Abstract: Frequency segmentation is important to the quality of encoding spectral data. Segmentation involves breaking the spectral data into units called sub-bands or vectors. Homogeneous segmentation may be suboptimal. Various features are described for providing spectral data intensity dependent segmentation. Finer segmentation is provided for regions of greater spectral variance and coarser segmentation is provided for more homogeneous regions. Sub-bands which have similar characteristics may be merged with very little effect on quality, whereas sub-bands with highly variable data may be better represented if a sub-band is split. Various methods are described for measuring tonality, energy, or shape of a sub-band. These various measurements are discussed in light of making decisions of when to split or merge sub-bands to provide variable frequency segmentation.

Abstract: An apparatus and method for encoding video frames is provided. The video frames are divided into blocks for encoding. Encoding of the video blocks utilizes motion detection, motion estimation and adaptive compression, to obtain the desired compression for a particular bit rate. Adaptive compression includes intra compression (without regard to other frames) and inter compression (with regard to other frames). Intra compression, inter compression with motion detection, and inter compression with motion estimation are performed on a block by block basis, as needed. Segmentation is provided to compare encoding of a block with encoding of its sub-blocks, and to select the best block size for encoding.

Abstract: For encoding of mixed-mode images containing text and continuous-tone content, the pixels in the image that form the text content are detected and separated. Text detection classifies pixels as text or continuous tone content by accumulating pixel counts for groups of contiguous, non-smooth pixels with the same color. Groups whose pixel count exceeds a threshold are classified as text. The text detection technique further reduces classification errors by testing for boundary dimensions and pixel density of the group characteristic of long straight lines or large borders. The text detection technique further searches the neighborhood of groups qualifying as text for pixels of the same color, so as to also detect pixels for isolated text marks like dots, accents or punctuation. The separated text and continuous-tone content can be encoded separately for efficient compression while preserving text quality, and the text again superimposed on the continuous tone content at decompression.

Abstract: For encoding of mixed-content images containing palettized and continuous-tone content, continuous tone content regions in the image are detected and separated. Continuous tone content segmentation classifies pixels as continuous tone content by counting a number of unique pixel values within a pixel neighborhood. Pixels whose count exceeds a threshold are classified as continuous tone content. The technique further scans the image for regions of high continuous tone pixel density. The segmented continuous-tone and palettized content can be encoded separately for efficient compression, and then reassembled at decompression.

Abstract: Transreceived packages use forward error correction (FEC) with matrix multiplication in a Galois field of size P (GF(P)) and contain at least a portion of K rows of matrix B having elements Bk,m in M columns. Packages include matrix C having elements Cn,m for FEC for the K rows. Matrix C has 0 to (N?1)th rows redundant with matrix B data. Elements Cn,m are computed by XOR'ing GExp[(GLog[An,k]+GLog[Bk,m]) mod (P?1)] for k from 0 to (K?1). Matrix A has elements An,k with N rows and K columns. GExp and GLog are one-dimensional arrays. Matrix A is chosen so up to N rows of B and C (total) can be lost, and B can be recovered. An inverse matrix D is computed from A with the rows of B and C. B is reconstructed from D and the received rows of B and C using another matrix multiplication.

Abstract: The present invention relates to regulating the quality and/or bitrate of content within mixed content video when the video is compressed subject to a bitrate constraint. For example, a screen capture encoder encodes palletized content within a frame of screen capture video. Subject to an overall bitrate constraint, the encoder then allocates bits for continuous tone content within the frame. By controlling the allocation of bits used to encode the continuous tone content, the encoder regulates bitrate for the continuous tone content. This in turn can allow the encoder to regulate spatial quality and/or overall temporal quality for the video. In one scenario, for screen capture video encoded to a (relatively) constant overall bitrate, the screen capture encoder reduces the bitrate (and quality) of the continuous tone content, instead spending bits to increase the overall frame rate of the video.

Abstract: A system and method for correcting errors and losses occurring during a receiver-driven layered multicast (RLM) of real-time media over a heterogeneous packet network such as the Internet. This is accomplished by augmenting RLM with one or more layers of error correction information. This allows each receiver to separately optimize the quality of received audio and video information by subscribing to at least one error correction layer. Ideally, each source layer in a RLM would have one or more multicasted error correction data streams (i.e., layers) associated therewith. Each of the error correction layers would contain information that can be used to replace lost packets from the associated source layer. More than one error correction layer is proposed as some of the error correction packets contained in the data stream needed to replace the packets lost in the associated source stream may themselves be lost in transmission.

Abstract: The present invention relates to regulating the quality and/or bitrate of content within mixed content video when the video is compressed subject to a bitrate constraint. For example, a screen capture encoder encodes palletized content within a frame of screen capture video. Subject to an overall bitrate constraint, the encoder then allocates bits for continuous tone content within the frame. By controlling the allocation of bits used to encode the continuous tone content, the encoder regulates bitrate for the continuous tone content. This in turn can allow the encoder to regulate spatial quality and/or overall temporal quality for the video. In one scenario, for screen capture video encoded to a (relatively) constant overall bitrate, the screen capture encoder reduces the bitrate (and quality) of the continuous tone content, instead spending bits to increase the overall frame rate of the video.

Abstract: An apparatus and method for encoding video frames is provided. The video frames are divided into blocks for encoding. Encoding of the video blocks utilizes motion detection, motion estimation and adaptive compression, to obtain the desired compression for a particular bit rate. Adaptive compression includes intra compression (without regard to other frames) and inter compression (with regard to other frames). Intra compression, inter compression with motion detection, and inter compression with motion estimation are performed on a block by block basis, as needed. Segmentation is provided to compare encoding of a block with encoding of its sub-blocks, and to select the best block size for encoding.

Abstract: A system and method for correcting errors and losses occurring during a receiver-driven layered multicast (RLM) of real-time media over a heterogeneous packet network such as the Internet. This is accomplished by augmenting RLM with one or more layers of error correction information. This allows each receiver to separately optimize the quality of received audio and video information by subscribing to at least one error correction layer. Ideally, each source layer in a RLM would have one or more multicasted error correction data streams (i.e., layers) associated therewith. Each of the error correction layers would contain information that can be used to replace lost packets from the associated source layer. More than one error correction layer is proposed as some of the error correction packets contained in the data stream needed to replace the packets lost in the associated source stream may themselves be lost in transmission.

Abstract: A system and method for correcting errors and losses occurring during a receiver-driven layered multicast (RLM) of real-time media over a heterogeneous packet network such as the Internet. This is accomplished by augmenting RLM with one or more layers of error correction information. This allows each receiver to separately optimize the quality of received audio and video information by subscribing to at least one error correction layer. Ideally, each source layer in a RLM would have one or more multicasted error correction data streams (i.e., layers) associated therewith. Each of the error correction layers would contain information that can be used to replace lost packets from the associated source layer. More than one error correction layer is proposed as some of the error correction packets contained in the data stream needed to replace the packets lost in the associated source stream may themselves be lost in transmission.

Abstract: Traditional audio encoders may conserve coding bit-rate by encoding fewer than all spectral coefficients, which can produce a blurry low-pass sound in the reconstruction. An audio encoder using wide-sense perceptual similarity improves the quality by encoding a perceptually similar version of the omitted spectral coefficients, represented as a scaled version of already coded spectrum. The omitted spectral coefficients are divided into a number of sub-bands. The sub-bands are encoded as two parameters: a scale factor, which may represent the energy in the band; and a shape parameter, which may represent a shape of the band. The shape parameter may be in the form of a motion vector pointing to a portion of the already coded spectrum, an index to a spectral shape in a fixed code-book, or a random noise vector. The encoding thus efficiently represents a scaled version of a similarly shaped portion of spectrum to be copied at decoding.

Abstract: An audio encoder performs adaptive entropy encoding of audio data. For example, an audio encoder switches between variable dimension vector Huffman coding of direct levels of quantized audio data and run-level coding of run lengths and levels of quantized audio data. The encoder can use, for example, context-based arithmetic coding for coding run lengths and levels. The encoder can determine when to switch between coding modes by counting consecutive coefficients having a predominant value (e.g., zero). An audio decoder performs corresponding adaptive entropy decoding.

