Vous êtes sur la page 1sur 4

Digital Signal Processing Questions and Answers

Ques. 1)
a) Compare and contrast FIR filters and IIR filters.
Ans. a)

FIR FILTER IIR FILTER

FIR stands for finite impulse response. IIR stands for infinite impulse response.

FIR filters are more powerful than IIR filters, IIR filters are less powerful than FIR filters,
but also require more processing power & require less processing power and less
and more work to set up the filters. work to set up the filters.

More flexibility and ability to finely adjust These are less flexible.
the response.

It can be used to correct frequency- It cannot be used to correct frequency-


response errors. response errors.

FIRs can be limited in resolution at low IIRs can provide good resolution even at
frequencies, and the success of applying FIR low frequencies.
filters depends greatly on the program that
is used to generate the filter coefficients.

Usage is generally more complicated and Usage is generally more easier than FIR
time-consuming than IIR filters. filters.

FIR filter uses only current and past input IIR filter uses current input sample value,
digital samples to obtain a current output past input and output samples to obtain
sample value. It does not utilize past output current output sample value.
samples.

Simple FIR equation is mention below. Simple IIR equation is mention below.
y(n)= b(0)x(n) + b(1)x(n-1) + b(2)x(n-2) + y(n)= b(0)x(n) + b(1)x(n-1) + b(2)x(n-2) +
b(3)x(n-3) + b(4)x(n-4) b(3)x(n-3)+ … + a(1)y(n-1) + a(2)y(n-2) +
a(3)y(n-3) + …

Transfer function of FIR filter will have only


zeros, need more memory.

SHIKHAR TYAGI
Transfer function of IIR filter will have both
zeros and poles and will require less
memory than FIR counterpart.

FIR filters are preferred due to its linear IIR filters are not stable as they are
phase response and also they are non- recursive in nature and feedback is also
recursive. Feedback is not involved in FIR, involved in the process of calculating
hence they are stable. output sample values.

FIR filters are less efficient. IIR filters are more efficient.

FIR filters are used as anti-aliasing, low pass IIR filters are used as notch(band
and baseband filters. stop),band pass functions.

FIR filter need higher order than IIR filter to IIR filter need lower order than FIR filter to
achieve same performance. achieve same performance.

It has lower sensitivity than IIR filter. It has higher sensitivity than FIR filter

𝟏
b) Consider the first order analog filter H(s) = 𝐬+𝟏. Determine H(z) using the impulse
invariant method for a sampling rate of 1Hz.
Ans. b)
1
Given : H(s) = , Fs(Sampling frequency) = 1Hz.
s+1

Using impulse invariant method, we will first find h(t) from H(s) using inverse laplace
transformation, then we will sample h(t) at the given sampling rate to get h(nTs), then we
will apply Z-transform to get H(z).
We have,
1
H(s) = s+1

Applying inverse laplace transform,


1
h(t) = L-1 { s+1 } = e-t u(t) ,

where u(t) is a unit step signal.


Now we will sample h(t) at a rate of Fs = 1Hz, to do this simply replace ‘t’ with ‘nTs’, where Ts
1 1
is = 1 Hz = 1 sec.
Sampling Frequency

SHIKHAR TYAGI
Hence replacing ‘t’ with ‘n’,
h(n) = e-n u(n)
Now apply Z-transform on h(n) to get H(z),

H(z) = Z{h(n)} = ∑𝑛=−∞ h(n) 𝑧 −𝑛

= ∑𝑛=−∞ 𝑒 −𝑛 u(n) 𝑧 −𝑛

= ∑𝑛=0 𝑒 −𝑛 (𝑧 −𝑛 )
1 e.z 𝟐.𝟕𝟏𝐳
= 1−e−1 𝑧 −1 = e.z − 1 = 𝟐.𝟕𝟏𝐳 − 𝟏 (Answer)

c) Draw a table for magnitude of vector (z-0.48) with frequency and comment on the
results you get in terms of nature of the filtering capability of the vector.
Ans. c)
Given H(z) = z-0.48
Substituting z = ejΩ, we have frequency response as,
H(ejΩ) = ejΩ - 0.48 = (cos(Ω)-0.48) + j sin(Ω)
|H(ejΩ)| = √((cos(Ω)-0.48)2 + (sin(Ω))2)

Ω (radians) |H(ejΩ)| |H(ejΩ)| (db)

0 0.52 -5.67
π/4 0.74268 -2.5839
π/2 1.10923 0.9004
3π/4 1.3817 2.8085
π 1.48 3.4052

Conclusion: From above response we can see that as we are increasing the digital frequency from 0
to π, the magnitude of the response increase from 0.52 to 1.48, which refers that the digital filter
represented by H(z) is suppressing the low frequencies and passing / allowing the high frequencies.
Hence exhibiting a high-pass filter characteristics. (Answer)

𝐓
d) Simpson numerical integration is defined by 𝒚(𝒏) = 𝒚(𝒏 − 𝟐) + (𝒙(𝒏) +
𝟑
𝟒𝒙(𝒏 − 𝟏) + 𝒙(𝒏 − 𝟐)) . Derive a mapping rule to convert H(s) to H(z).

Ans. d)
First we will find H(z), applying Z-transform to the given equation,
T
Y(z) = z-2 Y(z) + 3 ( 1 + 4z-1 + z-2 ) X(z)

SHIKHAR TYAGI
Y(z) T 1 + 4z−1 + z−2 T z2 +4z+1
H(z) = X(z) = 3 . =3.
1−z−2 z2 − 1

1
Comparing above with the transfer function of an ideal integrator HI(s) = s

Hence,
𝟑 𝐳𝟐 − 𝟏
s = 𝐓 . 𝐳𝟐 +𝟒𝐳+𝟏 (Answer)

Now for a given H(s) we can find H(z) by just replacing s with above derived expression.

SHIKHAR TYAGI

Vous aimerez peut-être aussi