CA2070827C - Device for coding digital signals correspsonding to television images and corresponding decoding device - Google Patents
Device for coding digital signals correspsonding to television images and corresponding decoding device Download PDFInfo
- Publication number
- CA2070827C CA2070827C CA002070827A CA2070827A CA2070827C CA 2070827 C CA2070827 C CA 2070827C CA 002070827 A CA002070827 A CA 002070827A CA 2070827 A CA2070827 A CA 2070827A CA 2070827 C CA2070827 C CA 2070827C
- Authority
- CA
- Canada
- Prior art keywords
- circuit
- stage
- signals
- output
- definition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/015—High-definition television systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/30—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
- H04N19/619—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding the transform being operated outside the prediction loop
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
A coding system for digital signals corresponding to televi-sion pictures of a determined definition, incorporating an image rebuilding stage comprising a first image rebuilding channel ac-cording to said determined definition and a second image rebuild.
ing channel according to a reduced definition, which is parallel on said first channel, and corresponding decoding system also incor-porating an image rebuilding stage giving said reduced definition from decoded signals, associated with a decoding stage, a motion compensating stage using additional motion information, and a prediction stage using the output signals of said image rebuilding stage. This system is for use on television receivers.
ing channel according to a reduced definition, which is parallel on said first channel, and corresponding decoding system also incor-porating an image rebuilding stage giving said reduced definition from decoded signals, associated with a decoding stage, a motion compensating stage using additional motion information, and a prediction stage using the output signals of said image rebuilding stage. This system is for use on television receivers.
Description
PHF 90-567C 1 ~ ~ ~ ~ $ ~ r~ 14.05.1992 Device for coding digital signals corresponding to television images and corresponding decoding device.
The invention relates to a device for coding digital signals corresponding to television images having a given definition, comprising:
(A) a stage for selecting the coding mode of signals to be coded from input current signals of the coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of the device, said stage supplying the signals to be coded and coding mode information;
(B) a coding stage;
(C) an image reconstruction stage in a first path for reconstructing the image in accordance with said given definition;
(D) a stage for estimating the motion between images, supplying motion information;
(E) a stage for predicting from output signals of said image reconstruction and motion estimation stages the output signals of said prediction stage constituting said predicted signals transmitted to said coding mode selection stage.
The invention also relates to a device for decoding coded digital signals previously transmitted and/or stored after treatment in a device for coding digital signals corresponding to television images of a given definition, said coding device comprising:
(a) a stage for selecting the coding mode of the signals to be coded from input current signals of said coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of said device, said preceding or .
current input signals being considered as blocks of a given size representing a subdivision of images and being treated independently;
(b) a coding stage;
(c) a stage for reconstructing the image in accordance with said given definition on the one hand and in accordance with a reduced definition on the other hand;
(d) a stage for estimating the motion between images;
PHF 90-567C 2 ~ ~ ~ ~ ~ ~ ~ 14.05.1992 (e) for supplying said predicted signals, a stage for predicting from the output signals of said motion estimation and image reconstruction stages the information components which are additional and relative to the estimated motion between images and to the selected coding mode being intended, like the coded signals, for transmission and/or storage.
This invention can be used to advantage in the field of television picture reception in accordance with two definition levels for ensuring the restitution of high-definition images of excellent quality which can nevertheless be received by conventional television receivers having a lower definition.
The transmission of high-definition digital television images is currently the subject of intensive research. However, the' industries concerned have realised at a very_early stage that this novel service would only be successful if the high-definition programs could be received by high-definition television receivers as well as by conventional receivers.
Such a transmission, which is referred to as "compatible", is effectively ensured if a fraction of the multitude of data corresponding to the high-definition program can easily be taken and used for the conventional receiver to supply the normal television images (compatible images: 625 lines, 50 Hz, 2:1, 16/9 frame). This technical solution provides the possibility of simultaneously transmitting conventional television programs and high-definition television programs with a certain economy of information output (the output thus economized is substantially that which would correspond to the transmission of TV images alone). ' The separation of the multitude of high-definition data into two parts corresponding to single compatible data (which will hereinafter be referred to as TV
information components) and to complementary, additional data (relating to the high-definition images and hereinafter referred to as HD or HDTV information components) constitutes a drawback at the level of the high-definition coder (or HD coder) upon transmission, because the entire multitude of data must comprise the information components with which compatible images can be subsequently reconstructed. The positioning of these information components may particularly lead to a degradation of the high-definition image quality.
The solution which currently seems to be the best to ensure this compatible transmission is based on the use of a coder employing an orthogonal 20'~082"~
PHF 90-567C 3 14.05.1992 transform such as a discrete cosine transform (DCT) which acts on each image which is divided into blocks. The reconstruction of the compatible images based on high-definition data is realised by means of a cut-up in the frequency domain: for each high-definition image block on which the orthogonal transform is realised only those coefficients resulting from this transform which correspond to the lowest frequencies are transmitted to the compatible image decoder (or TV decoder). The coefficients thus selected become the constituent coefficients of the new image blocks having dimensions which in this case are twice smaller than those of the preceding blocks in each horizontal and vertical direction.
The compatible images thus obtained keep their satisfactory quality as long as the TV decoders do not incorporate the motion compensation devices.
When such devices are provided, a prediction before the motion is necessary, which may be effected with the high-definition resolution for maintaining the image quality during decoding of the high-definition images. A certain disagreement is found between the contents of ti'~e blocks thus predicted during coding and high-definition decoding and of the blocks predicted during decoding compatible images is found. Effects degrading the quality of the compatible images are then produced, which effects are cumulative because of the recursivity of the device.
It is a first object of the invention to provide a device for coding digital signals corresponding to television images with which these drawbacks can be obviated.
To this end the invention relates to a coding device as defined in the opening paragraph and is characterized in that said digital signals are considered as blocks of a given size representing a subdivision of images and being treated independently and in that the image reconstruction stage also comprises a second image reconstruction path in accordance with a reduced definition, which path is situated parallel to said first path, and in that the prediction stage comprises n circuit for reconstituting the image by combining the output signals of said first and second reconstruction paths.
The document "Coding Television Signals at 320 and 64 kbit/s" G.
Kummerfeldt et al., Proceedings of the SPIE, December 1985, Cannes (France), vol.
594, Image Coding, pp. 119-128 describes, particularly with reference to Fig.
The invention relates to a device for coding digital signals corresponding to television images having a given definition, comprising:
(A) a stage for selecting the coding mode of signals to be coded from input current signals of the coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of the device, said stage supplying the signals to be coded and coding mode information;
(B) a coding stage;
(C) an image reconstruction stage in a first path for reconstructing the image in accordance with said given definition;
(D) a stage for estimating the motion between images, supplying motion information;
(E) a stage for predicting from output signals of said image reconstruction and motion estimation stages the output signals of said prediction stage constituting said predicted signals transmitted to said coding mode selection stage.
The invention also relates to a device for decoding coded digital signals previously transmitted and/or stored after treatment in a device for coding digital signals corresponding to television images of a given definition, said coding device comprising:
(a) a stage for selecting the coding mode of the signals to be coded from input current signals of said coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of said device, said preceding or .
current input signals being considered as blocks of a given size representing a subdivision of images and being treated independently;
(b) a coding stage;
(c) a stage for reconstructing the image in accordance with said given definition on the one hand and in accordance with a reduced definition on the other hand;
(d) a stage for estimating the motion between images;
PHF 90-567C 2 ~ ~ ~ ~ ~ ~ ~ 14.05.1992 (e) for supplying said predicted signals, a stage for predicting from the output signals of said motion estimation and image reconstruction stages the information components which are additional and relative to the estimated motion between images and to the selected coding mode being intended, like the coded signals, for transmission and/or storage.
This invention can be used to advantage in the field of television picture reception in accordance with two definition levels for ensuring the restitution of high-definition images of excellent quality which can nevertheless be received by conventional television receivers having a lower definition.
The transmission of high-definition digital television images is currently the subject of intensive research. However, the' industries concerned have realised at a very_early stage that this novel service would only be successful if the high-definition programs could be received by high-definition television receivers as well as by conventional receivers.
Such a transmission, which is referred to as "compatible", is effectively ensured if a fraction of the multitude of data corresponding to the high-definition program can easily be taken and used for the conventional receiver to supply the normal television images (compatible images: 625 lines, 50 Hz, 2:1, 16/9 frame). This technical solution provides the possibility of simultaneously transmitting conventional television programs and high-definition television programs with a certain economy of information output (the output thus economized is substantially that which would correspond to the transmission of TV images alone). ' The separation of the multitude of high-definition data into two parts corresponding to single compatible data (which will hereinafter be referred to as TV
information components) and to complementary, additional data (relating to the high-definition images and hereinafter referred to as HD or HDTV information components) constitutes a drawback at the level of the high-definition coder (or HD coder) upon transmission, because the entire multitude of data must comprise the information components with which compatible images can be subsequently reconstructed. The positioning of these information components may particularly lead to a degradation of the high-definition image quality.
The solution which currently seems to be the best to ensure this compatible transmission is based on the use of a coder employing an orthogonal 20'~082"~
PHF 90-567C 3 14.05.1992 transform such as a discrete cosine transform (DCT) which acts on each image which is divided into blocks. The reconstruction of the compatible images based on high-definition data is realised by means of a cut-up in the frequency domain: for each high-definition image block on which the orthogonal transform is realised only those coefficients resulting from this transform which correspond to the lowest frequencies are transmitted to the compatible image decoder (or TV decoder). The coefficients thus selected become the constituent coefficients of the new image blocks having dimensions which in this case are twice smaller than those of the preceding blocks in each horizontal and vertical direction.
The compatible images thus obtained keep their satisfactory quality as long as the TV decoders do not incorporate the motion compensation devices.
When such devices are provided, a prediction before the motion is necessary, which may be effected with the high-definition resolution for maintaining the image quality during decoding of the high-definition images. A certain disagreement is found between the contents of ti'~e blocks thus predicted during coding and high-definition decoding and of the blocks predicted during decoding compatible images is found. Effects degrading the quality of the compatible images are then produced, which effects are cumulative because of the recursivity of the device.
It is a first object of the invention to provide a device for coding digital signals corresponding to television images with which these drawbacks can be obviated.
To this end the invention relates to a coding device as defined in the opening paragraph and is characterized in that said digital signals are considered as blocks of a given size representing a subdivision of images and being treated independently and in that the image reconstruction stage also comprises a second image reconstruction path in accordance with a reduced definition, which path is situated parallel to said first path, and in that the prediction stage comprises n circuit for reconstituting the image by combining the output signals of said first and second reconstruction paths.
The document "Coding Television Signals at 320 and 64 kbit/s" G.
Kummerfeldt et al., Proceedings of the SPIE, December 1985, Cannes (France), vol.
594, Image Coding, pp. 119-128 describes, particularly with reference to Fig.
