52 s · ends with a 5 s card naming every source. I cannot hear; neither could the critics. Sound is unjudged until a person listens.
A man does up his collar and tie, first in profile and then facing us. Other people's small gestures begin and do not finish: a woman's hand rises to her mouth, a driver looks at his watch and away, a boy finds a loose button. The film returns to the first man, hands at the knot, and holds.
10searches
16filmstrips looked at
11shots
1review rounds
1 The brief
People hesitate, adjust their clothes, check a wrist, and look away. Their small unfinished gestures become a performance no one planned together.
First proof. Make a seventy-five-second sequence from six situated gesture excerpts. Begin with one complete collar adjustment, including the pause afterward. Let another person begin a similar movement, then preserve their different outcome. Build a brief rhythm across three further gestures before allowing one participant to stop halfway. Return to the first setting after its action has ended. Use clothing, breath, and room sounds to carry the rhythm without assigning a shared emotional state.
Development boundary. The comparison concerns visible actions. It must preserve source circumstances rather than infer identical motives or a universal account of humanity.
The operator’s acceptance criteria
The critic can describe what changed within each microevent.
At least one held aftermath feels necessary rather than merely slow.
Pacing varies for identifiable reasons.
The return to an earlier location activates memory of its event.
The film remains specific when music is removed.
2 The prompt
The method prompt for this run, followed by this film's inputs. Paste it into an agent with a shell, ffmpeg, a speech recognizer and image reading. It is a reconstruction of what the run did, written afterwards; the run itself began from a short request.
1,139 words
You are making one proof film for ALLWAYS (a cinema practice within PAELEO) from the Moving Image Archive (movingimagearchive.com; public-domain films, which the site says are free to reuse). You work as a hand editor.
"By hand" means: you search, you look at what came back, you listen, and you write every in-point, out-point, order, gain and reason yourself, in an edit list. Software renders the list; it never chooses a shot, a trim or a sound. Do not generate footage. Do not pick shots by similarity score. Do not run a pipeline that edits for you.
## Input
- `[FILM]` the title, concept, first proof and development boundary (below).
- `[OPERATOR]` the primary SCION operator and its acceptance criteria (below; the full prompt is at `operator.html#s<N>`).
- `[SOURCE SET]` the archive's search API: `POST https://www.movingimagearchive.com/api/search` with `{"query": "..."}` returns clips with `videoUrl`, `thumbnailUrl`, `startSeconds`, `endSeconds`, `sourceTitle`, `sourceYear`, `sourceSlug`. One request at a time, about 1.4 s apart; it rate-limits.
- `[DELIVERY]` one film, 4:3, with sound, no longer than the proof asks, plus the case-study folder described in `RUNS.md`.
## Process (do all of it, in order; keep what you make, including what fails)
1. **Read the proof.** Write the three to six specific things a viewer must be able to notice. These are what you will judge your own cut against.
2. **Forage.** Write 8 to 14 searches in plain visible-action language ("a woman adjusts her hat, looking at herself"), not themes. Run each, save the results, and make one numbered contact sheet per query (thumbnail, #number, duration, year, title). Look at every sheet.
3. **Look closer.** For each candidate, fetch the clip and make a time-coded filmstrip with its waveform. Find where the action starts, completes, and what follows it. Note the native size and whether it has sound.
4. **Listen.** Transcribe candidates with word timings. Speech recognition invents words in hiss ("Thank you."). Check against the waveform and loudness before believing a word. An invented word is never evidence. If you cannot hear audio, say so in the case study and treat sound decisions as unverified.
5. **Write the edit list.** For each shot: clip number, in, out, gain, any sound taken from another clip (with its source), speed, hold, caption, and one sentence of reason that a reader could disagree with. Sound or captions from another source go in the ledger as such.
6. **Render, then check the render.** Make a contact sheet of every shot's head, middle and tail; transcribe the render; measure loudness per shot. Fix what you find. Known traps: a clip's closing cut can show up in your trim; holding a last frame can freeze the next shot of the source.
7. **Disclose derived layers.** If you remove, extend, generate, retime, mirror, or re-voice anything: keep the original, mark the derived layer on screen or in the ledger, and say what is unknown. Authored narration is labelled invented and synthetic.
8. **Blind critic.** Give a fresh agent only: a timecoded contact sheet of the final film, a transcript with loudness per 2 s, and the film's brief and acceptance criteria. Not your edit list, ledger, reasons or summaries. It names one specific event before it reads the brief. It returns a verdict (pass / reference stronger / inconclusive), the largest gap, up to three repairs each with a timecode and a checkable condition, and one strength to protect. It cannot hear: take no sound verdicts from it, and check its claims against the cut before obeying them.
