4: Linear Time Invariant Systems

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

• Overlap Add • Overlap Save • Summary • MATLAB routines

4: Linear Time Invariant Systems

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 1 / 15

LTI Systems
4: Linear Time Invariant Systems

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 2 / 15

LTI Systems
4: Linear Time Invariant Systems

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

Linear Time-invariant (LTI) systems have two properties:

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 2 / 15

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r The behaviour of an LTI system is completely defined by its impulse response: h[n] = H (δ [n]) • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r The behaviour of an LTI system is completely defined by its impulse response: h[n] = H (δ [n]) Proof: We can always write x[n] = ∞ r =−∞ x[r ]δ [n • Overlap Add • Overlap Save • Summary • MATLAB routines − r] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r The behaviour of an LTI system is completely defined by its impulse response: h[n] = H (δ [n]) Proof: We can always write x[n] = Hence ∞ r =−∞ x[r ]δ [n • Overlap Add • Overlap Save • Summary • MATLAB routines − r] − r] H (x[n]) = H ∞ r =−∞ x[r ]δ [n DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r The behaviour of an LTI system is completely defined by its impulse response: h[n] = H (δ [n]) Proof: We can always write x[n] = Hence ∞ r =−∞ x[r ]δ [n • Overlap Add • Overlap Save • Summary • MATLAB routines − r] − r] H (x[n]) = H = ∞ r =−∞ x[r ]δ [n ∞ r =−∞ x[r ]H (δ [n − r ]) DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r The behaviour of an LTI system is completely defined by its impulse response: h[n] = H (δ [n]) Proof: We can always write x[n] = Hence ∞ r =−∞ x[r ]δ [n • Overlap Add • Overlap Save • Summary • MATLAB routines − r] − r] r ]) H (x[n]) = H = = ∞ r =−∞ x[r ]δ [n ∞ r =−∞ x[r ]H (δ [n − ∞ r =−∞ x[r ]h[n − r ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

LTI Systems 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Linear Time-invariant (LTI) systems have two properties: Linear: H (αu[n] + βv [n]) = αH (u[n]) + β H (v [n]) Time Invariant: y [n] = H (x[n]) ⇒ y [n − r ] = H (x[n − r ]) ∀r The behaviour of an LTI system is completely defined by its impulse response: h[n] = H (δ [n]) Proof: We can always write x[n] = Hence ∞ r =−∞ x[r ]δ [n • Overlap Add • Overlap Save • Summary • MATLAB routines − r] − r] r ]) H (x[n]) = H = = ∞ r =−∞ x[r ]δ [n ∞ r =−∞ x[r ]H (δ [n − ∞ r =−∞ x[r ]h[n − r ] = x [n ] ∗ h [n ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 2 / 15 .

Convolution Properties
4: Linear Time Invariant Systems

Convolution: x[n] ∗ v [n] =

∞ r =−∞ x[r ]v [n

− r]

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 3 / 15

Convolution Properties
4: Linear Time Invariant Systems

Convolution: x[n] ∗ v [n] =

∞ r =−∞ x[r ]v [n

− r]

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

Convolution obeys normal arithmetic rules for multiplication:

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 3 / 15

Convolution Properties
4: Linear Time Invariant Systems

Convolution: x[n] ∗ v [n] =

∞ r =−∞ x[r ]v [n

− r]

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n]

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 3 / 15

Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .

Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .

Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] ⇒ x[n] ∗ v [n] ∗ w[n] is unambiguous DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .

s x[n − r ]v [r − s]w [s] = p. q = n − s (i) DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] ⇒ x[n] ∗ v [n] ∗ w[n] is unambiguous Proof: r.q x[p]v [q − p]w [n − q ] (i) substitute p = n − r .

q x[p]v [q − p]w [n − q ] (i) substitute p = n − r . q = n − s (i) Distributive over +: x[n] ∗ (αv [n] + βw[n]) = (x[n] ∗ αv [n]) + (x[n] ∗ βw[n]) DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .s x[n − r ]v [r − s]w [s] = p.Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] ⇒ x[n] ∗ v [n] ∗ w[n] is unambiguous Proof: r.

