Abstract: In a system in which a plurality of modules have different operational rates and a common clock controlling data delivery to the modules, the rate at which data is delivered to the system can be maximized using a return clock signal to prevent the loss of synchronization of the modules. A clocking error signal may be produced when the clock signal makes a transition to a logic state that may cause loss of synchronization between the modules.

Abstract: An image filtering method, apparatus and system, wherein the method comprising the steps of detecting at least one portion of an edge, wherein the portion of the edge provides an indication that ringing artifact are probable; subjecting at least one portion of a pixel, related to the at least a portion of the edge, to a low pass filter to produce a filtered pixel; and blending the filtered pixel with a value relating to the filter prior to filtering to produce a filtered image.

Abstract: A method for adapting acoustic models used for automatic speech recognition is provided. The method includes estimating noise in a portion of a speech signal, determining a first estimated variance scaling vector using an estimated 2-order polynomial and the noise estimation. The estimated 2-order polynomial represents a prior knowledge of a dependency of a variance scaling vector on noise, determining a second estimated variance scaling vector using statistics from prior portions of the speech signal, determining a variance scaling factor using the first estimated variance scaling vector and the second estimated variance scaling vector, and using the variance scaling factor to adapt an acoustic model.

Abstract: A method of video encoding is provided that includes computing spatial variance for video data in a block of a video sequence, estimating a first bit-rate based on the spatial variance, a transform coefficient threshold, and variance multiplicative factors empirically determined for first transform coefficients, and encoding the block based on the first bit-rate.

Abstract: A method and apparatus of enhancing an image, the method comprising applying at least one multi-scale filter bank to at least a portion of an image to detect at least one edge at different scales, and combining the detected edges with said image to yield an enhanced to at least a portion of the image.

Abstract: In a test and debug system in which a plurality of selectable modules under test have different operational rates, a selection unit associated with each module is used to control the application of the RCLK signal from the module to the combiner unit, the combiner unit providing a composite RCLK signal. Each selection unit has output signals of RCLK_NE and RCLK_PE signals which are applied to an combiner unit to form the composite RCLK signal. In response to the SELECT signal, the RCLK_NE and RCLK_PE are synchronized with the module RCLK signal. When the SELECT signal is removed, the RCLK_NE and RCLK_PE signals are continuously applied to the combiner unit.

Abstract: An embodiment of the present invention uses estimates of delay spreads of transmissions from user equipments (UEs) to a NodeB to determine a set of transmission parameters for the UEs reference signals. In an exemplary embodiment, the transmission parameters for reference signals include cyclic shifts. Thus, embodiments include a set of allocated cyclic shift values that are tailored to the delay spreads. The set of allocated cyclic shift values are used by a corresponding set of UE being served by a NodeB to form references signals. Each UE uses the allocated cyclic shift to form its reference signal by applying the cyclic shift to a modified reference sequence. The modified reference sequence can be generated from a Constant-Amplitude-Zero-Auto Correlation (CAZAC) sequence. The set of allocated cyclic shift values can be updated periodically to account for changes of delay spreads, which can be caused by physical movements of the set of UEs.

Abstract: Audio loudspeaker virtualizers and cross-talk cancellers and methods use a combination of interaural intensity difference and interaural time difference to define virtualizing filters. This allows enlargement of a listener's sweet spot based on psychoacoustic effects.

Abstract: Methods, digital systems, and computer readable media are provided for determining a gain reduction parameter level for loudspeaker equalization by determining a noise score, an equalization effectiveness score, and an equalization non-effectiveness score for a candidate gain reduction parameter level, determining a composite quality score using the three scores, and designing a compensating filter for the loudspeaker using the candidate gain reduction parameter level if the composite quality score is better than composite quality scores of all other candidate gain reduction parameter levels.

