example structure of a digital signal conversion device according to a first embodiment of the present invention.

Figure 2 shows the principle of resolution conversion in the orthogonal transform domain.

Fig. 3 shows the principle of resolution conversion in the orthogonal transform domain.

4A to 4C schematically show a state in which a DV video signal is converted into an MPEG video signal by digital signal conversion according to a first embodiment of the present invention.

Fig. 5 shows the relationship between the DV format and the MPEG format.

Fig. 6 shows the basic calculation procedure of the resolution conversion process.

7A and 7B show "static mode" and "dynamic mode" of the DV format.

Fig. 8 shows the procedure of switching processing in "static mode".

9A to 9C are block diagrams showing an exemplary structure of a digital signal conversion device according to a second embodiment of the present invention.

FIG. 10 shows the procedure of conversion processing in enlargement of an image.

FIG. 11 is a block diagram showing an example structure of a digital signal conversion device according to a third embodiment of the present invention.

FIG. 12 is a block diagram showing an example structure of a digital signal conversion device according to a fourth embodiment of the present invention.

Fig. 13 is a block diagram showing an example structure of a digital signal conversion device according to a fifth embodiment of the present invention.

FIG. 14 is a block diagram showing an example structure of a digital signal conversion device according to a sixth embodiment of the present invention.

Fig. 15 is a block diagram showing an example structure of a digital signal conversion device according to a seventh embodiment of the present invention.

16 is a flowchart showing a basic procedure for setting the quantizer scale of each macroblock (MB) of each frame when a DV video signal is converted into an MPEG signal in the seventh embodiment of the present invention picture.

FIG. 17 is a flowchart showing a basic process of applying feedback to the next frame by using a predetermined quantizer scale in the seventh embodiment of the present invention.

FIG. 18 is a block diagram showing an example structure of a conventional digital signal conversion device for converting an MPEG video signal into a DV video signal.

FIG. 19 is a block diagram showing an example structure of a digital signal conversion device according to an eighth embodiment of the present invention.

FIG. 20 is a block diagram showing an example structure of a digital signal conversion device according to a ninth embodiment of the present invention.

FIG. 21 is a block diagram showing an example structure of a digital signal conversion device according to a tenth embodiment of the present invention.

Fig. 22 is a block diagram showing an example structure of a digital signal conversion device according to an eleventh embodiment of the present invention.

FIG. 23 is a block diagram showing an example structure of a digital signal conversion device according to a twelfth embodiment of the present invention.

Fig. 24 shows motion compensation and motion estimation processing in the orthogonal transform domain in the twelfth embodiment of the present invention, showing a state where a macroblock B is stretched over a plurality of macroblocks of a reference picture.

Fig. 25 shows motion compensation and motion estimation processing in the orthogonal transform domain in the twelfth embodiment of the present invention, showing conversion processing of reference macroblocks.

Fig. 26 shows motion compensation and motion estimation processing in the orthogonal transform domain in the twelfth embodiment of the present invention, showing the conversion process of reference macroblocks.

27 is a block diagram showing an example structure of a digital signal conversion device according to a thirteenth embodiment of the present invention.

Fig. 28 is a block diagram showing an example structure of a conventional digital signal conversion device.

Best Mode for Realizing the Invention

Preferred embodiments of the present invention are described below with reference to the accompanying drawings.

First, the structure of the digital signal conversion device according to the present invention will be described, and then the digital signal conversion method according to the present invention will be described with reference to the structure.

FIG. 1 shows an example structure of a main part of a digital signal conversion device as a first embodiment of the present invention. Although signal conversion is exemplified by resolution conversion, signal conversion is actually not limited to resolution conversion, and various types of signal processing such as format conversion and filter processing may be employed.

In this digital signal conversion device, the video signal of the above-mentioned so-called "DV format" (hereinafter referred to as DV video signal) is input as the first digital signal, and the video signal of the format conforming to the MPEG (Moving Picture Experts Group) standard (hereinafter referred to as is an MPEG video signal) is output as a second digital signal.

The deframing section 11 is for deframing of the DV video signal. In this deframing section 11, a DV video signal framed in a predetermined format (so-called DV format) is restored into variable-length codes.

A variable length decoding (VLD) section 12 performs variable length decoding on the video signal restored to a variable length code by the deframing section 11 .

An inverse quantization (IQ) section 13 performs inverse quantization on the video signal decoded by the variable length decoding section 12 .

The inverse weighting (IW) section 14 performs inverse weighting which is an inverse operation of the weighting performed on the video signal inversely quantized by the inverse quantization section 13 .

If resolution conversion is performed, as an example of signal conversion processing, the resolution conversion section 16 performs required resolution conversion on the video signal inversely weighted by the inverse weighting section 14 in the orthogonal transform domain (frequency domain).

A weighting (W) section 18 weights the video signal processed by resolution conversion.

The quantization (Q) section 19 quantizes the video signal weighted by the weighting section 18 .

Then, a variable length coding (VLC) section 20 performs variable length coding on the video signal quantized by the quantization section 19, and outputs the resultant signal as an MPEG video signal.

The structure of each part of the above-mentioned digital signal conversion device according to the present invention shown in FIG. 1 can be made similar to the structure of each part of the conventional digital signal conversion device shown in FIG. 28 .

However, the digital signal conversion device according to the present invention is different from the conventional digital signal conversion device in that an inverse discrete cosine transform (IDCT) section and a discrete cosine transform (DCT) section are not provided before and after the resolution conversion section 16 .

That is, in the conventional digital signal conversion apparatus, the orthogonal transform coefficients of the input digital signal of the first format are inversely orthogonally transformed to be restored as data in the space domain (on the frequency axis), and then required transform operations are performed. Therefore, orthogonal transformation is performed again in order to restore the data to orthogonal transformation coefficients.

On the contrary, in the digital signal conversion apparatus according to the present invention, the required conversion operation of the orthogonal transform coefficients of the input digital signal of the first format is performed in the orthogonal transform coefficient domain (frequency domain), which is used for performing such as resolution The inverse orthogonal transform means and the orthogonal transform means are not provided before and after the transform processing means of the transform.

The principle of the resolution conversion process in the resolution conversion section 16 will now be described with reference to FIGS. 2 and 3 .

In FIG. 2 , an input orthogonal transformation matrix generation section 1 generates an inverse matrix Ts (k ₎ ^-1 representing an orthogonal transformation matrix Ts _(k) that has been subjected to an orthogonal transformation of an input digital signal 5 in advance, and the inverse matrix Send to the transformation matrix generating part 3. The output orthogonal transformation matrix generating section 2 generates an orthogonal transformation matrix Td (L) corresponding to an inverse transformation matrix Td _{( L)} ⁻¹ representing an inverse orthogonal transformation to be performed on an output digital signal, and sends the orthogonal transformation matrix to Generate part 3 for the transformation matrix. The transformation matrix generation section 3 generates a transformation matrix D for conversion processing such as resolution conversion in the frequency domain, and sends the transformation matrix to the signal conversion section 4 . The signal conversion section 4 converts the input digital signal 5 that has been converted into the frequency domain by the orthogonal transform while maintaining the orthogonal transform domain such as the frequency domain, generating an output digital signal 6 .

Particularly, as shown in Fig. 3, by using the orthogonal transformation matrix Ts _(k) , the signal (original signal) A of the original time domain (or space domain) is transformed into the frequency domain to generate the frequency signal B ₁ (corresponding to In the input digital signal 5). This frequency signal _B1 is reduced to N/L (or amplified) by the signal conversion section 4 to generate a frequency signal _B2 (corresponding to the output digital signal 6). The frequency signal B ₂ is inverse-orthogonally transformed using an inverse transform matrix Td _(L) ^-1 to generate a signal C in the time domain.

In the example in Figure 3, each transform block of the one-dimensional original signal A is orthogonally transformed, and the length of each transform block is k, and the adjacent blocks of m units of the obtained transform block in the frequency domain have L A continuous frequency signal with a length of (=k×m) is converted into a block with a length of N (N<L), that is, reduced to N/L as a whole.

In the following description, a matrix (orthogonal transformation matrix) having orthogonal transformation basis vectors e ₁ , e ₂ , ..., e _n of length n arranged in corresponding rows is denoted as T _(n) , whose inverse transformation The matrix is denoted T _(n) ^-1 . In this description, x represents an x vector. In this case, each matrix is a forward matrix of order n. For example, a one-dimensional DCT transformation matrix T ₍₈₎ of n=8 is represented by the following equation (1).

In FIG. 3, when the size of the orthogonally transformed block of the input digital signal 5 that has been orthogonally transformed into the frequency domain by using the orthogonally transformed matrix Ts _(k) is k, that is, when the base length is equal to k, the input The orthogonal transform matrix generating section 1 generates an inverse orthogonal transform matrix Ts _(k) ^-1 , and the output orthogonal transform matrix generating section 2 generates an orthogonal transform matrix Td _{(L) having a base length L (=k×m)} .

Here, the inverse orthogonal transform matrix Ts _(k) ^-1 generated by the input orthogonal transform matrix generating section 1 corresponds to the inverse process of the orthogonal transform process in generating the input digital signal 5, and the output orthogonal transform matrix The orthogonal transform matrix Td _(L) generated by the generating section 2 corresponds to inverse processing of the inverse orthogonal transform processing of converting the signal into the time domain in decoding the output digital signal converted by the signal converting section 4 . Both of these orthogonal transformation matrix generating parts 1 and 2 can generate base vectors of arbitrary length.

The orthogonal transformation matrix generation sections 1 and 2 may be the same orthogonal transformation matrix generation section. In this case, the orthogonal transformation matrices Ts _(k) and Td _(L) become the same type of orthogonal transformation matrix except that their base lengths are different from each other. Every different orthogonal transform system has an orthogonal transform matrix generating part.

Next, the transformation matrix generation section 3 generates the L-order forward matrix A by arranging the inverse orthogonal transformation matrix Ts _(k) ^-1 of m units generated by the input orthogonal transformation matrix generation section 1 on the diagonal, as follows Equation (2) shows. When the base length of the output digital signal 6 is equal to N, the transformation matrix generation part 3 takes out the low-frequency basis vectors of N elements of the orthogonal transformation matrix Td _(L) , and generates a matrix B with N rows and L columns.