q x[p]v [q − p]w [n − q ] (i) substitute p = n − r .s x[n − r ]v [r − s]w [s] = p. q = n − s (i) Distributive over +: x[n] ∗ (αv [n] + βw[n]) = (x[n] ∗ αv [n]) + (x[n] ∗ βw[n]) Proof: r x[n − r ] (αv [r ] + βw [r ]) = α r x [n − r ]v [r ] + β r x [n − r ]w [r ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] ⇒ x[n] ∗ v [n] ∗ w[n] is unambiguous Proof: r.

q x[p]v [q − p]w [n − q ] (i) substitute p = n − r .Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] ⇒ x[n] ∗ v [n] ∗ w[n] is unambiguous Proof: r. q = n − s (i) Distributive over +: x[n] ∗ (αv [n] + βw[n]) = (x[n] ∗ αv [n]) + (x[n] ∗ βw[n]) Proof: r x[n − r ] (αv [r ] + βw [r ]) = α r x [n − r ]v [r ] + β r x [n − r ]w [r ] Identity: x[n] ∗ δ [n] = x[n] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 3 / 15 .s x[n − r ]v [r − s]w [s] = p.

q = n − s (i) Distributive over +: x[n] ∗ (αv [n] + βw[n]) = (x[n] ∗ αv [n]) + (x[n] ∗ βw[n]) Proof: r x[n − r ] (αv [r ] + βw [r ]) = α r x [n − r ]v [r ] + β r x [n − r ]w [r ] Identity: x[n] ∗ δ [n] = x[n] Proof: DSP and Digital Filters (2012-2446) r δ [r ]x [n − r ] = x [n ] (i) (i) all terms zero except r = 0.s x[n − r ]v [r − s]w [s] = p. LTI Systems: 4 – 3 / 15 .q x[p]v [q − p]w [n − q ] (i) substitute p = n − r .Convolution Properties 4: Linear Time Invariant Systems Convolution: x[n] ∗ v [n] = ∞ r =−∞ x[r ]v [n − r] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution obeys normal arithmetic rules for multiplication: Commutative: x[n] ∗ v [n] = v [n] ∗ x[n] Proof: r x [r ]v [n − r ] = p x [n − p ]v [p ] (i) substitute p = n − r (i) • Overlap Add • Overlap Save • Summary • MATLAB routines Associative: x[n] ∗ (v [n] ∗ w[n]) = (x[n] ∗ v [n]) ∗ w[n] ⇒ x[n] ∗ v [n] ∗ w[n] is unambiguous Proof: r.

y [n] • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 .BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. x[n] ⇒ Bounded Output.

x[n] ⇒ Bounded Output.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. y [n] The following are equivalent: (1) An LTI system is BIBO stable • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 .

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∞ n=−∞ |h [n ]| < ∞ • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 . y [n] The following are equivalent: (1) An LTI system is BIBO stable (2) h[n] is absolutely summable.e. i.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. x[n] ⇒ Bounded Output.

x[n] ⇒ Bounded Output. i. y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable. • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 .BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1.e.

i.e. y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1. x[n] ⇒ Bounded Output. Proof (1) ⇒ (2): Define x[n] = • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines 1 h[−n] ≥ 0 −1 h[−n] < 0 then y [0] = x[0 − n]h[n] = |h [n ]|. DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 .

n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1. Proof (1) ⇒ (2): • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines 1 h[−n] ≥ 0 −1 h[−n] < 0 then y [0] = But |x[n]| ≤ 1∀n so BIBO ⇒ y [0] = Define x[n] = x[0 − n]h[n] = |h[n]| < ∞. DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 .e. y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. x[n] ⇒ Bounded Output. |h [n ]|. i.

|h [n ]|.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1. y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable.e. i. Proof (1) ⇒ (2): • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines 1 h[−n] ≥ 0 −1 h[−n] < 0 then y [0] = But |x[n]| ≤ 1∀n so BIBO ⇒ y [0] = Define x[n] = x[0 − n]h[n] = |h[n]| < ∞. x[n] ⇒ Bounded Output. Proof (2) ⇒ (1): Suppose |h[n]| = S < ∞ and |x[n]| ≤ B is bounded. DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 .

