A class of linear cyclic-invariant filter for parallel filtering of discrete-time signals

In this paper, a block-based filtering technique for discrete-time signals is introduced. A Block-Mirrored signal model which reduces the error caused by the signal truncation due to the block-based processing is proposed and an overlap-add technique is used to reduce blocking effects. Since data blocks can be processed independently, parallel filtering can be used to enhance the processing speed and to reduce the hardware requirement. Computer simulations and some restrictions of using this technique are also discussed