Quality Scalable Video Data Stream

Abstract

An apparatus for generating a quality-scalable video data stream includes a coder for coding a video signal using block-wise transformation to obtain transform blocks of transformation coefficient values for a picture of the video signal, a predetermined scan order with possible scan positions being defined among the transformation coefficient values within the transform blocks so that in each transform block, for each possible scan position, at least one of the transformation coefficient values within the respective transform block belongs to the respective possible scan position; and a generator for forming, for each of a plurality of quality layers, a video sub-data stream containing scan range information indicating a sub-set of the possible scan positions, and transform coefficient information on transformation coefficient values belonging to the sub-set of possible scan positions such that the sub-set of each quality layer includes at least one possible scan position not included by the sub-set of any other of the plurality of quality layers.

1 - 42 . (canceled) 
     
     
         43 . An apparatus for generating a quality-scalable video data stream, comprising:
 a coder for coding a video signal using block-wise transformation to acquire transform blocks of two-dimensionally arranged transformation coefficient values for a picture of the video signal, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values; and   a generator for forming, for each of a plurality of quality layers, a video sub-data stream comprising scan range information indicating a sub-set of the possible scan positions, such that the sub-set of each of the plurality of quality layers comprises at least one possible scan position not comprised by the sub-set of any other of the plurality of quality layers and one of the possible scan positions is comprised by more than one of the sub-sets of the quality layers, and transform coefficient information on transformation coefficient values belonging to the sub-set of possible scan positions of the respective quality layer, comprising a contribution value per possible scan position of the sub-set of possible scan positions of the respective quality layer, such that the transform coefficient value of the one possible scan position is derivable based on a sum of the contribution values for the one possible scan position of the more than one of the sub-sets of the quality layers.   
     
     
         44 . An apparatus for reconstructing a video signal from a quality-scalable video data stream comprising, for each of a plurality of quality layers, a video sub-data stream, comprising:
 a parser for parsing the video sub-data streams of the plurality of quality layers, to acquire, for each quality layer, a scan range information and transform coefficient information on two-dimensionally arranged transformation coefficient values of different transform blocks, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values, and the scan range information indicates a sub-set of the possible scan positions;   a constructor for, using the scan range information, for each quality layer, constructing the transform blocks by associating the transformation coefficient values of the respective transform I blocks from the transform coefficient information to the sub-set of the possible scan positions; and   a reconstructor for reconstructing a picture of the video signal by a back-transformation of the transform blocks, wherein the parser is configured such that the transform coefficient information of more than one of the quality layers may comprise a contribution value relating to one transformation coefficient value, and wherein the constructor is configured to derive the value for the one the transform coefficient value based on a sum of the contribution values relating to the one transformation coefficient value.   
     
     
         45 . The apparatus according to  claim 44 , wherein the constructor is configured to use the scan range information as if same would indicate a first scan position among the possible scan positions within the sub-set of possible scan positions in the predetermined scan order. 
     
     
         46 . The apparatus according to  claim 44 , wherein the constructor is configured to use the scan range information as if same would indicate a last scan position among the possible scan positions within the sub-set of possible scan positions in the predetermined scan order. 
     
     
         47 . The apparatus according to  claim 44 , wherein among the transform blocks, at least one first transform block is comprised of 4×4 transformation coefficient values and at least one second transform block is comprised of 8×8 transformation coefficient values, and wherein the parser is configured to expect that the scan range information for each of the plurality of quality layers indicates the sub-set of possible scan positions for the at least one first transform block in an accuracy enabling a distinction between all 16 possible scan positions of the at least one first transform block, and for the at least one second transform block in an accuracy enabling a distinction between all 16 quadruples of consecutive transform coefficients of the at least one second transform block. 
     
     
         48 . The apparatus according to  claim 44 , wherein the parser is configured to expect the transformation coefficient values belonging to the sub-set of possible scan positions to be block-wise coded into the transform coefficient information so that the transformation coefficient value of a predetermined transform block form a continuous portion of the transform coefficient information. 
     
     
         49 . The apparatus according to  claim 48 , wherein the parser is configured to decode the consecutive portion by
 decoding a significance map specifying positions of the transformation coefficient values being unequal to zero and belonging to the sub-set of possible scan positions in the predetermined transform block into the video sub-data stream, and subsequently   in a reverse scan order reversed relative to the predetermined scan order—starting with the last transformation coefficient value being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block—decoding the transform coefficient values being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block.   
     
