the signal language. a signal is a waveform — a continuous function sampled at discrete points
| Op | Action |
|---|---|
fft(x) |
Fast fourier transform |
ifft(X) |
Inverse FFT |
convolve(a, b) |
Convolution of two signals |
lowpass(x, cutoff) |
Low-pass filter |
resample(x, rate) |
Resample to new rate |
correlate(a, b) |
Cross-correlation |
energy(x) |
Signal energy |
peak_detect(x) |
Find peaks |
use cases: sensor data processing, audio, seismic, environmental monitoring. and — critically — Goldilocks homomorphic encryption: polynomial multiplication in R_q = Z_q[X]/(Xⁿ+1) is negacyclic convolution of coefficient vectors. the same butterfly network, the same NTT engine. Wav is the language of sensing at all scales — from physical waveforms to encrypted computation over polynomial rings. FHE is not a separate concern; it is sensing at the algebraic level, where the signal is a ciphertext. Wav extended with noise budget types becomes the FHE compiler: noise tracking, modulus ladder management, PBS scheduling. compiles to Trident for proof of correct R_q operations
see cyb/languages for the complete language set. see cyb/multiproof for the proving architecture