Then |y [n]| = ∞ r =−∞ x [n − r ]h [r ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 . y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable.e. x[n] ⇒ Bounded Output. n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1. Proof (2) ⇒ (1): Suppose |h[n]| = S < ∞ and |x[n]| ≤ B is bounded. i.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. Proof (1) ⇒ (2): • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines 1 h[−n] ≥ 0 −1 h[−n] < 0 then y [0] = But |x[n]| ≤ 1∀n so BIBO ⇒ y [0] = Define x[n] = x[0 − n]h[n] = |h[n]| < ∞. |h [n ]|.

BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. Proof (1) ⇒ (2): • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines 1 h[−n] ≥ 0 −1 h[−n] < 0 then y [0] = But |x[n]| ≤ 1∀n so BIBO ⇒ y [0] = Define x[n] = x[0 − n]h[n] = |h[n]| < ∞. |h [n ]|. Proof (2) ⇒ (1): Suppose |h[n]| = S < ∞ and |x[n]| ≤ B is bounded. y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable. n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1. x[n] ⇒ Bounded Output.e. Then |y [n]| = ∞ r =−∞ x[n − r ]h[r ] ∞ r =−∞ |x[n − r ]| |h[r ]| ≤ DSP and Digital Filters (2012-2446) LTI Systems: 4 – 4 / 15 . i.

Proof (2) ⇒ (1): Suppose |h[n]| = S < ∞ and |x[n]| ≤ B is bounded. Proof (1) ⇒ (2): • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines 1 h[−n] ≥ 0 −1 h[−n] < 0 then y [0] = But |x[n]| ≤ 1∀n so BIBO ⇒ y [0] = Define x[n] = x[0 − n]h[n] = |h[n]| < ∞. i. Then |y [n]| = ∞ r =−∞ x[n − r ]h[r ] ∞ r =−∞ |x[n − r ]| |h[r ]| B ∞ r =−∞ |h[r ]|≤ BS < ≤ ≤ DSP and Digital Filters (2012-2446) ∞ LTI Systems: 4 – 4 / 15 . x[n] ⇒ Bounded Output.e.BIBO Stability 4: Linear Time Invariant Systems BIBO Stability: Bounded Input. n=−∞ |h[n]| < ∞ (3) H (z ) region of convergence includes |z | = 1. y [n] The following are equivalent: (1) An LTI system is BIBO stable ∞ (2) h[n] is absolutely summable. |h [n ]|.

H (z ) evaluated at z = ejω .e. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i.

e. H (z ) evaluated at z = ejω .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. Example: h[n] = 1 1 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .

Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i.e. H (z ) evaluated at z = ejω .

H (z ) evaluated at z = ejω .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i.e. Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .

Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. H (z ) evaluated at z = ejω . Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .e.

Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i.e. H (z ) evaluated at z = ejω .

e. Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 2 3 ω DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. H (z ) evaluated at z = ejω .

H (z ) evaluated at z = ejω .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | cos ω ) ∠H (ejω ) = −ω + π 1−sgn(1+2 2 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 2 3 ω DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .e.

H (z ) evaluated at z = ejω . Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | cos ω ) ∠H (ejω ) = −ω + π 1−sgn(1+2 2 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 0 ∠ 1 ω 2 3 -1 -2 0 1 ω 2 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .e.Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i.

Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | cos ω ) ∠H (ejω ) = −ω + π 1−sgn(1+2 2 Sign change in (1 + 2 cos ω ) at ω = 2. 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 0 ∠ 1 ω 2 3 -1 -2 0 1 ω 2 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .e. H (z ) evaluated at z = ejω .1 gives (a) gradient discontinuity in |H (ejω )| (b) an abrupt phase change of ±π .

H (z ) evaluated at z = ejω .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. ∠ 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 0 -1 -2 0 1 ω 1 d ∠H (ejω ) − dω ω 2 3 2 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 . : gives delay of the Group delay is modulation envelope at each ω . Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | cos ω ) ∠H (ejω ) = −ω + π 1−sgn(1+2 2 Sign change in (1 + 2 cos ω ) at ω = 2.1 gives (a) gradient discontinuity in |H (ejω )| (b) an abrupt phase change of ±π .e.

