3d-video codec supporting inter-component prediction

3D video encoder comprising

• a video encoding core configured to encode a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• an inter-component prediction switch configured to
• insert a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture, and indicating, for a current picture of a current layer which represents texture, a texture reference layer set of layers representing texture and a depth reference layer set of layers representing depth,
• insert a second parameter set into the data stream, the second parameter set relating to the current picture or a portion of the current picture, and indicating, for coding units within the current picture or a portion of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set,
• if an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand, equals the potentially available set, then deciding whether the depth-to-texture prediction is enabled or disabled for the current picture or the portion of the picture and inserting a flag into the data stream, the flag relating to the current picture or portion of the current picture and indicating whether the depth-to-texture prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• if the intersection is unequal to the potentially available set, suppressing inserting the flag relating to the coding units within the current picture or the portion of the current picture and disabling the depth-to-texture prediction for the coding units,
• wherein the video encoding core is configured to, depending on the depth-to-texture prediction being enabled or disabled for the current coding unit, apply or not apply depth-to-texture prediction for a current coding unit among the coding units within the current picture or the portion of the current picture.

3D video decoder comprising

• a video decoding core configured to decode a sequence of layers of a video from a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for a layer representing depth of a current view, prediction from layers representing depth of a different view and texture-to-depth prediction from a layer representing texture,
• an inter-component prediction switch configured to
• read a first parameter set from the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture of the layer representing depth of the current view,
• if an intersection of a potentially available set of one or more layers representing texture of one or more views the depth of which is represented by the layer representing depth of the current view, on the one hand and a texture reference layer set of layers representing texture determined by the first parameter set, on the other hand equals the potentially available set, read a flag from the data stream, the flag relating to the current picture or a portion of the current picture and indicating whether the texture-to-depth prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture,
• if the intersection is unequal to the potentially available set, then infer the flag relating to the coding units within the current picture or the portion of the current picture as indicating that the texture-to-depth prediction is disabled for the coding units,
• wherein the video decoding core is configured to be responsive to the flag in order to, depending on the flag, apply or not apply texture-to-depth prediction for a current coding unit among the coding units within the current picture or the portion of the current picture.

3D video encoder comprising

• a video encoding core configured to encode a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing depth of a current view, prediction from layers representing depth of a different view and texture-to-depth prediction from a layer representing depth of the current layer,
• an inter-component prediction switch configured to
• insert a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture of the layer representing depth of the current view, and indicating, for the current picture, a texture reference layer set of layers representing texture,
• if an intersection of a potentially available set of one or more layers representing texture of one or more views the depth of which is represented by the layer representing depth of the current view, on the one hand and the texture reference layer set, on the other hand equals the potentially available set, deciding whether the texture-to-depth prediction is enabled or disabled for the coding units within the current picture or a portion of the current picture, and inserting a flag into the data stream, the flag relating to the current picture or portion of the current picture and indicating whether the texture-to-depth prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• wherein the video encoding core is configured to, depending on the depth-to-texture prediction being enabled or disabled for the current coding unit, apply or not apply texture-to-depth prediction for a current coding unit among the coding units within the current picture or the portion of the current picture.

Method for 3D video decoding, comprising

• decoding, performed by a video decoding core , a sequence of layers of a video from a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view,
• reading a first parameter set from the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture, and derive therefrom, for a current picture of a current layer which represents texture, a texture reference layer set of layers representing texture and a depth reference layer set of layers representing depth,
• reading a second parameter set from the data stream, the second parameter set relating to the current picture or a portion of the current picture, and derive therefrom, for coding units within the current picture or the portion of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set,
• if an intersection of
• a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and a depth reference layer set of layers representing depth determined by the first parameter set, on the other hand
• equals the potentially available set, then reading a flag from the data stream, the flag relating to the current picture or the portion of the current picture and indicating whether the depth-to-texture prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• if the intersection is unequal to the potentially available set, then inferring the flag relating to the coding units within the current picture or the portion of the current picture as indicating that the depth-to-texture prediction is disabled for the coding units,
• wherein the video decoding core, depending on the flag, applies or not applies depth-to-texture prediction for a current coding unit among the coding units within the current picture or the portion of the current picture.

