JPWO1999023834A1 - Digital signal conversion method and digital signal conversion device - Google Patents

Abstract

(57) [Abstract] An input digital signal (DV video signal) in a first format is deframed by a deframing unit (11) and converted back into a variable-length code, decoded by a variable-length decoding (VLD) unit (12), inverse-quantized by an inverse quantization (IQ) unit (13), and inverse-weighted by an inverse weighting (IW) unit (14). A resolution conversion unit (16) then performs a required resolution conversion on the inverse-weighted video signal in the orthogonal transform domain (frequency domain). The resolution-converted video signal is then weighted by a weighting (W) unit (18), quantized by a quantization (Q) unit (19), variable-length coded by a variable-length coding (VLC) unit (20), and output as a digital signal (MPEG video signal) in a second format.

Description

[Detailed Description of the Invention] Digital Signal Conversion Method and Digital Signal Conversion Device Technical Field The present invention relates to the conversion processing of digital signals compressed and encoded using orthogonal transforms such as the discrete cosine transform (DCT), and more particularly to a digital signal conversion method and digital signal conversion device for converting the resolution between compressed video signals of different formats.

BACKGROUND ART Conventionally, the discrete cosine transform (DCT), a type of orthogonal transform coding, has been used as a coding method for efficiently compressing still image data, video data, and other data. When handling digital signals that have undergone such orthogonal transformation, it is sometimes necessary to change the resolution or transform base.

For example, when converting a first orthogonally transformed digital signal with a resolution of 720 x 480 pixels, which is one of the formats for home digital video, into a second orthogonally transformed digital signal with a resolution of 360 x 240 pixels, which is the so-called MPEG-1 format, the first signal is subjected to an inverse orthogonal transform to restore it to a spatial domain signal, after which any necessary conversion processes such as interpolation and thinning are performed, and the signal is again orthogonally transformed to restore it to the second signal.

In this way, the orthogonally transformed digital signal is often first inversely transformed to return it to the original signal, after which the required transformation operation is performed, and then the signal is orthogonally transformed again.

FIG. 28 shows an example of the configuration of a conventional digital signal processing device for performing the above-described resolution conversion on a DCT-transformed digital signal.

This conventional digital signal conversion device receives a video signal in the so-called "DV format," which is one of the formats for home digital video signals (hereinafter referred to as a DV video signal), as a digital signal in a first format, and outputs a video signal conforming to the Moving Picture Experts Group (MPEG) format (hereinafter referred to as an MPEG video signal) as a digital signal in a second format.

The deframing unit 51 is for deframing the DV video signal. In this deframing unit 51, the DV video signal, which is framed in accordance with the so-called DV format, is restored to a variable-length code.

The variable-length decoding (VLD) unit 52 performs variable-length decoding on the video signal converted back into variable-length code by the deframing unit 51. Compressed data in the DV format is compressed at a fixed rate so that the data volume is approximately 1/5 of the original signal, and is variable-length coded to improve data compression efficiency. The variable-length decoding unit 52 performs decoding according to this variable-length coding.

The inverse quantization (IQ) unit 53 inverse quantizes the video signal decoded by the variable length decoding unit 52.

The inverse weighting (IW) unit 54 performs inverse weighting, which is the inverse operation of the weighting applied to the inversely quantized video signal by the inverse quantization unit 53.

Here, weighting refers to the process of taking advantage of the fact that human vision is less sensitive to distortion in high frequencies, so that the higher the frequency components of a video signal, the smaller the DCT coefficient values. This increases the number of high-frequency coefficients that become zero, improving the efficiency of variable-length coding. This can also potentially reduce the amount of computation required for DCT transformation.

An inverse discrete cosine transform (IDCT) unit 55 performs an inverse DCT (inverse discrete cosine transform) on the video signal inversely weighted by the inverse weighting unit 54, converting the DCT coefficients back into spatial domain data, i.e., pixel data.

Then, in the resolution conversion unit 56, the video signal converted back into pixel data by the inverse discrete cosine transform unit 55 is subjected to the required resolution conversion.

Next, the discrete cosine transform (DCT) unit 57 performs a discrete cosine transform (DCT) on the video signal whose resolution has been converted by the resolution conversion unit 56, and converts it back into orthogonal transform coefficients (DCT coefficients).

The weighting (W) unit 58 weights the resolution-converted video signal converted into DCT coefficients, as described above.

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

A variable length coding (VLC) unit 60 then variable-length codes the video signal quantized by the quantization unit 59 and outputs it as an MPEG video signal.

The aforementioned "MPEG" is an abbreviation for the Moving Picture Image Coding Experts Group (MPEG) of the International Organization for Standardization/International Electrotechnical Commission, Joint Technical Committee 1/Subcommittee 29 (ISO/IEC JTC1/SC29). The MPEG1 standard is IS011172, and the MPEG2 standard is IS013818. Among these international standards, ISO11172-1 and ISO13818-1 cover multimedia multiplexing, ISO11172-2 and ISO13818-2 cover video, and ISO11172-3 and ISO13818-3 cover audio.

The image compression coding standards ISO 11172-2 and ISO 13818-2 compress and code image signals on a picture (frame or field) basis, taking advantage of the temporal and spatial correlations of the image. The spatial correlation is achieved by using DCT coding.

Orthogonal transforms such as DCT are also widely used in various image compression coding formats, such as JPEG (Joint Photographic Coding Experts group).

Generally, orthogonal transforms convert original signals from the time or spatial domain into an orthogonally transformed domain such as the frequency domain, enabling highly efficient and reproducible compression and coding.

The aforementioned "DV format" compresses the data volume of digital video signals to approximately one-fifth of their original size for component recording on magnetic tape, and is used in some home and professional digital video equipment. This DV format achieves efficient compression of video signals by combining discrete cosine transform (DCT) and variable-length coding (VLC).

However, orthogonal transforms such as the discrete cosine transform (DCT) and their inverse transforms typically require a large amount of computation, which can hinder efficient resolution conversion of video signals as described above. Furthermore, the increased computational complexity can lead to accumulated errors, resulting in signal degradation.

Disclosure of the Invention The present invention addresses these problems by providing a digital signal conversion method and device that can efficiently perform conversion processes such as resolution conversion by reducing the amount of data calculation required for converting signals into different formats, while minimizing signal degradation.

To solve the above-mentioned problems, the digital signal conversion method according to the present invention comprises a data extraction step of extracting a portion of the orthogonal transform coefficients from each block of a digital signal in a first format, the block consisting of orthogonal transform coefficient blocks of a predetermined unit, to form a partial block; an inverse orthogonal transform step of inversely orthogonally transforming the orthogonal transform coefficients of each partial block on a partial block basis; a partial block concatenation step of concatenating the inverse orthogonally transformed partial blocks to form a new block of the predetermined unit; and an orthogonal transform step of orthogonally transforming the new block on a block basis to generate a digital signal in a second format consisting of the new orthogonal transform block of the predetermined unit.

The present invention proposed to solve the above-mentioned problems comprises an inverse orthogonal transform step of inversely orthogonally transforming a first-format digital signal consisting of orthogonal transform coefficient blocks of a predetermined unit on a block-by-block basis; a block division step of dividing each block of the inverse orthogonally transformed first-format digital signal; an orthogonal transform step of orthogonally transforming the orthogonal transform coefficients constituting each divided block on a block-by-block basis; and a data expansion step of interpolating orthogonal transform coefficients of a predetermined value into each orthogonally transformed block to form a second-format digital signal in the predetermined unit.

To solve the above-mentioned problems, the digital signal conversion device according to the present invention comprises: decoding means for decoding a digital signal of a first format consisting of orthogonal transform coefficients of a predetermined unit; inverse quantization means for inverse quantizing the decoded digital signal; resolution conversion means for extracting a portion of the orthogonal transform coefficients from adjacent blocks of the orthogonal coefficient block of the dequantized digital signal to form partial blocks and convert the resolution; quantization means for quantizing the resolution-converted digital signal; and encoding means for encoding the quantized digital signal into a digital signal of a second format.

The present invention proposed to solve the above problems comprises decoding means for decoding a digital signal of a first format that has been compression-coded using an orthogonal transform; inverse quantization means for inverse quantizing the decoded digital signal; resolution conversion means for converting the resolution of each of the predetermined blocks of the inverse quantized digital signal by interpolating orthogonal transform coefficients of predetermined values into each of the predetermined blocks, thereby constituting each of the blocks into the predetermined units; quantization means for quantizing the resolution-converted digital signal; and encoding means for encoding the quantized digital signal to generate a digital signal of a second format.

Furthermore, the present invention, proposed to solve the above-mentioned problems, provides a digital signal conversion method for converting a digital signal in a first format consisting of orthogonal transform coefficient blocks of a predetermined unit into a digital signal in a second format consisting of new orthogonal transform coefficient blocks of another predetermined unit, characterized in that the data volume of the digital signal in the second format is controlled using data volume information contained in the digital signal in the first format.

The present invention, proposed to solve the above-mentioned problems, provides a digital signal conversion device for converting a digital signal in a first format consisting of orthogonal transform coefficient blocks in a predetermined unit into a digital signal in a second format consisting of new orthogonal transform coefficient blocks in another predetermined unit, the device comprising: decoding means for decoding the digital signal in the first format; inverse quantization means for inverse quantizing the decoded digital signal; signal conversion means for performing signal processing involving format conversion of the inverse quantized digital signal; quantization means for quantizing the processed digital signal; data amount control means for controlling the data amount in the quantization means; and encoding means for encoding the quantized digital signal, the data amount of which is controlled by the data amount control means, into a digital signal in the second format.

The present invention, proposed to solve the above-mentioned problems, provides a digital signal conversion method for converting a digital signal of a first format into a digital signal of a second format, the method comprising: a decoding step for decoding the digital signal of the first format; a signal conversion step for converting the decoded digital signal of the first format into a digital signal of the second format; an encoding step for encoding the digital signal of the second format; and a weighting step for simultaneously performing inverse weighting on the decoded digital signal of the first format and weighting on the digital signal of the second format.

The present invention, proposed to solve the above-mentioned problems, provides a digital signal conversion device for converting a digital signal of a first format into a digital signal of a second format, the device comprising: decoding means for decoding the digital signal of the first format; signal conversion means for converting the decoded digital signal of the first format into a digital signal of the second format; encoding means for encoding the second digital signal; and weighting means for simultaneously performing inverse weighting on the digital signal of the first format and weighting on the digital signal of the second format.

In addition, the present invention proposed to solve the above problems involves decoding an input information signal that has been compression-coded with motion estimation and with motion compensation, subjecting this decoded signal to signal conversion processing, and then subjecting this converted signal to compression-coding processing with motion estimation based on motion vector information of the input information signal.

In addition, the present invention proposed to solve the above problems involves partially decoding an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding to obtain a decoded signal in the orthogonal transform domain, performing signal transformation on this decoded signal in the orthogonal transform domain, and then performing compression coding on this transformed signal with motion compensation prediction using motion estimation based on the motion vector information of the input information signal.

In addition, in the present invention proposed to solve the above-mentioned problems, an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding is partially decoded to obtain a signal in the orthogonal transform domain, this signal is subjected to signal transform processing, and motion vector information converted based on the motion vector information of the input information signal is added to this transformed signal, followed by compression coding.

In addition, the present invention proposed to solve the above problems involves decoding a digital signal of a first format to which motion mode/still mode information has been added in advance, performing signal conversion processing on the decoded signal, determining whether to perform inter-frame differential encoding on each predetermined block of the converted signal based on the motion mode/still mode information, and encoding the converted signal based on the result of this determination, thereby outputting a digital signal of a second format that has been encoded using inter-frame differentials.

In addition, the present invention proposed to solve the above problem involves partially decoding a digital signal in a first format to obtain a signal in an orthogonal transform domain, performing signal transform processing on this orthogonal transform domain signal, and determining whether to perform inter-frame differential encoding on this transformed signal for each predetermined block unit based on the maximum absolute value of the inter-frame difference of the transformed signal, and encoding the transformed signal based on the result of this determination to output a digital signal in the second format.

In addition, the present invention proposed to solve the above-mentioned problems involves performing an inverse orthogonal transform on the intraframe-coded signal and the forward-predictive-coded signal, which are digital signals of a first format that are composed of an intraframe-coded signal that has been subjected to intraframe coding and a forward-predictive-coded signal and a bidirectionally predictive-coded signal that have been subjected to forward and bidirectional interframe predictive coding with motion estimation. Based on the inverse orthogonal transform output, a motion compensation output is generated for addition to the partially decoded forward-predictive-coded signal and the bidirectionally predictive-coded signal. The motion compensation output is then orthogonally transformed, the orthogonal transform output is added to the partially decoded forward-predictive-coded signal and the bidirectionally predictive-coded signal, and a signal based on the addition output is compression-coded to output a digital signal of a second format.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a block diagram showing an example configuration of a digital signal conversion device according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating the principle of resolution conversion in the orthogonal transform domain.

FIG. 3 is a diagram illustrating the principle of resolution conversion in the orthogonal transform domain.

4A to 4C are diagrams illustrating the process of converting a DV video signal into an MPEG video signal by digital signal conversion according to a first embodiment of the present invention.

Figure 5 illustrates the relationship between the DV format and the MPEG format.

Figure 6 illustrates the basic calculation procedure for resolution conversion.

7A and 7B are diagrams for explaining the "still mode" and "motion mode" of the DV format.

FIG. 8 is a diagram for explaining the procedure of the conversion process in the "still mode."

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

FIG. 10 is a diagram for explaining the procedure of the conversion process when enlarging an image.

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

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

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

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

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

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

FIG. 17 is a flowchart showing the basic steps for applying feedback to the next frame using a set quantization scale in the seventh embodiment of the present invention.

FIG. 18 is a block diagram showing an example of the configuration of a conventional digital signal conversion device that converts an MPEG video signal into a DV video signal.

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

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

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

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

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

Figure 24 is a diagram for explaining motion compensation and motion estimation processing in the orthogonal transform domain in the twelfth embodiment of the present invention, showing how macroblock B spans multiple macroblocks in the reference image.

FIG. 25 is a diagram illustrating motion compensation and motion estimation processing in the orthogonal transform domain in the twelfth embodiment of the present invention, showing the transformation processing of reference macroblocks.

FIG. 26 is a diagram illustrating motion compensation and motion estimation processing in the orthogonal transform domain in the twelfth embodiment of the present invention, showing the transformation procedure for reference macroblocks.

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

FIG. 28 is a block diagram showing an example of the configuration of a conventional digital signal conversion device.

BEST MODE FOR CARRYING OUT THE INVENTION A preferred embodiment of the present invention will now be described with reference to the drawings.

In the following, the configuration of a digital signal conversion device according to the present invention will be described first, and then the digital signal conversion method according to the present invention will be described with reference to that configuration.

FIG. 1 shows an example of the configuration of the main components of a digital signal conversion device according to a first embodiment of the present invention. While resolution conversion is shown as an example of signal conversion, it is not limited to this, and various other signal conversions, such as format conversion and filtering, can of course be applied.

This digital signal conversion device receives as input a video signal in the aforementioned "DV format" (hereinafter referred to as a DV signal) as a first digital signal, and outputs as an output a video signal conforming to the Moving Picture Experts Group (MPEG) format (hereinafter referred to as an MPEG video signal) as a second digital signal.

The deframing unit 11 is used to deframe the DV video signal. In this deframing unit 11, the DV video signal, which has been framed in accordance with a predetermined format (the so-called DV format), is restored to a variable-length code.

The variable length decoding (VLD) unit 12 performs variable length decoding on the video signal that has been converted back into a variable length code by the deframing unit 11.

The inverse quantization (IQ) unit 13 inverse quantizes the video signal that has been variable-length decoded by the variable-length decoding unit 12.

The inverse weighting (IW) unit 14 performs inverse weighting, which is the inverse operation of the weighting applied to the inversely quantized video signal by the inverse quantization unit 14.

When performing resolution conversion as an example of signal conversion processing, the resolution conversion unit 16 performs the required resolution conversion in the orthogonal transform domain (frequency domain) on the video signal that has been inverse weighted by the inverse weighting unit 14.

The weighting (W) unit 18 weights the video signal after the resolution conversion.

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

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

The configuration of each part of the digital signal conversion device according to the present invention illustrated in FIG. 1 as described above can be the same as each part of the conventional digital signal conversion device illustrated in FIG. 28.

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

That is, conventional digital signal conversion devices perform the required conversion operation after performing an inverse orthogonal transform on the orthogonal transform coefficients of an input digital signal in a first format to convert it back to data in the spatial domain (on the frequency axis), and then perform another orthogonal transform to convert it back to orthogonal transform coefficients.

In contrast, the digital signal conversion device according to the present invention is characterized in that it performs the required conversion operation on the orthogonal transform coefficients of the input digital signal in the first format in the orthogonal transform coefficient domain (frequency domain), and does not include inverse orthogonal transform means or orthogonal transform means before or after the means for performing conversion processing such as resolution conversion.

Next, the principle of the resolution conversion process in the resolution conversion unit 16 will be explained with reference to FIGS. 2 and 3.

2, an input orthogonal transform matrix generation unit 1 generates an inverse matrix Ts _(k) ^-1 of an orthogonal transform matrix Ts _(k) representing an orthogonal transform previously performed on an input digital signal 5, and sends this to a transform matrix generation unit 3. An output orthogonal transform matrix generation unit 2 generates an orthogonal transform matrix Td _{(L ) corresponding to the inverse transform matrix Td(} L) ^-1 representing an inverse orthogonal transform to be performed on an output digital signal, and sends this to the transform matrix generation unit 3. The transform matrix generation unit 3 generates a transform matrix D for performing conversion processing such as resolution conversion in the frequency domain, and sends this to a signal conversion unit 4. This signal conversion unit 4 converts the input digital signal 5, which has been converted into, for example, the frequency domain by an orthogonal transform, while still in the orthogonally transformed domain, such as the frequency domain, to generate an output digital signal 6.

That is, as shown in FIG. 3, the original time domain (or space domain) signal (original signal A) is transformed, for example, into the frequency domain using the orthogonal transformation matrix Ts _(k) to obtain a frequency signal _B1 (corresponding to the input digital signal 5), which is then reduced (or expanded) by the signal conversion unit 4, for example, to N/L to obtain a frequency signal _B2 (corresponding to the output digital signal 6), and this frequency signal _B2 is then inversely orthogonally transformed using the inverse transformation matrix Td _(L) ^-1 to obtain a time domain signal C.

The example shown in Figure 3 illustrates the case where a one-dimensional source signal A is orthogonally transformed in transform blocks of length k, and m adjacent blocks of the resulting frequency-domain transform blocks, i.e., a continuous frequency signal of length L (= k × m), are transformed into a single block of length N (where N < L), thereby reducing the entire signal to N/L.

In the following explanation, a matrix (orthogonal transform matrix) _{in which} orthogonal transform basis vectors < e1 , e2 , ..., _en > of length n are arranged in each row will be described as T _(n) , and its inverse transform matrix will be described as T _(n) ^-1 . Note that x indicates the vector representation of x. In this case, both matrices are n-th order square matrices. As an example, the one-dimensional DCT transform matrix T ₍₈₎ when n=8 is shown in the following equation (1).

In FIG. 3, for an input digital signal 5 that has already been orthogonally transformed into the frequency domain using an orthogonal transform matrix Ts _(k) , when the size of the orthogonal transform block, i.e., the length of the base, is k, the input orthogonal transform matrix generator 1 generates an inverse orthogonal transform matrix Ts _(k) ^-1 , and the output orthogonal transform matrix generator 2 generates an orthogonal transform matrix Td _(L) whose length of the base is L (= k × m).

At this time, the inverse orthogonal transformation matrix Ts generated by the input orthogonal transformation matrix generation unit 1 is_(k) ^-1 corresponds to the orthogonal transform processing (the inverse processing) used to generate the input digital signal 5, and the orthogonal transform matrix Td generated by the output orthogonal transform matrix generator 2_(L)corresponds to the inverse orthogonal transform process (the inverse process) performed when decoding the output digital signal converted by the signal conversion unit 14, i.e., when converting it to the time domain. Both of these orthogonal transform matrix generation units 1 and 2 are capable of generating basis vectors of any length.

Note that these orthogonal transform matrix generators 1 and 2 may be the same orthogonal transform matrix generator, in which case the orthogonal transform matrices Ts _(k) and Td _(L) are the same type of orthogonal transform matrices that differ only in the length of the base. An orthogonal transform matrix generator exists for each different orthogonal transform method.

Next, the transform matrix generator 3 arranges m inverse orthogonal transform matrices Ts _(k) ^-1 generated by the input orthogonal transform matrix generator 1 on the diagonal as shown in the following equation (2) to create an L-th order square matrix A. Furthermore, when the length of the basis of the output digital signal 6 is N, N low-frequency basis vectors of the orthogonal transform matrix Td _(L) are extracted to create a matrix B consisting of N rows and L columns.

Here, e ₁ , e ₂ , . . . , e _n are N low frequency components extracted when Td _(L) is expressed as a basis vector as follows:

Then, the following is calculated to create a matrix D with N rows and L columns: D = α B A (5) This matrix D becomes the conversion matrix that converts the resolution to the reduction ratio (or enlargement ratio) N/L. Note that α is a scalar or vector value, and is a coefficient for level correction, etc.

In the signal conversion unit 4 shown in Fig. 2, m blocks of the frequency domain input digital signal _B1 are grouped together and divided into metablocks of size L (1 metablock = m blocks), as shown in Fig. 3. If the length of the input digital signal _B1 is not a multiple of L, the signal is supplemented, for example, by stuffing with dummy data such as 0, to make it a multiple of L. The metablocks created in this way are designated Mi (i = 0, 1, 2, ...).

The principles of the above resolution conversion process are described in detail in PCT/JP98/02653, filed by the present applicant on June 16, 1998.

Next, the digital signal conversion method of the first embodiment will be described with reference to the configuration of the digital signal conversion device described above.

4A-4C are schematic diagrams showing the process of converting a DV video signal into an MPEG video signal by digital signal conversion according to an embodiment of the present invention. This process is mainly performed in resolution conversion unit 16 in the digital signal processing device according to the embodiment of the present invention shown in FIG.

Note that the following explanation uses a one-dimensional DCT coefficient block as an example, but the processing for two-dimensional DCT coefficients is similar.

First, as shown in FIG. 4A , four low-frequency DCT coefficients are extracted from adjacent blocks (i) and (i+1), each of which consists of eight DCT coefficients, of a first-format digital signal. That is, of the eight DCT coefficients a0, a1, a2, a3, ..., a7 of block (i), only the four low-frequency DCT coefficients a0, a1, a2, and a3 are extracted to create a partial block with half the number of DCT coefficients. Similarly, of the eight DCT coefficients b0, b1, b2, b3, ..., b7 of block (i+1), only the four low-frequency DCT coefficients b0, b1, b2, and b3 are extracted to create a partial block with half the number of DCT coefficients. Extracting the low-frequency DCT coefficients here is based on the fact that when a video signal is frequency-converted, energy is concentrated in the low DC and AC frequencies.

Each of the subblocks, each consisting of four DCT coefficients, is then subjected to a 4-point inverse discrete cosine transform (4-point IDCT) to obtain reduced pixel data, which are shown as pixel data p0, p1, p2, p3 and pixel data p4, p5, p6, p7 in FIG. 4B, respectively.

Next, each partial block consisting of the reduced pixel data that has been subjected to the inverse discrete cosine transform is combined to generate a block of the same size as the original block. That is, pixel data p0, p1, p2, and p3 are combined with pixel data p4, p5, p6, and p7 to generate a new block consisting of eight pixel data.

Then, an 8-point discrete cosine transform (8-point DCT) is applied to the new block of eight pixel data, generating a block (j) consisting of eight DCT coefficients c0, c1, c2, c3, ..., c7, as shown in Figure 4C.

Using the above procedure, the number of orthogonal transform coefficients (DCT coefficients) per block can be halved to convert the signal to a video signal with a different resolution format. For example, if you want to reduce the number of DCT coefficients to one-quarter, you can achieve this by performing the above process twice in succession.

The above-described resolution conversion process can be applied, for example, when converting from the DV format to the MPEG1 format.

Here, with reference to FIG. 5, the relationship between the DV format and the MPEG format and the format conversion between them will be described.

That is, when the video signal is in the NTSC format as shown in Figure 5, the DV format is a compressed video signal with a resolution of 720 x 480 pixels and a 4:1:1 ratio between the sampling frequency of the luminance signal and the two color difference signals, while the MPEG1 format is a compressed video signal with a resolution of 360 x 240 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the two color difference signals. Therefore, in this case, the resolution conversion process according to the present invention described above simply reduces the number of horizontal and vertical DCT coefficients of the luminance (Y) signal by half, and the number of vertical DCT coefficients of the color difference (C) signal by a quarter.

Note that 4:2:0 is shown as a representative value because odd and even lines alternate between 4:2:0 and 4:0:2.

Furthermore, when the video signal is in the PAL system, the DV format is a compressed video signal with a resolution of 720 pixels x 576 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the two color difference signals, while the MPEG1 format is a compressed video signal with a resolution of 360 pixels x 288 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the two color difference signals. Therefore, in this case, the resolution conversion process according to the present invention described above simply halves the number of horizontal and vertical DCT coefficients of the Y signal, and halves the number of horizontal and vertical DCT coefficients of the C signal.

The above-described resolution conversion process can also be applied to, for example, converting from the DV format to the MPEG2 format.

If the video signal is NTSC, the MPEG2 format is a compressed video signal with a resolution of 720 x 480 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the two color-difference signals. Therefore, in this case, no conversion is performed on the Y signal, and the number of vertical DCT coefficients of the C signal is halved and the number of horizontal DCT coefficients of the C signal is doubled. This expansion method will be described later.

Also, if the video signal is in the PAL system, the MPEG-2 format is a compressed video signal with a resolution of 720 x 576 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the two color-difference signals. Therefore, in this case, no conversion processing is required for either the Y or C signal.