9. **Repair what is true; leave what is wrong and say so.** If you changed the film, ask a new critic. Stop when you can say honestly what still falls short of the brief.
10. **Package** the case study per `RUNS.md`: searches and sheets, strips, the edit list with reasons, the ledger, the critic packets and verbatim reports, repairs, and what is open.
## Rules
- Status is not self-awarded. A film is "proof in progress" unless a blind critic passed it; "released" is for a person to say.
- Unknown dates and names stay unknown. Never imply two clips recorded the same event.
- Do not make assembled speech look like a speaker's own statement.
- Say what you could not perceive. A maker that cannot hear, or cannot see motion, must say so, because it changes what the cut proves.
## Tools (in `tools/`; Python 3, ffmpeg, Pillow, numpy)
- `forage.py <film> "query" ...` runs searches one at a time, writes `forage.json` and a contact sheet per query.
- `look.py <film> <n>[:step[:in:out]] ...` makes a time-coded filmstrip with waveform for clip #n.
- `hear.py <film> <n> ...` word timings (needs `mlx_whisper`, Apple silicon; substitute any word-timing ASR).
- `beats.py <film> <n> <in> <out>` motion-energy peaks, a helper for the eye.
- `cut.py <film>` renders `edl.json`; it chooses nothing. `pairs.py <film>` renders side-by-side pairs from `pairs.json`.
- `reel.py <film>` contact sheet of the rendered film, shot by shot.
- `packet.py <film>` builds what the critic gets.
- `erase.py` + `erase_plate.py` person removal for NOBODY HOME (SAM 2.1 tracked mask; replacement pixels from the same shot; unknowns listed). Set `SAM2_ROOT` to a folder with `models/sam2.1_hiera_small.pt`.
- `speak.py` authored narration with Kokoro-82M. `adopt.py` brings clips found in a local word index into a film's forage (set `CINEOSIS_LAB`).
---
## This film
**Title:** ALMOST HUMAN (Unfinished gestures)
**Concept:** People hesitate, adjust their clothes, check a wrist, and look away. Their small unfinished gestures become a performance no one planned together.
**First proof:** Make a seventy-five-second sequence from six situated gesture excerpts. Begin with one complete collar adjustment, including the pause afterward. Let another person begin a similar movement, then preserve their different outcome. Build a brief rhythm across three further gestures before allowing one participant to stop halfway. Return to the first setting after its action has ended. Use clothing, breath, and room sounds to carry the rhythm without assigning a shared emotional state.
**Development boundary:** The comparison concerns visible actions. It must preserve source circumstances rather than infer identical motives or a universal account of humanity.
**Primary operator:** SCION 57 · Moments — Microevent Attention. Full operator prompt: https://hartswf0.github.io/bolas/paeleo/allways/operator.html#s57
**Supporting:** SCION 84 · Pickpocket — Object Transfer as Dramatic Grammar, SCION 85 · Jeanne Dielman — Routine as a Sensitive Measuring Instrument
**Operator's creative theorem:** Make a brief ordinary event feel complete enough to matter, then let its duration change the viewer's attention to the next event. The unit is a beginning, small change and aftermath, not a picturesque fragment.
**Acceptance criteria (from the operator):**
- The critic can describe what changed within each microevent.
- At least one held aftermath feels necessary rather than merely slow.
- Pacing varies for identifiable reasons.
- The return to an earlier location activates memory of its event.
- The film remains specific when music is removed.
3 What was searched
One search at a time, in plain visible-action language. Each sheet is what came back: numbered thumbnails with duration, year and title. These are the images the maker looked at (or, where it could not, see the note below).
a man straightens his necktie and collar in a mirror · 24 resultsa woman adjusts her hat, looking at herself · 24 resultsa man checks his wristwatch and looks away · 24 resultsa person hesitates in a doorway before going in · 24 resultsa nervous man tugs at his collar · 24 resultsa woman smooths her dress and waits · 24 resultsa man buttons his jacket before going in · 24 resultsa woman touches her hair nervously · 24 resultssomeone starts to speak and stops, close up face · 24 resultsa man waits on a street corner looking at his watch · 24 results
4 What was looked at closely
A time-coded filmstrip and waveform for each candidate. Outlined strips are in the cut.
