Created
December 2, 2021 22:32
-
-
Save gorbatschow/67ecf4173853c00b4a43297a5f2a864e to your computer and use it in GitHub Desktop.
Matlab etude demonstrates spectrum amplitude error when frequency is not a product of DF by integer
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
% Spectral analysis etudes | |
% This etude demonstrates spectrum amplitude error when frequency is not a | |
% product of df by integer | |
% - Try to change f1 variable to see wrong amplitude on spectrum | |
% - Try to change NFFT variable to see zero-padding effect. | |
% code by [email protected] | |
clear | |
% PARAMETERS | |
% ------------------------------------------------------------------------- | |
% Sample rate (Hz) | |
Fs = 48e3; | |
% Time resolution (secs) | |
dt = 1/Fs; | |
% Number of signal samples | |
NSIG = 2^10; | |
% Number of FFT input samples (zero padded signal samples) | |
NFFT = 2^10; | |
% Actual frequency resolution (Hz) | |
df = Fs/NSIG; | |
% Interpolated frequency resolution (Hz) | |
dfi = Fs/NFFT; | |
% Central frequency (Hz) | |
f0 = 100*df; | |
% Frequency difference between neighbor tones (Hz) | |
f1 = 2.1*df; | |
% Number of tone signals | |
nf = 15; | |
% Tones frequencies (Hz) | |
fsig = (f0-(nf-1)/2*f1):f1:(f0+nf/2*f1); | |
% Tones amplitudes 0-Pk (Volts) | |
asig = ones(size(fsig)); % uniform | |
%asig = (1:numel(fsig)) / numel(fsig); % stairs | |
% GENERATOR | |
% ------------------------------------------------------------------------- | |
% Generate tone signals | |
s = zeros(nf, NSIG); | |
t = (0:NSIG-1)*dt; | |
for i = 1:nf | |
s(i,1:NSIG) = asig(i)*sin(2*pi*fsig(i)*t); | |
end | |
% PROCESSOR | |
% ------------------------------------------------------------------------- | |
S = zeros(nf, NFFT/2-1); | |
for i = 1:nf | |
% Calculate FFT magnitude | |
x = zeros(1,NFFT); | |
x(1:NSIG) = s(i,:); | |
X = fft(x, NFFT); | |
X = X(2:NFFT/2); % X(1) is DC | |
X = abs(X); | |
% Multiply by 2 for 0-Pk amplitude | |
X = X/NSIG*2; | |
% Store | |
S(i,:) = X; | |
end | |
% RESULTS | |
% ------------------------------------------------------------------------- | |
% Interpolated spectrum frequnecies (Hz) | |
f = (1:NFFT/2-1)*dfi; | |
f_khz = f*1e-3; | |
fsig_khz = fsig*1e-3; | |
% Find max for each spectrum | |
apk = zeros(1,nf); | |
fpk = zeros(1,nf); | |
for i = 1:nf | |
[~, j] = max(S(i,:)); | |
apk(i) = S(i,j); | |
fpk(i) = f_khz(j); | |
end | |
[fpk, sidx] = sort(fpk); | |
apk = apk(sidx); | |
figure; | |
plot(f_khz, S', '-'); | |
hold on; | |
plot(f_khz, S', '.', 'Color', [0.5 0.5 0.5]); | |
plot(fpk, apk, 'LineWidth', 2, 'Color', [0 0 0]); | |
fb = max(fsig_khz) - min(fsig_khz); | |
fc = mean(fsig_khz); | |
xlim([fc-fb fc+fb]); | |
ylim([0 1.1*max(asig)]); | |
xlabel('kHz'); ylabel('V 0-Pk'); | |
grid on; grid minor; | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment