Someone hands you an IQ recording you've never seen. No flowgraph. No documentation. No modulation parameters. Learn how to observe the signal, form hypotheses, test them and recover how the radio communicates.
Course 03 worked because you knew both sides of the link. Course 04 removes that advantage. Start only with observations and work backward from an unknown signal toward symbols, packets and information.
Use spectrum, waterfall, time-domain and I/Q views to build a signal profile before attempting to demodulate anything.
Ask what magnitude, frequency and phase reveal about how the unknown radio may be representing symbols.
Use controlled experiments to determine which packet fields respond when a known input or behavior changes.
Change one variable. Observe what changes. Form a hypothesis. Test it again. Keep observations separate from assumptions until the evidence supports the model.
Course 03 — Program Radios, or equivalent comfort with modulation, synchronization, framing and packet recovery.
Reverse the direction of the design process. Instead of choosing the waveform and packet structure, infer them from evidence.
Each stage answers the next question that appears when all you have is an unfamiliar capture.
Characterize center frequency, occupied bandwidth, burst behavior, repetition, power behavior and visible spectral features. Build the first signal dossier.
Use the Course 02 magnitude–frequency–phase mental model in reverse. Inspect amplitude behavior, instantaneous frequency and phase/IQ geometry for modulation clues.
Look for transitions, repetition and candidate symbol periods. Estimate symbol rate and determine where individual symbol decisions may belong.
Recover symbol/bit sequences and hunt for repeated patterns, common burst beginnings, packet boundaries and fields that stay constant or change.
Capture controlled transmissions while changing one known variable at a time. Compare captures and build a defensible packet map from evidence.
Test the inferred modulation, timing and packet model against additional recordings. Revise assumptions that fail and document confidence in each conclusion.
Start with an unfamiliar IQ recording and produce a signal/protocol explanation that leads all the way to recovered information.
The labs progressively remove documentation and known parameters. Your job is to decide what measurement or experiment will reveal the next useful fact.
Inspect an unknown capture in spectrum, waterfall, time and I/Q views. Record only what the evidence supports.
Compare magnitude, frequency and phase behavior and develop a modulation hypothesis from the observed signal.
Estimate a candidate symbol period, test it against transitions and determine whether recovered symbol states become stable.
Compare multiple bursts and expose repeating prefixes, constant regions and changing fields that suggest packet structure.
Alter one known transmitter input, capture again and identify exactly which recovered bits or symbols respond.
Turn repeated observations into a tentative packet map and label each field as observed, inferred or still unknown.
Apply the model to captures you did not use while developing it. A good hypothesis should predict what the new transmission contains.
Receive only an unfamiliar IQ capture. Characterize it, recover the symbols, infer the structure and extract the information.
You've learned to reason about and reverse engineer relatively simple radio waveforms. What happens when the signal is wideband and appears to contain many closely spaced carriers at once?
Course 04 turns the concepts from the first three courses into an investigative toolbox for radios you didn't design.
Join the course