Figure 6 shows the basic calculation procedure for the resolution conversion process described above.

That is, a block of eight DCT coefficients is created by concatenating four DCT coefficients a0, a1, a2, a3 and four DCT coefficients b0, b1, b2, b3 extracted from two adjacent blocks of the input digital signal in the first format. The resulting block is then multiplied by an (8x8) matrix containing two inverse discrete cosine transform matrices (IDCT4), each expressed as a (4x4) matrix, on the diagonal and with the other components set to zero.

These products are then multiplied by a discrete cosine transform matrix (DCT8), given as an (8x8) matrix, to obtain a new block of eight DCT coefficients c0, c1, c2, c3,..., c7.

In the digital signal conversion method according to the present invention, the resolution conversion process is performed in the DCT domain (frequency domain), eliminating the need for the inverse DCT and DCT before and after the resolution conversion process. Furthermore, the amount of computation can be effectively reduced by calculating in advance the product of the (8x8) discrete cosine transform matrix D, which is an (8x8) matrix containing the two (4x4) inverse discrete cosine transform matrices (DCT4) on its diagonal.

Next, we will explain in more detail the process of converting a DV video signal, which is a digital signal in the first format, into an MPEG-1 video signal, which is a digital signal in the second format.

The DV format has a "still mode" and a "motion mode" that can be switched depending on the results of motion detection. These modes are determined by motion detection before DCT of each (8x8) matrix in a video segment, and the DCT is performed in one of two modes depending on the results. Various methods for motion detection are possible, including comparing the sum of the absolute values of inter-field differences with a predetermined threshold.

The "still mode" is the basic mode of the DV format, and (8x8) DCT is performed on (8x8) pixels in a block.

Each 8x8 block consists of one DC component and 63 AC components.

Furthermore, "motion mode" is used to prevent a decrease in compression efficiency due to energy dispersion caused by interlaced scanning when performing DCT on a moving subject. In this motion mode, an (8x8) block is divided into a (4x8) block in the first field and a (4x8) block in the second field, and a (4x8) DCT is performed on the pixel data of each (4x8) block. This suppresses an increase in vertical high-frequency components and prevents a decrease in compression ratio.

Each of the above (4x8) blocks consists of one DC component and 31 AC components.

As described above, in the DV format, the block configuration differs between still mode and motion mode. To ensure that subsequent processing is the same, for motion mode blocks, after each (4 x 8) DCT, the sum and difference of coefficients of the same order for each block are calculated to form an (8 x 8) block. This process allows motion mode blocks to be considered to be composed of one DC component and 63 AC components, just like still mode blocks.

When converting a DV format video signal to an MPEG1 format video signal, it is necessary to separate only one of the fields, since the MPEG1 format only handles 30 frames per second video signals and does not have the concept of fields.

FIG. 7A shows a schematic diagram of the process of separating fields when converting DCT coefficients in the "motion mode (2x4x8 DCT mode)" of the DV format into DCT coefficients in the MPEG-1 format.

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

Therefore, by adding the upper half (4x8) block 31a of the (8x8) DCT coefficient block 31 to the lower half (4x8) block 31b and halving the sum, we can obtain a (4x8) block 35a consisting only of DCT coefficients of the first field (A). Similarly, by subtracting the (4x8) block 31a from the lower half (4x8) block 31b and halving the difference, we can obtain a (4x8) block 35b consisting only of discrete cosine coefficients of the second field (B). In other words, the above process allows us to obtain an (8x8) block 35 with separated fields.

Then, the DCT coefficients of one of these fields, for example the first field, are subjected to the resolution conversion process described above.

FIG. 7B illustrates the process of separating fields in "still mode (8x8 DCT mode)."

The (8x8) DCT coefficient block 32 is a mixture of DCT coefficients from the first field (A) and the second field (B). Therefore, by separating the fields using the process described below, a (4x8) block 35a consisting only of the first field (A) can be obtained. Similarly, a transform process is required to subtract the (4x8) block 31a from its lower half, the (4x8) block 31b, to obtain a (4x8) block 35b consisting only of the second field (B).

FIG. 8 shows the procedure for field separation processing in "still mode."

First, the input consisting of eight DCT coefficients d0, d1, d2, d3, ..., d7 is multiplied by an eighth-order inverse discrete cosine transform matrix (IDCT8) to convert it back to pixel data.

Next, an 8x8 matrix is applied for field separation, dividing the top and bottom of the 8x8 block into a first field and a second field of 4x8 blocks, respectively.

This is then multiplied by an 8x8 matrix whose diagonal contains two discrete cosine transform matrices (DCT4), each of which is given as a 4x4 matrix.

This results in eight DCT coefficients: four DCT coefficients e0, e1, e2, and e3 from the first field, and four DCT coefficients f0, f1, f2, and f3 from the second field.

Then, the DCT coefficients of one of these fields, for example the first field, are subjected to the resolution conversion process described above.

In the digital signal conversion method according to the present invention, resolution conversion is performed in the DCT domain (frequency domain), eliminating the need for the inverse DCT and DCT before and after the resolution conversion. Furthermore, by calculating in advance the product of the (8 × 8) discrete cosine transform matrix described above with an (8 × 8) matrix containing the two (4 × 4) inverse discrete cosine transform matrices (IDCT4) shown in Figure 6 on its diagonal, the amount of calculations can be effectively reduced.

The resolution conversion process described above applies to reducing an image. Below, we will explain the resolution conversion process for enlarging an image as a second embodiment.

9A-9C are schematic diagrams illustrating how a DV video signal is converted into an MPEG-2 video signal using the digital signal conversion method of the present invention.

In the following explanation, one-dimensional DCT coefficients will be used as an example, but two-dimensional DCT coefficients can also be processed in the same manner.

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

Next, the block of eight pixels is divided into two parts, generating two partial blocks of four pixels each.

Next, a 4-point DCT is applied to each of the two subblocks, each consisting of four DCT coefficients, to generate two subblocks (i0-i3 and j0-j3), each consisting of four DCT coefficients.

Then, as shown in Figure 9C, for each of the two partial blocks consisting of the four pixel data, four DCT coefficients are padded with zeros on the high-frequency side to generate blocks (v) and (v+1), each consisting of eight DCT coefficients.

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

FIG. 10 shows the procedure of the conversion process at this time.

First, the input consisting of eight DCT coefficients g0, g1, g2, g3, ..., g7 is multiplied by an eighth-order inverse discrete cosine transform (IDCT) matrix to return eight pixel data.

Next, the block of eight pixels is divided into two parts, generating two partial blocks of four pixels each.

Next, each of the two subblocks, each consisting of four DCT coefficients, is multiplied by a 4-point discrete cosine transform matrix (given as a (4 × 4) matrix) and a (4 × 8) matrix containing zero matrices above and below (given as a (4 × 4) matrix), to generate two subblocks (i0–i7 and j0–j7) each containing eight DCT coefficients.

By processing in this way, the DCT coefficients of two blocks are obtained from one block, thereby increasing the resolution in the frequency domain.

In the case of NTSC, when converting from DV format to MPEG2 format, there is no need to convert the luminance signal Y in the horizontal or vertical directions, as shown in Figure 5. However, the color difference signal C must be expanded by two times in the horizontal direction and reduced by half in the vertical direction. Therefore, the expansion process described above is used to convert the horizontal resolution of the color difference signal C when converting from DV format to MPEG2 format.

FIG. 11 shows an example of the configuration of the main components of a digital signal conversion device according to a third embodiment of the present invention. Note that the same reference numbers are used for components that are the same as those in the first embodiment. The difference from FIG. 1 is that the weighting unit 18 and the inverse weighting unit 14 are combined into a weighting processing unit 21.

In other words, the weighting processing (IW*W) unit 21 performs both inverse weighting, which is the inverse operation of the weighting applied to the DV video signal, which is the input digital signal in the first format, and weighting for the MPEG video signal, which is the output digital signal in the second format.

According to this configuration, the inverse weighting process for the input video signal in the first format and the weighting process for the output video signal in the second format can be performed simultaneously, thereby reducing the amount of calculations required compared to when the inverse weighting process and the weighting process are performed separately.

In the digital signal conversion device of the third embodiment shown in FIG. 11, the weighting processing unit 21 is arranged after the resolution conversion unit 16, but the weighting processing unit may also be arranged before the resolution conversion unit 16.

FIG. 12 shows a digital signal conversion device according to a fourth embodiment of the present invention, in which the weighting processor 22 is arranged before the resolution converter 16. The configuration of each component of the digital signal processing device shown in this figure can be similar to that of the digital signal conversion device shown in FIG.

The weighting process, which combines the inverse weighting of the first format digital signal with the weighting of the second format digital signal, and the weighting process can be performed either before or after an orthogonal transform such as a discrete cosine transform (DCT), are based on the fact that these computational operations are linear.

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

As shown in FIG. 13, this digital video signal conversion device comprises a decoder 8 that decodes the DV video signal, a resolution converter 16 that performs resolution conversion processing for format conversion on the decoded output from the decoder 8, a determination unit 7 that determines whether to perform forward interframe differential encoding on the converted output from the resolution converter 16 for each predetermined block in accordance with the motion mode/still mode information, and an encoder 9 that encodes the converted output from the resolution converter 16 based on the determination result from the determination unit 7 and outputs the MPEG video signal.

The following description will be made of a digital video signal conversion device configured from these components, but it goes without saying that each component performs the respective steps of the digital signal conversion method according to the present invention.

In the DV video signal input to this digital video signal conversion device, a mode flag (e.g., 1 bit) indicating the still mode/motion mode is added to each DCT block in advance.

In this digital video signal conversion device, the decision unit 7 determines, based on this mode flag, whether or not to perform forward interframe differential coding for each predetermined block of the converted output from the resolution conversion unit 16. This operation will be described in detail later.

The deframing unit 11 extracts the mode flag indicating the still mode/motion mode and supplies it to the determination unit 7.

The deshuffling unit 15 undoes the shuffling that was performed on the DV encoding side to equalize the amount of information in the video segment, which is a unit of fixed length.

The decision unit 7 comprises an adder 27 and an I (I picture)/P (P picture) decision and determination unit 28. The adder 27 adds reference DCT coefficients stored in a frame memory (FM) unit 24 (described later) as negative values to the resolution conversion output. The I/P decision and determination unit 28, to which the added output from the adder 27 is supplied, also receives a mode flag indicating the still mode/motion mode from the deframing unit 11.

The operation of the I/P judgment and determination unit 28 will now be described in detail. The conversion output from the resolution conversion unit 16 is in units of 8 x 8 DCT coefficients. Four of these 8 x 8 DCT coefficient blocks are allocated to the luminance signal and two to the color difference signal, making a total of six DCT coefficient blocks that make up one of the above-mentioned predetermined blocks. This predetermined block will be called a macroblock.

Incidentally, P pictures are based on the premise that the difference from the previous frame is simply taken. In the case of a still image, taking the difference reduces the amount of information, but in the case of a moving image, taking the difference increases the amount of information. Therefore, if the mode flag indicating still mode/motion mode is checked and it is determined that there is movement, the amount of information increases, so the macroblock remains an I picture; if it is determined to be still, the difference is taken and it becomes a P picture, allowing for efficient encoding.

For example, if the mode flags sent from the deframing unit for the six DCT coefficient blocks all indicate the motion mode, the I/P judgment and decision unit 28 determines the macroblock to be an I picture. Also, if, for example, only one flag indicating a motion mode is detected among the six DCT blocks, the I/P judgment and decision unit 28 determines the macroblock to be a P picture.

Furthermore, if four or more of the six DCT blocks have a motion flag attached, the macroblock may be an I picture. If all six DCT blocks have a still mode flag attached, the macroblock may be a P picture.

The DCT coefficients of each macroblock determined to be an I/P picture by the I/P judgment and determination unit 28 are supplied to the encoding unit 9.

The encoding unit 9 includes a weighting (W) unit 18, a quantization (Q) unit 19, an inverse quantization (IQ) unit 26, an inverse weighting (IW) unit 25, an FM unit 24, a variable length coding (VLC) unit 20, a buffer memory 23, and a rate control unit 29.

The weighting (W) unit 18 weights the DCT coefficients, which are the transform output from the transform unit 16 via the determination unit 7.

The quantization (Q) unit 19 quantizes the DCT coefficients weighted by the weighting (W) unit 18. The variable length coding (VLC) unit 20 then variable-length codes the DCT coefficients quantized by the quantization unit 19 and supplies the coded data to the buffer memory 23 as MPEG coded data.

The buffer memory 23 maintains a constant transfer rate for the MPEG-encoded data and outputs it as a bitstream. The rate control unit 29 controls the amount of information generated by the quantization (Q) unit 19, i.e., the quantization step, based on information about changes in the buffer capacity of the buffer memory 23.

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

As described above, in the digital video signal conversion device shown in FIG. 13, the decision unit 7 uses the I/P decision and determination unit 28 to determine whether each macroblock is an I or P picture in response to the mode flag indicating motion mode/still mode sent from the deframing unit 11. This allows a DV signal originally consisting only of I pictures to be converted into an MPEG picture using either I or P pictures, thereby taking advantage of the improved compression rate characteristic of MPEG video signals.

