# Experiment objective

Estimate the frequency-dependent transfer function between two synchronized voltage signals. The generated output equals 2.5 times the input, so H1 and H2 magnitude should be 2.5 and coherence should approach 1 at 64, 256 and 640 Hz.

# Cross-spectral transfer experiment

1. Import `raw/cross_spectral_transfer_demo.csv` as a time-series Dataset. Set `time_s` to independent elapsed time in seconds and both voltage columns to measured variables in V.
2. Create a Temporal Window Set using one-second windows with a one-second step. The source contains 12 complete windows.
3. Enable **Cross-Spectral & Transfer Function Analysis**. Bind Time to `time_s`, Input X to `input_V`, and Output Y to `output_V` from the same Dataset.
4. Select the Temporal Window Set, Hann weighting, mean detrending, automatic NFFT and serial or automatic workers. Run the analysis.
5. Inspect the published spectrum Dataset and the public plot.dataset graphs. At 64, 256 and 640 Hz, compare H1 and H2 magnitude with the known gain 2.5, phase with 0 rad, and coherence with 1.
6. Change only a plot title, axis range or variant and rerender. Verify that the scientific Dataset is reused. Then change a scientific FFT setting and explain why the cached result must be recalculated.

The source is deterministic synthetic teaching data, generated by this package. It is not a measured instrument recording. Preserve the source generator version and ground-truth file with the Session.