Abstract: An image/audio information recording and reproducing apparatus using a semiconductor memory. The image/audio information recording and reproducing apparatus includes a semiconductor memory, a signal processor for processing image and audio information so that the image and audio information can be recorded on and reproduced from the semiconductor memory, and a control portion coupled between the semiconductor memory and the signal processor, for storing individual image information and individual audio information corresponding to the individual image information in the semiconductor memory and reading the stored image and audio information from the semiconductor memory, using a start address of the individual information and both a start address and an end address of the individual audio information.

Abstract: A digital signal proces sing system for receiving and processing a high definition television signal, eg., in a Pulse Amplitude Modulation (PAM) format, includes a filter for rejecting interferer signals such as a narrowband continuous wave NTSC co-channel interferer. In an illustrated embodiment the rejection filter is a digital FIR filter (18) located after a (de)rotator (16) in a carrier recovery network (16, 20, 24, 26, 28). The filter exhibits a zero throughput delay bandpass response with attenuation notches for rejecting NTSC co-channel picture and chrominance baseband carrier frequencies.

Abstract: A video interface unit for transferring image data between video memory and video processing components interfaces with a video bus. The video interface unit has a partionable data buffer. The partionable buffer enables data to be accessed without redundant fetches of image data and for associated processing of data to be interleaved.

Abstract: Temporarily filtered video samples are formed as the sum of a previous frame output sample and a non-linear function of the difference between current and previous samples. The function is non-linear only for absolute values of differences less than 50% (preferably 10%) of the maximum range of the difference, and a smaller lookup table is used for implementing the non-linear function than would otherwise be the case.