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

The digital video signal conversion device according to the sixth embodiment is a digital video signal conversion device in which the determination unit 7 shown in FIG. 13 is replaced with a determination unit 30 shown in FIG. 14.

That is, the system includes a decoder 8 that partially decodes the DV signal to obtain an orthogonal transform domain signal, such as DCT coefficients; a converter 16 that performs signal conversion processing for format conversion on the DCT coefficients from the decoder 8; a decision unit 30 that determines whether to apply forward interframe differential coding to the converted output from the converter 16 for each predetermined block, based on the maximum absolute value of the interframe difference of the converted output; and an encoder 9 that encodes the converted output from the converter 16 based on the decision result of the decision unit 30, and outputs the MPEG video signal.

The decision unit 30 determines the maximum absolute value of the AC coefficients obtained by subtracting the transformed DCT coefficients, which are the output from the conversion unit 16, from the reference DCT coefficients from the FM unit 24, compares this maximum value with a predetermined threshold, and assigns an I/P picture to each macroblock based on the comparison result.

The determination unit 30 includes a difference calculation unit 31, a maximum value detection unit 32, a comparison unit 33, and an I/P determination unit 35.

The difference calculation unit 31 calculates the difference between the transformed DCT coefficients from the transform unit 16 and the reference DCT coefficients from the FM unit 24. The difference output from the difference calculation unit 31 is supplied to the maximum value detection unit 32 and also to the I/P determination unit 35.

The maximum value detector 32 detects the maximum absolute value of the AC coefficients of the differential output.

Basically, if the amount of information converted into the DCT coefficients is large, the AC coefficients will also be large, and if the amount of information is small, the AC coefficients will be small.

Comparator 33 compares the maximum absolute value from maximum value detector 32 with a predetermined threshold value supplied from terminal 34. If this threshold value is appropriately selected, the amount of information converted into the DCT coefficients can be determined based on the magnitude of the maximum absolute value of the AC coefficients.

The I/P determination unit 35 uses the comparison result from the comparison unit 33 to determine whether the difference in DCT coefficients from the difference calculation unit 31, i.e., the difference in information amount, is large or small. If it determines that the difference is large, it allocates the macroblock consisting of the transformed DCT coefficient block from the transformation unit 16 to an I picture; if it determines that the difference is small, it allocates the macroblock from the difference calculation unit 31 to a P picture.

That is, if the absolute value of the maximum value is greater than the threshold value, it is determined that the amount of information of the difference is large, and the macroblock is designated as an I picture. On the other hand, if the absolute value of the maximum value is smaller than the threshold value, it is determined that the amount of information of the difference is small, and the macroblock is designated as a P picture.

As a result, this sixth embodiment of the digital video signal conversion device can also convert a DV signal, which originally consisted only of I pictures, into an MPEG picture that uses either I or P pictures, thereby taking advantage of the improved compression rate that is a characteristic of MPEG video signals.

The digital video signal conversion device shown in FIGS. 13 and 14 inputs and outputs NTSC DV signals and MPEG1 video signals, but may also be applied to PAL signals.

The above-described resolution conversion process can also be applied to, for example, converting from the DV format to the MPEG2 format.

Furthermore, while the resolution conversion process performed by the converter 16 has been described as primarily reducing the resolution, it is also possible to increase the resolution. In other words, generally, by adding high-frequency components to the input digital signal in the frequency domain, the resolution can be increased by any factor.

For example, when applying MPEG-2 video signals to digital broadcasting services, the signals are classified by profile (function) and level (resolution). Resolution expansion can be applied, for example, when converting the DV signal to a Main Profile/High Level (MP@HL) video signal used in US digital HDTV.

The processing of the sixth embodiment may be performed by software.

Next, a digital signal conversion method and apparatus according to a seventh embodiment of the present invention will be described with reference to FIG. 15. Note that the same reference numbers are used to designate the same components as those in the above-described embodiments.

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

FIG. 16 shows the basic procedure for setting the quantization scale for each macroblock (MB) of each frame when a DV video signal is converted into an MPEG video signal by the digital signal conversion method of the seventh embodiment.

In step S1, a quantization number (Q_NO) and a class number (Class) are first obtained for each macroblock. The quantization number (Q_NO) is a value ranging from 0 to 15 and is common to all six DCT blocks in the macroblock. The class number (Class) is a value ranging from 0 to 3 and is assigned to each of the six DCT blocks.

Next, in step S2, the quantization parameter (q param) is calculated.

Quantization table q_table[4] = {9, 6, 3, 0} Quantization parameter q_param = Q_NO + q_table[class] In other words, the quantization table has four possible values (9, 6, 3, 0), which correspond to class numbers 0, 1, 2, and 3. For example, if the class number is 2 and the quantizer number is 8, the quantization table value 3 corresponding to class number 2 and the quantizer number 8 are added together, resulting in a quantization parameter of 11.

Next, in step S3, the average of the quantization parameters (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 calculated using 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_pram] In other words, the quantization table has 25 values (32 to 2), each corresponding to the quantization parameter calculated as described above. That is, the quantization table corresponding to a quantization parameter value of 0 is 32, the quantization table corresponding to a quantization parameter value of 1 is 16, and the quantization table corresponding to a quantization parameter value of 5 is 8. For example, if the average quantization parameter calculated as described above is 10, then the quantizer scale value of 4, corresponding to a quantization parameter value of 10, becomes the quantizer scale value. Through the above procedure, the MPEG quantizer scale (quantizer_scale), which depends on the target rate, is calculated for each macroblock in each frame based on the quantization parameter (q_param). Note that the correspondence between class numbers and quantization tables and the relationship between quantization parameters and quantization tables are empirically determined.

In the digital signal conversion apparatus according to the present invention shown in FIG. 15, the above processing is performed in the rate control unit 40 based on the quantization number (Q_NO) and class number (Class) sent from the deframing unit 11.

FIG. 17 shows the basic procedure for applying feedback to the next frame using the quantization scale set by the procedure described above.

In step S11, first, a target number of bits per frame is set at the bit rate set by the procedure described above.

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

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

Then, in step S14, the quantization scale is adjusted based on the results of the above calculations.

The calculations in each step above are expressed as follows:

diff = cont * diff (cont: constant) q_param = q_param ± f(diff) quantizer_scale = q_table[q_param] In other words, normalization is performed by multiplying the difference value diff calculated in step S13 by the constant cont. This normalized difference value is multiplied by an empirically determined function, and the result is added or subtracted from the quantization parameter to obtain the quantization parameter. The value corresponding to this quantization parameter value is selected from the quantization table with 25 possible values mentioned above and used as the quantizer scale for the next frame.

Through the above procedure, a new quantizer scale (quantizer_scale) is calculated based on the adjusted quantizer parameter (q_param), and inter-frame feedback is performed to use this new quantizer scale for the next frame.

Next, as an eighth embodiment, a digital signal conversion method and a digital signal conversion apparatus according to the present invention will be described. While the above embodiment illustrates an example of converting from DV format to MPEG format, the following embodiment will illustrate an example of converting from MPEG format to DV format.

First, a conventional device for converting from the MPEG format to the DV format will be described with reference to FIG.

The digital video signal conversion device shown in FIG. 18 comprises an MPEG decoder 70 that decodes MPEG2 video data and a DV encoder 80 that outputs DV video data.

In the MPEG decoder 70, a parser 71 receives the MPEG2 video data bitstream, detects the header of the quantized DCT coefficient bitstream framed in accordance with the MPEG2 format, and supplies the variable-length coded quantized DCT coefficients to a variable-length decoding (VLD) unit 72. The parser also extracts motion vectors (mv) and supplies them to a motion compensation (MC) unit 77.

The variable length decoding (VLD) unit 72 variable length decodes the variable length coded quantized DCT coefficients and supplies them to the inverse quantization (IQ) unit 73.

The inverse quantization unit 73 performs inverse quantization by multiplying the quantized DCT coefficients decoded by the variable-length decoding unit 72 by the quantization step used on the encoding side to obtain DCT coefficients, which are then supplied to the inverse discrete cosine transform (IDCT) unit 74.

The inverse discrete cosine transform unit 74 performs inverse DCT on the DCT coefficients from the inverse quantization unit 73, converting the DCT coefficients back into spatial domain data, i.e., pixel data. Specifically, the inverse DCT calculates pixel values (luminance Y, color differences Cr, Cb) for each 8x8 pixel block. Note that while the pixel values used here are the actual pixel values for I-pictures, they are the difference values between corresponding pixel values for P-pictures and B-pictures.

The motion compensation unit 77 generates a motion compensation output using the image information stored in the two frame memories FM of the frame memory (FM) unit 76 and the motion vector mv extracted by the parser 71, and supplies this motion compensation output to the adder 75.

The adder 75 adds the motion compensation output to the difference value from the inverse discrete cosine transform unit 74, and supplies the decoded image data to a discrete cosine transform (DCT) unit 81 and a frame memory unit 76 of the DV encoder 80.

In the DV encoder 80, the discrete cosine transform unit 81 performs DCT processing on the decoded image data to convert it back into data in the orthogonal transform domain, i.e., DCT coefficients, and supplies the data to the quantization (Q) unit 82.

The quantization unit 82 quantizes the DCT coefficients using a matrix table that takes into account visual characteristics, and supplies the quantized DCT coefficients to the variable length coding (VLC) unit 83 as an I-picture in the DV format.

The variable-length coding unit 83 performs variable-length coding on the I-pictures in the DV format to compress them and supply them to the framing unit 84.

The framing unit 84 frames the DV format data that has been subjected to the variable-length coding process, and outputs it as a bit stream of DV video data.

However, orthogonal transforms such as the discrete cosine transform (DCT) and their inverse transforms typically require a large amount of computation, which can hinder efficient format conversion of video data as described above. Furthermore, the increased computational complexity can lead to accumulated errors, resulting in signal degradation.

To solve this problem, a digital video signal conversion device will be described as an eighth embodiment with reference to FIG.

The digital signal conversion device shown in FIG. 19 receives an MPEG video signal conforming to the MPEG format described above as a first digital signal and outputs a DV signal as a second digital signal.

Parser 111 references the header of the MPEG video signal, a digital signal in the first format, input as a bitstream, and extracts image motion information such as the motion vector mv and quantization scale.

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

The variable length decoding (VLD) unit 112 performs variable length decoding on the bitstream of the MPEG video signal from which the necessary information has been extracted by the parser 111.

The inverse quantization (TQ) unit 113 inverse quantizes the MPEG video signal decoded by the variable length decoding unit 112.

The MPEG video signal dequantized by the dequantization unit 113 is then input to the adder 125. The result of motion compensation for the motion vector mv from the parser 111 is also input to this adder 125 from the motion compensation unit 115.

The output from the adder 125 is sent to the signal converter 116, which will be described later, and is also input to the motion compensation unit 115 via the frame memory 114. The signal converter 116 performs required signal conversion processing, such as resolution conversion, in the orthogonal transform domain (frequency domain) on the video signal input via the adder 125.

The video signal that has undergone the required signal conversion processing in the signal conversion unit 116 is then shuffled in the shuffling unit 117 and sent to the buffer 118 and the classification unit 122.

The video signal sent to the buffer 118 is sent to a quantization (Q) unit 119 where it is quantized, then variable-length coded by a variable-length coding (VLC) unit 120, and then framed by a framing unit 121, and output as a bitstream of a DV video signal.

Meanwhile, the classifying unit 122 classifies the video signals shuffled by the shuffling unit 117 and sends the results to the evaluating unit 123 as class information.

The evaluation unit 123 determines the quantization number to be used by the quantization unit 119 based on the class information from the classifier unit 122 and the quantizer scale (quantizer_scale) from the parser 111.

With this configuration, the data volume of the DV video signal output as the second format video signal can be determined based on the data volume information contained in the MPEG video signal input as the first format video signal, thereby simplifying the process of determining the data volume for the second format video signal generated by signal conversion.

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

Next, a digital signal conversion method and a digital signal conversion device according to the ninth embodiment of the present invention will be described with reference to FIG.

This is a digital video signal conversion device that converts MPEG video data conforming to the MPEG2 format into DV video data conforming to the DV format, both of which are intended for PAL system data.

When the video signal is in the PAL system, the MPEG-2 and DV formats are compressed video signals with a resolution of 720 x 576 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the two color difference signals, so no special resolution conversion processing is required for either the Y or C signal.

20, the MPEG decoder 100 includes a parser (Parse) 111, a variable length decoding (VLD) unit 112, an inverse quantization (IQ) unit 113, an adder 125, an inverse discrete cosine transform (IDCT) unit 131, a frame memory (FM) unit 132, a motion compensation (MC) unit 115, and a discrete cosine transform (DCT) unit 130. Here, the frame memory FM unit 132 is configured to be used as two prediction memories.

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

The overall operation is explained below. First, the parser 111 references the header of the MPEG-2 video data input as a bitstream, and converts the quantized DCT coefficients that have been framed according to the MPEG-2 format back into variable-length codes and supplies them to the variable-length decoding unit 112. It also extracts motion vectors (mv) and supplies them to the motion compensation unit 115.