#1 Speech: Platform Posture and Appearance (1949) · in the cut (shot 1, 2, 11)#1 Speech: Platform Posture and Appearance (1949) · in the cut (shot 1, 2, 11)#1 Speech: Platform Posture and Appearance (1949) · in the cut (shot 1, 2, 11)#2 Act Your Age (1949) · looked at, not used#18 Control Your Emotions (1950) · in the cut (shot 6)#18 Control Your Emotions (1950) · in the cut (shot 6)#53 Two Pay Days (Part II) (1927) · in the cut (shot 4)#63 Dial Comes To Town (undated) · looked at, not used#97 Weather Spotter In Action (Dramatization) (1950) · looked at, not used#100 How to Be Well Groomed (1949) · in the cut (shot 5)#104 Something to Sing About (1937) · in the cut (shot 9)#108 Two Pay Days (Part I) (1927) · in the cut (shot 7)#110 Going Steady? (1951) · in the cut (shot 8)#112 How Quiet Helps at School (1953) · looked at, not used#165 Fitting Faces (1940) · looked at, not used#166 Family Life (1949) · in the cut (shot 3, 10)
5 The cut
The finished film, shot by shot (first, middle and last frame of each):
The final cut as a contact sheet
Every shot, where it came from, and why it is here. Source seconds are in the original film. Links go to the archive's own copy of the clip.
a woman's hand rises to her mouth; the source film is about managing emotions. Different outcome, different life. Source's spoken line removed from under the gesture so no feeling is assigned to it (repair after critic).
sound: Speech: Platform Posture and Appearance (1949) 171.3–172.8
sound: Speech: Platform Posture and Appearance (1949) 171.3–172.8
sound: Speech: Platform Posture and Appearance (1949) 171.3–172.8
163.4–164.1
return to the first setting after the action has ended: same wall, hands still at the knot, last frame held.
6 Review
A fresh agent was given only a timecoded contact sheet of the film, a transcript with loudness, and the brief. Not the edit list, the ledger, or these reasons. It cannot hear.
Round 1 inconclusive, leaning reference stronger
independent critic (fresh Claude Sonnet 5.5 agent). Its report, verbatim:
The report
VERDICT: inconclusive, leaning reference stronger. I can't hear the film, so I couldn't judge the room-sound rhythm.
Before the brief: at 0-10s a man in suspenders adjusts his tie (0-5s, profile; 6-10s, frontal). The collar gesture reads as complete and the repetition is held. Then at 11-15s a woman cries into her fist, with the voice asking "Isn't there a better way to manage ourselves?" at 12.4s. That line pins an emotional meaning on a gesture, which is the opposite of the film's stated intent. The film looks like a montage of period-instruction footage, narrated by one lecturing voice.
Largest gap: no memory-activating return. The sheet shows no return to the striped-wall setting after 10s. It ends on a standing man (44-47s) that has not appeared before, and the 43.1s line "It's worth a try" closes the film on a moral. The brief's 75s was not met either; the sheet runs 48s.
Repairs:
1. Add the return to the striped wall (0-10s) after the last gesture, at about 40-48s, with the man absent or his tie finished. Check that the same frame reappears after the 38s shot.
2. Cut or move the 12.4-15.1s line over the woman. Check that 11-15s has no speech.
3. The wristwatch (18-19s) is a 2s hold on an object with no visible outcome. Extend it with the next gesture's start, or cut to 20s with the driver's head turn. Check that something changes across the cut.
Strength to protect: the 0-10s collar adjustment. It is long and complete, and the profile-to-frontal shift stays specific without music.
True on the line pinned to a gesture and on the missing return; both repaired. The film is still shorter than the 75 s the proof asks for.
7 What happened
Repairs after review
Removed the spoken line from under the woman's gesture.
Replaced the closing shot with a return to the first man at the same wall, last frame held 3.6 s.
What worked
Ten searches written as actions ('a man checks his wristwatch and looks away'), not emotions. Instruction and etiquette films answered best.
The long uncut collar sequence is a source film's own 'appearance' lesson, used almost whole because its pause after the adjustment is the point.
The 1927 silent clips carry raised hiss on purpose, so the cut never goes quiet.
What failed
I could not see this film's contact sheets or filmstrips while cutting it: every image read returned a request-limit error. I cut from search titles, Whisper word timings, durations and loudness numbers. After the cut I checked each reason against frames of the final film and corrected five that described things the frames do not show (a hand 'climbing' to a collar, a slip 'pocketed', hands 'in the air').
A line from the source film ('Isn't there a better way to manage ourselves?') had been left over a woman's gesture and assigned her a feeling. The critic was right; I rebuilt the audio under that shot from the source's own sound around the line.
The first cut ended on a man not seen before and on a moral line. It now returns to the first man.
What is still open
48 s against 75 s.
The proof wants one gesture whose outcome differs between two people inside a pair of shots; the cut shows differences in sequence, not as a pair.
8 Files
edl.jsonedit list: every in, out, gain and reason, as written
ledger.jsonledger: each shot traced to its source seconds
Argon