Abstract: A signal-processing method for performing a coordinate transformation operation and a quantization operation on an input data to obtain quantization outputs, and a signal-processing device therefor. A search strategy based on the characteristic of the input data is established. By using the search strategy, the End of Block (EOB) is predicted and the calculation structure is determined by the predicted EOB.

Abstract: A local nonlinear feedback loop (LNFL) technique and a sigma-delta modulator (SDM) for promoting highly stable oversampling using the LNFL technique. Not only can gain mismatch of the sigma-delta modulator be minimized, but the overload of the later stage sigma-delta modulator can also be resolved without sacrificing the signal noise ratios (SNR) of the overall system. Moreover, using the local nonlinear feedback loop technique of this invention attenuates the voltage signal of the sigma-delta modulator, therefore decreasing the power consumption of the overall sigma-delta modulator system.

Abstract: The present invention discloses a method and system to encrypt an object code using a non-inversible transform and dynamic keys in an encryption circuit and an decryption circuit. The encryption circuit operates to encrypt an object code to an encrypted code using a sequence of dynamic keys which possess non-inversible transform characteristics. Similarly, the decryption circuit operates to decrypt the encrypt code using a sequence of dynamic keys which possess non-inversible transform characteristics. The sequence of dynamic keys change in accordance to a timing index in which a different key is generated for a different timing cycle. In one embodiment, the decryption circuit contains a transform circuit U to transform the encrypted code back to the object code. In another embodiment, the transform circuit U is partitioned into a resident device and mobile device.

Abstract: The present invention discloses a distributed data transfer system for transferring data among several processing units and an integrated data storage means, e.g., a memory sub-system. The distributed data transfer system includes a plurality of distributed data transfer means for connecting to the processing units. The distributed data transfer system further includes a distribution control means connected to the distributed data transfer means and the integrated data storage means. The distributed data transfer means, under the control of the distribution control means, is capable of transferring the data in divisible portions over a plurality of scheduled time periods. In a preferred embodiment, the distributed data transfer means further includes a plurality of data bus branches and a bus trunk connecting to the data bus branches. The distribution control means further includes a programmable control means for down-loading of control programs for controlling the distributed data transfers.

Abstract: The present invention relates to variable length coding (VLC). Variable length coding is a widely-used method in data compression, especially, in the applications of video data communication and storage. Many international standards have adopted this technique in video data compression, for example, JPEG, MPEG, CCITT H.261 and so on. Various implementation methods have been proposed for variable length coding. Most of those methods implement the coding with two-field codes. That is, a code is represented by a code word and a code length. However, in this invention, a three-field representation is used for each code. In comparison with the two-field method, this method not only reduces the storage requirements of the code-book but also reduces the hardware or cost to implement a variable length encoder. Moreover, variable length coding can be implemented using both parallel VLC encoders and serial VLC encoders.

Abstract: An improved video encoder performs quantization after run length conversion. This is in contrast to the conventional encoder wherein quanization takes place before run-level conversion. The inventive encoder design provides a unique look forward process for controlling the number of run-level pairs and thus the number of coded bits. In addition, the stringent speed requirements on the quantizer are reduced because only run-level pairs rather than all DCT coefficients are quantized.