However, in the expression, when Td _(L) is expressed by the following basis vectors, e ₁ , e ₂ , . . . , e _N are low-frequency basis vectors of N units.

Then, calculate the equation

D＝α · B · A ... (5)

To generate a matrix D with N rows and L columns. This matrix D is a conversion matrix for converting resolution at a reduction rate (or enlargement rate) of N/L. In this equation, α is a scalar value or a vector value, and is a coefficient for level correction.

The signal conversion part 4 of Fig. 2 collects m blocks of the input digital signal _B1 in the frequency domain as a group, and divides the signal into meta-blocks (meta-blocks) with a size of L (wherein a meta-block includes m blocks ),As shown in Figure 3. If the length of the input digital signal _B1 is not a multiple of L, the signal is supplemented and filled with dummy data such as 0, becoming a multiple of L. The metablocks thus generated are denoted by Mi (where i=0, 1, 2, . . . ).

The principle of the resolution conversion process described above is described in detail in PCT/JP98/02653 filed on June 16, 1998 by the present assignee.

The digital signal conversion method of the first embodiment will be described below with reference to the structure of the above-mentioned digital signal conversion apparatus.

4A to 4C schematically illustrate a process in which a DV video signal is converted into an MPEG video signal by digital signal conversion according to an embodiment of the present invention. This processing is mainly performed by the resolution conversion section 16 in the digital signal conversion device of the embodiment of the present invention shown in FIG. 1 .

In the following description, a one-dimensional block of DCT coefficients is used as an example. However, the processing of the two-dimensional DCT coefficients can be done similarly.

First, four DCT coefficients on the low-frequency side are taken out from each of adjacent blocks (i) and (i+1), each of which includes 8 bits of the digital signal in the first format DCT coefficients, as shown in Figure 4A. That is, among the 8 DCT coefficients a0, a1, a2, a3, ..., a7 of the block (i), only the 4 DCT coefficients a0, a1, a2, a3 on the low frequency side are taken out, resulting in the number of DCT coefficients having Partial blocks reduced to 1/2, similarly, among the 8 DCT coefficients b0, b1, b2, b3, ..., b7 of the block (i+1), only the 4 DCT coefficients b0, b1 on the low frequency side , b2, b3 are taken out, resulting in partial blocks with the number of DCT coefficients reduced to 1/2. The operation of extracting DCT coefficients on the low frequency side is based on the characteristic that energy is concentrated in DC and AC low frequencies when a video signal is frequency converted.

Then, 4-point inverse discrete cosine transform (4-point IDCT) is performed on each partial block including 4 DCT coefficients, thus obtaining reduced pixel data. These pixel data are represented as pixel data p0, p1, p2, p3 and pixel data p4, p5, p6, p7 in Fig. 4B.

Next, each partial block including the reduced pixel data processed by the inverse discrete cosine transform is concatenated to generate a block having the same size as the original block. That is, pixel data p0, p1, p2, p3 and pixel data p4, p5, p6, p7 are concatenated to generate a new block including 8 pixel data.

Then, an 8-point discrete cosine transform is performed on a new block comprising 8 pixel data, thus generating a block (j) comprising 8 DCT coefficients c0, c1, c2, c3, ..., c7, as shown in Figure 4C .

Through the above-described process, a video signal can be converted into a video signal of a different resolution format by thinning out the number of orthogonal transform coefficients (DCT coefficients) per predetermined block unit to half. When the number of DCT coefficients is to be thinned down to 1/4, the above-described processing is continuously performed twice.

For example, the above-described resolution conversion processing can be applied to conversion from the DV format to the MPEG1 format.

The relationship between the DV format and the MPEG format and format conversion between these formats will now be described with reference to FIG. 5 .

In particular, for the case of a video signal conforming to the NTSC system shown in FIG. 5, the video signal in DV format is a compressed video signal with a resolution of 720×480 pixels, the sampling frequency of the luminance signal and the sampling frequency of the two color difference signals The ratio is equal to 4:1:1. The video signal in the MPEG1 format is a compressed video signal with a resolution of 360×240 pixels, and the ratio of the sampling frequency of the luminance signal to the sampling frequency of the two color difference signals is equal to 4:2:0. Thus, in this case, by the above-mentioned resolution conversion process according to the present invention, the number of DCT coefficients in the horizontal and vertical directions of the luminance (Y) signal can be reduced to 1/2, and the number of DCT coefficients in the vertical direction of the color difference (C) signal can be reduced to 1/2. The number of DCT coefficients for a direction can be reduced to 1/4.

Since the odd and even rows alternately take values of 4:2:0 and 4:0:2, the ratio of 4:2:0 represents the value of either the odd or even rows.

On the other hand, in the case of a video signal conforming to the PAL system, the video signal of the DV format is a compressed video signal having a resolution of 720×576 pixels, and the ratio of the sampling frequency of the luminance signal to the sampling frequency of the two color difference signals is equal to 4 :2:0. The video signal in the MPEG1 format is a compressed video signal with a resolution of 360×288 pixels, and the ratio of the sampling frequency of the luminance signal to the sampling frequency of the two color difference signals is equal to 4:2:0. Thus, in this case, by the above-mentioned resolution conversion process according to the present invention, the number of DCT coefficients in the horizontal and vertical directions of the Y signal can be reduced to 1/2, and the number of DCT coefficients in the horizontal and vertical directions of the C signal The number can be reduced to 1/2.

The resolution conversion processing described above can be similarly applied to conversion from the DV format to the MPEG2 format.

For the case of a video signal conforming to the NTSC standard, the video signal in the MPEG2 format is a compressed video signal with a resolution of 720×480 pixels, and the ratio of the sampling frequency of the luminance signal to the sampling frequency of the two color difference signals is equal to 4:2:0 . Thus, in this case, the number of DCT coefficients in the vertical direction of the C signal can be reduced to 1/2, and the number of DCT coefficients in the horizontal direction of the C signal can be doubled without performing conversion processing of the Y signal . The method of such amplification will be described later.

For the case of a video signal conforming to the PAL system, the video signal in the MPEG2 format is a compressed video signal with a resolution of 720×576 pixels, and the ratio of the sampling frequency of the luminance signal to the sampling frequency of the two color difference signals is equal to 4:2:0 . Therefore, in this case, conversion processing of the Y signal or the C signal is not required.

FIG. 6 shows the basic calculation procedure of the above-mentioned resolution conversion processing.

Specifically, four DCT coefficients a0, a1, a2, a3 and four DCT coefficients b0, b1, b2, b3 obtained by connecting two adjacent blocks of the input digital signal of the first format include 8 A block of DCT coefficients is multiplied by a (8×8) matrix consisting of two inverse discrete cosine transform matrices (IDCT4) on the diagonal, each of which is a (4×4) matrix and the other components are 0 .

The product is multiplied again by a discrete cosine transform matrix (DCT8) which is a (8×8) matrix, and a new block including 8 DCT coefficients c0, c1, c2, c3, . . . , c7 is generated.

In the digital signal conversion method according to the present invention, since resolution conversion processing is performed in the DCT domain (frequency domain), inverse DCT before resolution conversion and DCT after resolution conversion are unnecessary. In addition, by obtaining in advance the product of (8×8) matrix and (8×8) discrete cosine transform matrix including two (4×4) inverse discrete cosine transform matrices (IDCT4) on the diagonal as transform matrix D, The amount of arithmetic operations can be effectively reduced.

Processing for converting a DV video signal which is a digital signal of the first format into an MPEG1 video signal which is a digital signal of the second format will be described in further detail.

The DV format has "static mode" and "dynamic mode", which are switched according to the motion detection result of the picture. For example, these modes are distinguished by motion detection before DCT of each (8×8) matrix of video segments, and DCT is performed in a certain mode according to the result of the distinction. Various methods of motion detection can be considered. Specifically, a method of comparing the sum of absolute values of inter-field differences with a predetermined threshold may be used.

"Static mode" is a basic mode of the DV format in which (8x8) DCT is performed on (8x8) pixels in one block.

The (8×8) block is composed of one DC component and 63 AC components.

"Dynamic mode" is used to avoid a situation where if DCT is performed while the object is in motion, the compression efficiency is reduced by energy dispersion due to interlacing. In this dynamic mode, the (8×8) block is divided into the (4×8) block of the first field and the (4×8) block of the second field, and the pixel data of each (4×8) block is processed (4×8) DCT. In this way, an increase in high-frequency components in the vertical direction is suppressed, and a decrease in the compression ratio can be prevented.

Each of the above-mentioned (4×8) blocks is composed of one DC component and 31 AC components.

Thus, in the DV format, the block structure differs between the static mode and the dynamic mode. Therefore, in order to perform similar processing on the subsequent processing, after the DCT of each (4×8) block, the sum and difference of the coefficients of the same order in each block are calculated to form the (8× 8) Blocks. Through this process, the dynamic mode block can be regarded as composed of one DC component and 63 AC components, similar to the static mode block.

Meanwhile, in converting a video signal in DV format to a video signal in MPEG1 format, only one field must be separated, since the MPEG1 format only handles a video signal of 30 frames/second, and there is no concept of field.

FIG. 7A schematically shows processing for separating fields in converting DCT coefficients of "dynamic mode (2x4x8DCT mode)" in DV format into DCT coefficients of MPEG1 format.

The upper half (4×8) block 31a of the (8×8) DCT coefficient block 31 is the sum (A+B) of the coefficients of the first field and the coefficient of the second field, and the (8×8) DCT coefficient block The lower half (4x8) block 31b of 31 is the difference (A-B) of the coefficients of the two fields.

Thus, by adding the upper half (4×8) block 31a and the lower half (4×8) block 31b of the (8×8) DCT coefficient block 31, and then dividing the sum by 2, it is possible to obtain A (4x8) block 35a of DCT coefficients for field (A). Similarly, the ( 4x8) block 35b. That is, through this processing, (8×8) blocks 35 with separation fields can be obtained.

Then, the above-mentioned resolution conversion processing is performed on the DCT coefficients of one of these fields, for example, the first field.

