The waterfall (time cascade)

The spectrum alone shows only the present instant. The waterfall adds time: each new FFT becomes a coloured horizontal line (colour = the power at each frequency), and lines scroll downward. You read frequency in X, time in Y.

The bestiary: learning to read shapes

It's the most powerful tool for understanding a signal, because it reveals its behaviour over time:

 frequency →                          what it is
 ┃                                    stable carrier
 ┃          ▌ ▌    ▌                  intermittent bursts (sensors, LoRa)
 ┃         ╱  ╱   ╱                   diagonal chirps (LoRa)
 ┃      ▖ ▘▗  ▘ ▖ ▗ ▘                 frequency hopping (Bluetooth)
 ┃   ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓                  wide grainy block (WiFi/OFDM, drone)
 ↓ time

• A continuous vertical line = a stable carrier (an FM station, a beacon).
• Short, spaced dashes = intermittent bursts (ISM sensors, remote controls).
• Brief diagonals = LoRa chirps.
• Dots hopping across the band = frequency hopping (Bluetooth, some drones).
• A wide, fuzzy band = a wideband signal like WiFi/OFDM or a drone video link.

Setting the scale: where everything is decided

The choice of palette and scale (min/max in dB) changes everything: well set, a weak signal leaps out; badly set, it drowns.

• Min: pin it just below the noise floor. Too low and everything is dark, you lose contrast; too high and you erase weak signals.
• Max: just above your strongest peak. The min–max gap is your "contrast window".
• On this site the scale adjusts automatically around the current floor and peak — in other software (SDR++, GQRX), taming this manual setting is the first skill to learn.

The three-question reading method

1. Width? Narrow (kHz) = voice/telemetry; medium (100–200 kHz) = FM broadcast/LoRa; very wide (MHz) = OFDM, video.
2. Continuity? Permanent = broadcast/beacon; periodic = sensor (count the seconds between bursts!); erratic = human activity.
3. Motion? Fixed frequency = assigned channel; slow drift = cheap oscillator or Doppler (satellites!); jumps = FHSS.

With those three answers you can recognise a modulation by eye before decoding anything. A sensor's period, a satellite's Doppler drift, a drone's signature: it's all in this drawing.

Your turn

1. A fine line shows for 1 s exactly every 60 s. Hypothesis? (A fixed-period sensor — meter, weather station.)
2. A line drifts slowly downward over 10 minutes. Hypothesis? (Doppler from a passing satellite, or a warming oscillator.)
3. The whole waterfall is bright orange. First reflex? (Badly pinned scale or too much gain — reframe min/max before concluding.)

👉 Watch a live waterfall in any mission, e.g. Catch an FM station