29 DT Fir Filters

EECE 301
Signals & Systems

Prof. Mark Fowler
Note Set #29
D-T Systems: FIR Filters
1/16
FIR Filters (Non-Recursive Filters)

FIR (Non-Recursive) filters are certainly the most widely used DT filters.
There are several reasons for this:
Simple & effective design methods exist
Designs are always stable (since only poles at origin)
Linear phase designs can be easily achieved
However, the only real downside of FIR filters is that they can be
computationally complex
Well look at 3
methods
So its easy to
meet part of the
requirements for
ideal filters!
High-quality frequency response requires long filters (M > 100)

That means your computer will have to a lot of computation
M+1 Multiplies & M Additions
So The goal in FIR
design is to get the Freq.
Resp. specs you want
with the smallest M!!!
2/16
FIR LPF Design Windowed sinc Method (fir1)

bm
We looked at a truncated sinc as a way to design the bm:

But we saw that this gave only mediocre designs
in particular, it gave poor stop-band performance.
M/2
We could use math to explain why but the reason is pretty easy to see by recalling
that the frequency response of an FIR filter is the DTFT of the bm coefficients:
H () DTFT {bm } b0 b1e j ... bM e jM

10
0
-10
-20
|H ( ) |(d B )
Now we can reason that:

The truncation causes
discontinuities at the edges
of the bm sequence
Discontinuities (i.e., rapid
changes) require high
frequencies
Thus the poor stopband!!!
-30
Red: M = 30
Blue: M = 120
-40
-50
-60
-70
-80
-90
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
Normalized DT Frequency /
0.4
0.6
0.8
3/16
This reasoning also gives us an idea as to how to fix it:

Multiply the sinc-generated bm by a tapering window
Then we can try different shaped windows to see which is best
window
sinc
bm
1
M
M/2
M/2
M/2
MATLAB has a function that lets you easily design filters this way.
Thats a one
The command is called fir1 and here is how it is used:

>> b = fir1(M,omega_c,window(M+1,opt))
fir1 has many other options and

can be used to design more than
just LPF see MATLAB help.
Filter Order
Normalized Cutoff
Frequency
between 0 & 1
Specified Window
(length = M+1)
Some have optional parameters!
4/16
20
0
-20
M = 30
|H()| (dB)
-40
Blue: No Window (Rectangle Win)

Red: Hamming Window
Green: Chebyshev 60 dB
Magenta: Chebyshev 100 dB
-60
-80
-100
-120
-140
-160
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Note: Putting 60 (dB) in here gets a

stopband of about -70dB
>> b_rect=fir1(30,0.6,rectwin(31));
>> b_hamm=fir1(30,0.6,hamming(31));
>> b_chebwin=fir1(30,0.6,chebwin(31,60));
>> b_cheb_100=fir1(30,0.6,chebwin(31,100));
Note: Putting 100 (dB) in here gets a

stopband of about -120dB
A General Trend: For a fixed M choosing the window to get more stopband
attenuation (good!) causes the transition band to widen (bad!!).
5/16
A General Trend: For a fixed window choosing Larger M narrows transition

band (GOOD!).
20
0
-20
|H()| (dB)
-40
-60
Window: Chebyshev 100 dB
-80
-100
-120
-140
-160
-180
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
So the general design approach is guided trial and error:
Pick window to get desired stopband level (Chebyshev window helps with this)
Increase order (M = order) to get desired transition band width
Note that passband is VERY flat (good!)
6/16
FIR LPF Design Frequency Sampling Method (fir2)

Still uses windowing of a non-causal impulse response
But it comes from IFFTing a user specified a desired frequency response
This a simplified version of MATLABs fir2
function b = fir2_min(nn, ff, aa)

command the real one does not have this
% nn = order of the filter... must be an integer...
constraint and has many options for the design
%
Must be less than 1024!!!
% ff = row vector of frequency "break points" relative to 1 (which corresponds to Fs/2)
%
First element MUST be 0. Last element MUST be 1. Elements must NOT decrease
% aa = row vector of amplitudes desired (last element MUST be 0)
% Work with filter length instead of filter order
nn = nn + 1; %% nn as imput was order now nn is length
% Set some parameters needed
npt = 512; %%% npt sets the FFT size used.... This value is good as long as filter order < 1024
lap = fix(npt/25); % set # of points to use as transition band if no transition band is spec'd
wind = hamming(nn); % use Hamming window
1
%%% Convert from rows to columns

