Provenance classes#
| Badge | Meaning | Examples |
|---|---|---|
| MEASURED | Recorded directly by NOESIS. | Microphone, camera frames, accelerometer, RSSI, raw IQ, timestamps. |
| DERIVED | Computed from measured data by a stated algorithm. | Spectra, envelopes, RF PSD and features, band tracks, detections, Synthesis features. |
| INFERRED | An estimate that depends on assumptions. | Gravity direction (Plumb Line), pulse candidate. |
| ABSTRACT MAPPING | An interpretive visual or audio mapping of recorded data. | EM Fog, Backlight Aura, Mass-Warp, sonification. |
The NOESIS classification of each stream is kept verbatim in the Inspector (NOESIS class) next to the display class.
Missing data#
- Values that were not recorded read Not recorded; timeline gaps are drawn as breaks; Synthesis cells are hatched.
- Nothing is interpolated, carried forward or zero-filled for display.
- Views and detectors that need unrecorded data say exactly which input is missing.
Timing uncertainty#
Each stream reports what its timestamps mark and how precise they are, for example:
| Stream | Typical uncertainty |
|---|---|
| Microphone | Block timestamp jitter (sub-millisecond, relative precision only) |
| Camera | About one frame interval (host arrival after decode) |
| Polled telemetry | About one poll interval |
| Window input | About one GUI frame |
| RF windows | Each window's own acquisition span; sweeps are sequential |
Correlation results combine the uncertainty of both streams. See the Inspector's Timing section for any stream.
Data quality#
Quality is reported per modality rather than as one score: dropped blocks, gaps, clipping, constant channels, jitter, missing RF windows, duplicate windows, incomplete sweeps, missing Synthesis cells and more. Warnings appear as ⚠ counts on tracks and streams and in full on the Inspector's Stream tab and in reports.
Simulated sessions#
Sessions recorded with NOESIS Simulation are labelled SIMULATED SESSION — synthetic data in the viewer toolbar, the burned-in timestamp, the Session tab and every report.