摘要:
A display mixer exports a first set of functions adapted to be called by one or more object libraries to create one or more generic objects. A surface/attribute manager exports a second set of functions adapted to be called by an application to manipulate the generic objects. The display mixer and the surface/attribute manager cause the generic objects to be rendered into a common draw surface. The display mixer and the surface/attribute manager manage a set of common attributes shared by the generic objects and manage attachments between the generic objects. A first object library calls a create-object function in the display mixer to create one or more first generic objects. The first object library generates display data for the first generic objects and the display mixer coordinates the display of the display data for the first generic objects.
摘要:
Statistics are generated for a set of signals, where each signal is represented by a palette index. A remap table is generated based on frequency of occurrence of each palette index in the set of signals. The remap table is applied to the palette indices to map the set of signals into a set of remapped values. An encoded bitstream is generated based on the remapped values. In a preferred embodiment, a differencing operation is applied to the remapped values to generate difference signals for the set of signals. Statistical encoding is applied to the difference signals to generate variable-length codes for the set of signals. The encoded bitstream is generated for the set of signals from the variable-length codes. The encoded bitstream is decoded by reversing the encode processing, using the remap table.
摘要:
Digital video data is transmitted by sending significant pixel data differences, column by column, within the same frame rather than differences between corresponding pixel data of adjacent frames; a "no change" signal is transmitted when a difference between corresponding pixel data of adjacent frames is not significant (i.e. less than a fixed minimum threshold value). This method eliminates the necessity of storing a previous frame's data at the receiver while computing the subsequent frame's data. The technique also eliminates "bleed-through" artifacts sometimes experienced in the prior art during scene transitions.
摘要:
First-pass processing is applied to a video sequence comprising a plurality of video frames to generate first-pass processing results for the video sequence. After completing the first-pass processing, subsequent-pass processing is applied to the video sequence using the first-pass processing results to generate subsequent-pass processing results. An encoded bitstream is generated for the video sequence based on the subsequent-pass processing results.
摘要:
Graphical, video, and audio dam is integrated into a single processing environment. The present invention employs an integrated graphics/video controller (IVC) which interfaces with application software through a graphics API and a video API. The IVC receives graphics commands through the graphics API and video commands through the video API. A mask driver produces information from the graphics commands including clipping information, graphics information and mask information. A blender uses the mask information, the graphics information, and the clipping information for combining or compositing graphics images with video images. The video commands of the video command stream provide functions for configuring the operation of the IVC. These functions include commands for loading software video decoders in a decoder block within the IVC. Video data transferred to the IVC via the video API may be encoded in a variety of different formats. The present invention provides a means for dynamically loading a plurality of different video decoders through a video command interface to a video decode block both within the IVC. Each of the independent decoders within the decode block contain processing logic for decoding a particular type of video data to produce a uniform type of decoded video data which is provided to the blender. The blender receives the decoded video data and combines the video data with graphics data as defined by the mask information, the clipping information, and the graphics information.
摘要:
One or more streams of signals (e.g., audio/video sequences) are associated with the different possible processing paths of processing intersections of an interactive computer application. For example, in a computer-based video game, the flow of processing may approach an intersection where the user may select any one of a number of different paths, each path being associated with a different audio/video sequence corresponding to that path. As the flow of the application progresses towards the intersection, the different audio/video sequences associated with the different paths of that intersection are preprocessed. Preprocessing may include preloading the audio/video sequences and optionally partially decompressing the audio/video sequences. When the flow of the application reaches the intersection, one of the possible paths is selected based on the actions taken by the user. The application causes the audio/video sequence associated with the selected path to be played and the other sequences to be dropped. In this way, the interactive application is provided with smooth transitions at processing intersections. The invention avoids the delays that would otherwise result at a processing intersection from opening the audio/video file associated with the selected path and beginning to decompress the compressed signals contained in that file.