OravontOravont Systems LLP

Open tools and data

Code and data released in the open

Oravont publishes what can be published: a classical DEMON toolkit, a physics-grounded synthetic dataset, and the artefacts behind the 2026 preprint — each with a licence that says exactly what may be done with it.

Enquire about the engineering tools
Toolkit · Python · code MIT, documentation and samples CC BY 4.0

DEMON-Analysis

Detection of Envelope Modulation on Noise for underwater acoustics: recovers a vessel's shaft rate and blade rate from radiated noise by demodulating cavitation. The pipeline has four steps — bandpass filtering in a cavitation-dominated band, Hilbert envelope extraction, AC-coupling to remove drift, and an FFT of the envelope into a low-frequency modulation spectrum — followed by reading the harmonic comb for shaft and blade rate. Input: audio clips as NumPy arrays; output: modulation-spectrum plots.

DEMON-Analysis output: modulation spectrum of a vessel recording, blade-rate harmonics marked What the toolkit produces: the DEMON spectrum of a recording, blade-rate harmonics marked.

github.com/Oravont/DEMON-Analysis · Cite as: Tyagi S. DEMON Analysis: Reading a Vessel's Propulsion from Its Own Noise. Oravont Systems LLP, 2026.

Dataset · CC BY 4.0

Underwater-Acoustic-Synthetic-Dataset

12,000 five-second clips at 16 kHz across four vessel classes — tanker, cargo ship, fishing vessel, small craft — and ambient ocean noise, generated from first-principle acoustic models with full-factorial parameter coverage. All quantities are maintained in physical units (pascals, dB re 1 µPa) and verified against classical references. Built for self-supervised pretraining of acoustic encoders; the generation physics and verification calculations are documented in the accompanying article.

Synthetic Underwater Acoustic Dataset: 12,000 physics-grounded clips for self-supervised learning 12,000 physics-grounded clips, four vessel classes plus ambient noise.

Hugging Face · DOI 10.57967/hf/11191 · GitHub repository (CC BY 4.0) · Article

Research artefacts · Zenodo · CC BY 4.0 with an MIT carve-out

SKANN preprint artefacts: checkpoint, embeddings, transit map, per-query outputs

Everything needed to recompute the numbers in the 2026 preprint: the model checkpoint with SHA-256 digests, validation embeddings, the audio-adjudicated transit-deduplication map and per-query evaluation outputs. The adjudicator script is released under MIT; the rest under CC BY 4.0. Source corpora are credited in the record: IARA (CC BY-NC 4.0), ShipsEar (distributed by the Universidad de Vigo on request) and Ocean Networks Canada, Oceans 3.0.

The evaluation these artefacts reproduce: protocol, 40-hull gallery, rank-1 and ranking AUC The numbers these files let you recompute (arXiv:2609.07399). Click to enlarge.

doi:10.5281/zenodo.22160138 (version) · doi:10.5281/zenodo.22160137 (concept) · Preprint: Tyagi, arXiv:2609.07399 · SKANN in brief

SKANN is published research, not an off-the-shelf software product; Oravont builds SKANN-based models for an organisation's own recordings as a service, including offline deployment with no data egress (Using SKANN on your own recordings).

Tutorial · Python · on suniltyagi.in

Python examples for underwater sound levels and spectra

Underwater acoustics for machine-learning engineers: pressure, levels and spectra, with Python (Sunil Tyagi, 27 October 2026): eight runnable examples (NumPy, SciPy, 16 kHz) for RMS pressure, sound pressure level in dB re 1 µPa, the power spectral density, band levels and spectral synthesis, shown with their outputs. Download the examples.

On enquiry

Engineering software and programme code

The stealth-engineering tools — SSR-Sim, Dhvanyakalan, ABNContour and SMS-Suite — are Oravont's own and are made available to defence organisations, shipyards and research institutions on enquiry. For the DISC5 programme, the enrolment-and-inference application is public on GitHub; the training code and its documentation are held privately.

Enquire

Work with Oravont. To discuss a project, write to info@oravontsystems.com, call +91 99222 42303 or use the contact page. Oravont Systems LLP, Noida, India.