: gives delay of the Group delay is modulation envelope at each ω . d ∠H (ejω ) − dω 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 0 ∠ 1 ω 2 3 -1 -2 0 1 ω 2 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 . Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | cos ω ) ∠H (ejω ) = −ω + π 1−sgn(1+2 2 Sign change in (1 + 2 cos ω ) at ω = 2.e.1 gives (a) gradient discontinuity in |H (ejω )| (b) an abrupt phase change of ±π . H (z ) evaluated at z = ejω .Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. Normally varies with ω but for a symmetric filter it is constant: in this case +1 samples.

Frequency Response 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution For a BIBO stable system Y (ejω ) = X (ejω )H (ejω ) where H (ejω )is the DTFT of h[n] i. : gives delay of the Group delay is modulation envelope at each ω . Discontinuities of ±kπ do not affect group delay.e. Normally varies with ω but for a symmetric filter it is constant: in this case +1 samples. H (z ) evaluated at z = ejω . Example: h[n] = 1 1 1 0 H (ejω ) = 1 + e−jω + e−j 2ω = e−jω (1 + 2 cos ω ) H (ejω ) = |1 + 2 cos ω | cos ω ) ∠H (ejω ) = −ω + π 1−sgn(1+2 2 Sign change in (1 + 2 cos ω ) at ω = 2. d ∠H (ejω ) − dω 3 2 1 0 • Overlap Add • Overlap Save • Summary • MATLAB routines 0 1 ω 2 3 10 5 0 -5 -10 0 1 0 ∠ 1 ω 2 3 -1 -2 0 1 ω 2 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 5 / 15 .1 gives (a) gradient discontinuity in |H (ejω )| (b) an abrupt phase change of ±π .

Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i.e. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 . output at time n depends only on inputs up to time n.

Formal definition: If v [n] = x[n] for n ≤ n0 then H (v [n]) = H (x[n]) for n ≤ n0 .e.Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i. output at time n depends only on inputs up to time n. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 .

The following are equivalent: (1) An LTI system is causal • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 . Formal definition: If v [n] = x[n] for n ≤ n0 then H (v [n]) = H (x[n]) for n ≤ n0 .e.Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i. output at time n depends only on inputs up to time n.

Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i. output at time n depends only on inputs up to time n. The following are equivalent: (1) An LTI system is causal (2) h[n] is causal ⇔ h[n] = 0 for n < 0 • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 . Formal definition: If v [n] = x[n] for n ≤ n0 then H (v [n]) = H (x[n]) for n ≤ n0 .e.

output at time n depends only on inputs up to time n.Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i. The following are equivalent: (1) An LTI system is causal (2) h[n] is causal ⇔ h[n] = 0 for n < 0 (3) H (z ) converges for z = ∞ • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 .e. Formal definition: If v [n] = x[n] for n ≤ n0 then H (v [n]) = H (x[n]) for n ≤ n0 .

The following are equivalent: (1) An LTI system is causal (2) h[n] is causal ⇔ h[n] = 0 for n < 0 (3) H (z ) converges for z = ∞ Any right-sided sequence can be made causal by adding a delay. output at time n depends only on inputs up to time n.Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i.e. Formal definition: If v [n] = x[n] for n ≤ n0 then H (v [n]) = H (x[n]) for n ≤ n0 . • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 .

Formal definition: If v [n] = x[n] for n ≤ n0 then H (v [n]) = H (x[n]) for n ≤ n0 . All the systems we will deal with are causal.e.Causality 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Causal System: cannot see into the future i. The following are equivalent: (1) An LTI system is causal (2) h[n] is causal ⇔ h[n] = 0 for n < 0 (3) H (z ) converges for z = ∞ Any right-sided sequence can be made causal by adding a delay. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 6 / 15 . output at time n depends only on inputs up to time n.

Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .

Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .

n=−∞ n=−∞ • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n].

Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n]. n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .

n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω Somewhat analogous to a circuit consisting of R. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 . L and C only.Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n].

A lossless system always preserves energy exactly: • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n]. n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω Somewhat analogous to a circuit consisting of R. L and C only.

Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n]. n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω Somewhat analogous to a circuit consisting of R. L and C only. A lossless system always preserves energy exactly: ∞ 2 ∞ 2 | y [ n ] | = | x [ n ] | for any finite energy input x[n]. n=−∞ n=−∞ • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .

n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω Somewhat analogous to a circuit consisting of R.Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n]. n=−∞ n=−∞ A lossless LTI system must have H (ejω ) = 1∀ω called an allpass system. A lossless system always preserves energy exactly: ∞ 2 ∞ 2 | y [ n ] | = | x [ n ] | for any finite energy input x[n]. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 . L and C only.

n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω Somewhat analogous to a circuit consisting of R. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 .Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n]. L and C only. n=−∞ n=−∞ A lossless LTI system must have H (ejω ) = 1∀ω called an allpass system. A lossless system always preserves energy exactly: ∞ 2 ∞ 2 | y [ n ] | = | x [ n ] | for any finite energy input x[n]. Somewhat analogous to a circuit consisting of L and C only.

Somewhat analogous to a circuit consisting of L and C only. n=−∞ n=−∞ A lossless LTI system must have H (ejω ) = 1∀ω called an allpass system. Passive and lossless building blocks can be used to design systems which are insensitive to coefficient changes • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 7 / 15 . n=−∞ n=−∞ A passive LTI system must have H (ejω ) ≤ 1∀ω Somewhat analogous to a circuit consisting of R. L and C only.Passive and Lossless 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution A passive system can never gain energy: ∞ ∞ 2 2 | y [ n ] | ≤ | x [ n ] | for any finite energy input x[n]. A lossless system always preserves energy exactly: ∞ 2 ∞ 2 | y [ n ] | = | x [ n ] | for any finite energy input x[n].

Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .

Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines h [n ] = h 1 [n ] + h 2 [n ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .

Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines h [n ] = h 1 [n ] + h 2 [n ] H ( z ) = H1 ( z ) + H2 ( z ) DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .

cascade): DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .a.Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines h [n ] = h 1 [n ] + h 2 [n ] H ( z ) = H1 ( z ) + H2 ( z ) Two LTI systems in series (a.k.

k. cascade): h [n ] = h 1 [n ] ∗ h 2 [n ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .a.Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines h [n ] = h 1 [n ] + h 2 [n ] H ( z ) = H1 ( z ) + H2 ( z ) Two LTI systems in series (a.

k.Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines h [n ] = h 1 [n ] + h 2 [n ] H ( z ) = H1 ( z ) + H2 ( z ) Two LTI systems in series (a.a. cascade): h [n ] = h 1 [n ] ∗ h 2 [n ] H ( z ) = H 1 ( z ) H2 ( z ) DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .

Parallel and Series 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Two LTI systems in parallel: • Overlap Add • Overlap Save • Summary • MATLAB routines h [n ] = h 1 [n ] + h 2 [n ] H ( z ) = H1 ( z ) + H2 ( z ) Two LTI systems in series (a.a.k. cascade): h [n ] = h 1 [n ] ∗ h 2 [n ] H ( z ) = H 1 ( z ) H2 ( z ) Proof: y [n] = h2 [n] ∗ (h1 [n] ∗ x[n]) = (h2 [n] ∗ h1 [n]) ∗ x[n] associativity of convolution DSP and Digital Filters (2012-2446) LTI Systems: 4 – 8 / 15 .

Convolution Complexity
4: Linear Time Invariant Systems

y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1]
x

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 9 / 15

Convolution Complexity
4: Linear Time Invariant Systems

y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1]
x

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution


M −1 r =0

Convolution sum:

y [n ] =

h [r ]x [n − r ]

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 9 / 15

Convolution Complexity
4: Linear Time Invariant Systems

y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1]
x

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution


M −1 r =0

y[0]

y[9]

Convolution sum:

y [n ] =

h [r ]x [n − r ]

• Overlap Add • Overlap Save • Summary • MATLAB routines

N = 8, M = 3 M + N − 1 = 10

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 9 / 15

Convolution Complexity 4: Linear Time Invariant Systems y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1] x • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∗ M −1 r =0 → y[0] y[9] Convolution sum: y [n ] = h [r ]x [n − r ] • Overlap Add • Overlap Save • Summary • MATLAB routines y [n] is only non-zero in the range 0≤n≤M +N −2 N = 8. M = 3 M + N − 1 = 10 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 9 / 15 .

