phosphoros

Every changehas a moment.

Phosphoros finds the moments when someone’s voice, face or pace departs from their own normal, and takes you to the exact second of the recording.

It shows what changed, not why. It doesn’t detect lies or emotions.

Explore the demoDiscuss the work

01 · Range

Start withwhat’s normal for them.

The soft ring is this person’s usual range, learned from steady stretches of this same recording.

02 · Change

Then watchwhat moves.

Read clockwise from the top, start to finish. Each fine line is one sampled moment. The longer the line, the bigger the change.

03 · Exchange

Keep every changein context.

The short arcs inside the ring are answers; a longer arc is a longer answer. Each change stays attached to the question and answer it happened in.

04 · Together

When sight and soundagree.

Amber marks answers where what the camera sees and what the microphone hears both moved beyond this person’s range at once.

05 · Inspect

Follow the change.Inspect the measurement.

Range built from Elevated samples Answers Sight and sound together

Four signals. One clock.

One change can be noise. Changes that arrive together are worth a look.

  1. 01

    Speech timing

    How fast they talk, how much they pause, and how long they take to answer.

    • speech rate
    • pause burden
    • response latency
  2. 02

    The face

    52 small facial movements, grouped into expressions, plus eye opening, gaze and head turns.

    • 52 blendshapes
    • 15 action families
    • 6 left–right asymmetry
    • microexpression
    • eye aperture
    • gaze
    • head
  3. 03

    The voice

    Pitch and loudness, and how steady, clear and bright the voice sounds.

    • pitch
    • energy
    • jitter
    • shimmer
    • harmonicity
    • spectrum
  4. 04

    The body

    Posture: leaning, shrugging, squaring up or turning away.

    • shoulder tilt
    • shoulder width
    • shoulder height
    • asymmetry
    • head-over-torso

How they line up

Shared time
Face, voice and body land on one 100-millisecond grid — anchored to words, turns and exchanges.
Whose change
One person is locked in as the subject. Mouth movement and the microphone together decide whose turn is whose.

A change from what?

Some people talk fast. Some barely move their face. So Phosphoros never compares anyone with an average person, only with themselves. It learns each person’s usual range from their own recording, separately for face, voice and words.

Speaking rate, words per minute

Speaker Atheir range · 120–163

Speaker A's samples rise above a horizontal scale of words per minute and Speaker B's hang below it. Each hairline is one slice of the scale, as tall as the number of that speaker's samples that landed there, and a band of light along the line covers the middle eighty per cent of them. A line dropped at 190 words per minute sits far above Speaker A's band and in the thick of Speaker B's.

190a departure

an ordinary moment

Speaker Btheir range · 155–200

Each hairline counts samples at that rate, in words per minute; the light band marks the middle 80% of a speaker’s samples. The same rate is a change for one speaker and an ordinary moment for the other.

How the range is built
qualify → pool → compare

A window qualifies when it holds enough clear observations of the person and a steady pattern inside it. Each range is built from at least two qualifying windows, so it carries the variation between them.

The instrument.
Open to inspection.

The Viewer places the recording, transcript and measured changes on the same timeline.

The Board opens one exchange in detail, with its readings and source moment.

The demo runs on a long, calm, publicly available podcast interview, chosen for its length and clarity. There’s nothing in it to uncover, which is true of most conversations.

Phosphoros runs entirely on your own computer; recordings never leave it. It isn’t available for outside use yet.

The Board on one exchange: facial blendshape channels, a reading per channel against the speaker’s own range, and audio scopes

Take the evidencewith you.

Export includes a readable report, measurements, coverage and provenance.

Phosphoros decides where to look first. The export lets anyone check why.

Open the report and file details
report.md
## Analysis contract

This package records exchange-level measurements of voice
and face against the subject’s own recorded range,
with the exchanges the instrument ranked most
departed first. Read it for what the channels did and
where they agreed.
  • measurements.csv

    One row per bin, per channel, per exchange.

  • timeseries.jsonl

    One record per exchange, including whether measurements were missing, unavailable or excluded.

  • manifest.json

    Schema, provenance, coverage and checksums for the other three.

Bias orintuition?

It started with a political interview and a feeling: everything sounded fine, but something bothered me. I couldn’t tell which it was.

Phosphoros is the second opinion I wanted, one that doesn’t share my politics, my mood, or my history with the speaker. It shows you when something changed from their own normal, and exactly where.

When your gut and the instrument agree, pay attention.When they disagree, pay double.

Let’s talk.

If this interests you, I’d like to hear from you.

Email
hello@phosphoros.xyz
X
@igxmut

If the demo recording is yours and you’d like it changed or removed, write to me.

phosphoros An instrument for recorded conversation Back to top ↑