Fig. 7B schematically shows the processing for separating fields in "static mode (8x8DCT mode)".

In the (8×8) DCT coefficient block 32, the DCT coefficients of the first field (A) and the DCT coefficients of the second field (B) are mixed. Thus, a conversion process must be performed, similarly by subtraction between the (4×8) block 35a and the (4×8) block 35b, by using the field separation process described below, for obtaining A (4x8) block 35a of , and a (4x8) block 35b including the second field (B).

Fig. 8 shows the procedure of field separation processing in "static mode".

First, an input including 8 DCT coefficients d0, d1, d2, d3, ..., d7 is multiplied by an 8-order inverse discrete cosine transform matrix (IDCT8) to restore pixel data.

Next, this data is multiplied by a (8×8) matrix for field separation, thereby dividing the (8×8) block into a first field on the upper side and a second field on the lower side, each of which is a (4×8 )piece.

This data is then multiplied by an (8x8) block comprising two discrete cosine transform matrices (DCT4) on the diagonal, each being a (4x4) matrix.

In this way, eight DCT coefficients including four DCT coefficients e0, e1, e2, e3 of the first field and four DCT coefficients f0, f1, f2, f3 of the second field are obtained.

Then, the above-mentioned resolution conversion processing is performed on the DCT coefficients of one of these fields, for example, the first field.

In the digital signal conversion method according to the present invention, since resolution conversion is performed in the DCT domain (frequency domain), inverse DCT before resolution conversion and DCT after resolution conversion are unnecessary. In addition, by preliminarily finding the product of an (8×8) matrix including two (4×4) inverse discrete cosine transform matrices (IDCT4) on the diagonal and a (8×8) discrete cosine transform matrix, the amount of arithmetic operations can be reduced is effectively reduced.

The above resolution conversion is to downscale the image. Hereinafter, resolution conversion processing for enlarging an image will be described as a second embodiment.

9A to 9C schematically show a state in which a DV video signal is converted into an MPEG2 video signal by the digital signal conversion method according to the present invention.

Also in the following description, one-dimensional DCT coefficients are used. However, similar processing can be performed on two-dimensional DCT coefficients.

First, an 8-point inverse discrete cosine transform (8-point IDCT) is performed on a block (u) including 8 orthogonal coefficients (DCT coefficients g0 to g7) shown in FIG. 9A, thus restoring 8 pixel data (h0 to h7) .

Next, the block including 8 pixel data is divided into two parts, thereby generating two partial blocks each including four pixel data.

Then, 4-point DCT is performed on the two partial blocks each including 4 DCT coefficients, thereby generating two partial blocks (i0 to i3 and j0 to j3) each including 4 DCT coefficients.

Then, as shown in FIG. 9C , the high-frequency side of each of the two partial blocks each including 4 pixel data is filled with 0 as four DCT coefficients. Thus, block (v) and block (v+1) each including 8 DCT coefficients are generated.

Following the above procedure, resolution conversion between compressed video signals of different formats is performed in the orthogonal transform domain.

FIG. 10 shows the procedure of conversion processing in this case.

First, an input including 8 DCT coefficients g0, g1, g2, g3, ..., g7 is multiplied by an 8-order inverse discrete cosine transform (IDCT) matrix to restore 8 pixel data.

Next, the block including 8 pixel data is divided into two parts, thereby generating two partial blocks each including four pixel data.

Then, each of the two partial blocks each including four DCT coefficients is multiplied by ( 4×8) matrix. In this way, two partial blocks (i0 to i7 and j0 to j7) including 8 DCT coefficients are generated.

Through such processing, two blocks of DCT coefficients are obtained from one block. Thus, the resolution can be enlarged in the frequency domain.

In the case of the NTSC system, in order to convert the DV format to the MPEG2 format, it is not necessary to perform horizontal and vertical conversion of the brightness signal Y, but the color difference signal C must be doubled in the horizontal direction, and the color difference signal C must be reduced to 1 in the vertical direction. /2, as shown in Figure 5. Thus, the above-described enlarging process is used for resolution conversion of the color difference signal C in the horizontal direction in converting the DV format into the MPEG2 format.

Fig. 11 shows an example configuration of main parts of a digital signal conversion apparatus according to a third embodiment of the present invention. The same structural parts as those of the first embodiment are denoted by the same reference numerals. The difference from the structure of FIG. 1 is that the weighting section 18 and the inverse weighting section 14 are integrated into a weighting processing section 21 .

In particular, the weighting processing (1W*W) section 21 integrally performs inverse weighting, which is an inverse operation of weighting performed on the DV video signal as the input digital signal of the first format, and weighting, which is used as the second format. Format of the output digital signal MPEG video signal.

With this structure, since the inverse weighting process of the input video signal of the first format and the weighting process of the output video signal of the second format can be integrally performed, compared with the case where the inverse weighting process and the weighting process are performed separately, the calculation can be reduced.

In the digital signal conversion apparatus of the third embodiment shown in FIG. However, a weighting processing section may be provided at a stage preceding the resolution conversion section 16 .

FIG. 12 shows a digital signal conversion apparatus according to a fourth embodiment of the present invention, in which a weighting processing section 22 is provided at a stage preceding the resolution converting section 16. Parts of the structure of the digital signal conversion device shown in FIG. 12 may be similar to corresponding parts of the digital signal conversion device of FIG. 11 .

The weighting process for integrally performing the inverse weighting of the digital signal of the first format and the weighting of the second digital signal and the above-mentioned weighting process may be performed before or after orthogonal transform such as discrete cosine transform (DCT). This is because its arithmetic operations are linear operations.

A digital signal conversion method and apparatus according to a fifth embodiment of the present invention will now be described with reference to FIG. 13 .

This digital signal conversion device has: decoding part 8, is used for decoding DV video signal; Resolution conversion part 16, is used for carrying out resolution conversion process to the format conversion of the decoding that outputs from decoding part 8; Differentiation part 7, is used for according to dynamic mode/static mode information to distinguish whether to perform forward interframe differential coding on each predetermined conversion block unit output from the resolution converting section 16; The converted output from the converting section 16 is encoded and output as an MPEG video signal, as shown in FIG. 13 .

In the following description, a digital video signal conversion device constituted by these parts is employed. As a matter of course, each section performs the processing of each step of the digital signal conversion method according to the present invention.

In the DV video signal input to the digital video signal conversion apparatus, a mode flag (for example, one bit) as information indicating static mode/dynamic mode is preliminarily added to each DCT block.

In this digital video signal conversion device, the discriminating section 7 discriminates whether or not forward interframe differential encoding is performed for each predetermined conversion block unit output from the resolution converting section 16 based on the mode flag. This operation will be described in detail later.

The deframing section 11 extracts a mode flag indicating static mode/dynamic mode information, supplies the mode flag to the discriminating section 7 .

The unshuffling part 15 cancels the shuffling, and the purpose of shuffling is to unify the amount of information in the video segment into one unit, which is used for the fixed length of the DV encoding side.

The discriminating section 7 includes an adder 27 and an I (I picture)/P (P picture) discriminating and determining section 28 . The adder 27 adds to the resolution conversion output reference DCT coefficients which are negative DCT coefficients stored in a frame memory (FM) section 24 which will be described later. The sum output from the adder 27 is supplied to the I/P distinguishing and determining section 28, which is also provided with a mode flag indicating the static mode/dynamic mode from the deframing section 11.

The operation of the I/P discrimination and determination section 28 will now be described in detail. The converted output from the resolution converting section 16 has 8 x 8 DCT coefficients as a unit. Four blocks of DCT coefficients each having 8x8 DCT coefficients are assigned to the luminance signal, and two blocks of DCT coefficients are assigned to the color difference signal, so that a predetermined block is constituted by six blocks of DCT coefficients in total. This predetermined block is called a macroblock.

Meanwhile, the difference between the P picture assumption and the previous frame is simply obtained. In the case of a still image, the amount of information is reduced when taking the difference. However, in the case of a moving image, the amount of information increases when taking the difference. Thus, if it is discriminated that the image is dynamic from the mode flag indicating static mode/dynamic mode, the macroblock is taken as an I picture so as not to increase the amount of information. If it is distinguished that the image is static, efficient encoding can be performed by taking the difference to make a P picture.

When all the mode flags on the six DCT coefficient blocks sent from the deframing section indicate the dynamic mode, the I/P discrimination and determination section 28 uses I pictures for the macroblock. On the other hand, when the flag indicating the dynamic mode can be detected only in one of the six DCT coefficient blocks, the I/P distinguishing and determining section 28 uses a P picture for the macroblock.

If a dynamic mode flag is added to four or more of the six DCT coefficient blocks, an I picture can be used as a macroblock. Also, when a flag indicating a static mode is added to all six DCT coefficient blocks, a P picture can be used as a macroblock.

The DCT coefficients determined as I/P pictures by the I/P distinguishing and determining section 28 on a macroblock basis are supplied to the encoding section 9 .

Encoding section 9 has weighting (W) section 18, quantization (Q) section 19, inverse quantization (IQ) section 26, inverse weighting (IW) section 25, FM section 24, variable length coding (VLC) section 20, buffer memory 23. Rate control part 29.

A weighting (W) section 18 weights the DCT coefficients that are the converted output supplied from the converting section 16 via the distinguishing section 7 .

The quantization (Q) section 19 quantizes the DCT coefficients weighted by the weighting (W) section 18 . Then, a variable length coding (VLC) section 20 performs variable length coding on the DCT coefficient quantized by the quantization (Q) section 19 , and supplies the MPEG-encoded data to the buffer memory 23 .

The buffer memory 23 fixes the transfer rate of the MPEG-encoded data, and outputs the MPEG-encoded data as a bit stream. The rate control section 29 controls the increase or decrease of the amount of information generated in the quantization (Q) section 19 , that is, the quantization step size, based on change information such as an increase or decrease in the buffer capacity of the buffer memory 23 .

The inverse quantization (IQ) section 26 inversely quantizes the DCT coefficients quantized from the quantization (Q) section 19 , and supplies the inverse quantized DCT coefficients to the inverse weighting (IW) section 25 . The inverse weighting (IW) section 25 inversely weights the DCT coefficients from the inverse quantization (IQ) section 26 , which is the inverse operation of weighting. The DCT coefficients inversely weighted by the inverse weighting (IW) section 25 are stored in the FM section 24 as reference DCT coefficients.

