Summary: 1570 IEEE TRANSACTIONS ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL. ASSP-31, NO. 6 , DECEMBER 1983
An Error Analysis for 2-D Block Implemented
MAHMOOD R. AZIMI-SADJADI AKD ROBERT A. KING
implemented 2-D digital filters. Expressions for error bound and mean-
square error for roundoff error accumulation are derived using fixed-
point arithmetic, and comparedwith the results obtained usingordinary
2-D difference equations.
attentioninrecentyears,appliedalmost exclusively to 1-D
digitalfilters [ 11. Theprimarymotivationbehindthisstudy
was to develop a parallel processing model for recursive digital
filters, with increased data-throughput rate and re.duced com-
putational effort. Subsequent work [21, [31 ,however, reveals
several other fruitful features of this method which stem from
its unique structure.
Inthe 1-D case,Barnes andShinnaka  have shownthat