2, a coding device (of variable length in this case) for reconstruction of the image and prediction, based on said reconstruction, of the signals to be coded. However, such a 2fl~082'~
PHF 90-567C 4 14.05.1992 device completely disregards the problem in the case of transmitting images having two definition levels, viz. the problem of supplying the information components which are simultaneously necessary for transmitting and reconstituting the television images at the receiver end in accordance with these two definition levels, notably high-definition television images and compatible television images. In contrast thereto, the structure proposed in this Application provides a satisfactory technical solution to this problem.
In a particular embodiment, in which essentially signals corresponding to pixels are used, the invention relates to a coding device in which the coding stage is a variable length coding chain comprising an orthogonal transform circuit, a scanning conversion circuit, a quantizing circuit, a variable length coding circuit, a memory circuit and a rate control circuit and in which the first path of the reconstruction stage comprises a series arrangement of:
(a) an inverse quantizing circuit;
(b) an inverse scanning conversion circuit;
(c) an inverse orthogonal transform stage;
(d) a circuit for reconstructing the signal in accordance with said given definition, as would be decoded after transmission, but for transmission errors;
characterized in that in said image reconstruction stage the second reconstruction path compnses:
(a) said inverse quantizing circuit;
(b) said inverse scanning conversion circuit;
(c) a clipper circuit for taking a given fraction of the signals at the output of said inverse quantizing circuit;
(d) an inverse orthogonal transform circuit;
(e) a circuit for reconstructing the signal having a reduced definition as would be decoded after transmission, but for transmission errors;
(f) a memory for storing said reconstructed signal; (g) a motion compensation circuit between the output of said memory and said circuit for reconstructing the reduced definition signal;
and in that said image reconstituting circuit comprises:
(h) a multiplier for multiplying the output signal of the first reconstruction path by a coefficient a between 0 and 1;
(i) a multiplier for multiplying the output signal of the second PHF 90-567C 5 ~ ~ ~ o $ 2 ~ 14.05.1992 reconstruction path by the coefficient (1-a), a phase restoring and resampling circuit being arranged between said output of the second reconstruction path and the corresponding input of the associated multiplier;
(j) an adder for adding the output signals of said multipliers;
(k) a weighting memory for storing the output signal of said adder;
(1) a prediction circuit for receiving the output signal of said weighting memory and the motion information components supplied by said motion estimation stage.
In a second embodiment, whose operation is essentially based on coefficients resulting from an orthogonal transform of said signals corresponding to pixels, the invention relates particularly to a coding device in which the coding stage is a variable length coding chain comprising a quantizing circuit, a variable length coding circuit, a memory circuit and a rate control circuit, in which the stage for selecting the coding mode comprises a series arrangement of:
(a) an orthogonal transform stage for said input current signals;
(b) an inter/intradecision circuit for also receiving said predicted signals;
and in which the first reconstruction path comprises a series arrangement of:
(c) an inverse quantizing circuit;
(d) a circuit for reconstituting the block in accordance with said given , definition;
(e) an inverse orthogonal transform stage;
(f) a memory for the information components of said given definition;
characterized in that in said reconstruction stage the second reconstruction path comprises:
(a) said inverse quantizing circuit;
(b) a clipper circuit for taking a given fraction of the signals present at the output of said inverse quantizing circuit;
(c) a circuit for reconstituting the block in accordance with said reduced definition;
(d) an inverse orthogonal transform stage;
(e) a memory for the information components of said reduced definition;
and in that the image reconstitution circuit comprises:
(f) at the output of said first path, a first reconstitution branch comprising 2070~2'~
PHF 90-567C fi 14.05.1992 a a first prediction circuit with motion compensation, a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the second path, a second reconstitution branch comprising a second prediction circuit with motion compensation and a second orthogonal transform circuit;
(h) a circuit for combining the output signals of said first and second branches for supplying said predicted signals transmitted to said coding mode selection stage.
In a perfected modification of this second embodiment, the coding device is characterized in that the image reconstitution circuit for the prediction stage comprises, between the output of said second reconstitution branch and the corresponding input of said combination circuit, a weighted mixing circuit comprising: , (a) a multiplier for multiplying the output signal of the first branch by a coefficient a between 0'and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit ;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient (1-a);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to the corresponding input of the combination circuit.
It is another object of the invention to provide a device for decoding digital signals corresponding to television images with which reduced definition images and high-definition images of good quality can be reconstituted.
To this end the invention relates to a decoding device as defined in the opening paragraph and is characterized in that it comprises, in association with a decoding stage:
(A) a stage for reconstructing the image in accordance with said reduced definition from the decoded signals;
(B) a stage for motion compensation based on said additional information components;
(C) a stage for prediction based on the output signals of said image reconstruction stage.
PHF 90-567C 7 2 ~ ~ ~ ~ ~ ~ 14.05.1992 In a particular embodiment the invention relates to a decoding device in which the decoding stage is a variable length decoding chain for coded digital signals previously transmitted and/or stored after treatment in a variable length coding chain, said decoding chain comprising a memory circuit, a variable length decoding circuit, an inverse quantizing circuit, an inverse normalization circuit, an inverse scanning conversion circuit and an inverse orthogonal transform circuit, and is characterized in that:
(A) the stage for reconstructing the image in accordance with said reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) an inverse orthogonal transform circuit;
(c) an adder whose first input receives the output signal of said inverse orthogonal transform circuit;
(d) a memory for storing the image reconstituted in accordance with the reduced definition and being present at the output of said adder;
(e) a first circuit for compensating motion, receiving the output signal of said memory and the motion information components, and having its output connected to the second input of said adder;
(B) the prediction stage comprises:
(f) a first multiplier for multiplying the output signals of said decoding device by a coefficient a between 0 and 1;
(g) a second multiplier for multiplying the output signal of said stage for reconstructing the image in accordance with said reduced definition by the coefficient (1-a), a phase restoring and resampling circuit being arranged between said output and the corresponding input of said second multiplier;
(h) an adder for adding the output signals of said first and second multipliers;
(i) a memory for weighted mixing of images in accordance with said given definition and in accordance with said reduced definition;
(C) the motion compensation stage comprises:
(j) a second motion compensation circuit receiving the output signal of said prediction stage and the additional motion and coding mode information PHF 90-567C f3 2 0 ~ o g ~ ~ 14.05.1992 components;
(k) an adder for adding the output signals of said decoding chain and said second motion compensation circuit.
In a second embodiment the invention relates to a decoding device in which the decoding stage is a variable length decoding chain for coded digital signals previously transmitted and/or stored after treatment in a variable length coding chain, said decoding chain comprising a memory circuit, a variable length decoding circuit, an inverse quantizing circuit, an inverse normalization circuit and an inverse orthogonal transform circuit and is characterized in that, with a first memory for storing .
information components in accordance with said given definition being arranged at the output of said decoding chain, (A) the stage for reconstructing the image in accordance with said-reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) a circuit for reconstituting the block in accordance with said reduced definition;
(c) an inverse orthogonal transform circuit;
(d) a second memory for storing information components in accordance with said reduced definition;
(B) the motion compensation~stage comprises first and second motion compensation circuits each receiving the output signal of one of said two memories for storing information components and for storing the coding mode and motion information components;
(C) the prediction stage comprises:
(f) at the output of that one of said motion compensation circuits which follows said memory for storing the information components in accordance with the given definition, a first reconstitution branch comprising a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the other one of said motion compensation circuits which follows said memory for storing the information components in accordance with PHF 90-567C 9 ~ ~ ~ ~ g ~ ~ 14.05.1992 the reduced definition, a second reconstitution branch comprising a second orthogonal transform circuit;
(h) at the output of said first and second parallel branches a circuit for combining the output signals of said first and second branches;
S (i) an adder for adding the output signals of said combination circuit and of the inverse quantizing circuit of the variable length decoding chain, arranged between said inverse quantizing circuit and the inverse orthogonal transform circuit following the last-mentioned circuit.
In a perfected modification of this second embodiment the decoding device is characterized in that the prediction stage also comprises, between the output of said second reconstitution branch and the corresponding input of said combination circuit, a weighted mixing circuit comprising:
(a) a multiplier for multiplying the output signal of said first branch by a coefficient a between 0 and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient ( 1-«);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to said corresponding input of the combination circuit.
These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter with reference to the accompanying drawings in which Fig. 1 shows a first embodiment of a coding device according to the invention;
Figs. 2A and 2B show scanning patterns of the coefficients of the traditional type and of the modified zigzag type;
Fig. 3 shows a second embodiment of a coding device according to the invention;
Figs. 4 and 5 show two embodiments of a decoding device according to the invention.
The coding device shown in a first embodiment in Fig. 1 comprises a .
stage for selecting the coding mode for the signals to be coded, which are digital signals 20'7082'7 PHF 90-567C 10 14.05.1992 considered as blocks of a fixed size representing a subdivision of images received (HDTV images, 1250 lines, 1:1, 50 Hz) and being treated independently. In this case the stage is an inter/intradecision circuit 100. This circuit 100 comprises an energy calculation circuit 11 which first calculates, from the input current signals of the device, the energy per image block in the components other than the DC component, the reference B denoting the current block. The circuit 100 also comprises a subtracter 12 which receives the samples corresponding to the current block B and the samples .
corresponding to a predicted block B' supplied by a prediction stage 400 which will be described hereinafter. With the subtraction (B-B') being effected, the energy calculation circuit 11 determines the energy contained in this block of samples constituted at the output of the subtracter 12 (always without a DC component) and subsequently, in accordance with the result of a comparison of the two energies thus calculated, it selects either the one or the other of the two coding modes referred to as interframe coding (B-B' is coded) and intraframe coding (B is coded) and abbreviated as intercoding and intracoding, respectively. An additional information component Mc specifying this coding mode is supplied by the circuit 100, because it will be necessary for reconstructing the images, particularly at the receiver end.
The coding mode selection stage is followed by a coding stage, in this case a variable length coding stage 200 constituting the coding chain (but in another type of coding stage, for example a fixed length coding stage may also be used). This coding chain comprises an orthogonal transform circuit 21 (a discrete cosine transform circuit for blocks comprising, for example 8 x 8 coefficients, referred to as DCT 8 x 8 is used here and in the following description without the type of orthogonal transform and the number of coefficients in the present case limiting the invention in any way), a scanning conversion circuit 22, a quantizing circuit 23, a variable length coding circuit 24, a memory circuit 25 and a rate control circuit 26 supplying a normalization signal which is applied to the circuit 23.
The scanning, i. e. the order of reading, of the coefficients supplied by the orthogonal transform circuit 21 has a zigzag modified pattern in this case.
Without this example being limitative, Fig. 2A shows the case where the size of the blocks is 8 x 8 for the high-definition images and 4 x 4 for the reduced definition images and the order of reading the coefficients in the case of a traditional zigzag scanning pattern, and Fig.
2B shows the order of reading in the case of a modified zigzag scanning pattern.
