Plot Error and Residual Between Two Signals

The standard way to visualize signal discrepancies is a synchronized two-panel plot: the top panel overlays both signals, and the bottom panel displays the residual error.

Method 1: Two-Panel Stacked Subplot (Signals + Residual)

% Generate sample signals (Reference vs Measured with noise)
t = linspace(0, 1, 1000);
y_true = 5 * sin(2 * pi * 5 * t);
y_meas = y_true + 0.3 * randn(size(t));

% Calculate error metrics
residual = y_true - y_meas;
mae = mean(abs(residual));
rmse = sqrt(mean(residual.^2));
% Create synchronized figure
figure('Color', 'w', 'Position', [100, 100, 750, 500]);
% Top Panel: Signal Comparison
ax1 = subplot(2, 1, 1);
plot(t, y_true, 'b-', 'LineWidth', 1.5, 'DisplayName', 'Reference');
hold on;
plot(t, y_meas, 'r--', 'LineWidth', 1.0, 'DisplayName', 'Measured');
ylabel('Amplitude');
title(sprintf('Signal Comparison (RMSE = %.3f, MAE = %.3f)', rmse, mae));
legend('Location', 'northeast');
grid on;
% Bottom Panel: Residual Error
ax2 = subplot(2, 1, 2);
plot(t, residual, 'k-', 'LineWidth', 1.0, 'DisplayName', 'Residual (True - Meas)');
hold on;
yline(0, 'r--', 'LineWidth', 1.2); % Zero reference line
xlabel('Time (s)');
ylabel('Error (e = y_{true} - y_{meas})');
title('Residual Error');
grid on;
% Synchronize X-axis zoom and pan
linkaxes([ax1, ax2], 'x');

Method 2: Shaded Error Band (Direct Fill)

This method fills the area between two curves to highlight divergence directly on a single plot.

t = linspace(0, 2*pi, 300);
sig1 = sin(t);
sig2 = sin(t) + 0.15 * cos(3*t);

figure('Color', 'w');
% Fill region between curves
fill([t, fliplr(t)], [sig1, fliplr(sig2)], [0.9 0.8 0.8], ...
    'EdgeColor', 'none', 'DisplayName', 'Discrepancy');
hold on;
plot(t, sig1, 'b-', 'LineWidth', 1.5, 'DisplayName', 'Signal 1');
plot(t, sig2, 'r-', 'LineWidth', 1.5, 'DisplayName', 'Signal 2');
xlabel('Time');
ylabel('Amplitude');
title('Error Highlighted with Shaded Fill');
legend('Location', 'northeast');
grid on;

Method 3: Stem Plot for Discrete/Sampled Residuals

t_sample = 0:0.02:0.5;
ref = sin(2*pi*4*t_sample);
est = ref + 0.1 * randn(size(t_sample));
res = ref - est;

figure('Color', 'w');
stem(t_sample, res, 'filled', 'Color', [0.8 0.2 0.2], 'LineWidth', 1.2);
yline(0, 'k-');
xlabel('Time (s)');
ylabel('Residual Error');
title('Discrete Residual Stem Plot');
grid on;