Preface
xv
co n v ers ion by a n ar b itra ry fa cto r. Se ve ral a p plic atio ns to m ultirate sig nal p ro c ess
ing are p re sen te d , inc lud ing th e im p le m e ntatio n of digital filters, subb a n d cod ing
o f sp ee ch sig nals, tra n sm u ltiplex ing, and oversa m p ling A /D an d D /A co n verte rs.
L in e a r predic tio n a nd o p tim u m lin ear (W ien er) filters a re treated in C h a p
te r 11. A lso inc lude d in this ch a p te r are desc ription s o f th e L e v in so n -D u rb in
alg o rith m a n d Schiir algo rithm fo r solving the n orm al e q u ation s , as w ell as th e
A R lattice a n d A R M A lattice -la d d er filters.
P o w er sp e ctru m e stim atio n is the m ain topic of C h a p te r 12. O ur cov era ge
inc ludes a d escription o f n o n p a ra m etric an d m ode l-b ase d (param etric ) m eth o d s.
A lso des crib ed a re e igen -d eco m p o sitio n -b as ed m eth o ds, including M U SIC an d
E S P R IT .
A t N o rth e aste r n U n iv ers ity, we have u se d the first six c h ap te rs o f th is b ook
for a o n e- se m es ter (ju n io r level) co urse in discrete system s a n d d igital signal p ro
cessing.
A o ne- se m es te r se n io r level co u rse for stu den ts w h o h av e h a d p rio r e xpo su re
to discre te system s c an u se th e m ateria l in C h a pters 1 th ro u g h 4 for a quick re view
a nd th e n p ro c eed to c ov er C h a p te r 5 throug h 8.
In a first-vear g ra d u a te level cou rse in dig ital signal p rocessing , the first five
cha p ters p ro v ide th e studen t with a g oo d review of d iscre te-tim e sy stem s. T h e
in stru ctor c an m ov e quickly th ro u g h m o st of th is m a terial a n d th e n c ove r C h apte rs
6 th ro u g h 9, follow ed by e ith e r C h a p te rs 10 and 11 o r by C h ap ters 11 an d 12.
W e hav e inclu d ed m any ex am p les thro u g h o u t the bo o k an d ap p ro x im ately
500 h o m e w o rk p ro b lem s. M an y o f the h o m ew o rk p ro b le m s can be so lve d n u m e r
ically on a co m pu te r, using a so ftw a re pack ag e such as M A T L A B © . T h ese p ro b
lem s a re id entifie d by a n asterisk. A p p en d ix D co ntains a list o f M A T L A B fu nc
tion s that th e stu d e n t can use in solvin g th ese p rob lem s. T h e in stru c to r m ay also
w ish to co n sid e r the use o f a su p p le m e n tary b oo k th a t conta in s c o m p u te r bas ed
exe rcises, su ch as th e b o o k s Digilal Signal Processing Using M A T L A B (P.W .S.
K en t, 1996) by V. K. Ing le and J. G . P roak is a nd Computer-Based Exercises for
Signal Processing Using M A T L A B (P re n tice H all, 1994) by C. S. B u rru s et al.
T h e au th o rs a r e in d e b te d to th e ir m an y facu lty co llea gue s w ho h av e p ro vide d
v alu ab le sug gestio ns th ro u gh rev iew s of the first an d sec on d ed itio n s o f this bo o k.
T h e se in clu de D rs. W . E. A lex a n der, Y. B resler, J. D eller, V. In gle, C. K eller,
H . Lev -A ri, L. M e rako s , W. M ik hae l, P. M onticciolo , C. N ikias, M . S c he tzen ,
H . T ru sse ll, S. W ilson , and M. Z olto w sk i. W e a re also in d e b ted to D r. R, Pric e fo r
re com m e nding th e inc lusion o f sp lit-rad ix F F T a lgorith m s and r elate d sugg estions.
F inally, w e w ish to ack n ow ledg e the su gg estions and c o m m e n ts o f m a ny fo rm er
g ra d u a te stud e n ts, an d espec ially th o se by A . L. K o k, J. L in an d S. Srin id hi w ho
as sisted in th e p rep a ra tio n o f severa l illustratio n s an d th e solu tion s m an u al.
Jo h n G. P roak is
D im itris G , M a nolakis