Solution

Voice and channel measurement for human review

Sonotheia measures physical properties of the voice and the channel in call audio and reports each one as nominal, deviated or unavailable, with the validity of the recording. A reviewer decides what to do with the result. Sonotheia does not decide whether a voice is synthetic.

A voice-driven request needs review in its operational context. Measurements add evidence beside the reviewer's judgment; an in-range reading does not establish that a caller or request is genuine.

What does Sonotheia measure?

The instrument measures acoustic properties such as speech duration, background consistency, and changes in pitch and energy. Each observation is read with its threshold and known limits. A channel change can explain a deviation without implying synthesis.

What does the benchmark establish?

The benchmark configuration is documented internally, with the corpus, partition and simulated codec chain identified. No headline performance figure is published here. A corpus benchmark does not establish performance on a customer's calls. Sonotheia's measurements are benchmarked on simulated wideband, G.711 and AMR-NB chains built from public spoofing corpora. They have not been validated on live telephone traffic. A real trunk adds handset, network and transcoding effects the simulation did not include, so it stays pending until measured in place. Zoom, Teams and WebRTC meeting audio are out of scope.

How does the record support review?

A structured measurement record with the instrument version can support a compliance team's own review and SAR narrative preparation. It is not a qualified SAR filing or a determination that a reporting requirement has been met. The institution and its legal reviewers decide what to file.

Frequently asked questions

How is audio and analysis data handled?
Confirm data handling separately for your deployment: audio storage, result retention, access controls, training use and subprocessors before use.
Can Sonotheia replace a detection vendor?
Sonotheia provides measurements for a reviewer. It does not supply a synthetic-voice determination or replace an identity or transaction-authorization control.
Does the instrument identify a particular synthesis engine?
No engine-specific detection coverage is claimed. A physical observation is not attribution to a generator.
What does the analysis API return?
The analysis endpoint returns a measurement frame: the validity of the audio, voice and channel readings, and a recommended human action. It does not return an identity or fraud determination. Evidence exports are separate from this response.
How long does analysis take?
Analysis runs on uploaded or recorded audio. Timing depends on clip length, channel and deployment. No streaming or sub-second performance is claimed here.
What is the intended workflow?
The intended use is review of a recorded request after the call and before a transfer. Any live or demonstration mode is scoped separately.

Ready to evaluate Sonotheia for your voice channel?

Request a demo