Method for 3D video encoding, comprising

• encoding, performed by a video encoding core, a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• inserting a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture, and indicating, for a current picture of a current layer which represents texture, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• inserting a second parameter set into the data stream, the second parameter set relating to the current picture or a portion of the current picture, and indicating, for coding units within the current picture or a portion of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set,
• if an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand, equals the potentially available set, then deciding whether the depth-to-texture prediction is enabled or disabled for the current picture or the portion of the picture and inserting a flag into the data stream, the flag relating to the current picture or portion of the current picture and indicating whether the depth-to-texture prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• if the intersection is unequal to the potentially available set, suppressing inserting the flag relating to the coding units within the current picture or the portion of the current picture and disabling the depth-to-texture prediction for the coding units,
• wherein the video encoding core, depending on the depth-to-texture prediction being enabled or disabled for the current coding unit, applies or not applies depth-to-texture prediction for a current coding unit among the coding units within the current picture or the portion of the current picture.

A non-transitory digital storage medium having a computer program stored thereon to perform the method for 3D video decoding, the method comprising

• decoding, performed by a video decoding core, a sequence of layers of a video from a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view,
• reading a first parameter set from the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture, and derive therefrom, for a current picture of a current layer which represents texture, a texture reference layer set of layers representing texture and a depth reference layer set of layers representing depth,
• reading a second parameter set from the data stream, the second parameter set relating to the current picture or a portion of the current picture, and derive therefrom, for coding units within the current picture or the portion of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set,
• if an intersection of
• a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and a depth reference layer set of layers representing depth determined by the first parameter set, on the other hand
• equals the potentially available set, then reading a flag from the data stream, the flag relating to the current picture or the portion of the current picture and indicating whether the depth-to-texture prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• if the intersection is unequal to the potentially available set, then inferring the flag relating to the coding units within the current picture or the portion of the current picture as indicating that the depth-to-texture prediction is disabled for the coding units,
• wherein the video decoding core, depending on the flag, applies or not applies depth-to-texture prediction for a current coding unit among the coding units within the current picture or the portion of the current picture,
• when said computer program is run by a computer.

A non-transitory digital storage medium having a computer program stored thereon to perform the method for 3D video encoding, the method comprising

• encoding, performed by a video encoding core, a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• inserting a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture, and indicating, for a current picture of a current layer which represents texture, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• inserting a second parameter set into the data stream, the second parameter set relating to the current picture or a portion of the current picture, and indicating, for coding units within the current picture or a portion of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set,
• if an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand, equals the potentially available set, then deciding whether the depth-to-texture prediction is enabled or disabled for the current picture or the portion of the picture and inserting a flag into the data stream, the flag relating to the current picture or portion of the current picture and indicating whether the depth-to-texture prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• if the intersection is unequal to the potentially available set, suppressing inserting the flag relating to the coding units within the current picture or the portion of the current picture and disabling the depth-to-texture prediction for the coding units,
• wherein the video encoding core, depending on the depth-to-texture prediction being enabled or disabled for the current coding unit, applies or not applies depth-to-texture prediction for a current coding unit among the coding units within the current picture or the portion of the current picture,
• when said computer program is run by a computer.

Digital storage medium having stored thereon a data stream generated by a method for 3D video encoding, the method comprising

• encoding, performed by a video encoding core, a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• inserting a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video relating to a timestamp of a current picture, and indicating, for a current picture of a current layer which represents texture, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• inserting a second parameter set into the data stream, the second parameter set relating to the current picture or a portion of the current picture, and indicating, for coding units within the current picture or a portion of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set,
• if an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand, equals the potentially available set, then deciding whether the depth-to-texture prediction is enabled or disabled for the current picture or the portion of the picture and inserting a flag into the data stream, the flag relating to the current picture or portion of the current picture and indicating whether the depth-to-texture prediction is enabled or disabled for the coding units within the current picture or the portion of the current picture, and
• if the intersection is unequal to the potentially available set, suppressing inserting the flag relating to the coding units within the current picture or the portion of the current picture and disabling the depth-to-texture prediction for the coding units,
• wherein the video encoding core, depending on the depth-to-texture prediction being enabled or disabled for the current coding unit, applies or not applies depth-to-texture prediction for a current coding unit among the coding units within the current picture or the portion of the current picture.

