Fir filter with quantized coefficients and coefficient quantization method
US5353026A · kind A · utility
Assignee
Inventor
Key dates
| Filing date | Dec 15, 1992 |
| Grant date | Oct 4, 1994 |
| Priority date | — |
| Expiry date | Dec 15, 2012 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH03H17/06
- WIPO fieldBasic communication processes
- WIPO sectorElectrical engineering
Abstract
The n-bit coefficients of a FIR filter are encoded into m-bit digital words that represent selected bits of the original coefficients, where m is less than n, together with shift words that represent the order of the selected bits within the original n-bit coefficients. A software .sigma.-.DELTA. modulator is used to implement the encoding. The reduction in coefficient bits allows for a corresponding reduction in hardware multiplier capacity required to convolve the encoded coefficients with input data. The quantization error from the encoding is shaped to a higher frequency range that can later be filtered out. The coefficients are preferably encoded into a format that coincides with the data format employed by the multiplier, such as by using a Robertson's algorithm to encode to a modified Booth format.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.