Spectrum Analysis
Spectrum Analysis turns a selected time region into a frequency-domain view. Use it after you isolate the relevant section in Graph View, especially when the goal is to find resonances, broadband noise, or a mismatch between command and response.
Available analysis types
| Type | Use it for |
|---|---|
| Frequency Spectrum | Quick amplitude-oriented FFT inspection and peak hunting |
| Power Spectral Density (PSD) | Energy distribution in dB for noise floor and signal separation |
| Freq. vs Throttle | Heatmap view to see how frequency content changes with throttle |
| PID Error vs Setpoint | Check how control error behaves across command magnitude |
Window functions
Use Hann or Blackman-Harris when you want to reduce spectral leakage. Use None (Rectangular) only when you explicitly accept stronger leakage in exchange for sharper bins.
Peak detection
Peak detection highlights dominant frequencies in the standard spectrum view. Treat it as an aid, not as a diagnosis. A peak becomes meaningful only when it lines up with your time-domain evidence and flight context.
Frequency vs throttle
This heatmap is especially useful for motor or frame resonance work. Sparse data can produce gaps, so use a representative flight segment with enough throttle variation before interpreting the result.
PID error vs setpoint
This view helps you see whether control error grows in specific command ranges. It is most useful when you compare similar maneuvers across multiple flights after a tuning change.
Workflow: Isolate the event in Graph View, then run FFT or PSD on that same range. Avoid whole-flight spectra unless the entire log represents one consistent operating condition.
Want to test an actual file? Upload a .bbl file and compare the selected range against this guide.