The variable-length decoding unit 112 performs variable-length decoding on the quantized DCT coefficients that have been converted back into variable-length codes, and supplies the dequantized DCT coefficients to the inverse quantization unit 113.

The inverse quantization unit 113 performs inverse quantization by multiplying the quantized DCT coefficients decoded by the variable-length decoding unit 112 by the quantization step used on the encoding side to obtain DCT coefficients, which are then supplied to the adder 125. The DCT coefficients obtained by the variable-length decoding unit 112 and the inverse quantization unit 113 are supplied to the adder 125 as output that is not subjected to inverse discrete cosine transform to be restored to pixel data, i.e., as partially decoded data.

The adder 125 also receives the motion compensation output from the motion compensation unit 115, which has been orthogonally transformed by the discrete cosine transform unit 130. The adder 125 then adds the motion compensation output to the partially decoded data in the orthogonal transform domain, and supplies this sum to the DV encoder 110 and also to the inverse discrete cosine transform unit 131.

The inverse discrete cosine transform unit 131 performs inverse discrete cosine transform on the I-picture and P-picture from the summed output, converting them into spatial domain data. This spatial domain data becomes reference image data used for motion compensation. This reference image data for motion compensation is stored in the frame memory unit 132.

The motion compensation unit 115 then generates a motion compensation output using the reference image data stored in the frame memory unit 132 and the motion vector mv extracted by the parser 111, and supplies this motion compensation output to the discrete cosine transform unit 130.

The discrete cosine transform unit 130 converts the motion compensation output processed in the spatial domain back into the orthogonal transform domain as described above, and then supplies it to the adder 125.

The adder 125 adds the DCT coefficients of the motion compensation output from the discrete cosine transform unit 130 to the DCT coefficients of the partially decoded P and B picture difference signals from the inverse quantization unit 113. The sum output from the adder 125 is then supplied to the DV encoder 110 and the inverse discrete cosine transform unit 131 as partially decoded data in the orthogonal transform domain.

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

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

The quantization unit 141 quantizes the decoded output of the I, P, and B pictures from the MPEG decoder 100 in the orthogonal transform domain, i.e., the DCT coefficients, and supplies the quantized output to the variable-length coding unit 142.

The variable-length coding unit 142 performs variable-length coding on the quantized DCT coefficients and supplies the result to the framing unit 143. The framing unit 143 frames the compressed and encoded data from the variable-length coding unit 142 and outputs it as a bit stream of DV video data.

In this way, when the MPEG-2 video data to be converted is an I-picture, the MPEG decoder 100 partially decodes the MPEG-2 video data up to the orthogonal transform domain using the variable-length decoding unit 112 and inverse quantization unit 113, and the DV encoder 110 partially encodes the data using the quantization unit 141 and variable-length coding unit 142. At the same time, the I-picture undergoes inverse discrete cosine transform in the inverse discrete cosine transform unit 131 and is stored in the frame memory unit 132 to be used as a reference image for the P/B-picture.

Furthermore, when the P and B pictures to be transformed are processed, as described above, only the process of generating the motion compensation output is performed in the spatial domain using the inverse discrete cosine transform unit 131, while the P and B picture differential signals partially decoded by the variable length decoding unit 112 and inverse quantization unit 113, plus the remaining components that make up the frame, are processed in the discrete cosine transform domain by the discrete cosine transform unit 130. The resulting signals are then partially encoded by the DV encoder 110.

In particular, for a P picture, the macroblock at the position indicated by the motion vector mv is brought in by the motion compensation unit 115 from the I picture that has been inversely discrete cosine transformed by the inverse discrete cosine transform unit 131. The macroblock is then subjected to discrete cosine transform by the discrete cosine transform unit 130, and added to the DCT coefficients of the P picture, which is the difference signal, in the discrete cosine transform domain using the adder 125. This is based on the fact that the result of performing a discrete cosine transform on the result of an addition in the spatial domain is equivalent to the result of adding two discrete cosine transforms together. The result is then partially encoded by the DV encoder 110. At the same time, for reference to the next B picture, the sum output from the adder 125 is subjected to an inverse discrete cosine transform by the inverse discrete cosine transform unit 131, and stored in the frame memory unit 132.

In the case of a B picture, the macroblock at the position indicated by the motion vector mv is taken from the P picture that has been inversely discrete cosine transformed by the inverse discrete cosine transform unit 131. Then, the macroblock is subjected to a discrete cosine transform by the discrete cosine transform unit 130, and the B picture DCT coefficients, which are the difference signal, are added in the discrete cosine transform domain.

In the case of bidirectional transmission, the average is calculated from two reference frames.

The results are then partially encoded by the DV encoder 110. Note that since B pictures are not reference frames, they are not subjected to inverse discrete cosine transform by the inverse discrete cosine transform unit 131.

According to the ninth embodiment described above, to decode an I-picture, both an inverse discrete cosine transform (IDCT) and a discrete cosine transform (DCT) were conventionally required. However, with the digital video signal conversion apparatus of this embodiment, only the IDCT is required for reference.

Furthermore, decoding a P picture requires both DCT and reference IDCT processing, whereas decoding a B picture conventionally requires both DCT and IDCT. In contrast, this method uses only DCT and does not require reference IDCT.

Taking typical MPEG-2 data, for example, with a GOP count of N = 15 and a forward prediction picture interval of M = 3, there is one I-picture, four P-pictures, and ten B-pictures. Assuming the computational complexity of the DCT and IDCT is roughly the same, the MPEG-2 data for the above 15 frames, omitting weighting, would be: 2 x DCT x (1/15) + 2 x DCT x (4/15) + 2 x DCT x (10/15) = 2 x DCT In contrast, with the digital video signal conversion device shown in Figure 20, the computational complexity would be: 1 x DCT x (1/15) + 2 x DCT x (4/15) + 1 x DCT x (10/15) = 1.2666 x DCT This significantly reduces the computational complexity. In this formula, DCT indicates the computational complexity.

In other words, the digital video signal conversion device shown in FIG. 20 can significantly reduce the amount of data calculation processing required for format conversion from MPEG2 video data to DV video data.

Next, another embodiment of the digital video signal conversion device according to the tenth embodiment will be described with reference to FIG.

This tenth embodiment is also a digital video signal conversion device that converts MPEG video data conforming to the MPEG2 format into DV video data conforming to the above-mentioned DV format, but the MPEG2 video data is assumed to be a high-resolution, compressed video signal of, for example, 1440 pixels x 1080 pixels.

For example, when applying MPEG-2 video signals to digital broadcasting services, the signals are classified by profile (function) and level (resolution). For example, the Main Profile/High Level (MP@HL) video signals used in U.S. digital HDTV are high resolution, as mentioned above, and are converted into the DV video data mentioned above.

For this reason, the digital video signal conversion device shown in Figure 21 includes a signal conversion unit 140 for performing the above conversion process between the MPEG decoder 100 and DV encoder 110 shown in Figure 20.

The signal transform unit 140 performs resolution transformation on the DCT coefficients in the DCT transform domain from the MPEG decoder 100 using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the DCT encoding of the MPEG encoded data and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used in the IDCT encoding to obtain a time-domain signal transform output signal.

The DCT coefficients, which are the resolution converted output from the signal conversion unit 140, are supplied to the DV encoder 110.

The DV encoder 110 performs quantization and variable-length coding on the DCT coefficients of the resolution-converted output, frames them, and then outputs the DV video data as a bit stream.

In this manner, this digital video signal conversion device converts the resolution of the Main Profile/High Level (MP@HL) video signal within the MPEG2 video signal in the signal conversion unit 140, and then encodes it in the DV encoder to generate DV video data.

At this time, as with the digital video signal conversion device shown in FIG. 20, both IDCT and DCT processing was conventionally required for I-pictures, but in the digital video signal conversion device of the tenth embodiment, only IDCT is performed for reference purposes.

For P pictures, DCT and IDCT for reference are performed, which is the same as before. However, for B pictures, while both DCT and IDCT were previously required, only DCT is performed, eliminating the need for IDCT for reference.

In other words, the digital video signal conversion device shown in FIG. 21 can also significantly reduce the amount of data calculation processing required to convert the format from high-resolution MPEG-2 video data to DV video data.

Although the resolution conversion process by the signal conversion unit 140 has been described as being primarily for reducing the image size, it is also possible to perform resolution conversion for enlarging the image size.

In other words, by adding high-frequency components to an input digital signal in the frequency domain, it is generally possible to increase the resolution by any desired factor. For example, this is the case when converting MPEG-1 video data to the above-mentioned DV video data format.

The above processing may also be performed by software.

The compression methods used in the MPEG and DV formats mentioned above use a hybrid compression coding method that combines orthogonal transform coding and predictive coding to efficiently compress and code still and moving image data.

However, when an input information signal compressed and coded using a hybrid compression coding method is subjected to resolution conversion, and then subjected to orthogonal transformation and predictive coding with motion compensation again, motion vectors must also be estimated in the process of performing the re-predictive coding.

If re-predictive coding were performed at the exact same resolution without any resolution conversion, the motion vectors used during predictive coding could be used. However, converting the resolution changes the conversion distortion, which in turn changes the motion vectors used in the re-predictive coding process.

Therefore, in the re-predictive coding process, it is necessary to estimate a motion vector, but the amount of calculation required for estimating this motion vector is extremely large.

The solution to this problem is the digital signal conversion device according to the eleventh embodiment. The digital signal conversion method and device according to the eleventh embodiment performs signal conversion processing, such as resolution conversion, in the time domain or the orthogonal transform domain on an input information signal that has been compression-coded using hybrid compression coding, which combines orthogonal transform coding and predictive coding, and then returns the signal to the orthogonal transform domain or re-compresses and encodes it in the orthogonal transform domain.

Specific examples of the hybrid compression coding include H.261 and H.263 recommended by the International Telecommunication Union-Telecommunication Standardization Sector (ITU-T), as well as coding standards such as MPEG and DV.

H.261 is a video coding standard for low bit rates, developed primarily for ISDN videoconferencing and videophones. H.263 is an improved version of H.261 for GSTN videophone systems.

An eleventh embodiment will now be described with reference to FIG. 22. This embodiment is a digital video signal conversion device that receives MPEG-formatted encoded data, performs resolution conversion on the MPEG-encoded data as a signal conversion process, and then outputs the resulting resolution-converted MPEG-encoded data.

As shown in FIG. 22, this digital video signal conversion device comprises a decoding unit 210 that performs motion compensation (MC) decoding on a bitstream of MPEG-encoded data that has been compression-encoded with motion vector (mv) detection; a resolution conversion unit 160 that performs resolution conversion on the decoded output from the decoding unit 210; and an encoding unit 220 that performs compression encoding on the converted output image from the resolution conversion unit 160 with motion detection based on the motion vectors (mv) added to the MPEG-encoded data, and outputs a bitstream of resolution-converted video-encoded data.

The following description will be made of a digital video signal conversion device configured from these components, but it goes without saying that each component performs the respective steps of the digital signal conversion method according to the present invention.

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

The VLD unit 112 decodes the MPEG-encoded data, i.e., the coded data in which the motion vectors and quantized DCT coefficients (which are additional information) are variable-length coded, according to the variable-length coding, and extracts the motion vectors mv. The IQ unit 113 performs inverse quantization by multiplying the quantized DCT coefficients decoded by the VLD unit 112 by the quantization step used in the coding process, thereby obtaining DCT coefficients.

The IDCT unit 150 performs an inverse DCT on the DCT coefficients from the IQ unit 113 to convert the DCT coefficients back into spatial domain data, i.e., pixel data. Specifically, the inverse DCT calculates pixel values (luminance Y, color differences Cr, Cb) for each 8x8 pixel block. Note that while the pixel values used here are the actual pixel values for I-pictures, they are the difference values between corresponding pixel values for P-pictures and B-pictures.

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

The adder 151 adds the motion compensation output from the MC unit 152 to the difference value from the IDCT unit 150 and outputs a decoded image signal. The resolution conversion unit 160 performs the required resolution conversion processing on the decoded image signal. The converted output from the resolution conversion unit 160 is supplied to the encoding unit 220.

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

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

The ME unit 172 uses the scale conversion information from the scale conversion unit 171 to search a narrow range of the converted output from the resolution conversion unit 160, thereby estimating the optimal motion vector at the converted resolution.

The motion vectors estimated by the ME unit 172 are used during motion compensation by the MC unit 174. The converted output image from the resolution conversion unit 160 used when the ME unit 172 estimated the motion vectors is supplied to an adder 173.

The adder 173 takes the difference between the reference image (described later) and the converted output from the resolution conversion unit 160, and supplies it to the DCT unit 175.

The DCT unit 175 performs a discrete cosine transform on the difference between the reference image obtained by motion compensation in the MC unit 174 and the above-mentioned converted output image, in 8x8 block sizes. Note that for I pictures, since intraframe coding is used, the DCT operation is performed directly without taking the difference between frames.

The quantization (Q) unit 176 quantizes the DCT coefficients from the DCT unit 175 using a matrix table that takes visual characteristics into consideration. The VLC unit 177 compresses the quantized DCT coefficients from the Q unit 176 using variable length coding.

The buffer memory 178 is a memory for maintaining a constant transfer rate for the coded data compressed by the VLC unit 177 using variable-length coding. The resolution-converted coded video data is output from the buffer memory 178 as a bitstream at a constant transfer rate.

The rate control unit 183 controls the increase or decrease in the amount of information generated in the Q unit 176, i.e., the quantization step, based on change information regarding the increase or decrease in buffer capacity in the buffer memory 178.

The IQ unit 179, together with the IDCT unit 180, forms a local decoding unit. It inversely quantizes the quantized DCT coefficients from the Q unit 176 and supplies the DCT coefficients to the IDCT unit 180. The IDCT unit 180 performs an inverse DCT transform on the DCT coefficients from the IQ unit 179, converting them back into pixel data and supplying it to the adder 181.

The adder 181 adds the motion compensation output from the MC unit 174 to the pixel data, which is the inverse DCT output from the IDCT unit 180. The image information that is the sum output from the adder 181 is supplied to the FM unit 182. The image information stored in this FM unit 182 is subjected to motion compensation processing by the MC unit 174.

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

As described above, the adder 173 takes the difference between the converted output image from the resolution conversion unit 160 and the reference image and supplies it to the DCT unit 175.

The DCT unit 175, Q unit 176, VLC unit 177, and buffer memory 178 operate as described above, and ultimately, the resolution-converted video encoded data is output from this digital video signal conversion device as a bitstream at a constant transfer rate.

In this digital video signal conversion device, when the ME unit 172 of the encoding unit 220 estimates motion vectors, rather than estimating them from a completely blank state, the motion vectors attached to the macroblocks of the original compressed video signal are scaled by the scale conversion unit 171 in accordance with the resolution conversion rate of the resolution conversion unit 160, and then a motion vector for motion compensation is estimated by searching a narrow range of the converted output image from the resolution conversion unit 160 based on the scale conversion information from the scale conversion unit 171. This significantly reduces the amount of calculation in the ME unit 172, thereby achieving a smaller device and shorter conversion processing time.

Next, a twelfth embodiment will be described. This embodiment is also a digital video signal conversion device that performs resolution conversion processing on an MPEG video signal and outputs the resulting signal.

As shown in FIG. 23, this digital video signal conversion device comprises a decoding unit 211 that performs only predictive decoding using MC on the hybrid-encoded MPEG-encoded data to obtain decoded data in the orthogonal transform domain that has been orthogonally transformed; a resolution conversion unit 260 that performs resolution conversion on the decoded data in the orthogonal transform domain from the decoding unit 211; and an encoding unit 221 that performs compression encoding with motion compensation on the converted output from the resolution conversion unit 260 using motion estimation based on motion vector information of the MPEG-encoded data.

The following description will be directed to a digital video signal conversion device configured from these components, and it goes without saying that each component performs the respective steps of the digital signal conversion method according to the present invention.

Compared to the device shown in FIG. 22, this digital video signal conversion device eliminates the need for IDCT unit 150 in decoding unit 210 and DCT unit 175 and IDCT unit 180 in encoding unit 220. In other words, this digital video signal conversion device performs resolution conversion on decoded data still in the DCT domain and encodes the converted output.

Orthogonal transforms such as DCT and their inverse transforms generally require a large amount of computation. This can make it difficult to efficiently convert the resolution described above. Furthermore, the increased computational complexity can lead to accumulated errors, which can result in signal degradation.

Therefore, the digital video signal conversion device shown in FIG. 23 omits the IDCT unit 150, DCT unit 175, and IDCT unit 150 shown in FIG. 22, and further changes the function of the resolution conversion unit 160.

Furthermore, in the DCT domain, a resolution (to be described later) is calculated from the converted DCT coefficients from the resolution conversion unit 160, and a resolution calculation unit 200 is used in place of the scale conversion unit 171 shown in FIG. 22 to estimate motion vectors using this resolution.

The resolution conversion unit 260 shown in FIG. 23 receives an addition output (DCT coefficients) obtained by adding, in an adder 251, the motion compensation output from the MC unit 252 to the DCT coefficients obtained by inverse quantizing, in an IQ unit 213, the quantized DCT coefficients decoded by the VLD unit 212.

The resolution conversion unit 260 performs resolution conversion processing on the DCT coefficients in the DCT transform domain from the decoding unit 211 using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the DCT encoding of the MPEG-encoded data and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used in the IDCT encoding to obtain a signal transform output signal in the time domain.

The DCT coefficients, which are the resolution conversion output from the resolution conversion unit 260, are supplied to the definition calculation unit 200. The definition calculation unit 200 calculates the spatial definition (activity) for each macroblock from the luminance components of the DCT coefficients from the resolution conversion unit 260. Specifically, the maximum AC value of the DCT coefficients is used to calculate the image characteristics. For example, a low number of high-frequency components indicates a flat image.

The ME unit 272 estimates an optimal motion vector at the converted resolution based on the resolution calculated by the resolution calculation unit 200. That is, the ME unit 272 converts the motion vector mv extracted by the VLD 212 based on the resolution calculated by the resolution calculation unit 200, estimates a motion vector mv, and supplies this estimated motion vector mv to the ME unit 272. Here, the ME unit 272 estimates the motion vector while keeping the orthogonal transform domain. ME in this orthogonal transform domain will be described later.

The resolution conversion DCT coefficients from the resolution conversion unit 260 are supplied to the adder 273 via the definition calculation unit 200 and the ME unit 272.

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

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

In addition, the rate control unit 283 controls the increase/decrease in the amount of information generated in the Q unit 276, i.e., the quantization step, based on the resolution information from the resolution calculation unit 200 and information on changes in the buffer capacity of the buffer memory 278.

The VLC unit 277 compresses and encodes the quantized DCT coefficients from the Q unit 276 using variable-length coding and supplies the compressed data to the buffer memory 278. The buffer memory 278 maintains a constant transfer rate for the encoded data compressed using variable-length coding by the VLC unit 277 and outputs the resolution-converted video encoded data as a bitstream at a constant transfer rate.

The IQ unit 279 inverse-quantizes the quantized DCT coefficients from the Q unit 276 and supplies the DCT coefficients to an adder 281. The adder 281 adds the motion compensation output from the MC unit 274 to the DCT coefficients, which are the inverse Q output from the IQ unit 279. The DCT coefficient information, which is the sum output from the adder 281, is supplied to an FM unit 282. The DCT coefficient information stored in the FM unit 282 is subjected to motion compensation processing in the MC unit 274.

The MC unit 274 performs motion compensation on the DCT coefficient information stored in the FM unit 282 using the optimal motion vector estimated by the ME unit 272, and supplies the motion compensation output, which becomes the reference DCT coefficient, to the adder 281.

As described above, the adder 273 takes the difference between the converted DCT coefficients from the resolution conversion unit 260 and the reference DCT coefficients and supplies it to the Q unit 276.

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

Here, the MC unit 274 uses the optimal motion vector estimated by the ME unit 272 and the reference DCT coefficients stored in the FM 282 to perform motion compensation in the orthogonal transform domain, just like the ME unit 272.

ME and MC in the orthogonal transform domain will be explained with reference to Figures 24 to 26. In Figure 24, solid lines represent macroblocks in image A to be compressed, and dotted lines represent macroblocks in reference image B. When image A to be compressed and reference image B are superimposed using motion vectors as shown in Figure 24, the macroblock boundaries may not coincide. In the case of Figure 24, macroblock B' currently being compressed straddles four macroblocks _B1 , _B2 , _B3 , and _B4 in reference image B. Therefore, there is no macroblock in reference image B that corresponds directly to macroblock B', and it is therefore impossible to obtain the DCT coefficients of reference image B where macroblock B' is located. Therefore, it is necessary to obtain the DCT coefficients of reference image B for the portion where macroblock B' is located by transforming the DCT coefficients of the four macroblocks in reference image B that are straddled by macroblock B'.

Figure 25 shows a schematic diagram of the procedure for this conversion process. The lower left portion of macroblock _B1 in reference image B overlaps with macroblock B', but the upper right portion overlaps from the perspective of macroblock B'. Therefore, macroblock _B13 is generated by converting the DCT coefficients of macroblock _B1 , as described below. Similarly, the lower right portion of macroblock _B2 in reference image B overlaps with macroblock B', but the upper left portion overlaps from the perspective of macroblock B'. Therefore, macroblock _B24 is generated by converting the DCT coefficients of macroblock _B2 , as described below. Similar processing is performed on macroblocks _B3 and _B4 to generate macroblocks _B31 and _B42 . By combining the four macroblocks _B13 , _B24 , _B31 , and _B42 generated in this way, the DCT coefficients of reference image B for the portion where macroblock B' is located can be obtained.

That is, it can be expressed as the following equations (6) and (7).

B' = _B13 + _B24 + _B31 + _B42 (6) DCT(B') = DCT( _B13 ) + DCT( _B24 ) + DCT( _B31 ) + DCT( _B42 ) (7) Next, the conversion of DCT coefficients of a macroblock will be explained with reference to Figure 26. Figure 26 shows a mathematical model for calculating a partial block _B42 from an original block, such as _B4 in the spatial domain. Specifically, _B4 on the upper left side is extracted, interpolated with 0, and moved to the lower right side. This shows _B42 obtained from block _B4 by calculation using the following equation (8).

Here, _Ih and _Iw are the identification codes of the matrices of size h x h and w x w, respectively, extracted from block B4, each consisting of h and w columns and rows. As shown in Figure 26, the pre-matrix _H1 , which is first combined with _B4 , extracts the first h columns and transforms them to the bottom, and _H2 , which is later combined with _B4 , extracts the first w rows and transforms them to the right.

Based on the above equation (8), the DCT of B ₄₂ can be calculated directly from the DCT of B ₄ using the following equation (9).

DCT(B ₄₂ )=DCT(H ₁ )×DCT(B ₄ )×DCT(H ₂ ) (9) Applying this to all sub-blocks and summing them up, the DCT coefficients of the new block B' can be obtained directly from the DCTs of the original blocks B ₁ to B ₄ , as shown in the following equation (10).

Here, the DCT of H _i1 and H _i2 may be calculated in advance and stored in a memory to form a table memory. In this way, ME and MC are possible even in the orthogonal transform domain.

In the encoding unit 221, when the ME unit 272 estimates motion vectors, it does not estimate them from a completely empty state, but rather estimates the motion vectors attached to the macroblocks of the original compressed video signal by searching a narrow range based on the resolution calculated by the resolution calculation unit 200 from the converted output of the resolution conversion unit 260.

As described above, the decoder 211 of this other embodiment of the digital video signal converter performs predictive decoding with motion compensation on MPEG-encoded data that has undergone hybrid coding, which combines motion estimation predictive coding and orthogonal transform coding. This involves performing IQ after VLD, where motion compensation is performed to obtain decoded data in the DCT domain. This DCT-domain decoded data is then subjected to resolution conversion. This allows resolution conversion to be performed directly in the orthogonally transformed domain, eliminating the need for decoding to the temporal or spatial domain (inverse orthogonal transform). This simplifies calculations and enables high-quality conversion with minimal calculation error. Furthermore, when the encoder 221 estimates motion vectors in the ME unit 272, the motion vectors are estimated by searching for motion vectors associated with macroblocks in the original compressed video signal within a narrow range in accordance with the resolution calculated from the resolution conversion output, rather than estimating them from scratch. This allows a significant reduction in the amount of calculations required in ME unit 272, thereby enabling the device to be made smaller and the conversion process to be shortened.

Next, a thirteenth embodiment will be described. This example is also a digital video signal conversion device that performs signal conversion processing such as resolution conversion processing on MPEG encoded data and outputs the resulting video encoded data.

As shown in FIG. 27, this digital video signal conversion device comprises a decoding unit 340 that partially decodes the hybrid-encoded MPEG-encoded data to obtain data in the orthogonal transform domain, a conversion unit 343 that performs resolution conversion on the orthogonal transform domain data from the decoding unit 340, and an encoding unit 350 that adds motion vectors based on the motion vector information of the MPEG-encoded data to the converted output from the conversion unit 343 and performs compression encoding on the converted output.

The decoding unit 340 comprises a VLD unit 341 and an IQ unit 342. The VLD unit 341 and IQ unit 342 have the same configuration as the VLD unit 112 and IQ unit 113 shown in FIG. 21 above, and operate in the same manner. A distinctive feature of this decoding unit 340 is that it does not perform MC.

That is, P and B pictures do not undergo MC, and the DCT coefficients, which are the difference information, undergo resolution conversion in the conversion unit 343. The converted DCT coefficients obtained through resolution conversion are quantized by the Q unit 345, whose rate is controlled by the rate control unit 348, and then variable-length decoded by the VLC unit 346, after which they are output at a constant rate in the buffer memory 347.

At this time, the motion vector conversion unit 344 of the encoding unit 350 rescales the motion vector mv extracted by the VLD unit 341 according to the resolution conversion rate and supplies it to the VLC unit 346.

The VLC unit 346 adds the rescaled motion vector mv to the quantized DCT coefficients from the Q unit 345, performs variable length coding on the resulting data, and supplies the coded data to a buffer memory 347.