ff = ff'; aa = aa';
0.8
0.6
aa
aa = [1 1 0 0]
ff = [0 0.5 0.6 1]
MATLABs fir2 command

allows any window
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
ff
0.6
0.7
0.8
0.9
7/16
Value of H
% The next few lines and the loop below interpolate breakpoints onto large grid for FFT...
H = zeros(1,npt);
nbrk=length(ff);
nint=nbrk-1;
df = diff(ff');
npt = npt + 1; % Length of [dc 1 2 ... nyquist] frequencies.
nb = 1;
H(1)=aa(1);
for i=1:nint
if df(i) == 0
nb = ceil(nb - lap/2);
ne = nb + lap;
else
ne = fix(ff(i+1)*npt);
end
if (nb < 0 || ne > npt)
error(generatemsgid('SignalErr'),'Too abrupt an amplitude change near end of frequency interval.')
end
j=nb:ne;
if nb == ne
1
inc = 0;
else
inc = (j-nb)/(ne-nb);
0.5
end
H(nb:ne) = inc*aa(i+1) + (1 - inc)*aa(i);
0
nb = ne + 1;
0
100
200
300
400
500
Index of H
end
600
%% You now have magnitude interpolated onto a fine grid over the positive frequencies
8/16
%% Now want to apply a linear phase response over these positive frequencies. The phase slope is related to
%% amount of delay of filter.... the delay is what is needed to make the filter causal (the delay is half the order)
dt = 0.5 .* (nn - 1); % set delay to half the order (remember that nn is now length and order is length - 1)
rad = -dt .* sqrt(-1) .* pi .* (0:npt-1) ./ (npt-1); % create j*phi(n) that is a line with desired slope
H = H .* exp(rad); % multiply magnitude (H) by exp(j*phi(n)) to get mag & phase over positive frequencies
% Now...Append correct values for the negative frequencies.... remember that FT at negative frequencies is just
% conjugate of at positive freqs Also... since you are putting them "above" the positive freqs things have
% to "run backwards"... They go "above" because we won't use fftshift when we do the ifft
conj(H(npt-1:-1:2))];
1
Abs Value of H
H = [H
0.5
200
400
600
800
Index of H
1000
1200
200
400
600
800
Index of H
1000
1200
Angle of H
Linear phase but

wrapped between - &
2
0
-2
-4
9/16
%%% OK... now have the frequency response spec'd at all freqs on a fine grid and the ordering is positive freqs
%% first then negative freqs.... and we are all set for using ifft without fftshift
ht = real(ifft(H)); %%% technically don't need the real( ) operation but...
%% roundoff causes the imaginary part to be non-zero (but small!) so.... apply real( ) just to be sure
0.6
0.6
0.5
0.5
0.4
Value of b
0.3
0.2
0.3
0.2
0.1
0.1
-0.1
-0.1
200
400
600
Index of ht
800
1000
10
15
20
Index of b
25
30
35
1200
%%% Now you've got an imp. Resp. but it is much longer than desired... so extract the first nn points:
b = ht(1:nn);
%%% But that abrupt truncation can cause some problems with the resulting frequency response
0.6
%% So we apply a window to smooth the
0.5
%%% discontinuities at the edges:
0.4
b = b .* wind(:).'; % Apply window.

To see the designed filter's
frequency response:
>> [H,w] = freqz(b,1,8192);
>> plot(w/pi,20*log10(abs(H)))
Value of b
Value of ht
0.4
0.3
0.2
0.1
0
-0.1
10
15
20
Index of b
25
30
35
10/16
B = fir2(M,[0 0.3 0.31 1],[1 1 0 0],chebwin(M+1,60));

0
M = 30
M = 60
M = 120
M = 220
-20
|H()|(dB)
-40
Longer filter gives

better passband edge
-60
-80
-100
-120
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Just because you ASK for a specific transition band does not mean youll get it!!!
You have to ensure you make your filter long enough to get it!
11/16
One advantage of fir2 over fir1: it is easy to get very non-standard filter shapes!!
B = fir2(220,[0 0.4 0.6 0.7 0.8 1],[0.01 10 0 0 1 1],chebwin(221,60));
Remember these are non-dB values!
20
|H ( ) | (n o n - d B )
10
0
-20
|H()|(dB)
-40
-60
0
-80
-100
0.2
0.4
0.6
0.8
-120
-140
-160
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
B = fir2(220,[0 0.3 0.4 0.7 0.8 1],[1 1 0.1 0.1 0 0],chebwin(221,60));
Note: When the last element of aa is non-zero (e.g. for highpass) then the order
MUST be specified as being even!!!
12/16
FIR LPF Design Parks-McClellan (firpm)