Abstract: An activity-based system for, and method of, reducing Gr-Gb gain imbalance and a digital camera incorporating the system or the method. In one embodiment, the system includes: (1) a sensor configured to provide a input Bayer pattern array containing amplitudes corresponding to Gr and Gb cells and (2) a processor coupled to the sensor and configured to (2a) compute for at least some of the Gr and Gb cells: activity measures for pluralities of adjacent, same-type cells, green precompensation factors based on the activity measures, averages for the pluralities of adjacent, same-type cells and averages for pluralities of adjacent, opposite-type cells and (2b) use the green precompensation factors, the averages for the pluralities of adjacent, same-type cells and the averages for the pluralities of adjacent, opposite-type cells to form an output Bayer pattern in which the Gr-Gb gain imbalance is reduced.

Abstract: Methods, digital systems, and computer readable media are provided for determining a predominant fundamental frequency of a frame of an audio signal by finding a maximum absolute signal value in history data for the frame, determining a number of bits for downshifting based on the maximum absolute signal value, computing autocorrelations for the frame using signal values downshifted by the number of bits, and determining the predominant fundamental frequency using the computed autocorrelations.

Abstract: Method and system for auto-focusing. The method includes checking for presence of a subject in a first frame. The method also includes determining a second frame indicative of an in-focus position of the subject if the presence of the subject is detected. The method further includes causing to focus on the subject based on the second frame. Further, the method includes checking for a change in at least one of location of the subject in the second frame, and sharpness and luminance in an area corresponding to the subject in the second frame. Moreover, the method includes causing to refocus on the subject if the change is detected.

Abstract: Audio loudspeaker and headphone virtualizers and cross-talk cancellers and methods use separate virtual speaker locations for different Bark frequency bands and a single reverberation filter for multi-channel virtualizer inputs.

Abstract: Redeye removal methods detect redeyes in a two-step procedure. In the first step it detects faces in the input image and in the second step it searches for redeyes in the detected faces. The methods include the introduction of an efficient skin tone detector that reduces the average complexity of a Viola-Jones based face. We also introduce a summed area table (SAT) based optimization in our eye detector stage that reduces the complexity in the eye detector by a factor of 5× when compared to an optimized direct search method.

Abstract: An integrated circuit includes a memory array having a plurality of SRAM memory cells arranged in a plurality of rows and columns, the array also having a plurality of word lines for accessing rows of cells and a plurality bit lines for accessing columns of cells. The plurality of memory cells include a plurality of asymmetric cells, each of the asymmetric cells configured with a strong side including a first inverter having a strong side latch node, and a strong side pass transistor coupled to the strong side latch node, and a weak side including a second inverter cross-coupled with the first inverter having a weak side latch node and a weak side pass transistor coupled to the weak side latch node. Separate ones of the plurality of word lines are coupled to a gate of the strong side pass transistor and a gate of the weak side pass transistor.

Abstract: Transcoding as from MPEG-2 SDTV to MPEG-4 CIF reuses motion vectors and downsamples in the frequency (DCT) domain with differing treatments of frame-DCT and field-DCT blocks, and alternatively uses de-interlacing IDCT with respect to the row dimension plus deferred column downsampling for reference frame blocks.

Abstract: Audio loudspeaker and headphone virtualizers and cross-talk cancellers and methods use separate virtual speaker locations for different Bark frequency bands and a single reverberation filter for multi-channel virtualizer inputs.

Abstract: Digital camera contrast enhancement with piecewise-linear transform with lower and upper cutoffs for the transform determined from histogram analysis with a green color conversion approximation for images and IIR parameter filtering for videos.

Abstract: Hands-free phones with echo cancellation using dual filters, one fast adapting filter and one slow adapting filter; the choice of filters includes switching with hysteresis from recent performance, and the fast adapting filter has step size control to limit relative filter update energy and the slow adapting filter limits filter divergence.

Abstract: Modify H.263-type quantization with an adaptive quantization parameter floor; this limits clipping of quantized DCT coefficients and consequent artifacts. The maximum absolute level of AC coefficients of a DCT transformed macroblock provides a minimum quantization parameter from integer division by 256 when the quantized levels are clipped to a range such as ?127 to +127.