     
         50 . The apparatus according to  claim 49 , wherein the parser is configured to decode the significance map by
 in the predetermined scan order, decoding a significance flag per transformation coefficient value belonging to the sub-set of possible scan positions from the first transformation coefficient value belonging to the sub-set of possible scan positions to the last transform coefficient value belonging to the sub-set of possible scan positions and being unequal to zero, with the significance flags depending on the respective transformation coefficient value being zero or unequal to zero, and   following each significance flag of a respective transformation coefficient value being unequal to zero, decoding a last-flag depending on the respective transformation coefficient value being the last transformation coefficient value belonging to the sub-sets of possible scan positions within the predetermined transform block being non-zero or not.   
     
     
         5 . The apparatus according to  claim 48 , wherein the parser is configured to decode the consecutive portion by
 decoding a significance information specifying the number of transformation coefficient values being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block as well as the number of consecutive trailing transformation coefficient values comprising an absolute value of one within the number of transformation coefficient values being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block;   decoding the signs of the consecutive trailing transformation coefficient values and the remaining transformation coefficient values being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block;   decoding the total number of transformation coefficient values being equal to zero and belonging to the sub-set of possible scan positions up to the last transformation coefficient value being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block;   decoding the number of consecutive transformation coefficient values being equal to zero and belonging to the sub-set of possible scan positions immediately preceding any of the number of transformation coefficient values being unequal to zero and belonging to the sub-set of possible scan positions within the predetermined transform block in a reversed scan order.   
     
     
         52 . The apparatus according to  claim 44 , wherein the predetermined scan order scans the transformation coefficient values of the transform blocks such that transformation coefficient values belonging to a higher scan position in the predetermined scan order relate to higher spatial frequencies. 
     
     
         53 . The apparatus according to  claim 44 , wherein the reconstructor is configured to reconstruct the video signal using motion-prediction based on motion information and by combining a motion-prediction result with a motion-prediction residual with acquiring the motion-prediction residual by a block-wise inverse transformation of the transform blocks of transformation coefficient values. 
     
     
         54 . The apparatus according to  claim 53 , wherein the parser is configured to expect each sub-data stream to comprise an indication indicating motion information existence or motion information non-existence for the respective quality layer, and that the sub-data stream of a first of the quality layers comprises the motion information and comprises the indication indicating motion information existence, or the indication within the sub-data stream of the first quality layer indicates the motion information non-existence, with a part of the quality-scalable video data stream other than the sub-data streams comprising the motion information, and to expect the sub-data stream(s) of the other quality layer(s) to comprise the indication indicating motion information non-existence. 
     
     
         55 . The apparatus according to  claim 54 , wherein the parser is configured to expect the sub-data stream of the first quality layer to comprise the indication indicating motion information existence, with the motion information being equal to the higher-quality motion information or equal to a refinement information allowing a reconstruction of the higher-quality motion information based on the lower-quality motion information, and that the part of the quality-scalable video data stream also comprises the lower-quality motion information. 
     
     
         56 . The apparatus according to  claim 54 , wherein the parser is configured such that the motion information and the indication relate to a macroblock of the picture. 
     
     
         57 . The apparatus according to  claim 44 , wherein the parser is configured to parse each sub-data stream individually independently—with regard to a parsing result—from the other sub-data stream(s). 
     
     
         58 . The apparatus according to  claim 57 , wherein the constructor is configured to associate the respective transform coefficient information with the transformation coefficient values, with the association result being independent of the other sub-data stream(s). 
     
     
         59 . The apparatus according to  claim 58 , wherein a layer order is defined among the quality layers, and the sub-data stream of a first quality layer in the layer order enables an association of the respective transform coefficient information with the transformation coefficient values independent of the sub-data stream(s) of the following quality layer(s), whereas the sub-data stream(s) of the following quality layers in layer order enable an association of the respective transform coefficient information with the transformation coefficient values merely in combination with the sub-data stream(s) of (a) quality layer(s) preceding the respective quality layer, wherein the constructor is configured to associate the transform coefficient information of a respective quality layer with the transformation coefficient values by use of the sub-data streams of the respective quality layer and quality layer(s) preceding the respective quality layer. 
     