M = 3 M + N − 1 = 10 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 9 / 15 .Convolution Complexity 4: Linear Time Invariant Systems y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1] x • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∗ M −1 r =0 → y[0] y[9] Convolution sum: y [n ] = h [r ]x [n − r ] • Overlap Add • Overlap Save • Summary • MATLAB routines y [n] is only non-zero in the range 0≤n≤M +N −2 Thus y [n] has only M + N − 1 non-zero values N = 8.

Convolution Complexity 4: Linear Time Invariant Systems y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1] x • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∗ M −1 r =0 → y[0] y[9] Convolution sum: y [n ] = h [r ]x [n − r ] • Overlap Add • Overlap Save • Summary • MATLAB routines y [n] is only non-zero in the range 0≤n≤M +N −2 Thus y [n] has only M + N − 1 non-zero values Algebraically: N = 8. M = 3 M + N − 1 = 10 x[n − r ] = 0⇒ 0 ≤ n − r ≤ N − 1 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 9 / 15 .

Convolution Complexity 4: Linear Time Invariant Systems y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1] x • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∗ M −1 r =0 → y[0] y[9] Convolution sum: y [n ] = h [r ]x [n − r ] • Overlap Add • Overlap Save • Summary • MATLAB routines y [n] is only non-zero in the range 0≤n≤M +N −2 Thus y [n] has only M + N − 1 non-zero values Algebraically: N = 8. M = 3 M + N − 1 = 10 x[n − r ] = 0⇒ 0 ≤ n − r ≤ N − 1 ⇒n+1−N ≤r ≤n DSP and Digital Filters (2012-2446) LTI Systems: 4 – 9 / 15 .

Convolution Complexity 4: Linear Time Invariant Systems y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1] x • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∗ M −1 r =0 → y[0] y[9] Convolution sum: y [n ] = h [r ]x [n − r ] • Overlap Add • Overlap Save • Summary • MATLAB routines y [n] is only non-zero in the range 0≤n≤M +N −2 Thus y [n] has only M + N − 1 non-zero values Algebraically: N = 8.n)) r =max(0.n+1−N ) h [r ]x [n − r ] DSP and Digital Filters (2012-2446) LTI Systems: 4 – 9 / 15 . M = 3 M + N − 1 = 10 x[n − r ] = 0⇒ 0 ≤ n − r ≤ N − 1 ⇒n+1−N ≤r ≤n Hence: y [n] = min(M −1.

M = 3 M + N − 1 = 10 x[n − r ] = 0⇒ 0 ≤ n − r ≤ N − 1 ⇒n+1−N ≤r ≤n Hence: y [n] = min(M −1.n)) r =max(0.n+1−N ) h [r ]x [n − r ] We must multiply each h[n] by each x[n] and add them to a total ⇒ total arithmetic complexity (× or + operations) ≈ 2M N DSP and Digital Filters (2012-2446) LTI Systems: 4 – 9 / 15 .Convolution Complexity 4: Linear Time Invariant Systems y [n] = x[n] ∗ h[n]: convolve x[0 : N − 1] with h[0 : M − 1] x • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ∗ M −1 r =0 → y[0] y[9] Convolution sum: y [n ] = h [r ]x [n − r ] • Overlap Add • Overlap Save • Summary • MATLAB routines y [n] is only non-zero in the range 0≤n≤M +N −2 Thus y [n] has only M + N − 1 non-zero values Algebraically: N = 8.

Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .

Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Convolution sum: y [n ] = N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .

Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution → y[0] y[7] Convolution sum: y [n ] = N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines N = 8. M = 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .

Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution → y[0] y[7] Convolution sum: y [n ] = N y N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines [n] has period N N = 8. M = 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .

M = 3 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution → y[0] y[7] Convolution sum: y [n ] = N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines y N [n] has period N ⇒ y N [n] has N distinct values N = 8.

M = 3 Only the first M − 1 values are affected by the circular repetition: DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution → y[0] y[7] Convolution sum: y [n ] = N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines y N [n] has period N ⇒ y N [n] has N distinct values N = 8.