PIiF 90-567C Il ~ ~ ~ ~ ~ ~ ~ 14.05.1992 The signals at the output of the quantizing circuit 23 are applied to an image reconstruction stage 300 formed in the following way. This stage 300 comprises an inverse quantizing circuit 31, and at the output thereof a first path for reconstructing an image in accordance with said given definition, which first path comprises an inverse scanning conversion circuit 32, an inverse orthogonal transform circuit 34a (inverse discrete cosine transform effected on blocks of 8 x 8 coefficients, referred to as inverse DCT 8 x 8) and a circuit 35a for reconstructing the high-definition signal such as would be decoded, but for transmission errors (this circuit 35a receiving the signals B' and M~
for said reconstruction).
The stage 300 also comprises, at the output of the inverse scanning conversion circuit 32, a second path for reconstructing an image of reduced definition (here the compatible TV image, where the image of a given definition is a high-definitaon television image), which second path comprises a clipper circuit 33 with which a given fraction of the block coefficients can be selected, in this case the sixteen first coefficients (i.e. a quarter) scanned in a modified zigzag pattern, an inverse orthogonal transform ciicuit 34b (inverse discrete cosine transform, referred to as inverse DCT 4 x 4 because the.new block thus constituted does not comprise more than 4 x 4 coefficients), a circuit 35b for reconstituting the compatible television signal as would be decoded, but for transmission errors, a memory 36b for storing the signal thus reconstituted and a circuit 37b for motion compensation. The output signals of the circuits 35a and 36b constitute the output signals of the stage 300.
These output signals are applied to the previously mentioned prediction stage 400. This stage 400 is a circuit for reconstituting the image by combining the output signals of said first and second reconstruction paths and comprises a multiplier 39a and a multiplier 39b for multiplying the output signals of said first and second paths by a coefficient a and a coefficient (1-«), respectively, (« being between 0 and 1), and an adder 39c for adding the output signals of said multipliers and a weighting memory 39d storing the output signal of said adder. A phase restoring and resampling circuit 38b is arranged upstream of the multiplier 39b for realigning the number of pixels constituting the images at the input of the multiplier and for restoring the phase of these images (for reasons of phase shifting due to the inverse DCT 4 x 4 orthogonal transform). The output signal of the memory 39d is applied to a prediction circuit 41.
This circuit 41, which is used for supplying the predicted blocks B', is provided for PHF 90-567C 12 '~ ~'~ o ~ 2 ~ 14.05.1992 receiving the output signal of the memory 39d and the motion information components supplied by a motion estimation stage 500. -' This stage 500 conventionally comprises a correlator of blocks between the current image present at the input of the device and the reconstituted image present v at the output of the memory 39d, the output signal of this memory 39d being also applied to said stage 500. Said motion information components are motion vectors indicating the displacement between a block of an image and the corresponding block of the next considered image. The prediction circuit 41 consists in this case of a circuit for addressing the memory 39d, which addressing is effected as a function of said motion vectors.
The coding device shown in a second embodiment in Fig. 3 does not act on signals corresponding to the pixels but on coefficients resulting from an orthogonal transform (for example a discrete cosine transform) of these signals corresponding to the pixels. This device shown in Fig. 3 likewise comprises a stage for selecting the coding mode of the signals to be coded, having the reference numeral 150 and does not only comprise an inter/intradecision circuit composed of an energy calculation circuit 151 and a subtracter 152 whose structure and operation are similar to those of circuit 11 and subtracter 12, respectively, but also, upstream of these two elements, an orthogonal transform stage 153 (discrete cosine transform) for the input current signals of the coding device.
As in the case of Fig. 1, the coding mode selection stage is followed by a coding stage denoted by the reference numeral 250 and being a variable length coding chain. This chain comprises a quantizing circuit 253, a variable length coding circuit 254, a memory circuit 255 and a rate control circuit 256 supplying a normalization signal which is applied to the circuit 253. The signals present at the output of said quantizing circuit 253 are applied to a reconstruction stage which is constituted in the following manner.
This reconstruction stage, denoted by the reference numeral 350, comprises a first path for reconstructing an image in accordance with said given definition, said first path comprising an inverse quantizing circuit 351, a circuit 354a for reconstituting the block in accordance with said given definition, an inverse orthogonal transform stage 355a (inverse DCT transform) and a memory 356a for information components in accordance with said given definition. The stage 350 also PHF 90-567C 13 2 0'~ 0 8 2 ? 14.05.1992 comprises, at the output of the quantizing circuit 253, a second path for reconstructing the image in accordance with said reduced cle~nition, said second path comprising the inverse quantizing circuit 351, a clipper circuit 353b for taking a given fraction of the signals present at the output of the circuit 351, a circuit 354b for reconstituting the block in accordance with said reduced definition, an inverse orthogonal transform stage 355b (inverse DCT transform) and a memory 356b for information components in accordance with said reduced definition. The output signals of the memories 356a and 356b will be used for reconstituting the image prior to the selection of the signals to be coded and prior to coding and to this end they are applied to the prediction stage 450.
This stage 450 comprises, at the output of the first reconstruction path, a first reconstitution branch comprising a first prediction circuit 451 with motion compensation, a first orthogonal transform circuit 452 (DCT transform) and a low-frequency clipper circuit 453 with which, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients corresponding to the lowest frequencies can be eliminated. In all the examples described, this fraction is equal to one quarter, but this value is not a non-limitative example. The stage 450 comprises, at the output of the second reconstruction path, a second reconstitution branch comprising a second prediction circuit 461 with motion compensation and a second orthogonal transform circuit 462 (DCT transform). This stage 450 also comprises, at the outputs of said first and second parallel branches, a circuit 480 for combining the output signals of said low-frequency clipper circuit 453 and of the second orthogonal transform circuit 462 (DCT transform). This circuit 480 supplies the predicted signals which are applied to the coding mode selection stage.
Similarly as described hereinbefore, the motion information components are supplied by a motion estimation stage 550 which is identical to stage 500 of Fig. 1 and which addresses to the circuits 451 and 461 the motion vectors D
authorizing the motion compensation in blocks during stages of constructing the predicted blocks B'.
The vectors D, like the information components M~, are also transmitted and/or stored, because they are necessary at the receiver end.
Reciprocally, when the digital signals have been coded in a coding device according to the invention, their decoding is also ensured within the scope of this invention. The device which authorizes this decoding operation comprises, in association with a decoding stage, a stage for reconstructing the image in accordance PHF 90-567C 14 2 D ~ ~ g ~ ~ 14.05.1992 with said reduced definition, which stage is provided for operation based on decoded signals which are available in said given definition and based on the transmitted coding mode and motion information components, and a stage for prediction based on said decoded signals and on the output signals of said image reconstruction stage, and, downstream of said image reconstruction stage, a motion compensation stage.
In the embodiment shown in Fig. 4 the decoding device more specifically comprises a variable length decoding stage 600 constituting said decoding chain. This stage 600 comprises a buffer memory 61 receiving the input signals of the decoding device, i.e. all the signals previously coded and transmitted (or stored).
These signals include coded digital signals obtained from initial high-definition television images and motion information components supplied by the motion estimation stage of the coding device and coding mode information components supplied by the coding mode selection stage of this coding device. The buffer memory 61 is followed by a variable length decoding circuit 62, an inverse quantizing circuit 63 (connected to the buffer memory 61 via an inverse normalization circuit 64), an inverse scanning conversion circuit 65 and an inverse orthogonal transform circuit 66 (here an inverse discrete cosine transform circuit for blocks of 8 x 8 coefficients, referred to as inverse DCT
8 x 8).
The circuit 65 realises a scanning operation in an inverse modified zigzag pattern (transform inverse to that described with reference to Fig. 2B). The variable length decoding stage 600 thus described thereby constitutes a first path for reconstructing a high-definition, image.
The decoding device also comprises a stage 700 for reconstructing the reduced definition image, here a compatible TV image. In said device this stage constitutes a second path for reconstructing the reduced definition image, which path comprises a clipper circuit 71 which, likewise as the circuit 33, enables a given fraction of the decoded signals to be selected, namely a given fraction of coefficients corresponding to a block (in this case the sixteen first coefficients in the order of modified zigzag scanning) and which is to this end provided at the output of the inverse scanning conversion circuit 65. This circuit 71 is followed by an inverse orthogonal transform circuit 72 (here an inverse discrete cosine transform, referred to as inverse DCT 4 x 4), and a memory 74 for storing the reconstituted reduced definition image.
The estimation of motion during coding and the selected coding mode are taken into account in this reconstitution. As has been described hereinbefore, the PHF 90-567C 15 2 0 ~ o ~ ~ ~ 14.05.1992 transmitted signals include motion information components resulting from this estimation and coding mode information components indicating the effected selection.
These information components are applied to the stage 700 and to a motion compensation stage 900. This stage 900 comprises a motion compensation circuit 97 and an adder 98. In fact, the circuit 97 is a circuit for addressing a memory 84 (described hereinafter) which addressing takes the block displacements indicated by said motion information components into account. In the stage 700 the motion and coding mode information components are received after their decoding by a motion compensation circuit 75 which is also a circuit for addressing the memory 74, taking the block displacements indicated by the motion information components into account.
The images thus reconstructed with motion compensation are then applied to the adder 98 (for the high-definition images) and to an adder 73 (for the reduced definition images), respectively. The adder 98, which thus receives the output signal of the circuit 97 at its first input, receives the output signal of the inverse orthogonal transform circuit 66 of the stage 600 at its second input and supplies the reconstituted high-definition HDTV images (1250 lines, 2:1, 50 Hz) which correspond to the original high-definition images. The adder 73, which receives the output signal of the motion compensation circuit 75 at its first input, is arranged between the inverse orthogonal transform circuit 72, from which it receives the output signal at its second input, and a memory 74 to which its output signal is applied.
The decoding device comprises a prediction stage 800, which stage comprises a first and a second multiplier 82a and 82b for mu .plying the output signals of the adder 98 and the circuit 81 by a coefficient « between 0 and 1 and by the complementary coefficient (1-a), respectively, an adder 83 for adding the output signals of these multipliers and a memory storing the output signal of this adder 83.
This memory is the memory 84 mentioned hereinbefore and, due to the action of the multipliers, comprises a weighted mix of the reconstituted high-definition and reduced definition images which, after performance of the motion compensation, are present at the outputs of stages 600 and 700. A phase restoring and resampling circuit 81 is arranged at the output of the memory 74 between this output and the second multiplier 82b.
The decoding device shown in a second embodiment in Fig. 5 likewise comprises a variable length decoding stage 650 constituting the decoding chain. This PHF 90-567C 16 ~ ~ ~ ~ ~ ~ ~ 14.05.1992 stage 650 comprises a buffer memory 661 receiving the input signals of the decoding device (the previously coded digital signals and the coding mode and motion information components), followed by a variable length decoding circuit 662, an inverse quantizing circuit 663 (connected to the buffer memory 661 via an inverse normalization circuit 664.) and an inverse orthogonal transform circuit 665 (inverse DCT).