As described above, in the digital video signal converting apparatus shown in FIG. A macroblock distinguishes an I-picture or a P-picture. Thus, a DV signal originally consisting solely of I pictures can be converted to an MPEG picture using I pictures or P pictures, and the advantage of improvement in compression ratio, which is a feature of MPEG video signals, can be utilized.

A digital signal conversion method and apparatus according to a sixth embodiment of the present invention will now be described.

The digital video signal conversion apparatus according to the sixth embodiment is a digital video signal conversion apparatus in which the distinguishing portion 7 shown in FIG. 13 is replaced by a distinguishing portion 30 shown in FIG. 14 .

Specifically, the digital video signal conversion device has: a decoding part 8, which is used to perform partial decoding processing on the DV signal and obtain signals in the orthogonal transform domain such as DCT coefficients; a conversion part 16, which is used for format conversion. The DCT coefficient of the signal conversion process; the distinguishing part 30 is used to distinguish whether to perform forward interframe differential coding on each predetermined transform block unit output from the transform part 16 according to the maximum value of the absolute value of the transform output interframe difference and an encoding section 9 for encoding the converted output from the converting section 16 based on the discrimination result of the discriminating section 30 and outputting an MPEG video signal.

When the difference between the converted DCT coefficient that is the converted output from the converting section 16 and the reference DCT coefficient from the FM section 24 is extracted, the distinguishing section 30 refers to the maximum value of the absolute value of the AC coefficient, sums the maximum value and A predetermined threshold is compared. The distinguishing section 30 assigns an I/P picture to each macroblock based on the result of the comparison.

The distinguishing section 30 has a difference calculating section 31 , a maximum value detecting section 32 , a comparing section 33 , and an I/P determining section 35 .

The difference calculation section 31 calculates the difference between the DCT coefficient converted from the conversion section 16 and the reference DCT coefficient from the FM section 24 . The difference output from the difference calculation section 31 is supplied to the maximum value detection section 32 and is also supplied to the I/P determination section 35 .

The maximum value detection section 32 detects the maximum value of the absolute value of the AC coefficient of the differential output. Basically, when a large amount of information is converted into DCT coefficients, the AC coefficients become large. On the other hand, when a small amount of information is converted into DCT coefficients, the AC coefficients become small.

The comparison section 33 compares the maximum value of the absolute value from the maximum value detection section 32 with a predetermined threshold value supplied from the terminal 34 . When the predetermined threshold is properly selected, the amount of information converted into DCT coefficients can be distinguished according to the maximum value of the absolute value of the AC coefficients.

I/P determination section 35 distinguishes whether the difference in DCT coefficients from difference calculation section 31 , that is, the difference in information amount is large or small by using the comparison result of comparison section 33 . When discriminating that the difference is large, the I/P determination section 35 allocates an I picture to a macroblock including the converted DCT coefficient block from the conversion section 16 . When distinguishing that the difference is small, the I/P determination section 35 assigns the P picture to the macroblock from the difference calculation section 31 .

That is, if the absolute value of the maximum value is greater than the threshold value, it means that the amount of information to be distinguished is large, and an I picture is used as a macroblock. On the other hand, if the absolute value of the maximum value is smaller than the threshold value, the amount of information for distinguishing business trips is small, and a P picture is used as a macroblock.

Thus, the digital video signal conversion apparatus according to the sixth embodiment can convert a DV signal originally including an I picture into an MPEG picture using an I picture or a P picture, and the improved compression rate which is a feature of the video signal of the MPEG signal can be utilized. advantage.

In the digital video signal conversion apparatus shown in FIGS. 13 and 14, a DV signal conforming to the NTSC system and an MPEG1 video signal are used as input and output, respectively. However, the digital video signal conversion device can also be applied to each signal of the PAL system.

The resolution conversion processing described above can be similarly applied to the conversion from the DV format to the MPEG2 format.

Since the resolution conversion processing is performed by the conversion section 16, the reduced resolution conversion has been mainly described above. However, zooming in is also possible. Specifically, in general, by adding a high-frequency component to an input digital signal in the frequency domain, the resolution can be amplified at an arbitrary magnification.

When an MPEG2 video signal is applied to a digital broadcast service, the signal is classified according to profile (function)/level (resolution). For example, the upscaling of resolution can be applied in the case where a Main Performance/High Level (MP@HL) video signal for digital HDTV in the United States is converted to a DV signal.

The processing of the sixth embodiment can also be performed by software.

Referring now to FIG. 15, a digital signal conversion method and apparatus according to a seventh embodiment of the present invention will be described. Components of the same structure as those of the above-mentioned embodiments are denoted by the same reference numerals.

The rate control section 40 controls the amount of data in the quantization section 19 according to the quantizer number (Q_NO) and the class number (Class) from the deframing section 11 .

FIG. 16 is a diagram showing a basic procedure of setting a quantizer scale for each macroblock (MB) of each frame in converting a DV video signal into an MPEG video signal by the digital signal conversion method of the seventh embodiment.

First, in step S1, a quantizer number (Q_NO) and a class number (Class) are obtained for each macroblock. The quantizer number (Q_NO) is expressed by a value of 0 to 15, common to all six DCT blocks in a macroblock. The class number (Class) is expressed by a value of 0 to 3, provided for each of the six DCT blocks.

Next, in step S2, a quantization parameter (q_param) is calculated for each DCT block according to the following procedure.

Quantization table q_table[4]={9, 6, 3, 0}

Quantization parameter q_param=Q_NO+q_table[class]

Specifically, the quantization table has four values (9, 6, 3, 0), corresponding to class numbers 0, 1, 2, and 3, respectively. For example, when the class number is 2 and the quantizer number is 8, the quantization table value 3 corresponding to the class number 2 and the quantizer number 8 are added to obtain the quantization parameter 11.

Next, at step S3, the average value of the quantization parameter (q_param) of the six DCT blocks in the macroblock is calculated.

Then, in step S4, the quantizer scale (quantizer_scale) of the MPEG macroblock is obtained according to the following procedure, and the process ends.

Quantization table q_table[25]={32, 16, 16, 16, 16, 8, 8, 8, 8, 4, 4, 4, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,

2, 2}

quantizer_scale=q_table[q_param]

Specifically, the quantization table has 25 values (32 to 2), corresponding to the quantization parameters calculated in the above-described manner. The quantization table corresponding to the quantization parameter value 0 is 32. The quantization table corresponding to the quantization parameter value 1 is 16. The quantization table corresponding to the quantization parameter value 5 is 8. For example, when the average value of the quantization parameters found in the above-described manner is 10, the value 4 corresponding to the quantization parameter value of 10 becomes the quantizer scale value. Through this process, a target rate-dependent MPEG quantizer scale (quantizer_scale) is calculated based on the quantization parameter (q_param) of each macroblock within each frame. The correspondence between class numbers and quantization tables and the relationship between quantization parameters and quantizer scales are found empirically.

In the digital signal conversion apparatus of the present invention shown in FIG.

Fig. 17 shows the basic procedure for applying feedback to the next frame by using the quantizer scale set according to the procedure described above.

First, in step S11, the number of target bits per frame is set at the bit rate set according to the above-described procedure.

Next, in step S12, the total number of bits generated per frame is totaled.

Next, in step S13, the difference (diff) between the number of target bits and the total number of generated bits is calculated.

Then, in step S14, the quantizer scale is adjusted according to the calculated result.

The calculation at each step is expressed as follows:

diff=cont*diff(cont:constant)

q_param=q_param+f(diff)

quantizer_scale=q_table[q_Param]

Specifically, normalization is performed by multiplying the difference diff obtained in step S13 by a constant cont. The normalized difference is multiplied by an empirically found function and added to or subtracted from the quantization parameter. The resulting value is used as a quantization parameter. The value corresponding to this quantization parameter value is selected from a quantization table of 25 values and used as the quantizer scale for the next frame.

Through the foregoing process, feedback between frames is performed by calculating a new quantizer scale (quantizer_scale) based on the adjusted quantization parameter (q_param) and using the new quantizer scale of the next frame.

A digital signal conversion method and a digital signal conversion apparatus according to an eighth embodiment of the present invention will now be described. Although the DV format is converted to the MPEG format in the foregoing embodiments, the MPEG format is converted to the DV format in the following embodiments.

Referring to FIG. 18, a conventional apparatus for converting an MPEG format into a DV format is first described.

The digital video signal converting apparatus shown in FIG. 18 is composed of an MPEG decoder 70 and a DV encoder 80 . The MPEG decoder 70 is used to decode MPEG video data, and the DV encoder 80 is used to output DV video data.

In the MPEG decoder 70, a syntax analyzer 71 supplied with a bit stream of MPEG2 video data detects the header of the bit stream of quantized DCT coefficients framed in accordance with the MPEG2 format, and supplies the quantized DCT coefficients encoded by variable length coding to Variable Length Decoding (VLD) section 72 . Also, the syntax analyzer 71 extracts a motion vector (mv), and supplies the extracted motion vector to a motion compensation (MC) section 77 .

A variable length decoding (VLD) section 72 performs variable length decoding on quantized DCT coefficients encoded by variable length coding, and supplies the result of the variable length decoding to an inverse quantization (IQ) section 73 .

The inverse quantization section 73 performs inverse quantization by multiplying the quantized DCT coefficient decoded by the variable length decoding section 72 by the quantization step used on the encoding side. In this way, the inverse quantization section 73 obtains DCT coefficients, and supplies the DCT coefficients to an inverse discrete cosine transform (IDCT) section 74 .

The inverse discrete cosine transform section 74 performs inverse DCT on the DCT coefficients from the inverse quantization section 73 , thus restoring the DCT coefficients to data of the spatial domain, ie, pixel data. Specifically, by inverse DCT, pixel values (brightness Y and color differences Cr, Cb) are calculated for each block including 8×8 pixels. In the case of an I picture, the pixel value is its actual pixel value. However, in the case of a P picture and a B picture, a pixel value is a difference between corresponding pixel values.

