Temperature Sonification and Environmental EMF Artifact Detection

Author: Troy McQuinn
Status: Experimental / Open Research
Keywords: sonification, EMF interference, geomagnetic storms, pareidolia, audio synthesis, environmental sensing

---

What This Is

This project explores the unexpected and oddly structured results of sonifying temperature data collected from a low-cost USB thermometer. By mapping each 5-minute temperature reading to a single 16-bit audio sample and playing the resulting waveform at 8,000 Hz, the dataset was time-compressed by a factor of 2.4 million.

Surprisingly, the resulting audio sometimes exhibits speech-like formants and syllabic rhythms particularly during periods of known geomagnetic activity.

---

Why It’s Weird

While it started as a quirky sonification experiment, cross-referencing the audio with NOAA space weather logs revealed a persistent correlation between structured audio artifacts and solar/geomagnetic storm windows.

Even after ruling out software artifacts via independent Python and PHP implementations, the effect remained. This suggests that the USB thermometer might be acting (unintentionally) as a crude EMF sensor due to poor shielding or internal analog quirks.

---

What’s Inside

• paper/ – Full write-up in ODT and PDF formats (with figures and event alignment)
• code/ – Sonification scripts in both PHP and Python
• audio_samples/ – WAV files of real and simulated data
• figures/ – Spectrograms and waveform plots of key audio segments
• data.zip – Temperature log data
• README.md – This file

---

Simulated Control Experiments

Several synthetic datasets were generated to test whether pareidolia alone could explain the perception of speech-like structure. These included:

• Simulated thermal cycles with realistic modulation
• Formant-band noise shaping
• Chaotic amplitude envelopes
• Transient consonant-like bursts
• Phrase pacing and pitch glides

Despite best efforts, none of the synthetic signals reproduced the same kind of uncanny speech illusion found in the real dataset, suggesting the phenomenon may involve real-world nonlinearities or subtle environmental-electronic interactions.

See Section 6 of the write-up for detailed results.

---

Reproduction Instructions

All code runs with standard PHP 7+ or Python 3.8+ with numpy, scipy, and matplotlib.

1. Clone the repo.
2. Unzip and place your .dat log files in data/ (format: <timestamp> <temp> <date> <time>).
3. Run the appropriate script in code/ to generate a .wav file.
4. Analyze audio in Audacity or spectrogram tools like SoX or matplotlib.

---

Reflections

This work walks a fine line between traditional signal analysis and what could be described as accidental instrumentation. Whether these structures are artifacts, interference, or something more exotic, they appear real, repeatable, and worthy of further exploration.

Questions? Feedback? Want to fork this into a haunted USB ghost detector? Go for it.

---

License

• Code: MIT License
• Write-up: Creative Commons BY-NC-SA 4.0