In the device of Fig. 5 the stage 650 thus described constitutes a first path for reconstructing a high-definition image whose output signal is memorized in a first memory 774 to be described hereinafter.
The decoding device also comprises a stage 750 for reconstructing the reduced definition image, here a compatible TV image. This stage 750 constitutes a second path for reconstructing a reduced detinition image, which path comprises a series arrangement of a clipper circuit 771 for selecting a given fraction (here a quarter) of the decoded signals, which circuit 771 is arranged at the output of the inverse quantizing circuit 663. This circuit 771 is followed by an adder 772 for reconstituting the block in accordance with the reduced definition, then an inverse orthogonal transform circuit 773 arid a second memory 775 for storing information components in accordance with the reduced definition. The stage 750 is completed by the first memory 774 for storing high-definition information components, which stage receives the decoded signals which are present at the output of the decoding stage 650, i.e. of the first reconstruction path.
As described hereinbefore, the additional transmitted and/or stored information components (coding mode and motion information components) are applied to the stage 750 and to a motion compensation stage 950. This stage 950 comprises two motion compensation circuits 967 and 968 receiving the output signal of the two memories 774 and 775 of said image reconstruction stage and said motion information components (after their decoding) and is actually a circuit for addressing the memories 774 and 775, respectively, taking the block displacements indicated by said motion information components into account.
The blocks thus reconstructed with motion compensation are then applied to a prediction stage 850 comprising a first and a second reconstitution branch of high definition and reduced definition, respectively, and, at their output, a circuit for combining the signals which they supply. The first reconstitution branch (for high definition) comprises a first orthogonal transform circuit 851 and a low-frequency PHF 90-567C 17 ~ ~ ~ ~ $ 2 ~ 14.05.1992 clipper circuit 852 for eliminating, from the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency, while the second reconstitution branch (for reduced definition) comprises a second orthogonal transform circuit 861. At the output of these first and second parallel branches, a circuit 880 derives the signals provided by the low-frequency clipper circuit 852 and the second orthogonal transform circuit 861 and combines them so as to constitute predicted signals. These output signals of the combination circuit 880 are applied to the second input of an adder 881 whose first input receives the output signals of the inverse quantizing circuit 663. This adder 881 is arranged between said circuit 663 and the circuit 665 of the decoding chain 650.
The present invention is not limited to the embodiments described hereinbefore, which embodiments may be modified without passing beyond the scope of the invention. In a second embodiment of the coding device of Fig. 3 described hereinbefore it is particularly possible to provide a weighting in the image reconstitution circuit, which weighting is analogous to that provided in the first embodiment of Fig. 1 at the output of the two reconstruction paths. As indicated by the solid lines in Fig. 3 (while in the absence of this modification the direct connection between the output of the second branch and the corresponding input of the combination circuit 480 is shown in broken lines in this Fig. 3) said image reconstitution circuit comprises circuits which are similar to those already described, viz a multiplier 471 for multiplying the output signal of the low-frequency clipper circuit 453 by a coefficient a between 0 and 1, which signal corresponds to said eliminated coefficients in the case of the first embodiment using a direct connection, a multiplier 472 for multiplying the output signal of the other branch by the coefficient (1-a) and an adder 473 for adding the output signals of said multipliers. The output signal of the adder 473 is applied to that input of the combination circuit 480 which, in the case of the direct connection shown in broken lines, would receive the output signal of the second branch.
In the second embodiment of the decoding device of Fig. 5 it is alternatively possible to provide a weighting in the prediction stage, which weighting is analogous to that which is provided also in the prediction stage in the first embodiment of Fig. 4. This weighting is likewise realised with the aid of two multipliers and one adder: as is indicated in solid lines in Fig. 5 (while in the absence of this modification the direct connection between the output of the second branch and the corresponding 2 0'7 0 8 2'~
PHF 90-567C 18 14.05.1992 input of the combination circuit 880 is shown in broken lines in the same Figure), a multiplier 877 receives the output signal of the first branch and multiplies it by a coefficient a between 0 and 1, a multiplier 878 receives the output signal of the other branch and multiplies it by the coefficient (1-a), and the adder 879 receives the output signals of these two multipliers and applies its output signal to that input of the combination circuit 880 which, in the case of direct connection, would receive the output signal of the second branch.
The memory 774 for storing high-definition information components is herein considered to be incorporated in the stage 750 but may alternatively be considered to form part of the stage 650. This modification is easy to understand and is not shown in Fig. ~.
PHF 90-567C 4 14.05.1992 device completely disregards the problem in the case of transmitting images having two definition levels, viz. the problem of supplying the information components which are simultaneously necessary for transmitting and reconstituting the television images at the receiver end in accordance with these two definition levels, notably high-definition television images and compatible television images. In contrast thereto, the structure proposed in this Application provides a satisfactory technical solution to this problem.
In a particular embodiment, in which essentially signals corresponding to pixels are used, the invention relates to a coding device in which the coding stage is a variable length coding chain comprising an orthogonal transform circuit, a scanning conversion circuit, a quantizing circuit, a variable length coding circuit, a memory circuit and a rate control circuit and in which the first path of the reconstruction stage comprises a series arrangement of:
(a) an inverse quantizing circuit;
(b) an inverse scanning conversion circuit;
(c) an inverse orthogonal transform stage;
(d) a circuit for reconstructing the signal in accordance with said given definition, as would be decoded after transmission, but for transmission errors;
characterized in that in said image reconstruction stage the second reconstruction path compnses:
(a) said inverse quantizing circuit;
(b) said inverse scanning conversion circuit;
(c) a clipper circuit for taking a given fraction of the signals at the output of said inverse quantizing circuit;
(d) an inverse orthogonal transform circuit;
(e) a circuit for reconstructing the signal having a reduced definition as would be decoded after transmission, but for transmission errors;
(f) a memory for storing said reconstructed signal; (g) a motion compensation circuit between the output of said memory and said circuit for reconstructing the reduced definition signal;
and in that said image reconstituting circuit comprises:
(h) a multiplier for multiplying the output signal of the first reconstruction path by a coefficient a between 0 and 1;
(i) a multiplier for multiplying the output signal of the second PHF 90-567C 5 ~ ~ ~ o $ 2 ~ 14.05.1992 reconstruction path by the coefficient (1-a), a phase restoring and resampling circuit being arranged between said output of the second reconstruction path and the corresponding input of the associated multiplier;
(j) an adder for adding the output signals of said multipliers;
(k) a weighting memory for storing the output signal of said adder;
(1) a prediction circuit for receiving the output signal of said weighting memory and the motion information components supplied by said motion estimation stage.
In a second embodiment, whose operation is essentially based on coefficients resulting from an orthogonal transform of said signals corresponding to pixels, the invention relates particularly to a coding device in which the coding stage is a variable length coding chain comprising a quantizing circuit, a variable length coding circuit, a memory circuit and a rate control circuit, in which the stage for selecting the coding mode comprises a series arrangement of:
(a) an orthogonal transform stage for said input current signals;
(b) an inter/intradecision circuit for also receiving said predicted signals;
and in which the first reconstruction path comprises a series arrangement of:
(c) an inverse quantizing circuit;
(d) a circuit for reconstituting the block in accordance with said given , definition;
(e) an inverse orthogonal transform stage;
(f) a memory for the information components of said given definition;
characterized in that in said reconstruction stage the second reconstruction path comprises:
(a) said inverse quantizing circuit;
(b) a clipper circuit for taking a given fraction of the signals present at the output of said inverse quantizing circuit;
(c) a circuit for reconstituting the block in accordance with said reduced definition;
(d) an inverse orthogonal transform stage;
(e) a memory for the information components of said reduced definition;
and in that the image reconstitution circuit comprises:
(f) at the output of said first path, a first reconstitution branch comprising 2070~2'~
PHF 90-567C fi 14.05.1992 a a first prediction circuit with motion compensation, a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the second path, a second reconstitution branch comprising a second prediction circuit with motion compensation and a second orthogonal transform circuit;
(h) a circuit for combining the output signals of said first and second branches for supplying said predicted signals transmitted to said coding mode selection stage.
In a perfected modification of this second embodiment, the coding device is characterized in that the image reconstitution circuit for the prediction stage comprises, between the output of said second reconstitution branch and the corresponding input of said combination circuit, a weighted mixing circuit comprising: , (a) a multiplier for multiplying the output signal of the first branch by a coefficient a between 0'and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit ;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient (1-a);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to the corresponding input of the combination circuit.
It is another object of the invention to provide a device for decoding digital signals corresponding to television images with which reduced definition images and high-definition images of good quality can be reconstituted.
To this end the invention relates to a decoding device as defined in the opening paragraph and is characterized in that it comprises, in association with a decoding stage:
(A) a stage for reconstructing the image in accordance with said reduced definition from the decoded signals;
(B) a stage for motion compensation based on said additional information components;
(C) a stage for prediction based on the output signals of said image reconstruction stage.
PHF 90-567C 7 2 ~ ~ ~ ~ ~ ~ 14.05.1992 In a particular embodiment the invention relates to a decoding device in which the decoding stage is a variable length decoding chain for coded digital signals previously transmitted and/or stored after treatment in a variable length coding chain, said decoding chain comprising a memory circuit, a variable length decoding circuit, an inverse quantizing circuit, an inverse normalization circuit, an inverse scanning conversion circuit and an inverse orthogonal transform circuit, and is characterized in that:
(A) the stage for reconstructing the image in accordance with said reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) an inverse orthogonal transform circuit;
(c) an adder whose first input receives the output signal of said inverse orthogonal transform circuit;
(d) a memory for storing the image reconstituted in accordance with the reduced definition and being present at the output of said adder;
(e) a first circuit for compensating motion, receiving the output signal of said memory and the motion information components, and having its output connected to the second input of said adder;
(B) the prediction stage comprises:
(f) a first multiplier for multiplying the output signals of said decoding device by a coefficient a between 0 and 1;
(g) a second multiplier for multiplying the output signal of said stage for reconstructing the image in accordance with said reduced definition by the coefficient (1-a), a phase restoring and resampling circuit being arranged between said output and the corresponding input of said second multiplier;
(h) an adder for adding the output signals of said first and second multipliers;
(i) a memory for weighted mixing of images in accordance with said given definition and in accordance with said reduced definition;
(C) the motion compensation stage comprises:
(j) a second motion compensation circuit receiving the output signal of said prediction stage and the additional motion and coding mode information PHF 90-567C f3 2 0 ~ o g ~ ~ 14.05.1992 components;
(k) an adder for adding the output signals of said decoding chain and said second motion compensation circuit.
