1. Physical noise and SDR sampling
Most random numbers generated in digital computers rely on mathematical algorithms (PRNG). Physical entropy sources (TRNG) instead measure nondeterministic natural phenomena. Software Defined Radio receivers sample radio bands with minimal artificial emissions, recording background radiation, diode thermal noise, and receiver quantum noise.
Big Bang Entropy nodes analyze signal power spectral density in I/Q mode and extract micro-fluctuations from the least significant bits (LSB) of the analog-to-digital converter (ADC).
2. Cryptographic whitening (SHA-512)
Raw radio noise can contain small statistical bias caused by hardware characteristics or ambient temperature. To obtain a flatter probability distribution, raw bytes are processed with SHA-512 and von Neumann-style mixing.
3. Quality tests: Dieharder, ENT, and rngtest
Every generated entropy block is automatically checked with rigorous statistical tests: monobit test, runs test, autocorrelation, and 0/1 balance according to FIPS 140-2 and the Dieharder suite.
View current statistical reports ->4. Mixing entropy with a system CSPRNG
A cryptographic best practice is to feed the local `/dev/urandom` pool with external entropy bytes:
curl -s "https://entropy.sparksome.pl/raw?bytes=512" > /dev/urandom