JamesDimachkie/legible-film

"owner": {     "name": "James Dimachkie",     "url": "https://github.com/JamesDimachkie"

"name": "James Dimachkie",     "url": "https://github.com/JamesDimachkie"   },

"url": "https://github.com/JamesDimachkie"

"author": {     "name": "James Dimachkie",     "url": "https://github.com/JamesDimachkie"

"name": "James Dimachkie",     "url": "https://github.com/JamesDimachkie"   },

"url": "https://github.com/JamesDimachkie"

},   "homepage": "https://github.com/JamesDimachkie/legible-film",   "repository": "https://github.com/JamesDimachkie/legible-film",

"homepage": "https://github.com/JamesDimachkie/legible-film",

"repository": "https://github.com/JamesDimachkie/legible-film",

Apache License                            Version 2.0, January 2004                         http://www.apache.org/licenses/

http://www.apache.org/licenses/

http://www.apache.org/licenses/LICENSE-2.0

Some of the Screening Room's colour and size values are adapted from Rune (https://github.com/anomalyco/rune,

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: http://scripts.sil.org/OFL

http://scripts.sil.org/OFL

Copyright 2016 The Inter Project Authors (https://github.com/rsms/inter) Inter-Italic[opsz,wght].ttf: Copyright 2016 The Inter Project Authors (https://github.com/rsms/inter)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: http://scripts.sil.org/OFL

http://scripts.sil.org/OFL

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

Copyright 2024 The Geist Project Authors (https://github.com/vercel/geist-font)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: https://openfontlicense.org

https://openfontlicense.org

Copyright (c) 2023, GitHub https://github.com/github/mona-sans

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: http://scripts.sil.org/OFL

http://scripts.sil.org/OFL

- Noto Sans Arabic: https://github.com/google/fonts/tree/main/ofl/notosansarabic

- Noto Sans Hebrew: https://github.com/google/fonts/tree/main/ofl/notosanshebrew

- Noto Sans Thai: https://github.com/google/fonts/tree/main/ofl/notosansthai

- Noto Sans Devanagari: https://github.com/google/fonts/tree/main/ofl/notosansdevanagari

Copyright 2022 The Noto Project Authors (https://github.com/notofonts/arabic)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: https://scripts.sil.org/OFL

https://scripts.sil.org/OFL

Copyright 2022 The Noto Project Authors (https://github.com/notofonts/devanagari)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: https://scripts.sil.org/OFL

https://scripts.sil.org/OFL

Copyright 2022 The Noto Project Authors (https://github.com/notofonts/hebrew)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: https://scripts.sil.org/OFL

https://scripts.sil.org/OFL

Copyright 2022 The Noto Project Authors (https://github.com/notofonts/thai)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: https://scripts.sil.org/OFL

https://scripts.sil.org/OFL

Executable file

Executable file

Executable file

Executable file

win.style.cssText = `position:absolute;left:${hole.mx}px;top:${hole.my}px;width:${hole.mw}px;height:${hole.mh}px;visibility:hidden;`;     win.innerHTML = `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${hole.vw} ${hole.vh}" width="${

win.innerHTML = `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${hole.vw} ${hole.vh}" width="${hole.mw}" height="${hole.mh}" style="display:block;overflow:visible"><path d="${hole.path}" fill="${nextGround}"/></svg>`;

const counterD = loopAround(outline, letter, hole.cx, hole.cy);       const face = node.color || this.typeOn(this.groundAt(c - d));       vec.innerHTML = `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${node.w} ${node.h}" width="${no

vec.innerHTML = `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${node.w} ${node.h}" width="${node.w}" height="${node.h}" style="display:block;overflow:visible"><path fill-rule="evenodd" d="${outline.d}" fill="${face}"/></svg>`;

if (counterD) win.innerHTML = `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 ${node.w} ${node.h}" width="${node.w}" height="${node.h}" style="display:block;overflow:visible"><path d="${counterD}" fill="${nextGround}" stroke="${nextGr