In a second embodiment the invention relates to a decoding device in which the decoding stage is a variable length decoding chain for coded digital signals previously transmitted and/or stored after treatment in a variable length coding chain, said decoding chain comprising a memory circuit, a variable length decoding circuit, an inverse quantizing circuit, an inverse normalization circuit and an inverse orthogonal transform circuit and is characterized in that, with a first memory for storing .
information components in accordance with said given definition being arranged at the output of said decoding chain, (A) the stage for reconstructing the image in accordance with said-reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) a circuit for reconstituting the block in accordance with said reduced definition;
(c) an inverse orthogonal transform circuit;
(d) a second memory for storing information components in accordance with said reduced definition;
(B) the motion compensation~stage comprises first and second motion compensation circuits each receiving the output signal of one of said two memories for storing information components and for storing the coding mode and motion information components;
(C) the prediction stage comprises:
(f) at the output of that one of said motion compensation circuits which follows said memory for storing the information components in accordance with the given definition, a first reconstitution branch comprising a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the other one of said motion compensation circuits which follows said memory for storing the information components in accordance with PHF 90-567C 9 ~ ~ ~ ~ g ~ ~ 14.05.1992 the reduced definition, a second reconstitution branch comprising a second orthogonal transform circuit;
(h) at the output of said first and second parallel branches a circuit for combining the output signals of said first and second branches;
S (i) an adder for adding the output signals of said combination circuit and of the inverse quantizing circuit of the variable length decoding chain, arranged between said inverse quantizing circuit and the inverse orthogonal transform circuit following the last-mentioned circuit.
In a perfected modification of this second embodiment the decoding device is characterized in that the prediction stage also comprises, between the output of said second reconstitution branch and the corresponding input of said combination circuit, a weighted mixing circuit comprising:
(a) a multiplier for multiplying the output signal of said first branch by a coefficient a between 0 and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient ( 1-«);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to said corresponding input of the combination circuit.
These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter with reference to the accompanying drawings in which Fig. 1 shows a first embodiment of a coding device according to the invention;
Figs. 2A and 2B show scanning patterns of the coefficients of the traditional type and of the modified zigzag type;
Fig. 3 shows a second embodiment of a coding device according to the invention;
Figs. 4 and 5 show two embodiments of a decoding device according to the invention.
The coding device shown in a first embodiment in Fig. 1 comprises a .
stage for selecting the coding mode for the signals to be coded, which are digital signals 20'7082'7 PHF 90-567C 10 14.05.1992 considered as blocks of a fixed size representing a subdivision of images received (HDTV images, 1250 lines, 1:1, 50 Hz) and being treated independently. In this case the stage is an inter/intradecision circuit 100. This circuit 100 comprises an energy calculation circuit 11 which first calculates, from the input current signals of the device, the energy per image block in the components other than the DC component, the reference B denoting the current block. The circuit 100 also comprises a subtracter 12 which receives the samples corresponding to the current block B and the samples .
corresponding to a predicted block B' supplied by a prediction stage 400 which will be described hereinafter. With the subtraction (B-B') being effected, the energy calculation circuit 11 determines the energy contained in this block of samples constituted at the output of the subtracter 12 (always without a DC component) and subsequently, in accordance with the result of a comparison of the two energies thus calculated, it selects either the one or the other of the two coding modes referred to as interframe coding (B-B' is coded) and intraframe coding (B is coded) and abbreviated as intercoding and intracoding, respectively. An additional information component Mc specifying this coding mode is supplied by the circuit 100, because it will be necessary for reconstructing the images, particularly at the receiver end.
The coding mode selection stage is followed by a coding stage, in this case a variable length coding stage 200 constituting the coding chain (but in another type of coding stage, for example a fixed length coding stage may also be used). This coding chain comprises an orthogonal transform circuit 21 (a discrete cosine transform circuit for blocks comprising, for example 8 x 8 coefficients, referred to as DCT 8 x 8 is used here and in the following description without the type of orthogonal transform and the number of coefficients in the present case limiting the invention in any way), a scanning conversion circuit 22, a quantizing circuit 23, a variable length coding circuit 24, a memory circuit 25 and a rate control circuit 26 supplying a normalization signal which is applied to the circuit 23.
The scanning, i. e. the order of reading, of the coefficients supplied by the orthogonal transform circuit 21 has a zigzag modified pattern in this case.
Without this example being limitative, Fig. 2A shows the case where the size of the blocks is 8 x 8 for the high-definition images and 4 x 4 for the reduced definition images and the order of reading the coefficients in the case of a traditional zigzag scanning pattern, and Fig.
2B shows the order of reading in the case of a modified zigzag scanning pattern.
PIiF 90-567C Il ~ ~ ~ ~ ~ ~ ~ 14.05.1992 The signals at the output of the quantizing circuit 23 are applied to an image reconstruction stage 300 formed in the following way. This stage 300 comprises an inverse quantizing circuit 31, and at the output thereof a first path for reconstructing an image in accordance with said given definition, which first path comprises an inverse scanning conversion circuit 32, an inverse orthogonal transform circuit 34a (inverse discrete cosine transform effected on blocks of 8 x 8 coefficients, referred to as inverse DCT 8 x 8) and a circuit 35a for reconstructing the high-definition signal such as would be decoded, but for transmission errors (this circuit 35a receiving the signals B' and M~
for said reconstruction).
The stage 300 also comprises, at the output of the inverse scanning conversion circuit 32, a second path for reconstructing an image of reduced definition (here the compatible TV image, where the image of a given definition is a high-definitaon television image), which second path comprises a clipper circuit 33 with which a given fraction of the block coefficients can be selected, in this case the sixteen first coefficients (i.e. a quarter) scanned in a modified zigzag pattern, an inverse orthogonal transform ciicuit 34b (inverse discrete cosine transform, referred to as inverse DCT 4 x 4 because the.new block thus constituted does not comprise more than 4 x 4 coefficients), a circuit 35b for reconstituting the compatible television signal as would be decoded, but for transmission errors, a memory 36b for storing the signal thus reconstituted and a circuit 37b for motion compensation. The output signals of the circuits 35a and 36b constitute the output signals of the stage 300.
These output signals are applied to the previously mentioned prediction stage 400. This stage 400 is a circuit for reconstituting the image by combining the output signals of said first and second reconstruction paths and comprises a multiplier 39a and a multiplier 39b for multiplying the output signals of said first and second paths by a coefficient a and a coefficient (1-«), respectively, (« being between 0 and 1), and an adder 39c for adding the output signals of said multipliers and a weighting memory 39d storing the output signal of said adder. A phase restoring and resampling circuit 38b is arranged upstream of the multiplier 39b for realigning the number of pixels constituting the images at the input of the multiplier and for restoring the phase of these images (for reasons of phase shifting due to the inverse DCT 4 x 4 orthogonal transform). The output signal of the memory 39d is applied to a prediction circuit 41.
This circuit 41, which is used for supplying the predicted blocks B', is provided for PHF 90-567C 12 '~ ~'~ o ~ 2 ~ 14.05.1992 receiving the output signal of the memory 39d and the motion information components supplied by a motion estimation stage 500. -' This stage 500 conventionally comprises a correlator of blocks between the current image present at the input of the device and the reconstituted image present v at the output of the memory 39d, the output signal of this memory 39d being also applied to said stage 500. Said motion information components are motion vectors indicating the displacement between a block of an image and the corresponding block of the next considered image. The prediction circuit 41 consists in this case of a circuit for addressing the memory 39d, which addressing is effected as a function of said motion vectors.
The coding device shown in a second embodiment in Fig. 3 does not act on signals corresponding to the pixels but on coefficients resulting from an orthogonal transform (for example a discrete cosine transform) of these signals corresponding to the pixels. This device shown in Fig. 3 likewise comprises a stage for selecting the coding mode of the signals to be coded, having the reference numeral 150 and does not only comprise an inter/intradecision circuit composed of an energy calculation circuit 151 and a subtracter 152 whose structure and operation are similar to those of circuit 11 and subtracter 12, respectively, but also, upstream of these two elements, an orthogonal transform stage 153 (discrete cosine transform) for the input current signals of the coding device.
As in the case of Fig. 1, the coding mode selection stage is followed by a coding stage denoted by the reference numeral 250 and being a variable length coding chain. This chain comprises a quantizing circuit 253, a variable length coding circuit 254, a memory circuit 255 and a rate control circuit 256 supplying a normalization signal which is applied to the circuit 253. The signals present at the output of said quantizing circuit 253 are applied to a reconstruction stage which is constituted in the following manner.
This reconstruction stage, denoted by the reference numeral 350, comprises a first path for reconstructing an image in accordance with said given definition, said first path comprising an inverse quantizing circuit 351, a circuit 354a for reconstituting the block in accordance with said given definition, an inverse orthogonal transform stage 355a (inverse DCT transform) and a memory 356a for information components in accordance with said given definition. The stage 350 also PHF 90-567C 13 2 0'~ 0 8 2 ? 14.05.1992 comprises, at the output of the quantizing circuit 253, a second path for reconstructing the image in accordance with said reduced cle~nition, said second path comprising the inverse quantizing circuit 351, a clipper circuit 353b for taking a given fraction of the signals present at the output of the circuit 351, a circuit 354b for reconstituting the block in accordance with said reduced definition, an inverse orthogonal transform stage 355b (inverse DCT transform) and a memory 356b for information components in accordance with said reduced definition. The output signals of the memories 356a and 356b will be used for reconstituting the image prior to the selection of the signals to be coded and prior to coding and to this end they are applied to the prediction stage 450.
This stage 450 comprises, at the output of the first reconstruction path, a first reconstitution branch comprising a first prediction circuit 451 with motion compensation, a first orthogonal transform circuit 452 (DCT transform) and a low-frequency clipper circuit 453 with which, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients corresponding to the lowest frequencies can be eliminated. In all the examples described, this fraction is equal to one quarter, but this value is not a non-limitative example. The stage 450 comprises, at the output of the second reconstruction path, a second reconstitution branch comprising a second prediction circuit 461 with motion compensation and a second orthogonal transform circuit 462 (DCT transform). This stage 450 also comprises, at the outputs of said first and second parallel branches, a circuit 480 for combining the output signals of said low-frequency clipper circuit 453 and of the second orthogonal transform circuit 462 (DCT transform). This circuit 480 supplies the predicted signals which are applied to the coding mode selection stage.
Similarly as described hereinbefore, the motion information components are supplied by a motion estimation stage 550 which is identical to stage 500 of Fig. 1 and which addresses to the circuits 451 and 461 the motion vectors D
authorizing the motion compensation in blocks during stages of constructing the predicted blocks B'.
The vectors D, like the information components M~, are also transmitted and/or stored, because they are necessary at the receiver end.
Reciprocally, when the digital signals have been coded in a coding device according to the invention, their decoding is also ensured within the scope of this invention. The device which authorizes this decoding operation comprises, in association with a decoding stage, a stage for reconstructing the image in accordance PHF 90-567C 14 2 D ~ ~ g ~ ~ 14.05.1992 with said reduced definition, which stage is provided for operation based on decoded signals which are available in said given definition and based on the transmitted coding mode and motion information components, and a stage for prediction based on said decoded signals and on the output signals of said image reconstruction stage, and, downstream of said image reconstruction stage, a motion compensation stage.
