ORAVYS has integrated 33 analysis engines built on techniques derived from 37+ expired patents. These engines implement proven acoustic analysis methods that were previously locked behind patent protection but are now freely available for commercial use.

Patent Mining

Our research team systematically reviewed thousands of expired patents in the fields of speech processing, voice biometrics, psychoacoustics, and signal analysis. We identified 37 patents whose techniques could be adapted to modern deep learning frameworks while maintaining the scientific rigor of the original inventions.

IP Protection

Every engine implementation is documented in our e-Soleau deposit with INPI (French National Institute of Industrial Property) and timestamped using Bitcoin's OpenTimestamps protocol. This provides cryptographic proof of our implementation dates, protecting our specific adaptations and novel combinations of these techniques.

Key Technologies

The expired patent engines cover areas including:

  • Formant tracking and vocal tract modeling
  • Pitch perturbation analysis (jitter, shimmer, HNR)
  • Spectral envelope decomposition
  • Glottal source parameter extraction
  • Microphone array beamforming for noise suppression