Executable file

Executable file

Executable file

Executable file

let facts = null;   if (spec.facts) {     const r = await fetch(new URL(spec.facts, filmUrl)); // lf-lint-ok: build time, before __lf.ready

if (spec.facts) {     const r = await fetch(new URL(spec.facts, filmUrl)); // lf-lint-ok: build time, before __lf.ready     if (!r.ok) throw new Error(`facts file ${spec.facts} could not be read (${r.status})`);

const r = await fetch(new URL(spec.facts, filmUrl)); // lf-lint-ok: build time, before __lf.ready

const url = new URL(src, filmUrl).href;     if (/\.svg(\?|$)/i.test(src) && sc.inlineSvg) {       const r = await fetch(url); // lf-lint-ok: build time, before __lf.ready

if (/\.svg(\?|$)/i.test(src) && sc.inlineSvg) {       const r = await fetch(url); // lf-lint-ok: build time, before __lf.ready       images.set(src, { svgText: await r.text(), url });

const r = await fetch(url); // lf-lint-ok: build time, before __lf.ready

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

asset_id,role,file,sha256,source_or_provider,model_or_version,created_on,plan_at_creation,licence_or_terms_url,permitted_use,cost,private_material_uploaded,approved_by,notes example-01,music,audio/music-take-b.wav,,provider name,model versi

example-01,music,audio/music-take-b.wav,,provider name,model version,YYYY-MM-DD,paid / free / owned,https://,commercial web and social,0,no,,replace this row

"label_beside_value": "batch 1",       "footnote": "Measured by ___ on ___ (source, date)",       "source": "https://example.com/benchmark",

"footnote": "Measured by ___ on ___ (source, date)",       "source": "https://example.com/benchmark",       "read_on": "YYYY-MM-DD",

"source": "https://example.com/benchmark",

Example: Adobe's editorial designation excludes advertising even with an Extended licence, and Pexels warns that its licence does not cover depicted brands ([Adobe Stock licence terms](https://stock.adobe.com/license-terms),

licence does not cover depicted brands ([Adobe Stock licence terms](https://stock.adobe.com/license-terms), [Pexels](https://help.pexels.com/hc/en-us/articles/360042295214-Can-I-use-the-photos-and-videos-for-a-commercial-project)).

([Adobe Stock licence terms](https://stock.adobe.com/license-terms),

[Pexels](https://help.pexels.com/hc/en-us/articles/360042295214-Can-I-use-the-photos-and-videos-for-a-commercial-project)).

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Pexels](https://www.pexels.com/license/) | atmospheric inserts and textures where release risk is low | free commercial use; no implied endorsement 

|---|---|---|---| | [Pexels](https://www.pexels.com/license/) | atmospheric inserts and textures where release risk is low | free commercial use; no implied endorsement or trademark use | free | | [Artlist footage](https://artlist.io/help-c

| [Pexels](https://www.pexels.com/license/) | atmospheric inserts and textures where release risk is low | free commercial use; no implied endorsement or trademark use | free |

| [Artlist footage](https://artlist.io/help-center/privacy-terms/artlist-license) | repeated production from one consistent library | Pro covers an individual's client work and paid promotion; companies and agencies are pointed to Business 

| [Stocksy](https://www.stocksy.com/legal/contentlicense) | a few naturalistic shots with a point of view | royalty-free, worldwide, perpetual; clip audio not licensed; AI/ML processing prohibited under the standard terms | $175 for 1080p, 

| [Filmsupply](https://www.filmsupply.com/license-terms) | a hero sequence whose light, casting or camera must carry the film | rights-managed per project, client, use and territory | web and social from $219 a clip; wider media by quote ([

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Unsplash](https://unsplash.com/license) | landscapes, objects and editorial-feeling atmosphere | free commercial use; releases not guaranteed ([rele

|---|---|---|---| | [Unsplash](https://unsplash.com/license) | landscapes, objects and editorial-feeling atmosphere | free commercial use; releases not guaranteed ([releases](https://help.unsplash.com/en/articles/2612329-releases-and-tradem

| [Unsplash](https://unsplash.com/license) | landscapes, objects and editorial-feeling atmosphere | free commercial use; releases not guaranteed ([releases](https://help.unsplash.com/en/articles/2612329-releases-and-trademarks)) | free |

| [Adobe Stock](https://stock.adobe.com/license-terms) | broad searches across photos, illustration and vectors | Standard licence covers social and web; audience limits and editorial restrictions apply | 10 credits for $29.99/month on an a

| [Stocksy](https://www.stocksy.com/pricing) | a small, coherent set with an individual point of view | Standard includes advertising, worldwide, unlimited term | $35 (2 MP) to $135 (12 MP and up) |

| [10M](https://www.10m.co/about) | public-domain art, maps, engravings, botanical plates and posters from 22 museums and archives, searchable by colour, mood, era, medium or artist | rights are stated per work: about 130,000 works carry re

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Incompetech](https://incompetech.com/music/royalty-free/licenses/) | a low-budget piece that can carry a credit | free with Creative Commons credit;

|---|---|---|---| | [Incompetech](https://incompetech.com/music/royalty-free/licenses/) | a low-budget piece that can carry a credit | free with Creative Commons credit; a paid licence removes the credit | free with credit | | [Epidemic Sou

| [Incompetech](https://incompetech.com/music/royalty-free/licenses/) | a low-budget piece that can carry a credit | free with Creative Commons credit; a paid licence removes the credit | free with credit |

| [Epidemic Sound](https://www.epidemicsound.com/pricing/) | online client work with a channel-clearing workflow | Pro covers digital ads and freelance client work; published content stays covered after cancellation | Pro $16.99/month, bill

| [Artlist music](https://artlist.io/help-center/privacy-terms/artlist-license) | frequent edits where stems matter | the Social plan excludes client projects and paid media; Pro and Business are broader | Music Pro $16.58/month billed annu

| [PremiumBeat](https://www.premiumbeat.com/license) | a one-off film with versions and stems, no subscription | per track and project; tiers widen the media covered | $59 Standard, $199 Premium, $999 Business |

| [Universal Production Music](https://www.universalproductionmusic.com/en-gb/licensing) | advertising or broadcast needing a specific agreement or commissioned music | licensed by region and use | by quote |

cut-downs, paid placement, the client's publishing accounts, territory and term. Synchronisation and public performance are separate questions; ask who handles cue sheets. "Free Creative Commons music" is not enough: CC treats synchronising

performance are separate questions; ask who handles cue sheets. "Free Creative Commons music" is not enough: CC treats synchronising music to picture as an adaptation, so ND licences forbid it and NC licences forbid commercial use ([Creativ

synchronising music to picture as an adaptation, so ND licences forbid it and NC licences forbid commercial use

([Creative Commons FAQ](https://creativecommons.org/faq/#when-is-my-use-considered-an-adaptation)).

**Content ID** can block, monetise or track an upload without knowing the music was licensed elsewhere ([YouTube](https://support.google.com/youtube/answer/2797370?hl=en)). Clear the client's channel through the library's

([YouTube](https://support.google.com/youtube/answer/2797370?hl=en)). Clear the client's channel through the library's

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Freesound](https://freesound.org/help/faq/) | specific real-world details, room tone, unusual foley | licence varies per sound (CC0, CC BY, CC BY-NC

|---|---|---|---| | [Freesound](https://freesound.org/help/faq/) | specific real-world details, room tone, unusual foley | licence varies per sound (CC0, CC BY, CC BY-NC); uploads can contain unauthorised material | free | | [Sonniss GDC bu

| [Freesound](https://freesound.org/help/faq/) | specific real-world details, room tone, unusual foley | licence varies per sound (CC0, CC BY, CC BY-NC); uploads can contain unauthorised material | free |

| [Sonniss GDC bundles](https://sonniss.com/gdc-bundle-license/) | a broad starting palette | commercial project use without credit; no standalone redistribution (the project team may share them); no AI training | free |

| [Epidemic Sound](https://www.epidemicsound.com/pricing/) | teams clearing music and effects together | included with music on the same plans | as music |

| [BOOM Library](https://www.boomlibrary.com/terms-conditions-regular/) | designed transitions and material recordings with good metadata | worldwide media licence including advertising; no AI/ML training; subscription terms differ from buy

motion, and the room or tail. Check the stack for phase cancellation, and never pile full-spectrum impacts on a small interface change. **Reject** clipped attacks, denoising warble, baked-in music, rips from films or games, voices hidden in

hidden in an ambience, and anything NC for advertising.

