摘要:
Innovations in the area of encoding or decoding of blocks using intra block copy (“BC”) prediction improve coding efficiency for intra-BC-predicted blocks in many situations. For example, some of the innovations relate to block flipping in which an intra-BC-predicted block is flipped relative to a reference region, which can be indicated by a block vector (“BV”) value. Other innovations relate to signaling of a skip mode in which a current intra-BC-predicted block uses a signaled BV differential but lacks residual data.
摘要:
Embodiments for reducing redundancy in video coding are disclosed. In accordance with at least one embodiment, video content is represented as a tree structure in which the nodes of the tree structure are associated with attributes of the video content. An encoder analyzes the nodes to determine whether there is redundancy associated with a last node of a divided node. In another embodiment, bidirectional prediction video coding techniques are applied to the video content. The encoder analyzes two lists of reference frames to determine whether a third mode of bidirectional prediction is redundant to a second mode of bidirectional prediction. The efficiency of video coding is improved by refraining from sending data symbols to the decoder in the event that the video coding techniques contain redundancy.
摘要:
Innovations are provided for encoding and/or decoding video and/or image content using dictionary modes. For example, some innovations predict current pixel values from previous pixel values stored in a 1-D dictionary. Other innovations predict current pixel values from previous pixel values using a pseudo 2-D dictionary mode. Yet other innovations predict current pixel values from previous pixel values in a reference picture using an inter pseudo 2-D dictionary mode. Pixel values can be predicted from previous pixel values (e.g., stored in a dictionary) that are identified by an offset and a length. Yet other innovations encode pixel values using hash matching of pixel values.
摘要:
Embodiments for reducing redundancy in video coding are disclosed. In accordance with at least one embodiment, video content is represented as a tree structure in which the nodes of the tree structure are associated with attributes of the video content. An encoder analyzes the nodes to determine whether there is redundancy associated with a last node of a divided node. In another embodiment, bidirectional prediction video coding techniques are applied to the video content. The encoder analyzes two lists of reference frames to determine whether a third mode of bidirectional prediction is redundant to a second mode of bidirectional prediction. The efficiency of video coding is improved by refraining from sending data symbols to the decoder in the event that the video coding techniques contain redundancy.
摘要:
Techniques for implementing video encoding enhancements may increase video encoding efficiency. One of the techniques may involve the use of an exception value and a maximum index value in inter-prediction coding lookup table to enhance the combined coding of an inter-prediction direction and a reference frame index for a coding unit. Another of the techniques may include adaptively ordering the encoding of a split flag value and a skip flag value for a coding unit. An additional technique may include providing a uniform approach to adaptively combined code the coded block flag (CBF) values for transform or coding units of images that are used with the Residue Quad-tree (RQT) information to represent residue information.
摘要:
Techniques for implementing video encoding enhancements may increase video encoding efficiency. One of the techniques may involve the use of an exception value and a maximum index value in inter-prediction coding lookup table to enhance the combined coding of an inter-prediction direction and a reference frame index for a coding unit. Another of the techniques may include adaptively ordering the encoding of a split flag value and a skip flag value for a coding unit. An additional technique may include providing a uniform approach to adaptively combined code the coded block flag (CBF) values for transform or coding units of images that are used with the Residue Quad-tree (RQT) information to represent residue information.
摘要:
Innovations in the area of encoding blocks using hash-based block matching improve coding efficiency in many situations. For example, some of the innovations relate to encoder-side decisions about when to use block flipping during prediction. Other innovations relate to encoder-side decisions about when to use skip mode for a block predicted using intra block copy (“BC”) prediction.
摘要:
Innovations in the areas of hash table construction and availability checking reduce computational complexity of hash-based block matching. For example, some of the innovations speed up the process of constructing a hash table or reduce the size of a hash table. This can speed up and reduce memory usage for hash-based block matching within a picture (for block vector estimation) or between different pictures (for motion estimation). Other innovations relate to availability checking during block vector estimation that uses a hash table.
摘要:
Architecture that improves error robustness in video coding and decoding. In particular, this can apply to motion vector prediction (MVP) such as a temporal MVP (TMVP). Flags can be used to indicate the use or non-use of a feature, such as to indicate whether the current slice uses or does not use TMVP, and to indicate in the slice header whether list prediction is allowed or not allowed. A flag can be signaled in sequence parameter set (SPS) or picture parameter set (PPS) as a way to enable an entire sequence to use or not use TMVP. TVMP can also be used to copy all the reference motion information to the current block. To address possible error problems, the full index of the TMVP can be recorded, and temporal information decoding refresh (TIDR) can be inserted into slices periodically.
摘要:
Innovations in encoder-side decisions for coding of screen content video or other video can speed up encoding in various ways. For example, some of the innovations relate to ways to speed up motion estimation by identifying appropriate starting points for the motion estimation in different reference pictures. Many of the encoder-side decisions speed up encoding by terminating encoding for a block or skipping the evaluation of certain modes or options when a condition is satisfied. For example, some of the innovations relate to ways to speed up encoding when hash-based block matching is used. Still other innovations relate to ways to identify when certain intra-picture prediction modes should or should not be evaluated during encoding. Other innovations relate to other aspects of encoding.