As described above, the digital video signal conversion device shown in FIG. 27 does not perform MC in the decoding unit 340 and the encoding unit 350, thereby simplifying calculations and reducing the hardware load.

Rate conversion may be performed using any of the digital video signal conversion devices described above. This may be applied when converting the transfer rate from 4 Mbps to 2 Mbps while keeping the resolution the same.

Although each of the above embodiments has been described as a device configuration, each of the above devices may be configured by using the digital signal conversion method according to the present invention as software.

According to the present invention, an input information signal that has been compression-coded with motion estimation is decoded with motion compensation, the decoded signal is subjected to signal conversion processing, and the converted signal is compression-coded with motion estimation based on motion vector information of the input information signal. When resolution conversion processing is applied as the signal conversion processing, the converted signal is subjected to compression-coding with motion compensation based on information obtained by scaling the motion vector information in accordance with the resolution conversion processing. In particular, the motion vector information required for compression coding is scaled in accordance with the resolution conversion rate and estimated by searching within a narrow range, thereby significantly reducing the amount of calculation required for motion vector estimation, thereby achieving a smaller device and a shorter conversion processing time.

The present invention also involves partially decoding an input information signal that has undergone compression coding, including predictive coding with motion estimation and orthogonal transform coding, to obtain a decoded signal in the orthogonal transform domain. Then, signal transformation is performed on the decoded signal in the orthogonal transform domain. This transformed signal is then subjected to compression coding with motion compensation prediction using motion estimation based on motion vector information of the input information signal. When resolution conversion is applied as the signal transformation process, the transformed signal is subjected to compression coding with motion compensation based on information obtained by converting the motion vector information according to the resolution obtained from the resolution conversion process. This allows for narrow search and estimation of the motion vector information required for compression coding, significantly reducing the amount of calculation required and enabling a more compact device and shorter conversion processing time. Furthermore, signal transformation processing in the orthogonal transform domain eliminates the need for inverse orthogonal transform processing and decoding to the time domain or spatial domain (inverse orthogonal transform). This simplifies calculations and enables high-quality conversion with minimal calculation error.

Furthermore, the present invention performs a partial decoding process on an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding to obtain a decoded signal in the orthogonal transform domain, performs a signal transform process on the decoded signal in the orthogonal transform domain, and performs a compression coding process on the transformed signal by adding motion vector information converted based on the motion vector information of the input information signal. Therefore, when a resolution conversion process is applied as the signal transform process, the transformed signal is subjected to a compression coding process to which information obtained by scaling the motion vector information in accordance with the resolution conversion process is added.

In other words, since the motion vector information added during compression encoding can be estimated by searching a narrow range, the amount of calculation required for motion vector estimation can be significantly reduced. Furthermore, since signal transformation processing can be performed in the orthogonal transform domain, inverse orthogonal transform processing is unnecessary. Furthermore, since motion compensation processing is not used during decoding or encoding, further reductions in calculation amount are possible.

─────────────────────────────────────────────────────── Continued from the front page (31) Priority Number: Patent Application No. Hei 9-305959 (32) Priority Date: November 7, 1997 (November 7, 1997) (33) Priority Country: Japan (JP) (31) Priority Number: Patent Application No. Hei 9-310721 (32) Priority Date: November 12, 1997 (November 12, 1997) (33) Priority Country: Japan (JP) (31) Priority Number: Patent Application No. Hei 9-314078 (32) Priority Date: November 14, 1997 (November 14, 1997) (33) Priority Country: Japan (JP) (81) Designated States EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE), AU, CN, JP, KR, US (Note) This publication is based on the publication published internationally by the International Bureau of Patents (WIPO). The effect of the international publication of the Japanese patent application (Japanese utility model registration application) related to this publication arises pursuant to Article 184-10, Paragraph 1 of the Patent Act (Article 48-13, Paragraph 2 of the Utility Model Act) and is unrelated to this publication.

Claims (76)

