Sei sulla pagina 1di 14

Digital Filters

By
Ch sai pavan kumar
Abinash sahoo
Digital filters
• A system that performs mathematical operations on a sampled,
discrete-time signal.

• It usually consists of

input output
ADC microprocessor DAC

*microprocessor-FPGA, AISC, DSP

• In DSP, We majorly use two categories of discrete-time filters called FIR


and IIR
FIR IIR
Finite Impulse Response Infinite impulse response

• 𝑦[n]=σ𝑀
𝑘=0 𝑏𝑘 𝑥[𝑛 − 𝑘]
• 𝑦[n]=σ𝑀
𝑘=0 𝑏𝑘 𝑥 𝑛 − 𝑘 − σ 𝑁
𝑘=1 𝑎𝑘 𝑦 𝑛 − 𝑘

σ𝑀
𝑘=0 𝑏𝑘 𝑧
−𝑘
• H[𝑧]=σ𝑀 𝑏
𝑘=0 𝑘 𝑧 −𝑘
• H(𝑧)= , 𝑎0 = 1
σ𝑁
𝑘=0 𝑎𝑘 𝑧
−𝑘
= 𝑏𝑜 ς𝑀
𝑘=1(1 − 𝑐𝑘 𝑧 −1
) 𝑏𝑜 ς𝑀 (1−𝑐 𝑧 −1 )
𝑘
= ς𝑁𝑘=1
𝑏𝑜 𝑘=1(1−𝑑𝑘 𝑧 −1 )
𝑏𝑛 , 0 ≤ 𝑛 ≤ 𝑀
• h[n]=
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 • h[n]=σ𝑁 𝑎 (𝑑 𝑛
𝑘=1 𝑘 𝑘 u[n]
)
Filter Designs
Desired frequency response-𝐻𝑑 (𝑒 𝑗𝜔 )

• FIR • IIR
σ𝑀 𝑏 𝑒 −𝑗𝑘𝜔
• H(𝑒 𝑗𝜔 )=σ𝑀
𝑘=0 𝑘𝑏 𝑒 −𝑗𝑘𝜔
• H(𝑒 𝑗𝜔 )=σ𝑁
𝑘=0 𝑘
, 𝑎0 = 1
𝑘=0 𝑎𝑘 𝑒 −𝑗𝑘𝜔

• Optimization based design • Transform analog filter design


• Arbitrary magnitude/phase
• Frequency selective gain
response
• Can obtain linear phase • Nonlinear phase, no control
• Can require large M • Fewer coefficients than FIR
Filter design
• The design of filters involves the
following stages
1. The specification of desired
properties of system
2. The approximation of the
specifications using a causal
discrete time system
3. The realization of the system
IIR filter design from an analogue prototype
• Step-1 Convert DF specs AF specs
• Step-2 Find the order(n) & cut-off frequency of prototype
filter H(S) (Butterworth / Chebyshev)
• Step-3 Frequency Transformations
(Prototype LPF H(s) Required Filter H’(s) )
• Step-4 Convert Analog T.F to Digital T.F
• Step-5 Filter Structure

• While converting it must meet following requirements


• Direct relationship between two frequency variables(in two domains).
• Stable analog filter after mapping must result in stable digital filter.
Design of IIR filter from analog filters

Approximation derivative
Impulse invariance transformation
Bilinear transformation
Approximation derivatives

1  z 1
s
T
Disadvantage of approximation derivative
• The possible location of digital filters are confined to small
frequencies.
• As consequence, the mapping is restricted to low pass and band pass
filter with relatively small resonant frequencies.
• It is not possible to transform a high pass analog filter into
corresponding high pass digital filter.
• In attempt to overcome this limitation, We go for more complex
substitution for derivative.
Impulse Invariance Technique

Complex-plane mapping in impulse invariance technique


Bilinear transformation
It is based on the relation ship:
Frequency warping

Potrebbero piacerti anche