In the embodiment shown in Fig. 4 the decoding device more specifically comprises a variable length decoding stage 600 constituting said decoding chain. This stage 600 comprises a buffer memory 61 receiving the input signals of the decoding device, i.e. all the signals previously coded and transmitted (or stored).
These signals include coded digital signals obtained from initial high-definition television images and motion information components supplied by the motion estimation stage of the coding device and coding mode information components supplied by the coding mode selection stage of this coding device. The buffer memory 61 is followed by a variable length decoding circuit 62, an inverse quantizing circuit 63 (connected to the buffer memory 61 via an inverse normalization circuit 64), an inverse scanning conversion circuit 65 and an inverse orthogonal transform circuit 66 (here an inverse discrete cosine transform circuit for blocks of 8 x 8 coefficients, referred to as inverse DCT
8 x 8).
The circuit 65 realises a scanning operation in an inverse modified zigzag pattern (transform inverse to that described with reference to Fig. 2B). The variable length decoding stage 600 thus described thereby constitutes a first path for reconstructing a high-definition, image.
The decoding device also comprises a stage 700 for reconstructing the reduced definition image, here a compatible TV image. In said device this stage constitutes a second path for reconstructing the reduced definition image, which path comprises a clipper circuit 71 which, likewise as the circuit 33, enables a given fraction of the decoded signals to be selected, namely a given fraction of coefficients corresponding to a block (in this case the sixteen first coefficients in the order of modified zigzag scanning) and which is to this end provided at the output of the inverse scanning conversion circuit 65. This circuit 71 is followed by an inverse orthogonal transform circuit 72 (here an inverse discrete cosine transform, referred to as inverse DCT 4 x 4), and a memory 74 for storing the reconstituted reduced definition image.
The estimation of motion during coding and the selected coding mode are taken into account in this reconstitution. As has been described hereinbefore, the PHF 90-567C 15 2 0 ~ o ~ ~ ~ 14.05.1992 transmitted signals include motion information components resulting from this estimation and coding mode information components indicating the effected selection.
These information components are applied to the stage 700 and to a motion compensation stage 900. This stage 900 comprises a motion compensation circuit 97 and an adder 98. In fact, the circuit 97 is a circuit for addressing a memory 84 (described hereinafter) which addressing takes the block displacements indicated by said motion information components into account. In the stage 700 the motion and coding mode information components are received after their decoding by a motion compensation circuit 75 which is also a circuit for addressing the memory 74, taking the block displacements indicated by the motion information components into account.
The images thus reconstructed with motion compensation are then applied to the adder 98 (for the high-definition images) and to an adder 73 (for the reduced definition images), respectively. The adder 98, which thus receives the output signal of the circuit 97 at its first input, receives the output signal of the inverse orthogonal transform circuit 66 of the stage 600 at its second input and supplies the reconstituted high-definition HDTV images (1250 lines, 2:1, 50 Hz) which correspond to the original high-definition images. The adder 73, which receives the output signal of the motion compensation circuit 75 at its first input, is arranged between the inverse orthogonal transform circuit 72, from which it receives the output signal at its second input, and a memory 74 to which its output signal is applied.
The decoding device comprises a prediction stage 800, which stage comprises a first and a second multiplier 82a and 82b for mu .plying the output signals of the adder 98 and the circuit 81 by a coefficient « between 0 and 1 and by the complementary coefficient (1-a), respectively, an adder 83 for adding the output signals of these multipliers and a memory storing the output signal of this adder 83.
This memory is the memory 84 mentioned hereinbefore and, due to the action of the multipliers, comprises a weighted mix of the reconstituted high-definition and reduced definition images which, after performance of the motion compensation, are present at the outputs of stages 600 and 700. A phase restoring and resampling circuit 81 is arranged at the output of the memory 74 between this output and the second multiplier 82b.
The decoding device shown in a second embodiment in Fig. 5 likewise comprises a variable length decoding stage 650 constituting the decoding chain. This PHF 90-567C 16 ~ ~ ~ ~ ~ ~ ~ 14.05.1992 stage 650 comprises a buffer memory 661 receiving the input signals of the decoding device (the previously coded digital signals and the coding mode and motion information components), followed by a variable length decoding circuit 662, an inverse quantizing circuit 663 (connected to the buffer memory 661 via an inverse normalization circuit 664.) and an inverse orthogonal transform circuit 665 (inverse DCT).
In the device of Fig. 5 the stage 650 thus described constitutes a first path for reconstructing a high-definition image whose output signal is memorized in a first memory 774 to be described hereinafter.
The decoding device also comprises a stage 750 for reconstructing the reduced definition image, here a compatible TV image. This stage 750 constitutes a second path for reconstructing a reduced detinition image, which path comprises a series arrangement of a clipper circuit 771 for selecting a given fraction (here a quarter) of the decoded signals, which circuit 771 is arranged at the output of the inverse quantizing circuit 663. This circuit 771 is followed by an adder 772 for reconstituting the block in accordance with the reduced definition, then an inverse orthogonal transform circuit 773 arid a second memory 775 for storing information components in accordance with the reduced definition. The stage 750 is completed by the first memory 774 for storing high-definition information components, which stage receives the decoded signals which are present at the output of the decoding stage 650, i.e. of the first reconstruction path.
As described hereinbefore, the additional transmitted and/or stored information components (coding mode and motion information components) are applied to the stage 750 and to a motion compensation stage 950. This stage 950 comprises two motion compensation circuits 967 and 968 receiving the output signal of the two memories 774 and 775 of said image reconstruction stage and said motion information components (after their decoding) and is actually a circuit for addressing the memories 774 and 775, respectively, taking the block displacements indicated by said motion information components into account.
The blocks thus reconstructed with motion compensation are then applied to a prediction stage 850 comprising a first and a second reconstitution branch of high definition and reduced definition, respectively, and, at their output, a circuit for combining the signals which they supply. The first reconstitution branch (for high definition) comprises a first orthogonal transform circuit 851 and a low-frequency PHF 90-567C 17 ~ ~ ~ ~ $ 2 ~ 14.05.1992 clipper circuit 852 for eliminating, from the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency, while the second reconstitution branch (for reduced definition) comprises a second orthogonal transform circuit 861. At the output of these first and second parallel branches, a circuit 880 derives the signals provided by the low-frequency clipper circuit 852 and the second orthogonal transform circuit 861 and combines them so as to constitute predicted signals. These output signals of the combination circuit 880 are applied to the second input of an adder 881 whose first input receives the output signals of the inverse quantizing circuit 663. This adder 881 is arranged between said circuit 663 and the circuit 665 of the decoding chain 650.
The present invention is not limited to the embodiments described hereinbefore, which embodiments may be modified without passing beyond the scope of the invention. In a second embodiment of the coding device of Fig. 3 described hereinbefore it is particularly possible to provide a weighting in the image reconstitution circuit, which weighting is analogous to that provided in the first embodiment of Fig. 1 at the output of the two reconstruction paths. As indicated by the solid lines in Fig. 3 (while in the absence of this modification the direct connection between the output of the second branch and the corresponding input of the combination circuit 480 is shown in broken lines in this Fig. 3) said image reconstitution circuit comprises circuits which are similar to those already described, viz a multiplier 471 for multiplying the output signal of the low-frequency clipper circuit 453 by a coefficient a between 0 and 1, which signal corresponds to said eliminated coefficients in the case of the first embodiment using a direct connection, a multiplier 472 for multiplying the output signal of the other branch by the coefficient (1-a) and an adder 473 for adding the output signals of said multipliers. The output signal of the adder 473 is applied to that input of the combination circuit 480 which, in the case of the direct connection shown in broken lines, would receive the output signal of the second branch.
In the second embodiment of the decoding device of Fig. 5 it is alternatively possible to provide a weighting in the prediction stage, which weighting is analogous to that which is provided also in the prediction stage in the first embodiment of Fig. 4. This weighting is likewise realised with the aid of two multipliers and one adder: as is indicated in solid lines in Fig. 5 (while in the absence of this modification the direct connection between the output of the second branch and the corresponding 2 0'7 0 8 2'~
PHF 90-567C 18 14.05.1992 input of the combination circuit 880 is shown in broken lines in the same Figure), a multiplier 877 receives the output signal of the first branch and multiplies it by a coefficient a between 0 and 1, a multiplier 878 receives the output signal of the other branch and multiplies it by the coefficient (1-a), and the adder 879 receives the output signals of these two multipliers and applies its output signal to that input of the combination circuit 880 which, in the case of direct connection, would receive the output signal of the second branch.
The memory 774 for storing high-definition information components is herein considered to be incorporated in the stage 750 but may alternatively be considered to form part of the stage 650. This modification is easy to understand and is not shown in Fig. ~.
Claims (10)
1. A device for coding digital signals corresponding to television images having a given definition, comprising:
(A) a stage for selecting the coding mode of signals to be coded from input current signals of the coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of the device, said stage supplying the signals to be coded and coding mode information;
(B) a coding stage;
(C) an image reconstruction stage in a first path for reconstructing the image in accordance with said given definition;
(D) a stage for estimating the motion between images, supplying motion information;
(E) a stage for predicting from output signals of said image reconstruction and motion estimation stages the output signals of said prediction stage constituting said predicted signals transmitted to said coding mode selection stage, characterized in that said digital signals are considered as blocks of a given size representing a subdivision of images and being treated independently and in that the image reconstruction stage also comprises a second image reconstruction path in accordance with a reduced definition, which path is situated parallel to said first path, and in that the prediction stage comprises a circuit for reconstituting the image by combining the output signals of said first and second reconstruction paths.
(A) a stage for selecting the coding mode of signals to be coded from input current signals of the coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of the device, said stage supplying the signals to be coded and coding mode information;
(B) a coding stage;
(C) an image reconstruction stage in a first path for reconstructing the image in accordance with said given definition;
(D) a stage for estimating the motion between images, supplying motion information;
(E) a stage for predicting from output signals of said image reconstruction and motion estimation stages the output signals of said prediction stage constituting said predicted signals transmitted to said coding mode selection stage, characterized in that said digital signals are considered as blocks of a given size representing a subdivision of images and being treated independently and in that the image reconstruction stage also comprises a second image reconstruction path in accordance with a reduced definition, which path is situated parallel to said first path, and in that the prediction stage comprises a circuit for reconstituting the image by combining the output signals of said first and second reconstruction paths.
