Can I specify the detrending option in the PWELCH function in Signal

John Williams answered . 2025-03-27 13:33:52

% Example signal
Fs = 1000;            % Sampling frequency (Hz)
t = 0:1/Fs:1;         % Time vector
x = cos(2*pi*50*t) + randn(size(t));  % Example signal with noise

% Detrend the signal (linear detrending by default)
x_detrended = detrend(x);

% Use pwelch with detrended signal
[pxx, f] = pwelch(x_detrended, [], [], [], Fs);

% Plot the result
figure;
plot(f, 10*log10(pxx)); % Plot in dB
xlabel('Frequency (Hz)');
ylabel('Power/Frequency (dB/Hz)');
title('Power Spectral Density (After Detrending)');
grid on;

Explanation:

• Detrending: The detrend(x) function removes a linear trend from the signal x, which is the default behavior for dflag = 'linear' in the older psd function.
• pwelch: After detrending the signal, you can call pwelch to compute the power spectral density.

Advanced Detrending:

If you need other types of detrending, such as removing a constant (mean removal) or a custom polynomial trend, you can use detrend(x, 'constant') or specify a higher-order polynomial with detrend(x, order).

Handling Multiple Types of Detrending:

• Constant Detrending: x_detrended = detrend(x, 'constant'); removes the mean of the signal.
• Polynomial Detrending: x_detrended = detrend(x, 2); removes a quadratic trend.

For most cases, detrending manually before applying pwelch is a flexible approach. Let me know if you need more specific details!

Not satisfied with the answer ?? ASK NOW