The motion compensation section 77 generates a motion compensation output by using the picture information stored in the two frame memories FM of the frame memory (FM) section 76 and the motion vector mv extracted by the syntax analyzer 71, which is supplied to the addition device 75.

Adder 75 adds the motion compensation output to the difference from inverse discrete cosine transform section 74 , and provides decoded picture data to discrete cosine transform (DCT) section 81 and frame memory section 76 of DV encoder 80 .

In the DV encoder 80, a discrete cosine transform section 81 performs DCT processing on decoded picture data to convert the decoded picture data again into data of an orthogonal transform domain, that is, DCT coefficients, and supplies the DCT coefficients to quantization (Q) Section 82.

The quantization section 82 quantizes DCT coefficients by using a matrix table in consideration of visual characteristics, and supplies the quantization result to a variable length coding (VLC) section 83 as an I picture in DV format.

The variable length encoding section 83 compresses the I picture in the DV format by performing variable length encoding processing, and supplies the compressed I picture to the framing section 84 .

The framing section 84 frames the DV format data on which variable length encoding processing has been performed, and outputs a bit stream of DV video data.

Meanwhile, orthogonal transforms such as discrete cosine transform (DCT) and its inverse transform generally require a large amount of calculations, thus posing a problem that format conversion of video data as described above cannot be efficiently performed. Signal degradation is also an issue as errors accumulate as computations increase.

Thus, in order to solve these problems, a digital video signal conversion apparatus according to an eighth embodiment of the present invention will be described with reference to FIG. 19. FIG.

In the digital signal converting apparatus shown in FIG. 19, the above-mentioned MPEG video signal conforming to the MPEG format is input as a first digital signal, and a DV signal is output as a second digital signal.

The syntax analyzer 111 refers to the header of the MPEG video signal which is a digital signal of the first format, and extracts motion information of an image such as a motion vector mv and a quantizer scale.

The motion vector mv is sent to a motion compensation (MC) section 115, where motion compensation is performed. The quantizer scale (quantizer_scale) is sent to the evaluation section 123, which will be described later.

A variable length decoding (VLD) section 112 performs variable length decoding on a bit stream of an MPEG video signal from which necessary information is extracted by a syntax analyzer 111 .

An inverse quantization (IQ) section 113 inverse quantizes the MPEG video signal decoded by the variable length decoding section 112 .

Then, the MPEG video signal dequantized by the dequantization section 113 is input to the adder 125 . The motion compensation result of the motion vector mv from the syntax analyzer 111 is also input to this adder 125 from the motion compensation section 115 .

The output from the adder 125 is sent to a signal converting section 116, which will be described later. This output is also input to the motion compensation section 115 through the frame memory 114 . The signal conversion section 116 performs necessary signal conversion processing such as resolution conversion in the orthogonal transform domain (frequency domain) on the video signal input through the adder 125 .

The video signal to be subjected to required signal conversion processing by the signal conversion section 116 is shuffled by the shuffling section 117 and then sent to the buffer 118 and the sorting section 122 .

The video signal sent to the buffer 118 is sent to a quantization (Q) section 119 by which it is quantized. Then, the video signal is variable length coded by a variable length coding (VLC) section 120 . In addition, the video signal is framed by the framing section 121, and output as a bit stream of a DV video signal.

On the other hand, the sorting section 122 sorts the video signal shuffled by the shuffling section 117, and sends the result of the sorting to the evaluating section 123 as sorting information.

The evaluation section 123 determines the quantization number at the quantization section 119 based on the classification information from the classification section 122 and the quantizer scale (quantizer_scale) from the syntax analyzer 111 .

With this structure, since the data amount of the DV video signal output as the video signal of the second format can be determined based on the data amount information included in the MPEG video signal input as the video signal of the first format, for determining the amount converted by the signal The processing of the data volume of the generated video signal of the second format can be simplified.

The seventh and eighth embodiments described above can also be applied to the case where one of the digital signal in the first format and the digital signal in the second format is an MPEG1 video signal and the other is an MPEG2 video signal.

A digital signal conversion method and a digital signal conversion apparatus according to a ninth embodiment of the present invention will now be described with reference to FIG. 20 .

The digital signal conversion device is a device for converting an MPEG video signal conforming to the MPEG2 format into a DV video signal conforming to the DV format. It is assumed that these data are data of the PAL system.

In the case that the video signal is a signal of the PAL system, the signal conforming to the MPEG2 format and the DV format has a resolution of 720×576 pixels, and the ratio of the sampling frequency of the luminance signal to the sampling frequency of the two color difference signals is equal to 4:2:0 . Therefore, resolution conversion processing is not necessary for the Y signal or the C signal.

In FIG. 20, the MPEG decoder 100 has a syntax analyzer 111, a variable length decoding (VLD) section 112, an inverse quantization (IQ) section 113, an adder 125, an inverse discrete cosine transform (IDCT) section 131, a frame memory ( FM) section 132 , motion compensation (MC) section 115 , and discrete cosine transform (DCT) section 130 . The frame memory (FM) section 132 is structured such that it is used as two prediction memories.

As will be described later in detail, the inverse discrete cosine transform section 131 performs inverse discrete cosine transform processing on the I picture and the P picture partially decoded by the variable length decoding section 112 and the inverse quantization section 113 . The motion compensation section 115 generates a motion compensated output based on the inverse discrete cosine transform output. The discrete cosine transform section 130 performs discrete cosine transform on the motion compensation output. The adder 125 adds the motion compensation output from the discrete cosine transform section 130 to the P picture and the B picture partially decoded by the variable length decoding section 112 and the inverse quantization section 113 .

The entire operation will be described below. First, the syntax analyzer 111 restores the quantized DCT coefficients framed in accordance with the MPEG2 format to a variable length code with reference to the header of the MPEG2 video data input as a bit stream, and supplies the variable length code to the variable length decoding section 112. . Also, the syntax analyzer 111 extracts the motion vector mv, and supplies the extracted motion vector to the motion compensation section 115 .

The variable length decoding section 112 performs variable length decoding on the quantized DCT coefficients restored to variable length codes, and supplies the result of the variable length decoding to the inverse quantization section 113 .

The inverse quantization section 113 performs inverse quantization processing by multiplying the quantized DCT coefficient decoded by the variable length decoding section 112 by the quantization step size used on the encoding side. In this way, the inverse quantization section 113 obtains a DCT coefficient, and supplies the DCT coefficient to the adder 125 . The DCT coefficient obtained by the variable length decoding section 112 and the inverse quantization section 113 is supplied to the adder 125 as an output, which is not restored as pixel data by the inverse discrete cosine transform, ie, as partially decoded data.

Adder 125 is also provided with a motion compensated output from motion compensation section 115 , which is orthogonally transformed by discrete cosine transform section 130 . Adder 125 then adds the motion compensated output to the partially decoded data in the orthogonal transform domain. The adder 125 supplies the addition output to the DV encoder 110 and also to the inverse discrete cosine transform section 131 .

The inverse discrete cosine transform section 131 performs inverse discrete cosine transform processing on the I picture or the P picture within the addition output, thereby generating data of the spatial domain. The data in the spatial domain is reference picture data used for motion compensation. The reference picture data used for motion compensation is stored in the frame memory section 132 .

The motion compensation section 115 generates a motion compensation output to the discrete cosine transform section 130 by using the reference picture data stored in the frame memory section 132 and the motion vector mv extracted by the syntax analyzer 111 .

The discrete cosine transform section 130 restores the motion compensation output processed in the spatial domain to the orthogonal transform domain as described above, and then supplies the motion compensation output to the adder 125 .

The adder 125 adds the DCT coefficients of the motion compensation output from the discrete cosine transform section 130 to the DCT coefficients of the differential signals of the partially decoded P and B pictures from the inverse quantization section 113 . Then, the added output from the adder 125 is supplied to the DV encoder 110 and the inverse discrete cosine transform section 131 as partially decoded data in the orthogonal transform domain.

Since the partially decoded I picture from the inverse quantization section 113 is an intra-coded image signal, motion compensation addition processing is not required. The partially decoded I picture is supplied to the inverse discrete cosine transform section 131 as it is, and is also supplied to the DV encoder 110 .

The DV encoder 110 includes a quantization (Q) section 141 , a variable length coding (VLC) section 142 and a framing section 143 .

The quantization section 141 quantizes the decoded output of I pictures, P pictures, and B pictures in the orthogonal transform domain from the MPEG decoder 100 , that is, DCT coefficients, and supplies the quantized DCT coefficients to the variable length encoding section 142 .

The variable length encoding section 142 performs variable length encoding processing on the quantized DCT coefficients, and supplies the encoded data to the framing section 143 . The framing section 143 frames the compression-encoded data from the variable-length encoding section 142, and outputs a bit stream of DV video data.

In this way, when the MPEG2 video data to be converted is an I picture, the MPEG decoder 100 makes the variable length decoding section 112 and the inverse quantization section 113 partially decode the MPEG2 video data into the orthogonal transform domain, and the DV encoder 110 makes the The quantization section 141 and the variable length coding section 142 partially encode video data. Meanwhile, the MPEG decoder 100 causes the inverse discrete cosine transform section 131 to perform inverse discrete cosine transform on the I picture, and stores the resulting I picture in the frame memory section 132 as a reference picture of the P/B picture.

On the other hand, when the MPEG2 video data to be converted is a P-picture or a B-picture, only processing for producing a motion-compensated output is performed in the spatial domain by using the inverse discrete cosine transform section 131 for composing The processing of the frame of the differential signal of the P picture or B picture partially decoded by the length decoding section 112 and the inverse quantization section 113 is performed in the discrete cosine transform domain by using the discrete cosine transform section 130, as described above. Thereafter, partial encoding is performed by the DV encoder 110 .

In particular, in the case of a P picture, the macroblock at the position indicated by the motion vector mv is processed from the I picture processed by the inverse discrete cosine transform by the inverse discrete cosine transform section 131 through the motion compensation processing of the motion compensating section 115 extract. Discrete cosine transform processing is performed on the macroblock by the discrete cosine transform part 130, and then added to the DCT coefficients of the P picture by the adder 25 as a differential signal in the discrete cosine transform domain. This processing is based on the fact that the result of discrete cosine transform performed on the addition result in the space domain is equivalent to the result of data addition processed by the discrete cosine transform. The result is partially encoded by the DV encoder 110 . Meanwhile, as a reference for the next B picture, an inverse discrete cosine transform is performed on the added output from the adder 125 by the inverse discrete cosine transform section 131 , and the resultant data is stored in the frame memory section 132 .

In the case of a B picture, the macroblock at the position indicated by the motion vector mv is extracted from the P picture processed by the inverse discrete cosine transform by the inverse discrete cosine transform section 131 . Then, discrete cosine transform processing is performed on the macroblock by the discrete cosine transform section 130, and the DCT coefficients of the B pictures as differential signals are added thereto in the discrete cosine transform domain. In the bidirectional case, macroblocks are extracted from two reference frames and their average value is used.

The result is partially encoded by the DV encoder 110 . Since the B picture does not become a reference frame, the inverse discrete cosine transform by the inverse discrete cosine transform section 131 is not required.

Conventionally, inverse discrete cosine transform (IDCT) and discrete cosine transform (DCT) processing are required for decoding an I picture, but the digital video signal conversion apparatus according to the ninth embodiment described above requires only IDCT for reference.

In order to decode a P picture, DCT and IDCT processing for reference is necessary. However, DCT and IDCT are conventionally required for decoding B pictures, whereas the digital video signal conversion apparatus according to this embodiment requires only DCT and no IDCT.

In the case of typical MPEG2 data with the number of GOPs N=15 and the forward prediction picture pitch M=3, one I picture, four P pictures and 10 B pictures are included. Assuming that the calculation amount of DCT is substantially equal to that of IDCT, when the weighting is omitted, every 15 frames of MPEG2 data is represented by the following formula in the case of conventional technology

2×DCT×(1/15)+2×DCT×(4/15)+2×DCT×(10/15)

=2×DCT

In the case of the digital video signal conversion device shown in Fig. 20, it is represented by the following formula

1×DCT×(1/15)+2×DCT×(4/15)+1×DCT×(10/15)

=1.2666×DCT

In this way, the amount of computation can be significantly reduced. In these equations DCT stands for computation.

That is, in the digital video signal conversion apparatus shown in FIG. 20, the amount of data calculation processing for format conversion from MPEG2 video data to DV video data can be remarkably reduced.

Referring now to FIG. 21, a digital video signal conversion apparatus according to a tenth embodiment of the present invention will be described.

In this tenth embodiment, digital video signal converting means for converting MPEG video data conforming to the MPEG2 format into DV video data conforming to the DV format is employed. However, it is assumed that MPEG2 video data is a compressed video signal of high resolution such as 1440×1080 pixels.

For example, when an MPEG2 video signal is applied to a digital broadcasting service, the signal is classified according to performance (function)/level (resolution). A main performance/high level (MP@HL) video signal for digital HDTV in the United States has a high resolution, which is converted to DV video data as described above.

Thus, the digital video signal conversion apparatus shown in FIG. 21 is structured such that a signal conversion section 140 for performing the above-mentioned conversion processing is provided between the MPEG decoder 100 and the DV encoder 110 of FIG. 20 .

The signal converting section 140 performs resolution conversion processing on DCT coefficients from the DCT transform domain of the MPEG decoder by using a transform matrix generated based on the inverse orthogonal transform matrix and the orthogonal transform matrix. The inverse orthogonal transform matrix corresponds to an orthogonal transform matrix for performing DCT encoding on MPEG-encoded data, the orthogonal transform matrix corresponds to an inverse orthogonal transform matrix for IDCT encoding for obtaining a signal conversion output signal in a time domain Alternate transformation matrix.

The DCT coefficients from this signal conversion section 140 are supplied to the DV encoder 110 as a resolution conversion output.

The DV encoder 110 quantizes and variable-length-codes the DCT coefficients output as resolution conversion, then frames the DCT coefficients, and outputs a bit stream of DV video data.

Thus, in the digital video signal conversion apparatus, a main performance/high level (MP@HL) video signal within an MPEG video signal is resolution-converted by the signal conversion section 140, and then encoded by a DV encoder to generate DV video data.

Similar to the digital video signal conversion device of FIG. 20 , with respect to I pictures, the digital video signal conversion device of the tenth embodiment only needs IDCT as a reference, whereas conventionally both IDCT and DCT processing are required.

Regarding the P picture, as in the conventional technology, DCT and IDCT are performed for reference. With regard to the B picture, the digital video signal conversion apparatus requires only the DCT and not the IDCT, whereas conventionally both the IDCT and the DCT are required.

That is, also in the digital video signal conversion apparatus shown in FIG. 21, the amount of data calculation processing for format conversion from high-resolution MPEG2 video data to DV video data can be remarkably reduced.

As the resolution conversion processing performed by the signal conversion section 140 , reduced resolution conversion has been mainly described. However, zooming in is also possible. Specifically, in general, resolution can be amplified at an arbitrary magnification ratio by adding high-frequency components to an input digital signal in the frequency domain. For example, format conversion from MPEG1 video signal to DV video signal is performed.

The above processing can also be performed by software.

Meanwhile, in the compression system of the above-mentioned MPEG format or DV format, in order to efficiently compress-code still picture data or moving picture data, a hybrid compression coding method using orthogonal transform coding combined with predictive coding may be employed.

After performing resolution conversion processing on the input information signal compression-coded by the hybrid compression coding method, when performing orthogonal transformation and predictive coding with motion compensation again, motion vectors must be estimated at the step of re-predictive coding processing.

If predictive encoding is performed again with exactly the same resolution without performing resolution conversion processing, motion vectors can be used at the time of predictive encoding. However, if the resolution is converted, conversion distortion is changed. Accordingly, the motion vector used in the repredictive encoding step is also changed.

Thus, motion vectors need to be estimated during the predictive encoding step. However, motion vector estimation requires arithmetic processing.

In order to eliminate this problem, the digital signal conversion device according to the eleventh embodiment is used. In the digital signal conversion method and apparatus according to the eleventh embodiment, from the input information signal compression-coded by hybrid compression coding using orthogonal transform coding combined with predictive coding, in the time domain or the orthogonal transform domain by resolution such as The signal conversion process of rate conversion is processed, and then restored to the orthogonal transform domain for recompression coding, or compression coding in the orthogonal transform domain.

Examples of the above hybrid compression coding are H.261 and H.263 recommended by ITU-T (International Telecommunication Union - Telecommunication Standardization Sector), and MPEG and DV coding standards.

The H.261 standard is an image coding standard with a low bit rate, and its development is mainly for remote conferences and videophones through ISDN. H.263 is an improved version of H.261 for GSTN video telephony system.

The eleventh embodiment is now described with reference to FIG. 22. In the digital video signal conversion apparatus of the eleventh embodiment, MPEG coded data conforming to the MPEG format is input, processed by resolution conversion processing as signal conversion processing, The resolution-converted MPEG encoded signal is output.

This digital video signal conversion device has: decoding part 210, is used to use motion compensation MC to decode the bit stream of the MPEG coded data that detects compression coding along with motion vector (mv); The decoding output of 210 performs resolution conversion processing; and an encoding section 220 for performing compression encoding processing and outputting the converted output image from resolution conversion section 160 following motion detection based on the motion vector mv added to the MPEG encoded data The bit stream of the resolution-converted video coded data is shown in Figure 22.

A digital video signal conversion apparatus constituted by these components will be described below. Of course each component performs the processing of each step of the digital signal conversion method according to the present invention.

The decoding section 210 includes a variable length decoding (VLD) section 112, an inverse quantization (IQ) section 113, an inverse discrete cosine transform (IDCT) section 150, an adder 151, a motion compensation (MC) section 152, and a frame memory (FM) section 153. The FM section 153 is composed of two frame memories FM, which are used as prediction memories.

The VLD section 112 decodes MPEG encoded data, that is, encoded data obtained by variable length encoding of a motion vector and quantized DCT coefficients as additional information, according to variable length encoding, and extracts a motion vector mv. The IQ section 113 performs inverse quantization processing by multiplying the quantized DCT coefficient decoded by the VLD section 112 by the quantization step used on the encoding side.

The IDCT part 150 performs inverse DCT on the DCT coefficients from the IQ part 113, thereby restoring the DCT coefficients to data in the spatial domain, ie, pixel data. Specifically, by inverse DCT, corresponding pixel values (brightness Y and color differences Cr, Cb) are calculated for each block including 8×8 pixels. In the case of an I picture, the pixel value is its actual pixel value. However, in the case of a P picture and a B picture, a pixel value is a difference between corresponding pixel values.

The MC section 152 performs motion compensation processing on the image information stored in the two frame memories of the FM section 153 by using the motion vector mv extracted by the VLD section 112 , and supplies the motion compensation output to the adder 151 .

The adder 151 adds the motion compensation output from the MC section 152 to the difference value from the IDCT section 150, thereby outputting a decoded image signal. The resolution conversion section 160 performs necessary resolution conversion processing on the decoded image signal. The converted output from the resolution converting section 160 is supplied to the encoding section 220 .

Encoding section 220 includes scale conversion section 171, motion estimation (ME) section 172, adder 173, DCT section 175, rate control section 183, quantization (Q) section 176, variable length coding (VLC) section 177, buffer memory 178 , IQ section 179, IDCT section 180, adder 181, FM section 182 and MC section 174.

The scaling section 171 scales the motion vector mv extracted by the VLD section 112 according to the resolution conversion rate used by the resolution converting section 160 . For example, if the resolution conversion rate used by the resolution conversion section 160 is 1/2, the motion vector mv is converted to a 1/2 scale.

The ME section 172 searches a narrow range of the converted output from the resolution converting section 160 by using the scaling information from the scaling converting section 171 , thereby estimating an optimum motion vector of the converted resolution.

The motion vector estimated by the ME section 172 is used in motion compensation by the MC section 174 . The converted output image used for estimation of the motion vector by the ME section 172 from the resolution converting section 160 is supplied to the adder 173 .