2. A coding device as claimed in Claim 1, in which the coding stage is a variable length coding chain comprising an orthogonal transform circuit, a scanning conversion circuit, a quantizing circuit, a variable length coding circuit, a memory circuit and a rate control circuit and in which the first path of the reconstruction stage comprises a series arrangement of:
(a) an inverse quantizing circuit;
(b) an inverse scanning conversion circuit;
(c) an inverse orthogonal transform stage;
(d) a circuit for reconstructing the signal in accordance with said given definition, as would be decoded after transmission, but for transmission errors;
characterized in that in said image reconstruction stage the second reconstruction path comprises:
(a) said inverse quantizing circuit;
(b) said inverse scanning conversion circuit;
(c) a clipper circuit for taking a given fraction of the signals at the output of said inverse quantizing circuit;
(d) an inverse orthogonal transform circuit;
(e) a circuit for reconstructing the signal having a reduced definition as would be decoded after transmission, but for transmission errors;
(f) a memory for storing said reconstructed signal; (g) a motion compensation circuit between the output of said memory and said circuit for reconstructing the reduced definition signal;
and in that said image reconstituting circuit comprises:
(h) a multiplier for multiplying the output signal of the first reconstruction path by a coefficient .alpha. between 0 and 1;
(i) a multiplier for multiplying the output signal of the second reconstruction path by the coefficient (I-.alpha.), a phase restoring and resampling circuit being arranged between said output of the second reconstruction path and the corresponding input of the associated multiplier;
(j) an adder for adding the output signals of said multipliers;
(k) a weighting memory .for storing the output signal of said adder;
(l) a prediction circuit for receiving the output signal of said weighting memory and the motion information components supplied by said motion estimation stage.
(a) an inverse quantizing circuit;
(b) an inverse scanning conversion circuit;
(c) an inverse orthogonal transform stage;
(d) a circuit for reconstructing the signal in accordance with said given definition, as would be decoded after transmission, but for transmission errors;
characterized in that in said image reconstruction stage the second reconstruction path comprises:
(a) said inverse quantizing circuit;
(b) said inverse scanning conversion circuit;
(c) a clipper circuit for taking a given fraction of the signals at the output of said inverse quantizing circuit;
(d) an inverse orthogonal transform circuit;
(e) a circuit for reconstructing the signal having a reduced definition as would be decoded after transmission, but for transmission errors;
(f) a memory for storing said reconstructed signal; (g) a motion compensation circuit between the output of said memory and said circuit for reconstructing the reduced definition signal;
and in that said image reconstituting circuit comprises:
(h) a multiplier for multiplying the output signal of the first reconstruction path by a coefficient .alpha. between 0 and 1;
(i) a multiplier for multiplying the output signal of the second reconstruction path by the coefficient (I-.alpha.), a phase restoring and resampling circuit being arranged between said output of the second reconstruction path and the corresponding input of the associated multiplier;
(j) an adder for adding the output signals of said multipliers;
(k) a weighting memory .for storing the output signal of said adder;
(l) a prediction circuit for receiving the output signal of said weighting memory and the motion information components supplied by said motion estimation stage.
3. A coding device as claimed in Claim 1, in which the coding stage is a variable length coding chain comprising a quantizing circuit, a variable length coding circuit, a memory circuit and a rate control circuit, in which the stage for selecting the coding mode comprises a series arrangement of:
(a) an orthogonal transform stage for said input current signals;
(b) an inter/intradecision circuit for also receiving said predicted signals;
and in which the first reconstruction path comprises a series arrangement of:
(c) an inverse quantizing circuit;
(d) a circuit for reconstituting the block in accordance with said given definition;
(e) an inverse orthogonal transform stage;
(f) a memory for the information components of said given definition;
characterized in that in said reconstruction stage the second reconstruction path comprises:
(a) said inverse quantizing circuit;
(b) a clipper circuit for taking a given fraction of the signals present at the output of said inverse quantizing circuit;
(c) a circuit for reconstituting the block in accordance with said reduced definition;
(d) an inverse orthogonal transform stage;
(e) a memory for the information components of said reduced definition;
and in that the image reconstitution circuit comprises:
(f) at the output of said first path, a first reconstitution branch comprising a first prediction circuit with motion compensation, a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the second path, a second reconstitution brand comprising a second prediction circuit with motion compensation and a second orthogonal transform circuit;
(h) a circuit for combining the output signals of said first and second branches for supplying said predicted signals transmitted to said coding mode selection stage.
(a) an orthogonal transform stage for said input current signals;
(b) an inter/intradecision circuit for also receiving said predicted signals;
and in which the first reconstruction path comprises a series arrangement of:
(c) an inverse quantizing circuit;
(d) a circuit for reconstituting the block in accordance with said given definition;
(e) an inverse orthogonal transform stage;
(f) a memory for the information components of said given definition;
characterized in that in said reconstruction stage the second reconstruction path comprises:
(a) said inverse quantizing circuit;
(b) a clipper circuit for taking a given fraction of the signals present at the output of said inverse quantizing circuit;
(c) a circuit for reconstituting the block in accordance with said reduced definition;
(d) an inverse orthogonal transform stage;
(e) a memory for the information components of said reduced definition;
and in that the image reconstitution circuit comprises:
(f) at the output of said first path, a first reconstitution branch comprising a first prediction circuit with motion compensation, a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the second path, a second reconstitution brand comprising a second prediction circuit with motion compensation and a second orthogonal transform circuit;
(h) a circuit for combining the output signals of said first and second branches for supplying said predicted signals transmitted to said coding mode selection stage.
4. A coding device as claimed in Claim 3, characterized in that the image reconstitution circuit for the prediction stage comprises, between the output of said second reconstitution branch and the corresponding input of said combination circuit, a weighted mixing circuit comprising:
(a) a multiplier for multiplying the output signal of the first branch by a coefficient a between 0 and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit ;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient (1-.alpha.);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to the corresponding input of the combination circuit.
(a) a multiplier for multiplying the output signal of the first branch by a coefficient a between 0 and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit ;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient (1-.alpha.);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to the corresponding input of the combination circuit.
5. A coding device as claimed in any one of Claims 2 to 4, characterized in that said given fraction is equal to one quarter.
6. A device for decoding coded digital signals previously transmitted and/or stored after treatment in a device for coding digital signals corresponding to television images of a given definition, said coding device comprising:
(a) a stage for selecting the coding mode of the signals to be coded from input current signals of said coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of said device, said preceding or current input signals being considered as blocks of a given size representing a subdivision of images and being treated independently;
(b) a coding stage;
(c) a stage for reconstructing the image in accordance with said given definition on the one hand and in accordance with a reduced definition on the other hand;
(d) a stage for estimating the motion between images;
(e) for supplying said predicted signals, a stage for predicting from the output signals of said motion estimation and image reconstruction stages the information components which are additional and relative to the estimated motion between images and to the selected coding mode being intended, like the coded signals, for transmission and/or storage, characterized in that it comprises, in association with a decoding stage:
(A) a stage for reconstructing the image in accordance with said reduced definition from the decoded signals;
(B) a stage for motion compensation based on said additional information components;
(C) a stage for prediction based on the output signals of said image reconstruction stage.
(a) a stage for selecting the coding mode of the signals to be coded from input current signals of said coding device on the one hand and from predicted signals on the other hand, based on the preceding input signals of said device, said preceding or current input signals being considered as blocks of a given size representing a subdivision of images and being treated independently;
(b) a coding stage;
(c) a stage for reconstructing the image in accordance with said given definition on the one hand and in accordance with a reduced definition on the other hand;
(d) a stage for estimating the motion between images;
(e) for supplying said predicted signals, a stage for predicting from the output signals of said motion estimation and image reconstruction stages the information components which are additional and relative to the estimated motion between images and to the selected coding mode being intended, like the coded signals, for transmission and/or storage, characterized in that it comprises, in association with a decoding stage:
(A) a stage for reconstructing the image in accordance with said reduced definition from the decoded signals;
(B) a stage for motion compensation based on said additional information components;
(C) a stage for prediction based on the output signals of said image reconstruction stage.
7. A decoding device as claimed in Claim 6, in which the decoding stage is a variable length decoding chain for coded digital signals previously transmitted and/or stored after treatment in a variable length coding chain, said decoding chain comprising a memory circuit, a variable length decoding circuit, an inverse quantizing circuit, an inverse normalization circuit, an inverse scanning conversion circuit and an inverse orthogonal transform circuit, characterized in that:
(A) the stage for reconstructing the image in accordance with said reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) an inverse orthogonal transform circuit;
(c) an adder whose first input receives the output signal of said inverse orthogonal transform circuit;
(d) a memory for storing the image reconstituted in accordance with the reduced definition and being present at the output of said adder;
(e) a first circuit for compensating motion, receiving the output signal of said memory and the motion information components, and having its output connected to the second input of said adder;
(B) the prediction stage comprises:
(f) a first multiplier for multiplying the output signals of said decoding device by a coefficient a between 0 and 1;
(g) a second multiplier for multiplying the output signal of said stage for reconstructing the image in accordance with said reduced definition by the coefficient (1-.alpha.), a phase restoring and resampling circuit being arranged between said output and the corresponding input of said second multiplier;
(h) an adder for adding the output signals of said first and second multipliers;
(i) a memory for weighted mixing of images in accordance with said given definition and in accordance with said reduced definition;
(C) the motion compensation stage comprises:
(j) a second motion compensation circuit receiving the output signal of said prediction stage and the additional motion and coding mode information components;
(k) an adder for adding the output signals of said decoding chain and said second motion compensation circuit.
(A) the stage for reconstructing the image in accordance with said reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) an inverse orthogonal transform circuit;
(c) an adder whose first input receives the output signal of said inverse orthogonal transform circuit;
(d) a memory for storing the image reconstituted in accordance with the reduced definition and being present at the output of said adder;
(e) a first circuit for compensating motion, receiving the output signal of said memory and the motion information components, and having its output connected to the second input of said adder;
(B) the prediction stage comprises:
(f) a first multiplier for multiplying the output signals of said decoding device by a coefficient a between 0 and 1;
(g) a second multiplier for multiplying the output signal of said stage for reconstructing the image in accordance with said reduced definition by the coefficient (1-.alpha.), a phase restoring and resampling circuit being arranged between said output and the corresponding input of said second multiplier;
(h) an adder for adding the output signals of said first and second multipliers;
(i) a memory for weighted mixing of images in accordance with said given definition and in accordance with said reduced definition;
(C) the motion compensation stage comprises:
(j) a second motion compensation circuit receiving the output signal of said prediction stage and the additional motion and coding mode information components;
(k) an adder for adding the output signals of said decoding chain and said second motion compensation circuit.
8. A decoding device as claimed in Claim 6, in which the decoding stage is a variable length decoding chain for coded digital signals previously transmitted and/or stored after treatment in a variable length coding chain, said decoding chain comprising a memory circuit, a variable length decoding circuit, an inverse quantizing circuit, an inverse normalization circuit and an inverse orthogonal transform circuit, characterized in that, with a first memory for storing information components in accordance with said given definition being arranged at the output of said decoding chain, (A) the stage for reconstructing the image in accordance with said reduced definition comprises a series arrangement of:
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) a circuit for reconstituting the block in accordance with said reduced definition;
(c) an inverse orthogonal transform circuit;
(d) a second memory for storing information components in accordance with said reduced definition;
(B) the motion compensation stage comprises first arid second motion compensation circuits each receiving the output signal of one of said two memories for storing information components and for storing the coding mode and motion information components;
(C) the prediction stage comprises:
(f) at the output of that one of said motion compensation circuits which follows said memory for storing the information components in accordance with the given definition, a first reconstitution branch comprising a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the other one of said motion compensation circuits which follows said memory for storing the information components in accordance with the reduced definition, a second reconstitution branch comprising a second orthogonal transform circuit;
(h) at the output of said first and second parallel branches a circuit for combining the output signals of said first and second branches;
(i) an adder for adding the output signals of said combination circuit and of the inverse quantizing circuit of the variable length decoding chain, arranged between said inverse quantizing circuit and the inverse orthogonal transform circuit following the last-mentioned circuit.