M = 3 Only the first M − 1 values are affected by the circular repetition: y N [n] = y [n] for M − 1 ≤ n ≤ N − 1 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution → y[0] y[7] Convolution sum: y [n ] = N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines y N [n] has period N ⇒ y N [n] has N distinct values N = 8.

then the circular repetition has no effect at all and: y N +M −1 [n] = y [n] for 0 ≤ n ≤ N + M − 2 DSP and Digital Filters (2012-2446) LTI Systems: 4 – 10 / 15 .Circular Convolution 4: Linear Time Invariant Systems y [n ] = x [n ] x N h[n]: circ convolve x[0 : N − 1] with h[0 : M − 1] N • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution → y[0] y[7] Convolution sum: y [n ] = N M −1 r =0 h[r ]x[(n − r )mod N ] • Overlap Add • Overlap Save • Summary • MATLAB routines y N [n] has period N ⇒ y N [n] has N distinct values N = 8. M = 3 Only the first M − 1 values are affected by the circular repetition: y N [n] = y [n] for M − 1 ≤ n ≤ N − 1 If we append M − 1 zeros (or more) onto x[n].

Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 .

Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 .

Frequency-domain convolution
4: Linear Time Invariant Systems

Idea: Use DFT to perform circular convolution - less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2)

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 11 / 15

Frequency-domain convolution
4: Linear Time Invariant Systems

Idea: Use DFT to perform circular convolution - less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2)

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ]
L

˜ [n ] h

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 11 / 15

Frequency-domain convolution
4: Linear Time Invariant Systems

Idea: Use DFT to perform circular convolution - less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2)

• LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain
convolution

˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ]
(4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2.
L

˜ [n ] h

• Overlap Add • Overlap Save • Summary • MATLAB routines

DSP and Digital Filters (2012-2446)

LTI Systems: 4 – 11 / 15

Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not).Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2. L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 .

Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2. Total operations: ≈ 12L log2 L ≈ 12 (M + N ) log2 (M + N ) L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 . Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not).

L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 . Total operations: ≈ 12L log2 L ≈ 12 (M + N ) log2 (M + N ) Beneficial if both M and N are >∼ 100 . Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not).Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2.

N = 104 : L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 . Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not). Example: M = 103 . Total operations: ≈ 12L log2 L ≈ 12 (M + N ) log2 (M + N ) Beneficial if both M and N are >∼ 100 .Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2.

Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not).Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2. Total operations: ≈ 12L log2 L ≈ 12 (M + N ) log2 (M + N ) Beneficial if both M and N are >∼ 100 . N = 104 : Direct: 2M N = 2 × 107 L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 . Example: M = 103 .

less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2. Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not). Total operations: ≈ 12L log2 L ≈ 12 (M + N ) log2 (M + N ) Beneficial if both M and N are >∼ 100 . N = 104 : Direct: 2M N = 2 × 107 with DFT: 12 (M + N ) log2 (M + N ) = 1.Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution .8 × 106 L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 . Example: M = 103 .

N = 104 : Direct: 2M N = 2 × 107 with DFT: 12 (M + N ) log2 (M + N ) = 1.less computation (1) Choose L ≥ M + N − 1 (normally round up to a power of 2) • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution ˜ [n ] (2) Zero pad x[n] and h[n] to give sequences of length L: x ˜[n] and h ˜ [k ] H ˜ [k]) = x (3) Use DFT: y ˜[n] = F −1 (X ˜ [n ] (4) y [n] = y ˜[n] for 0 ≤ n ≤ M + N − 2.8 × 106 But: (a) DFT may be very long if N is large (b) No output samples until all x[n] has been input. Example: M = 103 .Frequency-domain convolution 4: Linear Time Invariant Systems Idea: Use DFT to perform circular convolution . L ˜ [n ] h • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 11 / 15 . Total operations: ≈ 12L log2 L ≈ 12 (M + N ) log2 (M + N ) Beneficial if both M and N are >∼ 100 . Arithmetic Complexity: DFT or IDFT take 4L log2 L operations if L is a power of 2 (or 16L log2 L if not).