     
         60 . A method for generating a quality-scalable video data stream, comprising:
 coding a video signal using block-wise transformation to acquire transform blocks of two-dimensionally arranged transformation coefficient values for a picture of the video signal, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values; and   forming, for each of a plurality of quality layers, a video sub-data stream comprising scan range information indicating a sub-set of the possible scan positions, such that the sub-set of each of the plurality of quality layers comprises at least one possible scan position not comprised by the sub-set of any other of the plurality of quality layers and one of the possible scan positions is comprised by more than one of the sub-sets of the quality layers, and transform coefficient information on transformation coefficient values belonging to the sub-set of possible scan positions of the respective quality layer, comprising a contribution value per possible scan position of the sub-set of possible scan positions of the respective quality layer such that the transform coefficient value of the one possible scan position is derivable based on a sum of the contribution values for the one possible scan position of the more than one of the sub-sets of the quality layers.   
     
     
         61 . A method for reconstructing a video signal from a quality-scalable video data stream comprising, for each of a plurality of quality layers, a video sub-data stream, comprising:
 parsing the video sub-data streams of the plurality of quality layers, to acquire, for each quality layer, a scan range information and transform coefficient information on two-dimensionally arranged transformation coefficient values of different transform blocks, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values, and the scan range information indicates a sub-set of the possible scan positions;   using the scan range information, for each quality layer, constructing the transform blocks by associating the transformation coefficient values of the respective transform blocks from the transform coefficient information to the sub-set of the possible scan positions; and   reconstructing a picture of the video signal by a back-transformation of the transform blocks, wherein   the parsing the video sub-data streams is performed such that the transform coefficient information of more than one of the quality layers may comprise a contribution value relating to one transformation coefficient value, and the constructing the transform blocks comprises deriving the value for the one transform coefficient value based on a sum of the contribution values relating to the one transformation coefficient value.   
     
     
         62 . A quality-scalable Video data stream enabling a reconstruction of a video signal comprising, for each of a plurality of quality layers, a scan range information and transform coefficient information on two-dimensionally arranged transformation coefficient values of different transform blocks, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values, and the scan range information indicates a sub-set of the possible scan positions, wherein the transform coefficient information concerns transformation coefficient values belonging to the sub-set of possible scan positions, wherein the transform coefficient information of more than one of the quality layers comprises a contribution value relating to one transformation coefficient value, and the transform coefficient value of the one possible scan position is derivable based on a sum of the contribution values for the one possible scan position of the more than one of the sub-sets of the quality layers. 
     
     
         63 . A computer-program comprising a program code for performing, when running on a computer, a method for generating a quality-scalable video data stream, the method comprising:
 coding a video signal using block-wise transformation to acquire transform blocks of two-dimensionally arranged transformation coefficient values for a picture of the video signal, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values; and   forming, for each of a plurality of quality layers, a video sub-data stream comprising scan range information indicating a sub-set of the possible scan positions, such that the sub-set of each of the plurality of quality layers comprises at least one possible scan position not comprised by the sub-set of any other of the plurality of quality layers and one of the possible scan positions is comprised by more than one of the sub-sets of the quality layers, and transform coefficient information on transformation coefficient values belonging to the sub-set of possible scan positions of the respective quality layer, comprising a contribution value per possible scan position of the sub-set of possible scan positions of the respective quality layer such that the transform coefficient value of the one possible scan position is derivable based on a sum of the contribution values for the one possible scan position of the more than one of the sub-sets of the quality layers.   
     
     
         64 . A computer-program comprising a program code for performing, when running on a computer, a method for reconstructing a video signal from a quality-scalable video data stream comprising, for each of a plurality of quality layers, a video sub-data stream, the method comprising:
 parsing the video sub-data streams of the plurality of quality layers, to acquire, for each quality layer, a scan range information and transform coefficient information on two-dimensionally arranged transformation coefficient values of different transform blocks, wherein a predetermined scan order with possible scan positions orders the transformation coefficient values into a linear sequence of transformation coefficient values, and the scan range information indicates a sub-set of the possible scan positions;   using the scan range information, for each quality layer, constructing the transform blocks by associating the transformation coefficient values of the respective transform blocks from the transform coefficient information to the sub-set of the possible scan positions; and   reconstructing a picture of the video signal by a back-transformation of the transform blocks, wherein   the parsing the video sub-data streams is performed such that the transform coefficient information of more than one of the quality layers may comprise a contribution value relating to one transformation coefficient value, and the constructing the transform blocks comprises deriving the value for the one transform coefficient value based on a sum of the contribution values relating to the one transformation coefficient value.