Abstract: A digital video encoder is presented adapted for dynamically switching between sets of quantizer matrix tables without pausing encoding of a stream of video data. Two or more sets of quantizer matrix tables are held at the encoder's quantization unit and compressed store interface for dynamically switching between sets of quant matrix tables at a picture boundary of the sequence of video data, i.e., without stopping encoding of the sequence of video data. Further, while one set of matrix tables is being employed to quantize the stream of video data, the encoder can be updating or modifying another set of quantization matrix tables, again without stopping encoding of the sequence of video data.

Abstract: Temporal compression of a digital video data stream with hierarchically searching in at least one search unit for pixels in a reference picture to find a best match for the current macroblock. This is followed by constructing a motion vector between the current macroblock and the best match macroblock in the reference picture.

Abstract: A digital video encode system, method and computer program product are described wherein a picture is divided into multiple active areas that are parallel encoded using multiple encoders. The encoders associated with adjoining active areas of the picture are configured to exchange reference data to facilitate avoiding appearance of a seam between the adjoining active areas of a subsequent picture in a sequence of pictures due to separate encoding thereof by the multiple encoders. The reference data transferred from one encoder to an adjoining encoder is an encoded and decoded section of the picture and is used by the adjoining encoder as an expanded reference area for motion estimation when encoding the subsequent picture.

Abstract: Method and apparatus for detecting and correcting error in an uncompressed digital video image data stream. The method and apparatus identify error or partial picture scenarios. Each of the possible error or partial picture scenarios is identified in a Partial Picture Repair Unit, which causes error processing of the uncompressed video input stream, resulting in the creation of a repaired data stream on the repaired pixel bus. The repair begins upon detection of an erroneous picture, field, or line. Processing for that picture, field or line is stopped upon detection of an error in the synchronization signals, and processing is resumed when the end of line, end of field, or end of picture indicator is detected.

Abstract: A method of encoding interlaced, full motion digital video image data. The digital video image data has two interlaced fields per frame with luminance and chrominance components, where the chrominance components are uncorrelated between adjacent fields. The method encompasses applying one field of chrominance information to both luminance fields of a frame. This encodes chrominance components at one quarter the spatial resolution of the luminance components.

Abstract: A filter for downsampling input video having a first chrominance sampling to a second chrominance sampling. The input video has a Bitstream of data representing a sequence of picture frames, where each picture frame has a plurality of video lines. The filter has a means for choosing a downsampling mode from a list of available downsampling modes for the downsampling of the first chrominance sampling. The filter also has a means for dividing each picture frame into video line sets according to predetermined criteria for the particular downsampling mode chosen, where each video line set has at least two video lines. A means for assigning default downsampling coefficients for each of the video lines in the video line sets based upon the downsampling mode chosen is also provided in the filter. Lastly, the filter has a means for downsampling each video line set to a single video line having the second chrominance sampling.

Abstract: An MPEG2 compliant digital video encoder system employs an I-frame video encoder module having a quantization unit and variable length encoding to encode coefficients of each macroblock of a picture. Bit regulation schemes are presented for constraining selected coefficients of a macroblock if the bit rate of two prior macroblocks in a row exceeded a bit rate limit, or if the bit rate of the current nonintra macroblock is greater than a predetermined megabits per second (Mbps) limit. Pattern constraining is implemented within the quantization unit. For nonintra pictures, the conventional variable length coding process is modified to generate a Fixed Coded Block Pattern (FCBP) header for each macroblock of the picture for outputting in the bitstream prior to coding of the coefficients. The FCBP signifies that each block of the macroblock contains at least one nonzero coefficient.

Abstract: Method, system and computer program product are provided for adaptively encoding in hardware, software or a combination thereof a sequence of video frames in real-time. A first encoding subsystem analyzes the sequence of video frames to derive information on at least one characteristic thereof, such as motion statistics, non-motion statistics, scene change statistics, or scene fade statistics. The gathered information may be either an intraframe characteristic or an interframe characteristic. A control processor is coupled to the first encoding subsystem to automatically analyze the gathered information in real time and dynamically produce a set of control parameters. A second encoding subsystem, coupled to the control processor, then encodes each frame of the sequence of video frames employing the corresponding set of control parameters.

Abstract: Method and apparatus to avoid the code space and time overhead of the software-loop. Loops (repeatedly executed blocks of instructions) are often used in software and microcode. Loops may be employed for array manipulation, storage initialization, division and square-root interpretation, and microinterpretation of instructions with variable-length operands. Software creates loops by keeping an iteration count in a register or in memory. During each iteration of the code loop, software decrements the count, and then branches to the "top" of the loop is the count remains nonzero. This apparatus puts the decrement, compare, and branch-to-top into hardware, reducing the number of instructions in the loop and speeding loop execution. Hardware further speeds loop execution by eliminating the wait for the branch to the top-of-loop instruction. That is, it prefetches the top-of-loop instruction near the bottom of the loop.

Abstract: A method of encoding a picture in an MPEG2 compliant digital video encoder. The method calculates a contrast function,Contrast=.SIGMA..vertline.P(j)-P(j+1).vertline.and thereafter calculates a quantization adjustment function therefromM(i+1)=?C(i+1)/C(i)!M(i),where C=Contrast, P(j) is the luminance or chrominance of the j.sup.th pixel, and M(i)is the average quantization of the i.sup.th picture. The quantization or picture type is adjusted in response to the contrast function, C.