Overlap Add 4: Linear Time Invariant Systems If N is very large: • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] Each output chunk is of length K + M − 1 and overlaps the next chunk • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

K = 104 .Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N Example: M = 500. N = 107 Direct: 2M N = 1010 • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

N = 107 Direct: 2M N = 1010 single DFT: 12 (M + N ) log2 (M + N ) = 2.8 × 109 • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 . K = 104 .Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N Example: M = 500.

N = 107 Direct: 2M N = 1010 single DFT: 12 (M + N ) log2 (M + N ) = 2. K = 104 .Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N Example: M = 500.8 × 109 N × 8 (M + K ) log2 (M + K ) = 1.1 × 109 overlap-add: K • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

1 × 109 overlap-add: K Other advantages: (a) Shorter DFT • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 . N = 107 Direct: 2M N = 1010 single DFT: 12 (M + N ) log2 (M + N ) = 2. K = 104 .Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N Example: M = 500.8 × 109 N × 8 (M + K ) log2 (M + K ) = 1.

K = 104 .Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N Example: M = 500. N = 107 Direct: 2M N = 1010 single DFT: 12 (M + N ) log2 (M + N ) = 2.8 × 109 N × 8 (M + K ) log2 (M + K ) = 1.1 × 109 overlap-add: K Other advantages: (a) Shorter DFT (b) Can cope with N = ∞ • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 .

1 × 109 overlap-add: K Other advantages: (a) Shorter DFT (b) Can cope with N = ∞ (c) Can calculate y [0] as soon as x[K − 1] has been read: algorithmic delay = K − 1 samples • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 12 / 15 . N = 107 Direct: 2M N = 1010 single DFT: 12 (M + N ) log2 (M + N ) = 2.8 × 109 N × 8 (M + K ) log2 (M + K ) = 1. K = 104 .Overlap Add 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution If N is very large: N (1) chop x[n] into K chunks of length K (2) convolve each chunk with h[n] (3) add up the results Each output chunk is of length K + M − 1 and overlaps the next chunk N Operations: ≈ K × 8 (M + K ) log2 (M + K ) Computational saving if ≈ 100 < M K N Example: M = 500.

Overlap Save 4: Linear Time Invariant Systems Alternative method: • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 (2) K +M −1 each chunk with h[n] • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 (2) K +M −1 each chunk with h[n] The first M − 1 points of each output chunk are invalid • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 (2) K +M −1 each chunk with h[n] (3) discard first M − 1 from each chunk The first M − 1 points of each output chunk are invalid • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 (2) K +M −1 each chunk with h[n] (3) discard first M − 1 from each chunk (4) concatenate to make y [n] The first M − 1 points of each output chunk are invalid • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 (2) K +M −1 each chunk with h[n] (3) discard first M − 1 from each chunk (4) concatenate to make y [n] The first M − 1 points of each output chunk are invalid Operations: slightly less than overlap-add because no addition needed to create y [n] • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

Overlap Save 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution Alternative method: N overlapping (1) chop x[n] into K chunks of length K + M − 1 (2) K +M −1 each chunk with h[n] (3) discard first M − 1 from each chunk (4) concatenate to make y [n] The first M − 1 points of each output chunk are invalid Operations: slightly less than overlap-add because no addition needed to create y [n] Advantages: same as overlap add Strangely. rather less popular than overlap-add • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 13 / 15 .

group delay BIBO stable. frequency response. • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 14 / 15 .Summary 4: Linear Time Invariant Systems • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution • • • • LTI systems: impulse response. Lossless systems Convolution and circular convolution properties Efficient methods for convolution ◦ single DFT ◦ overlap-add and overlap-save For further details see Mitra: 4 & 5. Passive. Causal.

*fft(y))) • Overlap Add • Overlap Save • Summary • MATLAB routines DSP and Digital Filters (2012-2446) LTI Systems: 4 – 15 / 15 .MATLAB routines 4: Linear Time Invariant Systems fftfilt • LTI Systems • Convolution Properties • BIBO Stability • Frequency Response • Causality • Passive and Lossless • Parallel and Series • Convolution Complexity • Circular Convolution • Frequency-domain convolution x [n ] y [n ] Convolution using overlap add real(ifft(fft(x).