(a) a clipper circuit for taking a given fraction of the signals after decoding;
(b) a circuit for reconstituting the block in accordance with said reduced definition;
(c) an inverse orthogonal transform circuit;
(d) a second memory for storing information components in accordance with said reduced definition;
(B) the motion compensation stage comprises first arid second motion compensation circuits each receiving the output signal of one of said two memories for storing information components and for storing the coding mode and motion information components;
(C) the prediction stage comprises:
(f) at the output of that one of said motion compensation circuits which follows said memory for storing the information components in accordance with the given definition, a first reconstitution branch comprising a first orthogonal transform circuit and a low-frequency clipper circuit for eliminating, in the coefficients resulting from said orthogonal transform, a given fraction representing the coefficients of the lowest frequency;
(g) at the output of the other one of said motion compensation circuits which follows said memory for storing the information components in accordance with the reduced definition, a second reconstitution branch comprising a second orthogonal transform circuit;
(h) at the output of said first and second parallel branches a circuit for combining the output signals of said first and second branches;
(i) an adder for adding the output signals of said combination circuit and of the inverse quantizing circuit of the variable length decoding chain, arranged between said inverse quantizing circuit and the inverse orthogonal transform circuit following the last-mentioned circuit.
9. A decoding device as claimed in Claim 8, characterized in that the prediction stage also comprises, between the output of said second reconstitution branch and the corresponding input of said combination circuit, a weighted mixing circuit comprising:
(a) a multiplier for multiplying the output signal of said first branch by a coefficient a between 0 and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient (1-.alpha.);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to said corresponding input of the combination circuit.
(a) a multiplier for multiplying the output signal of said first branch by a coefficient a between 0 and 1, which signal corresponds to said coefficients eliminated by the low-frequency clipper circuit;
(b) a multiplier for multiplying the output signal of said second branch by the coefficient (1-.alpha.);
(c) an adder for adding the output signals of said multipliers, the output signal of said adder being applied to said corresponding input of the combination circuit.
10. A decoding device as claimed in any one of Claims 7 to 9, characterized in that said given fraction is equal to one quarter.
Applications Claiming Priority (9)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR9012413A FR2667753B1 (en) | 1990-10-09 | 1990-10-09 | DEVICE FOR CODING DIGITAL SIGNALS CORRESPONDING TO TELEVISION IMAGES. |
| FR9012413 | 1990-10-09 | ||
| FR9015823 | 1990-12-18 | ||
| FR9015823A FR2670646B1 (en) | 1990-12-18 | 1990-12-18 | DEVICE FOR DECODING DIGITAL SIGNALS PREVIOUSLY TRANSMITTED AND / OR STORED. |
| FR9107656 | 1991-06-21 | ||
| FR9107656A FR2678126B1 (en) | 1991-06-21 | 1991-06-21 | DEVICE FOR CODING DIGITAL SIGNALS CORRESPONDING TO TELEVISION IMAGES. |
| FR9108588 | 1991-07-09 | ||
| FR9108588A FR2679092A1 (en) | 1991-07-09 | 1991-07-09 | Device for decoding digital signals which are coded then transmitted and/or stored |
| PCT/NL1991/000191 WO1992006563A2 (en) | 1990-10-09 | 1991-10-08 | Coding system for digital signals corresponding to television pictures and corresponding decoding system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2070827A1 CA2070827A1 (en) | 1992-04-10 |
| CA2070827C true CA2070827C (en) | 2002-01-01 |
Family
ID=27446777
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002070827A Expired - Fee Related CA2070827C (en) | 1990-10-09 | 1991-10-08 | Device for coding digital signals correspsonding to television images and corresponding decoding device |
Country Status (7)
| Country | Link |
|---|---|
| EP (1) | EP0504376B1 (en) |
| JP (1) | JPH05503403A (en) |
| KR (1) | KR920704513A (en) |
| CA (1) | CA2070827C (en) |
| DE (1) | DE69116128T2 (en) |
| FI (1) | FI922648A0 (en) |
| WO (1) | WO1992006563A2 (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5436665A (en) * | 1992-03-03 | 1995-07-25 | Kabushiki Kaisha Toshiba | Motion picture coding apparatus |
| DE69322769T2 (en) * | 1992-03-03 | 1999-07-22 | Kabushiki Kaisha Toshiba, Kawasaki, Kanagawa | CODE FOR CHANGEABLE IMAGES |
| KR100283343B1 (en) * | 1992-06-25 | 2001-03-02 | 이데이 노부유끼 | Image signal encoding method and decoding method, image signal encoding apparatus and decoding apparatus |
| NL9201594A (en) * | 1992-09-14 | 1994-04-05 | Nederland Ptt | System comprising at least one encoder for encoding a digital signal and at least one decoder for decoding an encoded digital signal, and encoder and decoder for use in the system. |
| NL9201593A (en) * | 1992-09-14 | 1994-04-05 | Nederland Ptt | System comprising a first encoder for encoding a first digital signal, a second encoder for encoding a second digital signal and at least one decoder for decoding encoded digital signals, and an encoder and decoder for use in the system. |
| NL9201640A (en) * | 1992-09-22 | 1994-04-18 | Nederland Ptt | System comprising at least one encoder for encoding a digital signal and at least one decoder for decoding a digital signal, and an encoder and decoder for use in the system according to the invention. |
| JP3374989B2 (en) * | 1993-03-26 | 2003-02-10 | ソニー株式会社 | Image signal encoding method and image signal encoding device, image signal decoding method and image signal decoding device |
| EP1701552B1 (en) * | 1994-06-08 | 2008-07-30 | Matsushita Electric Industrial Co., Ltd. | Image conversion apparatus |
| JP3788823B2 (en) | 1995-10-27 | 2006-06-21 | 株式会社東芝 | Moving picture encoding apparatus and moving picture decoding apparatus |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2638926A1 (en) * | 1988-11-10 | 1990-05-11 | Labo Electronique Physique | Method and device for digital coding of high-definition television pictures; corresponding decoding device, picture transmission system equipped with a stage for transmission and one for reception including such a device for coding or decoding transmission and reception stages comprising such devices |
| FR2643531B1 (en) * | 1989-02-21 | 1996-04-26 | Thomson Csf | INFORMATION COMPRESSION METHOD AND DEVICE FOR COMPATIBLE DECODING OF A FAMILY OF INCREASING RESOLUTIONS TELEVISION SIGNALS |
| US5057918A (en) * | 1989-09-15 | 1991-10-15 | U.S. Philips Corporation | Arrangement for encoding two-dimensional information, grouped in periodical information clusters using motion vector processing in a hybrid DPCM encoder |
| US4969040A (en) * | 1989-10-26 | 1990-11-06 | Bell Communications Research, Inc. | Apparatus and method for differential sub-band coding of video signals |
-
1991
- 1991-10-08 EP EP91918795A patent/EP0504376B1/en not_active Expired - Lifetime
- 1991-10-08 KR KR1019920701356A patent/KR920704513A/en not_active Ceased
- 1991-10-08 CA CA002070827A patent/CA2070827C/en not_active Expired - Fee Related
- 1991-10-08 WO PCT/NL1991/000191 patent/WO1992006563A2/en active IP Right Grant
- 1991-10-08 JP JP3517157A patent/JPH05503403A/en active Pending
- 1991-10-08 DE DE69116128T patent/DE69116128T2/en not_active Expired - Fee Related
-
1992
- 1992-06-08 FI FI922648A patent/FI922648A0/en not_active Application Discontinuation
Also Published As
| Publication number | Publication date |
|---|---|
| DE69116128D1 (en) | 1996-02-15 |
| FI922648A7 (en) | 1992-06-08 |
| FI922648A0 (en) | 1992-06-08 |
| CA2070827A1 (en) | 1992-04-10 |
| WO1992006563A3 (en) | 1992-05-14 |
| DE69116128T2 (en) | 1996-08-01 |
| JPH05503403A (en) | 1993-06-03 |
| EP0504376A1 (en) | 1992-09-23 |
| WO1992006563A2 (en) | 1992-04-16 |
| EP0504376B1 (en) | 1996-01-03 |
| KR920704513A (en) | 1992-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6990241B2 (en) | Circuit and method for decoding an encoded version of an image having a first resolution directly into a decoded version of the image having a second resolution | |
| US6507615B1 (en) | Video signal processing apparatus and video signal processing method | |
| EP0579395B1 (en) | Spatial/frequency hybrid video coding facilitating the derivation of variable-resolution images | |
| EP0690392B1 (en) | Method and device for transcoding a sequence of coded digital signals | |
| EP0564597B1 (en) | Systems and methods for coding even fields of interlaced video sequences | |
| US5786859A (en) | Video coding/decoding apparatus with preprocessing to form predictable coded video signal portion | |
| EP0595599B1 (en) | Compression/decompression system for image data | |
| US6233277B1 (en) | Reduced-memory video decoder for compressed high-definition video data | |
| JP2570384B2 (en) | Video signal encoding / decoding method | |
| EP0608092A2 (en) | Apparatus for scalable encoding and decoding of a digital video signal | |
| EP1442606A1 (en) | Spatial scalable compression | |
| KR20010092797A (en) | Hybrid temporal-SNR fine granular scalability video coding | |
| US5528299A (en) | Coding system for digital signals corresponding to television pictures and corresponding decoding system | |
| US6950466B2 (en) | Apparatus for receiving moving pictures | |
| CA2070827C (en) | Device for coding digital signals correspsonding to television images and corresponding decoding device | |
| EP0734176A2 (en) | Motion compensation apparatus for use in a video encoding system | |
| KR19980017213A (en) | Image Decoding System with Compensation Function for Degraded Image | |
| JPH08280022A (en) | Method and apparatus for decoded video signal | |
| EP1641274A1 (en) | VIDEO CODING EMPLOYING An image DE-GRAINING LOOP-FILTER | |
| RU2154917C2 (en) | Improved final processing method and device for image signal decoding system | |
| JP4028900B2 (en) | Moving picture coding apparatus and moving picture decoding apparatus | |
| EP0651582B1 (en) | Device for transmitting television pictures and device for receiving said pictures | |
| JP2000036963A (en) | Image coder, image coding method and image decoder | |
| US6061401A (en) | Method and apparatus for selectively encoding/decoding a video signal | |
| JP3432886B2 (en) | Hierarchical encoding / decoding apparatus and method, and transmission / reception system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKLA | Lapsed |