[Claims] 1. A digital signal conversion method comprising: a data extraction step of extracting a portion of orthogonal transform coefficients from each block of a digital signal in a first format, the data extraction step constructing partial blocks by extracting a portion of the orthogonal transform coefficients from each block of a digital signal in a first format, the partial blocks being orthogonal transform coefficient blocks of a predetermined unit; an inverse orthogonal transform step of inverse orthogonally transforming the orthogonal transform coefficients of each partial block on a partial block basis; a partial block concatenation step of concatenating the inverse orthogonally transformed partial blocks to construct a new block of the predetermined unit; and an orthogonal transform step of orthogonally transforming the new block on a block basis to generate a digital signal in a second format, the new block consisting of a new block of orthogonal transform coefficients of the predetermined unit. 2. The digital signal conversion method according to claim 1, wherein the orthogonal transform is a discrete cosine transform, the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate using a variable-length code, and the digital signal of the second format is a video signal compressed and encoded at a variable rate. 3. The digital signal conversion method according to claim 1, wherein the data extraction step extracts low-frequency discrete cosine transform coefficients from each block of the digital signal in the first format, thereby reducing the number of discrete cosine transform coefficients for the horizontal component of the luminance signal, the number of discrete cosine transform coefficients for the horizontal component of the color difference signal, and the number of discrete cosine transform coefficients for the vertical component. 4. The digital signal conversion method of claim 1, wherein, when one frame of the digital signal in the first format is composed of two fields, the data extraction step performs field separation on the discrete cosine transform coefficients of the vertical component of the luminance signal, separating the discrete cosine transform coefficients constituting the lines of the odd field of the frame from the discrete cosine transform coefficients constituting the lines of the even field of the frame to generate a block consisting of only the discrete cosine transform coefficients of one field. 5. The digital signal conversion method according to claim 1, wherein the digital signal of the first format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:1:1 ratio of the sampling frequency of the luminance signal to the sampling frequency of the color difference signal, and the digital signal of the second format is a compressed video signal with a resolution of 360 pixels x 240 pixels and a 4:2:0 ratio of the sampling frequency of the luminance signal to the sampling frequency of the color difference signal. 6. The digital signal conversion method of claim 1, wherein the digital signal of the first format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal, and the digital signal of the second format is a compressed video signal with a resolution of 360 pixels x 240 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal. 7. The digital signal conversion method according to claim 1, wherein the data extraction step extracts low-frequency orthogonal transform coefficients from each block of the digital signal in the first format, thereby reducing the number of discrete cosine transform coefficients of the vertical components of the color difference signal by half. 8. The digital signal conversion method of claim 7, wherein the digital signal of the first format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:1:1 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal, and the digital signal of the second format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal. 9. A digital signal conversion method comprising: an inverse orthogonal transform step of inversely orthogonally transforming a first-format digital signal consisting of orthogonal transform coefficient blocks of a predetermined unit on a block-by-block basis; a block division step of dividing each block of the inverse orthogonally transformed first-format digital signal; an orthogonal transform step of orthogonally transforming the orthogonal transform coefficients constituting each divided block on a block-by-block basis; and a data expansion step of interpolating the orthogonal transform coefficients to the values of each orthogonally transformed block to form a digital signal of a second format in the predetermined unit. 10. The digital signal conversion method according to claim 9, wherein the orthogonal transform is a discrete cosine transform, the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate using a variable-length code, and the digital signal of the second format is a video signal compressed and encoded at a variable rate. 11. The digital signal conversion method according to claim 9, wherein in the data expansion step, the orthogonal transform coefficients of each divided block of the digital signal in the first format are arranged on the low-frequency side, and zeros are interpolated on the high-frequency side, thereby constituting each block into the predetermined unit. 12. The digital signal conversion method according to claim 9, wherein the digital signal of the first format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:1:1 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal, and the digital signal of the second format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal. 13. The digital signal conversion method according to claim 9, wherein the digital signal of the first format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal, and the digital signal of the second format is a compressed video signal with a resolution of 720 pixels x 480 pixels and a 4:2:0 ratio between the sampling frequency of the luminance signal and the sampling frequency of the color difference signal. 14. A digital signal conversion device comprising: decoding means for decoding a digital signal of a first format consisting of orthogonal transform coefficient blocks of a predetermined unit; inverse quantization means for inverse quantizing the decoded digital signal; resolution conversion means for extracting a portion of the orthogonal transform coefficients from adjacent blocks of the orthogonal transform coefficient blocks of the dequantized digital signal to form partial blocks and convert the resolution; quantization means for quantizing the resolution-converted digital signal; and encoding means for encoding the quantized digital signal into a digital signal of a second format. 15. The digital signal conversion device according to claim 14, wherein the resolution conversion means concatenates the partial blocks obtained by the inverse orthogonal transformation to form new blocks of the predetermined unit. 16. The digital signal conversion device according to claim 14, wherein the orthogonal transform is a discrete cosine transform, the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate using a variable-length code, and the digital signal of the second format is a video signal compressed and encoded at a variable rate. 17. The digital signal conversion device according to claim 16, wherein the resolution conversion means extracts low-frequency discrete cosine transform coefficients from each block of the digital signal in the first format and halves the number of discrete cosine transform coefficients. 18. A digital signal conversion device comprising: decoding means for decoding a digital signal of a first format consisting of orthogonal transform coefficient blocks of a predetermined unit; inverse quantization means for inverse quantizing the decoded digital signal; resolution conversion means for interpolating orthogonal transform coefficients of a predetermined value into each block of the dequantized digital signal of the predetermined unit to convert the resolution of each block into the predetermined unit; quantization means for quantizing the resolution-converted digital signal; and encoding means for encoding the quantized digital signal into a digital signal of a second format. 19. The digital signal conversion device according to claim 18, wherein the resolution conversion means interpolates zeros into the high-frequency side of the orthogonal transform coefficients of each divided block of the digital signal in the first format, thereby configuring each block into the predetermined unit. 20. A digital signal conversion method for converting a digital signal in a first format consisting of orthogonal transform coefficient blocks of a predetermined unit into a digital signal in a second format consisting of new orthogonal transform coefficient blocks of another predetermined unit, the method comprising: controlling the data volume of the digital signal in the second format using data volume information contained in the digital signal in the first format. 21. The digital signal conversion method according to claim 20, wherein the orthogonal transform is a discrete cosine transform, the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate, and the digital signal of the second format is a video signal compressed and encoded at a variable rate. 22. The digital signal conversion method of claim 20, wherein control of the data amount of the digital signal in the second format is performed in the orthogonal transform domain. 23. The digital signal conversion method according to claim 20, wherein control of the data amount of the digital signal in the second format is performed in the spatial domain. 24. The digital signal conversion method according to claim 20, further comprising: a quantization parameter calculation step for calculating a quantization parameter for each of the predetermined blocks of the digital signal in the first format based on the quantizer number and class information; a metablock quantization parameter calculation step for calculating a quantization parameter for a metablock consisting of a plurality of the blocks by averaging the quantization parameters calculated for each of the blocks; and a quantization scale calculation step for calculating a quantization scale for the digital signal in the second format from the quantization parameter for each of the metablocks, wherein each of the blocks is quantized using the calculated quantization scale. 25. The digital signal conversion method of claim 20, further comprising: a total generated bit calculation step for calculating a total number of generated bits for each frame of the digital signal in the first format; and a quantization parameter adjustment step for adjusting the quantization parameter using a value obtained by multiplying the difference between the total generated bit number and a target number of bits by a constant; wherein the adjusted quantization parameter is used to calculate a new quantization scale, and the new quantization scale is used for the next frame of the digital signal in the second format. 26. A digital signal conversion device that converts a digital signal in a first format consisting of orthogonal transform coefficient blocks in a predetermined unit into a digital signal in a second format consisting of new orthogonal transform coefficient blocks in another predetermined unit, comprising: a decoding means for decoding the digital signal in the first format; a dequantization means for dequantizing the decoded digital signal; a signal conversion means for performing signal processing involving format conversion of the dequantized digital signal; a quantization means for quantizing the digital signal after the signal processing; a data amount control means for controlling the data amount in the quantization means; an encoding means for encoding the quantized digital signal whose data amount has been controlled by the data amount control means into a digital signal in the second format. 27. The digital signal conversion device according to claim 26, wherein the orthogonal transform is a discrete cosine transform, the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate, and the digital signal of the second format is a video signal compressed and encoded at a variable rate. 28. The digital signal conversion device according to claim 26, wherein the signal conversion means controls the data amount of the digital signal of the second format in the orthogonal transform domain using data amount information contained in the digital signal of the first format. 29. The digital signal conversion device according to claim 26, wherein the signal conversion means controls the data amount of the digital signal of the second format in the spatial domain by utilizing data amount information contained in the digital signal of the first format. 30. The digital signal conversion device according to claim 26, wherein the signal conversion means calculates a quantization parameter for each block included in the digital signal of the first format based on the quantizer number and class information, averages the quantization parameters calculated for each block to calculate a quantization parameter for a metablock consisting of a plurality of the blocks, calculates a quantization scale for the digital signal of the second format from the quantization parameter for each metablock, and quantizes each block using the calculated quantization scale. 31. A digital signal conversion method for converting a digital signal of a first format into a digital signal of a second format, comprising: a decoding step for decoding the digital signal of the first format; a signal conversion step for converting the decoded digital signal of the first format into a digital signal of the second format; an encoding step for encoding the digital signal of the second format; and a weighting step for simultaneously performing inverse weighting on the decoded digital signal of the first format and weighting on the digital signal of the second format. 32. The digital signal conversion method according to claim 31, wherein the digital signal of the first format is an orthogonally transformed digital signal, and the weighting process is performed in the orthogonal transform domain. 33. The digital signal conversion method according to claim 31, wherein the digital signal of the first format is an orthogonally transformed digital signal, and the weighting process is performed in the spatial domain after inverse orthogonally transforming the orthogonally transformed digital signal. 34. The digital signal conversion method of claim 31, wherein the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate by discrete cosine transform, and the digital signal of the second format is a video signal compressed and encoded at a variable rate by discrete cosine transform. 35. The digital signal conversion method according to claim 31, wherein the weighting process is performed after inverse quantization of the digital signal in the first format and before the signal conversion process. 36. The digital signal conversion method of claim 31, wherein the weighting process is performed after the signal conversion process and before quantization into the second format. 37. A digital signal conversion device that converts a digital signal of a first format into a digital signal of a second format, comprising: decoding means for decoding the digital signal of the first format; signal conversion means for converting the decoded digital signal of the first format into a digital signal of the second format; encoding means for encoding the second digital signal; and weighting means for simultaneously performing inverse weighting on the digital signal of the first format and weighting on the digital signal of the second format. 38. The digital signal conversion device according to claim 37, wherein the digital signal of the first format is an orthogonally transformed digital signal, and the weighting processing means performs inverse weighting on the digital signal of the first format and weighting on the digital signal of the second format in the orthogonal transform domain. 39. The digital signal conversion device according to claim 37, wherein the digital signal of the first format is an orthogonally transformed digital signal, and the weighting processing means performs inverse weighting on the digital signal of the first format and weighting on the digital signal of the second format in the spatial domain. 40. The digital signal conversion device according to claim 37, wherein the digital signal of the first format is a video signal compressed and encoded at a predetermined fixed rate by discrete cosine transform, and the digital signal of the second format is a video signal compressed and encoded at a variable rate by discrete cosine transform. 41. The digital signal conversion device according to claim 37, wherein the weighting processing means is arranged before the signal conversion means, and performs inverse weighting on the digital signal of the first format and weighting on the digital signal of the second format after inverse quantization of the digital signal of the first format. 42. The digital signal conversion device according to claim 37, wherein the weighting processing means is disposed after the signal conversion means, and performs inverse weighting on the digital signal of the first format and weighting on the digital signal of the second format before quantization of the digital signal of the second format. 43. A digital signal conversion method comprising: a decoding step of performing decoding with motion compensation on an input information signal that has been compression-coded with motion estimation; a signal conversion processing step of performing signal conversion on the decoded signal from the decoding step; and an encoding processing step of performing compression-coding on the converted signal from the signal conversion processing step with motion estimation based on motion vector information of the input information signal. 44. The digital signal conversion method according to claim 43, wherein the signal conversion processing step applies resolution conversion processing to the decoded signal. 45. The digital signal conversion method according to claim 44, wherein the encoding process step performs compression encoding on the converted signal based on information obtained by scaling the motion vector information in accordance with the resolution conversion process. 46. The digital signal conversion method according to claim 43, wherein the signal conversion processing step performs rate conversion processing on the decoded signal. 47. A digital signal conversion device comprising: decoding means for performing decoding with motion compensation on an input information signal that has been compression-coded with motion estimation; signal conversion processing means for performing signal conversion processing on the decoded signal from the decoding means; and encoding means for performing compression-coding processing with motion estimation on the converted signal from the signal conversion processing means based on motion vector information of the input information signal. 48. A digital signal conversion method comprising: a decoding step of performing only predictive decoding with motion compensation on an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding, thereby obtaining a decoded signal in the orthogonal transform domain that has been orthogonally transformed; a signal conversion processing step of performing signal conversion on the decoded signal in the orthogonal transform domain from the decoding step; and an encoding processing step of performing compression coding with motion compensation prediction on the converted signal from the signal conversion processing step, using motion estimation based on motion vector information of the input information signal. 49. The digital signal conversion method according to claim 48, wherein the signal conversion processing step performs signal conversion processing on the decoded signal in the orthogonal transform domain from the decoding step using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the orthogonal transform coding performed on the input information signal and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used to obtain the signal conversion output signal in the time domain. 50. The digital signal conversion method according to claim 48, wherein the signal conversion processing step applies resolution conversion processing to the decoded signal in the orthogonal transform domain from the decoding step. 51. The digital signal conversion method according to claim 50, wherein the encoding process step performs compression encoding on the converted signal based on information obtained by scaling the motion vector information in accordance with the resolution conversion process. 52. The digital signal conversion method according to claim 48, wherein the signal conversion processing step applies rate conversion processing to the decoded signal in the orthogonal transform domain from the decoding step. 53. A digital signal conversion device comprising: decoding means for performing only predictive decoding with motion compensation on an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding, thereby obtaining a decoded signal in the orthogonal transform domain that has been orthogonally transformed; signal conversion processing means for performing signal conversion processing on the decoded signal in the orthogonal transform domain from the decoding means; and encoding processing means for performing compression coding with motion compensation prediction on the converted signal from the signal conversion processing means, using motion estimation based on motion vector information of the input information signal. 54. A digital signal conversion method comprising: a decoding step of partially decoding an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding to obtain a signal in the orthogonal transform domain; a signal conversion step of performing signal conversion on the orthogonal transform domain signal from the decoding step; and an encoding step of adding motion vector information converted based on the motion vector information of the input information signal to the converted signal from the signal conversion step and performing compression coding on the converted signal. 55. The digital signal conversion method according to claim 54, wherein the signal conversion processing step performs signal conversion processing on the decoded signal in the orthogonal transform domain from the decoding step using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the orthogonal transform coding performed on the input information signal and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used to obtain the signal conversion output signal in the time domain. 56. The digital signal conversion method according to claim 54, wherein the signal conversion processing step applies resolution conversion processing to the decoded signal in the orthogonal transform domain from the decoding step. 57. The digital signal conversion method according to claim 56, wherein the encoding process step performs a compression encoding process on the converted signal, to which information obtained by scaling the motion vector information in accordance with the resolution conversion process is added. 58. The digital signal conversion method according to claim 54, wherein the signal conversion processing step applies rate conversion processing to the decoded signal in the orthogonal transform domain from the decoding step. 59. A digital signal conversion device comprising: a decoding means for partially decoding an input information signal that has been subjected to compression coding including predictive coding with motion estimation and orthogonal transform coding to produce a signal in the orthogonal transform domain; a signal conversion processing means for performing signal conversion processing on the orthogonal transform domain signal from the decoding means; and an encoding processing means for adding motion vector information converted based on the motion vector information of the input information signal to the converted signal from the signal conversion processing means and performing compression coding processing on the converted signal. 60. A digital signal conversion method for converting a digital signal of a first format, to which motion mode/still mode information has been added in advance, into a digital signal of a second format involving inter-frame differential encoding, comprising: a decoding step for decoding the digital signal of the first format; a signal conversion step for performing signal conversion processing on the decoded signal from the decoding step; a determination step for determining whether or not to perform inter-frame differential encoding on the converted signal from the signal conversion step for each predetermined block unit in accordance with the motion mode/still mode information; and an encoding step for encoding the converted signal from the conversion step based on the determination result from the determination step, and outputting a digital signal of the second format. 61. The digital signal conversion method of claim 60, wherein the decoding step partially decodes the digital signal in the first format to output a signal in the orthogonal transform domain, and the signal conversion step performs signal conversion processing on the signal in the orthogonal transform domain. 62. The digital signal conversion method according to claim 61, wherein the orthogonal transform is a discrete cosine transform. 63. The digital signal conversion method of claim 60, wherein the signal conversion step performs signal conversion processing on the digital signal in the first format using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the orthogonal transform coding performed on the digital signal in the first format and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used to obtain the digital signal in the second format. 64. The digital signal conversion method of claim 60, wherein the determination step determines whether or not to perform inter-frame differential encoding for each macroblock of the converted signal from the signal conversion step. 65. A digital signal conversion device that converts a digital signal of a first format, to which motion mode/still mode information has been added, into a digital signal of a second format that involves inter-frame differential encoding, comprising: decoding means for decoding the digital signal of the first format; signal conversion means for performing signal conversion processing on the decoded signal from the decoding means; determination means for determining whether to perform inter-frame differential encoding on each predetermined block of the converted signal from the signal conversion means in accordance with the motion mode/still mode information; and encoding means for encoding the converted signal from the signal conversion means based on the determination result from the determination means, and outputting the digital signal of the second format. 66. A digital signal conversion method for converting a digital signal of a first format into a digital signal of a second format involving encoding using inter-frame differences, comprising: a decoding step for partially decoding the digital signal of the first format to obtain a signal in the orthogonal transform domain; a signal conversion step for performing signal conversion on the signal in the orthogonal transform domain from the decoding step; a determination step for determining whether to apply inter-frame difference encoding to the converted signal from the signal conversion step for each predetermined block unit, based on the maximum absolute value of the inter-frame differences of the converted signal; and an encoding step for encoding the converted signal from the signal conversion step based on the determination result from the determination step, thereby outputting a digital signal of the second format. 67. The digital signal conversion method according to claim 66, wherein the orthogonal transform is a discrete cosine transform. 68. The digital signal conversion method according to claim 66, wherein the signal conversion step performs signal conversion processing on the digital signal in the first format using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the orthogonal transform coding performed on the digital signal in the first format and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used to obtain the digital signal in the second format. 69. The digital signal conversion method according to claim 66, wherein the determination step determines whether or not to apply inter-frame differential encoding to each macroblock of the converted signal from the signal conversion step. 70. A digital signal conversion device that converts a digital signal of a first format into a digital signal of a second format involving inter-frame differential encoding, comprising: decoding means that partially decodes the digital signal of the first format to obtain a signal in the orthogonal transform domain; signal conversion means that performs signal conversion processing on the signal in the orthogonal transform domain from the decoding means; determination means that determines whether to apply inter-frame differential encoding to the converted signal from the signal conversion means for each predetermined block unit based on the maximum absolute value of the inter-frame differential of the converted signal; and encoding means that encodes the converted signal from the signal conversion means based on the determination result of the determination means and outputs the digital signal of the second format. 71. A digital signal conversion method comprising: an inverse orthogonal transform step of performing an inverse orthogonal transform on an intraframe-coded signal and a forward-predictively coded signal of a first format, the digital signal comprising an intraframe-coded signal that has been intraframe-coded, and a forward-predictively coded signal and a bidirectionally predictively coded signal that have been interframe-coded in both forward and bidirectional predictive coding with motion estimation; a motion compensation output generation step of generating, based on the transform output from the inverse orthogonal transform step, a motion compensation output to be added to a partially decoded forward-predictively coded signal and a bidirectionally predictively coded signal; an orthogonal transform step of performing an orthogonal transform on the motion compensation output from the motion compensation output generation step; an addition step of adding the orthogonal transform output from the orthogonal transform step to the partially decoded forward-predictively coded signal and the bidirectionally predictively coded signal; and an encoding step of performing compression encoding on a signal based on the output from the addition step to output a digital signal of a second format. 72. The digital signal conversion method according to claim 71, further comprising a conversion step between the adding step and the encoding step for performing signal conversion processing on the addition output, and wherein the encoding step performs the compression coding processing on the converted signal from this conversion step. 73. The digital signal conversion method according to claim 71, wherein the orthogonal transform is a discrete cosine transform. 74. The digital signal conversion method according to claim 72, wherein the converting step performs signal conversion on the output from the adding step using a transform matrix generated based on an inverse orthogonal transform matrix corresponding to the orthogonal transform matrix used in the orthogonal transform coding performed on the digital signal in the first format and an orthogonal transform matrix corresponding to the inverse orthogonal transform matrix used to obtain the digital signal in the second format. 75. A digital signal conversion device comprising: inverse orthogonal transform means for performing inverse orthogonal transform on an intraframe-coded signal and a forward-predictively coded signal of a first format, the digital signal comprising an intraframe-coded signal that has been intraframe-coded, and a forward-predictively coded signal and a bidirectionally predictively coded signal that have been subjected to forward and bidirectional interframe predictive coding with motion detection; motion compensation output generation means for generating, based on the transform output from the inverse orthogonal transform means, a motion compensation output to be added to the partially decoded forward-predictively coded signal and the bidirectionally predictively coded signal; orthogonal transform means for performing an orthogonal transform on the motion compensation output from the motion compensation output generation means; addition means for adding the orthogonal transform output from the orthogonal transform means to the partially decoded forward-predictively coded signal and the bidirectionally predictively coded signal; and encoding means for performing compression encoding on a signal based on the output from the addition means to output a digital signal of a second format. 76. The digital signal conversion device according to claim 75, further comprising: a conversion means for performing signal conversion processing on the added output between the addition means and the encoding means; and the encoding means for performing the compression encoding processing on the converted signal from the conversion means.