Scenario-defined labels
Labels come from the graph and scenario configuration, not from fragile post-hoc manual annotation.
RFML datasets
Synthetic IQ datasets with scenario metadata.
RFix turns RF scenario graphs into labeled recordings for RFML training, receiver regression, and controlled comparison. Build the scene visually, then reuse it for dataset variants and automated generation.
Dataset workflow
From graph to labeled RFML package.
The existing export folder and file model becomes the dataset story: recordings, sidecar metadata, and project state travel together.
Repeatable variants
Generate parameter sweeps, seeds, channel changes, and mixture changes from the same base scene.
Exportable packages
Package IQ, SigMF metadata, and RFix project files so datasets stay traceable during model work.
What the requester should receive
A compact sample package should include IQ data, SigMF metadata, scenario notes, and the RFix project used to produce the recording.
Why this matters for RFML
Model teams need labels, provenance, and repeatability as much as they need raw samples. RFix can make those part of the generation flow.
Evaluating options?
How RFix compares.
See RFix side by side with the tools RFML teams usually consider for synthetic dataset generation.