Abstract: Transreceived packages use forward error correction (FEC) with matrix multiplication in a Galois field of size P (GF(P)) and contain at least a portion of K rows of matrix B having elements Bk,m in M columns. Packages include matrix C having elements Cn,m for FEC for the K rows. Matrix C has 0 to (N−1)th rows redundant with matrix B data. Elements Cn,m are computed by XOR'ing GE xp[(G Log[An,k]+GLog[Bk,m]) mod (P−1)] for k from 0 to (K−1). Matrix A has elements An, k with N rows and K columns. G Exp and G Log are one-dimensional arrays. Matrix A is chosen so up to N rows of B and C (total) can be lost, and B can be recovered. An inverse matrix D is computed from A with the rows of B and C. B is reconstructed from D and the received rows of B and C using another matrix multiplication.

Abstract: For encoding of mixed-content images containing palettized and continuous-tone content, continuous tone content regions in the image are detected and separated. Continuous tone content segmentation classifies pixels as continuous tone content by counting a number of unique pixel values within a pixel neighborhood. Pixels whose count exceeds a threshold are classified as continuous tone content. The technique further scans the image for regions of high continuous tone pixel density. The segmented continuous-tone and palettized content can be encoded separately for efficient compression, and then reassembled at decompression.

Abstract: For encoding of mixed-mode images containing text and continuous-tone content, the pixels in the image that form the text content are detected and separated. Text detection classifies pixels as text or continuous tone content by accumulating pixel counts for groups of contiguous, non-smooth pixels with the same color. Groups whose pixel count exceeds a threshold are classified as text. The text detection technique further reduces classification errors by testing for boundary dimensions and pixel density of the group characteristic of long straight lines or large borders. The text detection technique further searches the neighborhood of groups qualifying as text for pixels of the same color, so as to also detect pixels for isolated text marks like dots, accents or punctuation. The separated text and continuous-tone content can be encoded separately for efficient compression while preserving text quality, and the text again superimposed on the continuous tone content at decompression.

Abstract: The present invention relates to motion estimation and compensation. For example, a screen capture encoder performs motion estimation that is adapted to screen capture video in various respects. For example, the motion estimation uses a distortion measure based upon the count of equal/unequal pixels in two regions, sub-samples the distortion measure to speed up motion estimation, and/or uses a search pattern that prioritizes types of motion common in screen capture video. Or, a screen capture decoder performs motion compensation that is adapted to screen capture video in various respects. For example, the decoder performs the motion compensation for pixels with different values at corresponding locations in a current frame and a reference frame, but not for all pixels of the current frame. Alternatively, an encoder/decoder performs the motion estimation/compensation to compress/decompress other kinds of content.

Abstract: The present invention relates to regulating the quality and/or bitrate of content within mixed content video when the video is compressed subject to a bitrate constraint. For example, a screen capture encoder encodes palletized content within a frame of screen capture video. Subject to an overall bitrate constraint, the encoder then allocates bits for continuous tone content within the frame. By controlling the allocation of bits used to encode the continuous tone content, the encoder regulates bitrate for the continuous tone content. This in turn can allow the encoder to regulate spatial quality and/or overall temporal quality for the video. In one scenario, for screen capture video encoded to a (relatively) constant overall bitrate, the screen capture encoder reduces the bitrate (and quality) of the continuous tone content, instead spending bits to increase the overall frame rate of the video.

Abstract: An apparatus and method for encoding video frames is provided. The video frames are divided into blocks for encoding. Encoding of the video blocks utilizes motion detection, motion estimation and adaptive compression, to obtain the desired compression for a particular bit rate. Adaptive compression includes intra compression (without regard to other frames) and inter compression (with regard to other frames). Intra compression, inter compression with motion detection, and inter compression with motion estimation are performed on a block by block basis, as needed. Segmentation is provided to compare encoding of a block with encoding of its sub-blocks, and to select the best block size for encoding.

Abstract: Correction of errors and losses in a receiver-driven layered multicast (RLM) of real-time media over a network is augmented using one or more layers of error correction information. Each receiver separately optimizes the quality of received information by subscribing to at least one error correction layer. Ideally, each source layer in a RLM has one or more associated multicasted error correction data streams. Each error correction layer contains information for replacing lost packets from the associated source layer. More than one error correction layer is proposed to correct for lost packets in other error correction layers. Error correction streams are preferably generated using a pseudo-Automatic Repeat Request (ARQ) wherein a broadcaster sends both the source packets in a primary stream and delayed versions thereof in one or more redundant streams. A hybrid technique combines the psuedo-ARQ method with an adaptation of Forward Error Correction (FEC) techniques.