The adder 173 calculates the difference between a reference picture described later and the conversion output from the resolution conversion section 160 , and supplies the difference to the DCT section 175 .

The DCT section 175 performs discrete cosine transform on the difference between the reference picture obtained through motion compensation by the MC section 174 and the conversion output picture by using a block of 8×8 size. Regarding the I picture, since inter-frame encoding is performed, DCT arithmetic operation is performed directly without calculating the difference between frames.

A quantization (Q) section 176 quantizes the DCT coefficients from the DCT section 175 by using a matrix table in consideration of visual characteristics. The VLC section 177 compresses the quantized DCT coefficients from the Q section 176 by using variable length coding.

The buffer memory 178 is a memory for maintaining a constant transfer rate of encoded data compressed by the VLC section 177 with variable length encoding. From this buffer memory 178, the resolution-converted video encoded data is output as a constant transmission rate bit stream.

The rate control section 183 controls the increase/decrease of the amount of information generated in the Q section 176, that is, the quantization step size, based on the change information on the increase/decrease of the buffer capacity of the buffer memory 178.

The IQ section 179 and the IDCT section 180 together constitute a local decoding section. The IQ section 179 inverse quantizes the quantized DCT coefficients from the Q section 176 and supplies the DCT coefficients to the IDCT section 180 . The IDCT section 180 performs inverse DCT on the DCT coefficients from the IQ section 179 to restore pixel data, which is supplied to the adder 181 .

The adder 181 adds the motion compensation output from the MC section 174 to the pixel data which is the inverse DCT output from the IDCT section 180 . The image information as the added output from the adder 181 is supplied to the FM section 182 . The image information stored in the FM section 182 is processed by the MC section 174 with motion compensation.

The MC section 174 performs motion compensation on the image information stored in the FM section 182 by using the optimal motion vector estimated by the ME section 172 , and supplies the motion compensation output as a reference picture to the adder 173 .

The adder 173 calculates the difference between the converted output picture from the resolution converting section 160 and the reference picture, supplies the difference to the DCT section 175, as described above.

The DCT section 175, Q section 176, VLC section 177, and buffer memory 178 operate as described above. Finally, the resolution-converted video encoded data is output as a bit stream at a constant transmission rate from the digital video signal conversion device.

In this digital video signal conversion device, when the motion vector is estimated by the ME section 172 of the encoding section 220, the motion vector added to the video signal macroblock initially compressed is determined by the scale conversion section 171 according to the resolution in the resolution conversion section 160. The rate conversion rate is scaled, and based on the scaling information from the scaling section 171, a narrow range of the converted output picture from the resolution converting section 160 is searched to estimate a motion vector for motion compensation instead of a motion vector in the absence of any information. estimate. In this way, since the amount of calculation in the ME section 172 can be significantly reduced, miniaturization of the device and reduction of conversion processing time can be achieved.

A twelfth embodiment will now be described. In this embodiment, a digital video signal conversion device for performing resolution conversion processing on an MPEG video signal and outputting a resolution-converted video signal is also employed.

This digital video signal conversion device has: decoding part 211, is used for carrying out the MPEG coded data of above-mentioned hybrid coding by only carrying out predictive decoding processing with MC, obtains the decoded data of orthogonal transformation domain; Resolution conversion part 260, is used for Resolution conversion processing is performed on decoded data of an orthogonal transform domain from a decoding section 211; and an encoding section 221 for motion detection by using motion vector information based on MPEG-encoded data, along with The motion compensation prediction of the converted output is subjected to compression encoding processing as shown in Fig. 23 .

A digital video signal conversion apparatus constituted by these components will be described below. Of course each component performs the processing of each step of the digital signal conversion method according to the present invention.

In this digital video signal conversion apparatus, compared with the apparatus shown in FIG. That is, in this digital video signal conversion device, resolution conversion processing is performed on decoded data in the DCT domain, and the converted output thereof is encoded.

Orthogonal transforms and inverse orthogonal transforms such as DCT usually require a lot of computation. Therefore, the resolution conversion described above cannot be performed efficiently. Also, since errors are accumulated as the amount of calculation increases, the signal may be degraded.

Thus, in the digital video signal conversion apparatus of FIG. 23, the IDCT section 150, DCT section 174, and IDCT section 180 of FIG. 22 are eliminated. The function of the resolution conversion section 160 is changed.

Also, in order to calculate the activity (activity) to be described later from the converted DCT coefficients from the resolution converting section 160 in the DCT domain and estimate the motion vector by using the activity, an activity calculating section 200 is used instead of the one of FIG. Scale conversion section 171 .

The resolution converting part 260 shown in FIG. It is obtained by inverse quantizing DCT coefficients obtained by quantizing DCT coefficients decoded by the VLD section 212 .

The resolution converting section 260 performs resolution converting processing on the DCT coefficients of the DCT transform domain from the decoding section 211 by using a transform matrix. The transformation matrix is based on an inverse orthogonal transformation matrix corresponding to an orthogonal transformation matrix used for DCT encoding performed on MPEG-encoded data and an inverse orthogonal transformation matrix corresponding to an IDCT encoding used to transform an output signal for obtaining a signal in the time domain. The orthogonal transformation matrix of the alternating transformation matrix is generated.

The DCT coefficients as the resolution conversion output from the resolution conversion section 260 are supplied to the activity calculation section 200 . The activity calculation section 200 calculates the spatial activity for each macroblock from the luminance component of the DCT coefficient from the resolution conversion section 260 . Specifically, the feature of the image is calculated by using the maximum value of the AC value of the DCT coefficient. For example, the presence of fewer high frequency components indicates a smooth image.

The ME section 272 estimates an optimum motion vector at the converted resolution based on the activity calculated by the activity calculation section 200 . Specifically, the ME section 272 converts the motion vector mv extracted by the VLD 212 based on the activity calculated by the activity calculation section 200 so as to estimate the motion vector mv, and supplies the estimated motion vector mv to the MC section 274. The ME section 272 estimates motion vectors in the orthogonal transform domain. This motion estimation in the orthogonal transform domain will be described later.

The resolution-converted DCT coefficients from the resolution conversion section 260 are supplied to the adder 273 through the activity calculation section 200 and the ME section 272 .

The adder 273 calculates the difference between the reference DCT coefficient to be described later and the converted DCT coefficient from the resolution conversion section 260 , and supplies the difference to the quantization (Q) section 276 .

The Q section 276 quantizes the difference values (DCT coefficients), and supplies the quantized DCT coefficients to the VLC section 277 and the IQ section 279 .

The rate control section 283 controls the increase/decrease of the amount of information generated in the Q section 276, that is, the quantization step size, based on the activity information from the activity calculation section 200 and the change information on the increase/decrease of the buffer capacity of the buffer memory 278. .

The VLC section 277 compression-codes the quantized DCT coefficients from the Q section 276 by using variable length coding, and supplies the compressed DCT coefficients to the buffer memory 278 . The buffer memory 278 holds a constant transfer rate of encoded data compressed by the VLC section 277 by variable length encoding, and outputs resolution-converted video encoded data at a constant transfer rate as a bit stream.

The IQ section 279 dequantizes the quantized DCT coefficients from the Q section 276 and supplies the DCT coefficients to the adder 281 . An adder 281 adds the motion compensation output from the MC section 274 to the DCT coefficients from the IQ section 279 as an inverse quantization output. The DCT coefficient information from the adder 281 is supplied to the FM section 282 as an added output. The DCT coefficient information stored in the FM section 282 is processed by the MC section 274 through motion compensation.

The MC section 274 performs motion compensation on the DCT coefficient information stored in the FM section 282 by using the optimum motion vector estimated by the ME section 272, and supplies the motion compensation output to the adder 281 as a reference DCT coefficient.

The adder 273 calculates the difference between the converted DCT coefficient from the resolution converting section 260 and the reference DCT coefficient, supplies the difference to the Q section 276, as described above.

The Q section 276, VLC section 277 and buffer memory 278 operate as described above. Finally, the resolution-converted video encoded data is output at a constant transfer rate from the digital video signal conversion device.

The MC section 274 performs motion compensation in the orthogonal transform domain similarly to the ME section 272 by using the optimal motion vector estimated by the ME section 272 and the reference DCT coefficients stored in the FM section 282 .

Motion estimation and motion compensation in the orthogonal transform domain will now be described with reference to FIGS. 24 to 26 . In FIG. 24, the solid line indicates the macroblock of the picture A to be compressed, and the broken line indicates the macroblock of the reference picture B. In FIG. When the picture A to be compressed and the reference picture B overlap each other by using the motion vector, as shown in FIG. 24 , the boundaries of macroblocks may not coincide. In the case of Fig. 24, the macroblock B' to be compressed is partially stretched over the four macroblocks _B1 , _B2 , _B3 and _B4 of the reference picture B. Therefore, the macroblocks without the reference picture B correspond to the macroblock B' one by one, and the DCT coefficients of the reference picture B at the position of the macroblock B' cannot be obtained. Therefore, it is necessary to obtain the DCT coefficients of the reference picture B of the part where the macroblock B' is located by converting the DCT coefficients of the four macroblocks of the reference picture B on which the macroblock B' partially extends.

FIG. 25 schematically shows the procedure of this conversion process. Since the lower left part of the macroblock _B1 of the reference picture B overlaps the upper right part of the macroblock B', the macroblock _B13 is generated by converting the DCT coefficients of the macroblock _B1 , as described later. Similarly, since the lower right part of the macroblock _B2 of the reference picture B overlaps the upper left part of the macroblock B', the macroblock _B24 is generated by converting the DCT coefficients of the macroblock _B2 , as described later. Similar processing is performed on macroblocks _B3 and _B4 , resulting in macroblocks _B31 and _B42 . By combining the four macroblocks B ₁₃ , B ₂₄ , B ₃₁ , and B ₄₂ thus generated, the DCT coefficients of the reference picture B of the portion where the macroblock B' is located can be obtained.

In short, this processing can be expressed by the following equations (6) and (7).

