SLOP3255 · Week 05 · Time–frequency data

The noisy Sun

A bright streak can be a solar burst, local interference, an instrument effect—or not enough evidence.

A dynamic spectrum has three variables

xobservation time
yradio frequency channel
colourreceived power or intensity

Audio unfolds in one time dimension. We must decide how the second axis enters the listening experience.

Read the picture before mapping it

High radio frequency

Middle radio frequency

Low radio frequency

Ask whether a feature is narrowband, broadband, vertical, drifting or persistent. Geometry suggests—but does not prove—cause.

Different shapes imply different questions

Visible patternPossible readingAuditory test
fast frequency driftchanging emission regiondoes pitch sweep over time?
vertical broadband lineshort event or impulsive RFIdoes it appear across channels together?
persistent narrow bandlocal transmitter candidatedoes one pitch remain fixed?
blank rectanglemissing data or receiver stateis absence explicitly marked?

Pattern recognition generates hypotheses. Metadata and comparison test them.

Not every bright feature is solar

Radio-frequency interference

  • may remain at fixed local frequencies
  • may begin or end abruptly
  • can overwhelm astrophysical structure

Solar candidate

  • compares across observatories
  • fits plausible drift behaviour
  • survives instrument-quality checks

The sonification should not turn a classification guess into certainty.

Three routes through the same matrix

RowsHear one frequency channel over time.
ColumnsHear the spectrum at each time step.
PeaksTrack only strongest energy by time.
LayersAssign bands to distinct timbres.

Each route discards structure. The legend must say which structure survives.

Silence can mislead

Zero

The instrument measured no signal above its defined baseline.

Missing

No valid measurement exists for this interval.

Muted

The designer chose not to render a valid value.

These states may all sound quiet, but they make different claims.

Make the quality channel audible

Data statePrimary soundQuality cue
valid samplemapped tonenone
suspected RFImapped tonedry click or timbre marker
missing intervalno mapped toneexplicit gap cue
clipped receiverbounded tonewarning layer

Quality should not compete with the main scientific pattern, but it cannot disappear.

Listening Lab 05 · Output: classification card

Noise or event?

Task

Compare three synthetic dynamic spectra and their row-scan audio. Classify each feature with a confidence level.

Evidence

  • shape
  • quality flags
  • cross-station check

Reference: CSIRO Space Weather Services · solar radio observations