Abstract: Method and apparatus for detecting and correcting error in an uncompressed digital video image data stream. The method and apparatus identify error or partial picture scenarios. Each of the possible error or partial picture scenarios is identified in a Partial Picture Repair Unit, which causes error processing of the uncompressed video input stream, resulting in the creation of a repaired data stream on the repaired pixel bus. The repair begins upon detection of an erroneous picture, field, or line. Processing for that picture, field or line is stopped upon detection of an error in the synchronization signals, and processing is resumed when the end of line, end of field, or end of picture indicator is detected.

Abstract: A method for spatial compression of a digital video picture to obtain the quantizer step size so as to avoid over "lossy" reconstruction and loss of detail. The first step is dividing the picture into a plurality of macroblocks, for example, 16.times.16 macroblocks, each macroblock having luminance or chrominance pixel blocks, for example four 8.times.8 pixel blocks. This is followed by multiplying each luminance pixel block by a modified frequency ordered Hadamard matrix to yield a first dimension of each luminance pixel block. The first dimension of each pixel block is then multiplied by the inverse of the modified frequency ordered Hadamard matrix to yield a second dimension of each luminance pixel block. The second dimension of the pixel luminance block is then weighted against a weight matrix, and the individual weighted terms are summed for each pixel block. The minimum of the weighted terms is selected. This minimum is used to detect the edge or texture of the macroblock, e.g.

Abstract: A scalable architecture MPEG2 compliant digital video encoder system having an I-frame only video encoder module with a Discrete Cosine Transform processor, a quantization unit, a variable length encoder, a FIFO buffer, and a compressed store interface, for generating an I-frame containing bitstream. For IPB bitstreams the system includes a second processor element with a reference memory interface, motion estimation and compensation capability, inverse quantization, and inverse discrete cosine transformation, and motion compensation means; and at least one third processor element motion estimation. The system can be in the form of a single integrated circuit chip, or a plurality of integrated circuit chips, that is one for each processor, the I-frame video encoder module, the second processor element, and the third processor element. There can be one or more of the third processor units.

Abstract: Method and apparatus for encoding a digital video image stream in an encoder. The encoding includes spatial compression of still images in the digital video image stream and temporal compression between the still images. The spatial compression is carried out by converting a time domain image of a macroblock to a frequency domain image of the macroblock, taking the discrete cosine transform of the frequency domain image, transforming the discrete cosine transformed macroblock image by a quantization factor, and run length encoding the quantized discrete cosine transformed macroblock image. The temporal compression is carried out by reconstructing the run length encoded, quantized, discrete cosine transformed image of the macroblock, searching for a best match macroblock, and constructing a motion vector between them. This forms a bitstream of runlength encoded, quantized, discrete cosine transformed macroblocks and of motion vectors.

Abstract: A method of Discrete Cosine Transform compression of a digital video image. In the method the Field Variance and Frame Variance are calculated. When the Field Variance is less than the Frame Variance, Field Discrete Cosine Transform type compression is performed. Alternatively, when the Frame Variance is less than the Field Variance, then a Frame Discrete Cosine Transform compression is performed.

Abstract: Method and apparatus for calculating motion vectors. The method and apparatus calculate a motion trajectory value and initial base weights such that when x is the horizontal offset of the current macroblock from the left edge of the search window, and y is the vertical offset of the current macroblock from the top edge of the search window, and this pair of x and y values are used as the initial base weights, the motion estimation provides the best matched macroblock that has the shortest distance from the current macroblock. When the initial base weights are set to 0's, then the motion estimation will produce the first best matched macroblock encountered by the search circuitry. A scheme is also provided to set the initial base weights so that the best matched macroblock closest to the motion trajectory is selected for the motion estimation.

Abstract: Disclosed is a method of forming a bidirectionally coded picture, i.e., a B frame, from two reference pictures, e.g. I or P frame pictures. The method utilizes a single memory fetch of each reference picture, and interpolation of the estimated motion of each picture. This is accomplished by identifying a full pixel closest match from three full pixel boundary searches, calculating half pixel reference picture data therefrom, interpolating the half pixels to form bidirectionally coded pictures, and finding the closest match at the half pixel boundary. The intermediate results are stored in an 18.times.18.times.11 bit buffer holding a 7 bit partial sum for each pixel, said partial sum formed by adding the six most significant bits of corresponding I and P frame pixels. In this buffer four bits of each word in the buffer are the two Least Significant Bits of the corresponding I and P frame pixels.

Abstract: Disclosed is a digital video encoder processor for discrete cosine transform encoding. The discrete cosine transform encoding includes the encoding steps of (1) determining the discrete cosine transform field or frame type, (2) addressing individual pixels as either (i) vertically adjacent pixels on consecutive Odd and Even field lines, or (ii) vertically adjacent pixels on consecutive Odd field lines, then consecutive Even field lines; or (iii) vertically adjacent pixels on consecutive Even field lines, then consecutive Odd field lines. These subtractions may be performed between (i) consecutive lines, (ii) odd lines, or (iii) even lines. The next step is finding the smallest variance of the above subtractions to determine the discrete cosine transform coding type. The subtractions are carried out in a dynamically partitionable processor having a plurality of datapaths.