3D video decoder comprising

• a video decoding core configured to decode a sequence of layers of a video from a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• an inter-component prediction switch configured to
• read a first parameter set from the data stream, the first parameter set relating to a temporal portion of the video, and derive therefrom, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture,
• read a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and derive therefrom, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp, and, on the basis of the selected texture reference layer set, an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and a depth reference layer set of layers representing depth determined by the first parameter set, on the other hand,
• read, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video decoding core is configured to be responsive to the flags for the layers of the intersection in order to, depending on the flags, apply or not apply depth-to-texture prediction using one or more of the layers of the intersection as depth-to-texture prediction reference for the current coding unit.

3D video encoder comprising

• a video encoding core configured to encode a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video encoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• an inter-component prediction switch configured to
• insert a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video, and indicating, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• insert a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and indicating, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp,
• derive an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand, insert, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video encoding core is configured to, depending on whether the depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference is enabled or disabled for the current coding unit, apply or not apply depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference for the current coding unit.

Method for 3D video decoding, comprising

• decoding, performed by a video decoding core, a sequence of layers of a video from a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• reading a first parameter set from the data stream, the first parameter set relating to a temporal portion of the video, and derive therefrom, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• reading a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and derive therefrom, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp, and, on the basis of the selected texture reference layer set, an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand,
• reading, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video decoding core depending on the flags for the layers of the intersection applies or not applies depth-to-texture prediction using one or more of the layers of the intersection as depth-to-texture prediction reference for the current coding unit.

Method for 3D video encoding, comprising

• Encoding, performed by a video encoding core, a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video encoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• inserting a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video, and indicating, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• inserting a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and indicating, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp,
• deriving an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand,
• inserting, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video encoding core, depending on whether the depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference is enabled or disabled for the current coding unit, applies or not applies depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference for the current coding unit.

A non-transitory digital storage medium having a computer program stored thereon to perform the method for 3D video decoding, the method comprising

• decoding, performed by a video decoding core, a sequence of layers of a video from a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video decoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• reading a first parameter set from the data stream, the first parameter set relating to a temporal portion of the video, and derive therefrom, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• reading a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and derive therefrom, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp, and, on the basis of the selected texture reference layer set, an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand,
• reading, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video decoding core depending on the flags for the layers of the intersection applies or not applies depth-to-texture prediction using one or more of the layers of the intersection as depth-to-texture prediction reference for the current coding unit,
• when said computer program is run by a computer.

A non-transitory digital storage medium having a computer program stored thereon to perform the method for 3D video encoding, the method comprising

• encoding, performed by a video encoding core, a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video encoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• inserting a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video, and indicating, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• inserting a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and indicating, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp,
• deriving an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand,
• inserting, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video encoding core, depending on whether the depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference is enabled or disabled for the current coding unit, applies or not applies depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference for the current coding unit,
• when said computer program is run by a computer.

Digital storage medium having stored thereon a data stream generated by a method for 3D video encoding, the method comprising

• encoding, performed by a video encoding core, a sequence of layers of a video into a data stream using intra-layer prediction, each layer representing depth or texture of a respective one of a plurality of views, the video encoding core supporting, for layers representing texture, inter-view texture prediction from layers representing texture of a different view and depth-to-texture prediction from layers representing depth,
• inserting a first parameter set into the data stream, the first parameter set relating to a temporal portion of the video, and indicating, for a current picture of a current layer which represents texture, the timestamp of which is comprised within the temporal portion, a texture reference layer set of layers representing texture and a depth reference layer set of layers relating to depth,
• inserting a second parameter set from the data stream, the second parameter set relating to a portion of the current picture or a portion of the current picture, and indicating, for a current coding unit of the current picture, a selected texture reference layer set of layers representing texture from the texture reference layer set, each layer of the selected texture reference layer set comprising a picture at the current picture's timestamp,
• deriving an intersection of a potentially available set of layers representing depth of views the texture of which is represented by any of the selected texture reference layer set, on the one hand and the depth reference layer set, on the other hand,
• inserting, per layer of the intersection, a flag from the data stream, the flag relating, at least, to the current coding unit and indicating whether the depth-to-texture prediction using the respective layer as depth-to-texture prediction reference is enabled or disabled for the current coding unit, and
• wherein the video encoding core, depending on whether the depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference is enabled or disabled for the current coding unit, applies or not applies depth-to-texture prediction using the layers of the intersection as depth-to-texture prediction reference for the current coding unit.