B'=B ₁₃ +B ₂₄ +B ₃₁ +B ₄₂ ... (6)

DCT(B')=DCT(B ₁₃ )+DCT(B ₂₄ )+DCT(B ₃₁ )+DCT(B ₄₂ )...(7)

Conversion of DCT coefficients of macroblocks will now be described with reference to FIG. 26 . Fig. 26 shows a mathematical model for deriving a partial macroblock B ₄₂ in the spatial domain by computation from the original block B ₄ and so on. Specifically, on the upper left side B ₄ is extracted, 0 is inserted, shifted to the lower right side. That is, B ₄₂ obtained from the block B ₄ by calculation of the following equation (8) is shown.

B ₄₂ =H ₁ ×B ₄ ×H ₂

${\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="h"\; filenames="A20051008944900431.png,A20051008944900451.png"\; state="\{\{state\}\}">h</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="Ih"\; filenames="A20051008944900471.png,A20051008944900491.png"\; state="\{\{state\}\}">Ih</figure-callout></span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right]<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="h"\; filenames="A20051008944900431.png,A20051008944900471.png"\; state="\{\{state\}\}">h</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="Iw"\; filenames="A20051008944900471.png,A20051008944900491.png"\; state="\{\{state\}\}">Iw</figure-callout></span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right]<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

In this equation, Ih and Iw are the identification codes of a matrix with h×h size including h rows and h columns and the identification of a matrix with w×w size including w rows and w columns extracted from block _B4 code. As shown in FIG. 26 , for a pre-matrix H ₁ synthesized first with B ₄ , the first h columns are extracted and converted to the bottom. For _H2 subsequently synthesized with _B4 , the first w rows are extracted and shifted to the right.

Based on Equation (8), the DCT coefficients of B ₄₂ can be directly calculated from the DCT coefficients of B ₄ according to Equation (9) below.

DCT(B ₄₂ )=DCT(H ₁ )×DCT(B ₄ )×DCT(H ₂ )...(9)

This equation is applied to all subblocks and the sum is calculated. In this way, the DCT coefficients of the new block B' can be directly obtained from the DCT coefficients of the original blocks _B1 to _B4 , as represented by Equation (10) below.

$\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; B</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 4</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

The DCT coefficients of H _i1 and H _i2 may be calculated and pre-stored in a memory to constitute a table memory. In this way, motion estimation and motion compensation can be performed even in the orthogonal transform domain.

Then, in the encoding section 221, when the motion vector is estimated by the ME section 272, the motion vector added to the macroblock of the originally compressed video signal, based on the activity calculated by the activity calculation section 200, is obtained from the resolution conversion section 260 by searching A narrow range of the transformed output is estimated instead of an estimate of the motion vector in the absence of any information.

As described above, in the decoding section 211 of the digital video signal conversion apparatus of this embodiment, predictive decoding processing with motion compensation is performed on MPEG-encoded data including predictive encoding with motion detection and Hybrid coding of orthogonal transform coding is performed, that is, inverse quantization is performed after variable length decoding. Then, motion compensation is performed to obtain decoded data kept in the DCT domain, and resolution conversion is performed on the decoded data in the DCT domain. Therefore, resolution conversion can be performed directly in the orthogonal transform domain, and decoding (inverse orthogonal transform) to the time or space domain is unnecessary. In this way, calculations are simplified and high-quality conversions with fewer calculation errors can be achieved. Furthermore, in the encoding section 221, when the motion vector is estimated by the ME section 272, the motion vector attached to the macroblock of the originally compressed video signal is estimated by searching a narrow range based on the activity calculated from the resolution conversion output, instead of the missing No information is available when the motion vector is estimated. Thus, since the amount of calculation of the ME section 272 can be significantly reduced, miniaturization of the device and reduction of conversion processing time can be achieved.

A thirteenth embodiment will now be described. In this embodiment, a digital video signal conversion device for performing signal conversion processing such as resolution conversion processing on MPEG coded data and outputting video coded data is also employed.

This digital video signal conversion device has: a decoding part 340, which is used to perform partial decoding processing on the MPEG encoded data that performs the above-mentioned hybrid encoding, to obtain data in the orthogonal transform domain; The data in the alternating transform domain performs resolution conversion processing; and the encoding section 350 for adding a motion vector based on the motion vector information of the MPEG-encoded data performs compression encoding processing on the converted output from the converting section 343, as shown in FIG. 27 Show.

The decoding section 340 includes a VLD section 341 and an IQ section 342 . These VLD section 341 and IQ section 342 have structures similar to the VLD section 112 and IQ section 113 of FIG. 21, respectively, and operate similarly. The characteristic of this decoding section 340 is that no motion compensation is performed.

Specifically, regarding the P picture and the B picture, resolution conversion is performed on DCT coefficients as difference information by the conversion section 343, and motion compensation is not performed. The converted DCT coefficients obtained by the resolution conversion are quantized by the Q section 345 , and the Q section 345 is rate-controlled by the rate control section 348 . The DCT coefficients are variable-length coded by the VLC section 346 and then output from the buffer memory 347 at a constant rate.

In this case, the motion vector conversion section 344 of the encoding section 350 rescales the motion vector mv extracted by the VLD section 341 according to the resolution conversion rate, and supplies the rescaled motion vector to the VLC section 346 .

The VLC section 346 adds the refresh rate motion vector mv to the quantized DCT coefficients from the Q section 345, and performs variable length coding processing. The VLC section 346 then supplies the encoded data to the buffer memory 347 .

As described above, in the digital video signal conversion apparatus shown in FIG. 27, since motion compensation is not performed in the decoding section 340 and the encoding section 350, calculation can be simplified and the burden on hardware can be reduced.

In the digital video signal conversion apparatus described above, rate conversion can be performed. In short, the digital video signal conversion device can be applied to the conversion of the transmission rate from 4Mbps to 2Mbps without changing the resolution.

Although the configuration of the device has been described in the above embodiments, the corresponding device can be configured as software by using the digital signal conversion method of the present invention.

According to the present invention, decoding with motion compensation is performed on an input information signal compressed and encoded with motion detection, and signal conversion processing is performed on the decoded signal. The converted signal is subjected to compression encoding processing with motion detection based on motion vector information of the input information signal. When resolution conversion processing is applied as this signal conversion processing, compression encoding processing with motion compensation based on information obtained by scaling motion vector information according to the resolution conversion processing is performed on the converted signal. In particular, motion vector information required at the time of compression encoding is scaled according to the resolution conversion rate, and a narrow range is searched. Thus, the amount of calculation at the time of motion vector estimation can be significantly reduced, miniaturization of the device and reduction of conversion processing time can be achieved.

Also, according to the present invention, an input information signal subjected to compression encoding including predictive encoding with motion detection and orthogonal transform encoding is partially decoded to obtain a decoded signal in an orthogonal transform domain. Then, signal conversion processing is performed on the decoded signal in the orthogonal transform domain. The converted signal is subjected to compression coding processing with prediction using motion compensation using motion detection based on motion vector information of the input information signal. When resolution conversion processing is applied as this signal conversion processing, compression encoding processing with motion compensation based on information obtained by converting motion vector information from activity obtained from resolution conversion processing is performed on the converted signal. Thus, motion vector information required at the time of compression encoding can be estimated by searching a narrow range, and the amount of calculation can be significantly reduced. In this way, miniaturization of the device and reduction of conversion processing time can be realized. Also, since signal conversion processing can be performed in the orthogonal transform domain, inverse orthogonal transform processing is not required, and decoding (inverse orthogonal transform) to the time domain or space domain is not required. Thus, calculations are simplified, and high-quality conversion with fewer calculation errors can be performed.

Also, according to the present invention, an input information signal subjected to compression coding including predictive coding with motion detection and orthogonal transform coding is partially decoded to obtain a decoded signal in an orthogonal transform domain. Then, signal conversion processing is performed on the decoded signal in the orthogonal transform domain. Compression coding is performed on the converted signal by adding motion vector information converted based on the motion vector information of the input information signal. Thus, when resolution conversion processing is applied as this signal conversion processing, the converted signal is subjected to compression encoding processing by adding information obtained by scaling the motion vector information based on the resolution conversion processing.

That is, since motion vector information added at the time of compression encoding can be estimated by searching a narrow range, the amount of calculation at the time of motion vector estimation can be significantly reduced. Also, since signal conversion processing can be performed in the orthogonal transform domain, inverse orthogonal transform processing is not required. In addition, since motion compensation processing is not used during decoding and encoding, the calculation amount can be further reduced.

Claims (5)

1. digital signal transfer unit, the digital signal of first form that is used for comprising the orthogonal transform coefficient piece of scheduled unit is converted to the digital signal of second form of the new orthogonal transform coefficient block that comprises another scheduled unit, and this device comprises:

Decoding device, the digital signal of described first form that is used to decode;

The re-quantization device is used for the digital signal of described decoding is carried out re-quantization;

Chromacoder is used to follow the format conversion of the digital signal of described re-quantization, carries out signal processing;

Quantization device is used to quantize the described digital signal of being handled by signal processing;

The data volume control device is used to control the data volume of described quantization device; With

Code device, its data volume that is used to encode is controlled and the digital signal of quantification by described data volume control device, thereby produces the digital signal of described second form.

2. digital signal transfer unit as claimed in claim 1, wherein said orthogonal transform is discrete cosine transform, the digital signal of described first form is the vision signal with pre-determined constant speed compressed encoding, and the digital signal of described second form is the vision signal with the variable bit rate compressed encoding.

3. digital signal transfer unit as claimed in claim 1, wherein said chromacoder are included in the data amount information in the digital signal of described first form by utilization, control the data volume of the digital signal of described second form at orthogonal transform domain.

4. digital signal transfer unit as claimed in claim 1, wherein said chromacoder are included in the data amount information in the digital signal of described first form by utilization, control the data volume of the digital signal of described second form in spatial domain.

5. digital signal transfer unit as claimed in claim 1, wherein said chromacoder is to the relevant block in the digital signal that is included in described first form, based on quantizer number and category information, calculate quantization parameter, the average quantization parameter that relevant block is calculated is so that calculate the quantization parameter of the first piece that comprises a plurality of, calculate the quantizer ratio of the digital signal of described second form from the quantization parameter of each first piece,, quantize relevant block by using the quantizer ratio of calculating.

Images