DFT Bin Response

Below is the Matlab for Fig.6.3:

% Parameters (sampling rate = 1)
N = 16;               % DFT length
k = N/4;              % bin where DFT filter is centered
wk = 2*pi*k/N;        % normalized radian center-frequency
wStep = 2*pi/N;
w = [0:wStep:2*pi - wStep]; % DFT frequency grid

interp = 10;
N2 = interp*N; % Denser grid showing "arbitrary" frequencies
w2Step = 2*pi/N2;
w2 = [0:w2Step:2*pi - w2Step]; % Extra dense frequency grid
X = (1 - exp(j*(w2-wk)*N)) ./ (1 - exp(j*(w2-wk)));
X(1+k*interp) = N; % Fix divide-by-zero point (overwrite NaN)

% Plot spectral magnitude
clf;
magX = abs(X);
magXd = magX(1:interp:N2); % DFT frequencies only
subplot(2,1,1);
plot(w2,magX,'-'); hold on; grid;
plot(w,magXd,'*');         % Show DFT sample points
title('DFT Amplitude Response at k=N/4'); 
xlabel('Normalized Radian Frequency (radians per sample)'); 
ylabel('Magnitude (Linear)');
text(-1,20,'a)');

% Same thing on a dB scale
magXdb = 20*log10(magX);       % Spectral magnitude in dB
% Since the zeros go to minus infinity, clip at -60 dB:
magXdb = max(magXdb,-60*ones(1,N2));
magXddb = magXdb(1:interp:N2); % DFT frequencies only
subplot(2,1,2);
hold off; plot(w2,magXdb,'-'); hold on; plot(w,magXddb,'*'); 
xlabel('Normalized Radian Frequency (radians per sample)'); 
ylabel('Magnitude (dB)'); grid;
text(-1,40,'b)');
print -deps '../eps/dftfilter.eps';
hold off;

---

Order a Hardcopy of Mathematics of the DFT

Previous: DFT Sinusoids for
Next: DFT Matrix

---

written by Julius Orion Smith III
Julius Smith's background is in electrical engineering (BS Rice 1975, PhD Stanford 1983). He is presently Professor of Music and Associate Professor (by courtesy) of Electrical Engineering at Stanford's Center for Computer Research in Music and Acoustics (CCRMA), teaching courses and pursuing research related to signal processing applied to music and audio systems. See http://ccrma.stanford.edu/~jos/ for details.