This method is also called Optimal Equiripple Design
Unlike the other two methods the math here is quite complex so we wont study
HOW it does it but only how to apply it
This design method is pretty much the standard for FIR design these days,
Recall our earlier
visualization of how
we specify a lowpass
filter.
firpm and an
auxilliary command
allow us to specify
our desire for these
parameters and
then design a filter
to meet them!
13/16
% Lowpass Filter Design Specifications:

%
Passband cutoff frequency = 0.3 rad/sample
%
Stopband cutoff frequency = 0.31 rad/sample
%
At least 60 dB of stopband attenuation
%
No more than 1 dB passband ripple
Same as the LPF we

designed using fir2
About what we got
for our fir2 LPF
Our fir2 design gave
much lower PBR!!!
rp=1; rs=60; % specify passband ripple & stopband attenuation in dB

f_spec=[0.3 0.31]; % specify passband and stopband edges in normalized DT freq
AA=[1 0]; %%% specfies that you want a lowpass filter
dev=[(10^(rp/20)-1)/(10^(rp/20)+1) 10^(-rs/20)]; % parm. needed by design routine
Fs=2; % Fake value for Fs so our design is done in terms of normalized DT freq
[N,fo,ao,w]=firpmord(f_spec,AA,dev,Fs);
% estimates filter order and gives other parms needed to run firpm
b=firpm(N,fo,ao,w);
% Computes the designed filter coefficients in vector b
The resulting value for the order for this design is 385!!
14/16
firpm design
Order = 385
0
-20
|H()| (dB)
-40
-60
-80
-100
fir2 design
Order = 385
-120
-140
-160
0.1
0.2
0.3
0.4
0.5
0.6
0.8
0.9
firpm design has 1 dB of ripple.

Could reduce spec but would
need longer filter. E.g., for rp = 0.1
wed get Order = 544
5
0
-5
-10
|H()| (dB)
0.7
-15
-20
-25
-30
-35
0.24
0.25
0.26
0.27
0.28
0.29
0.3
0.31
0.32
firpm can design outstanding

filters but for the most
stringent design specs they can
be VERY long!
15/16
|H( )| (dB)
Lets look at pole-zero plot for a simpler firpm-designed filter

0
-20
-40
-60
<H( ) radians
-80
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0
-10
-20
-30
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Linear Phase
all designs by
firpm have this
very desirable
trait!!!
>> zplane(b,1)
Imaginary Part
1
0.5
28
0
-0.5
In Stopband: zeros placed right on UC

In Passband: zeros line the UC
-1
-1
-0.5
0.5
1.5
2
Real Part
2.5
3.5
16/16

29 DT Fir Filters

Uploaded by

Copyright:

Available Formats

29 DT Fir Filters

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

29 DT Fir Filters

Uploaded by

Copyright:

Available Formats

EECE 301

Signals & Systems

FIR Filters (Non-Recursive Filters)

High-quality frequency response requires long filters (M > 100)

FIR LPF Design Windowed sinc Method (fir1)

We looked at a truncated sinc as a way to design the bm:

H () DTFT {bm } b0 b1e j ... bM e jM

Now we can reason that:

This reasoning also gives us an idea as to how to fix it:

The command is called fir1 and here is how it is used:

fir1 has many other options and

Blue: No Window (Rectangle Win)

Note: Putting 60 (dB) in here gets a

Note: Putting 100 (dB) in here gets a

A General Trend: For a fixed window choosing Larger M narrows transition

Window: Chebyshev 100 dB

So the general design approach is guided trial and error:

FIR LPF Design Frequency Sampling Method (fir2)

function b = fir2_min(nn, ff, aa)

%%% Convert from rows to columns

MATLABs fir2 command

Linear phase but

b = b .* wind(:).'; % Apply window.

B = fir2(M,[0 0.3 0.31 1],[1 1 0 0],chebwin(M+1,60));

Longer filter gives

B = fir2(220,[0 0.3 0.4 0.7 0.8 1],[1 1 0.1 0.1 0 0],chebwin(221,60));

FIR LPF Design Parks-McClellan (firpm)

% Lowpass Filter Design Specifications:

Same as the LPF we

rp=1; rs=60; % specify passband ripple & stopband attenuation in dB

% Computes the designed filter coefficients in vector b

firpm design has 1 dB of ripple.

firpm can design outstanding

Lets look at pole-zero plot for a simpler firpm-designed filter

In Stopband: zeros placed right on UC

You might also like