| Route | Use it for | Usage notes | Price level, read 2026-10-28 | |---|---|---|---| | [Fiverr](https://www.fiverr.com/) (vetted Pro listings) | narrow casting with a dry audition that fits | terms are seller-specific; some listings exclud

|---|---|---|---| | [Fiverr](https://www.fiverr.com/) (vetted Pro listings) | narrow casting with a dry audition that fits | terms are seller-specific; some listings exclude paid media | listings vary; confirm the package covers the planned

| [Fiverr](https://www.fiverr.com/) (vetted Pro listings) | narrow casting with a dry audition that fits | terms are seller-specific; some listings exclude paid media | listings vary; confirm the package covers the planned use |

| [Voice123](https://support.voice123.com/hc/en-us/articles/205237644-What-is-the-rate-for-a-voice-actor-on-Voice123) | direct auditions and negotiation | fees negotiated; agree media, territory, term and any AI use in writing | by quote |

| [Voices](https://www.voices.com/help/knowledge/faq/Understanding-Licensing) | usage-led casting and budgeting | paid online ads are licensed like broadcast, by market and term | its online-ad jobs average $350-$1,125 ([rates](https://www.

| Union casting ([SAG-AFTRA](https://www.sagaftra.org/production-center/contract/802/getting-started)) | campaigns requiring union talent | a contract route with session, usage and digital-replica terms | session fees plus usage; a payroll 

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Inter](https://rsms.me/inter/) | clear interface labels, product numbers, contemporary titles | SIL Open Font License | free |

|---|---|---|---| | [Inter](https://rsms.me/inter/) | clear interface labels, product numbers, contemporary titles | SIL Open Font License | free | | [IBM Plex](https://github.com/IBM/plex) | a coordinated sans, serif and mono family for te

| [Inter](https://rsms.me/inter/) | clear interface labels, product numbers, contemporary titles | SIL Open Font License | free |

| [IBM Plex](https://github.com/IBM/plex) | a coordinated sans, serif and mono family for technical work | SIL Open Font License | free |

| [Source Serif 4](https://github.com/adobe-fonts/source-serif/blob/main/LICENSE.md) | editorial headings, a serif counterpoint to interface type | SIL Open Font License, with optical sizes | free |

| [Adobe Fonts](https://helpx.adobe.com/fonts/web/font-licensing/font-licensing.html) | art-direction exploration under an existing Creative Cloud plan | film, broadcast and online video permitted; no app embedding or sharing font files | w

| Foundry licences ([Monotype guide](https://www.monotypefonts.com/pages/content/resources-font-licensing-guide)) | brand type and negotiated scopes | desktop, web, app and embedding licensed separately | by quote |

The SIL Open Font License permits titling in commercial video without an artwork credit ([OFL FAQ](https://openfontlicense.org/ofl-faq/)). For a typeface from a directory, read the upstream designer's

([OFL FAQ](https://openfontlicense.org/ofl-faq/)). For a typeface from a directory, read the upstream designer's

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Poly Haven](https://polyhaven.com/license) | lighting environments, materials, supporting models | CC0, commercial use without attribution | free |

|---|---|---|---| | [Poly Haven](https://polyhaven.com/license) | lighting environments, materials, supporting models | CC0, commercial use without attribution | free | | [Fab](https://www.fab.com/eula) | specific props, environments, speci

| [Poly Haven](https://polyhaven.com/license) | lighting environments, materials, supporting models | CC0, commercial use without attribution | free |

| [Fab](https://www.fab.com/eula) | specific props, environments, specialist models | Standard licence allows commercial use in any tool; no standalone redistribution; CC BY items carry CC terms | per item |

| [Poliigon](https://help.poliigon.com/en/articles/8749749-asset-use-licensing) | consistent materials for repeated renders | advertising included; no sharing of extractable scene files; no AI/ML use; Individual plans are for one person | $

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | Procedural (FFmpeg [`noise`](https://ffmpeg.org/ffmpeg-filters.html#noise), [`lut3d`](https://ffmpeg.org/ffmpeg-filters.html#lut3d), an original shad

|---|---|---|---| | Procedural (FFmpeg [`noise`](https://ffmpeg.org/ffmpeg-filters.html#noise), [`lut3d`](https://ffmpeg.org/ffmpeg-filters.html#lut3d), an original shader) | subtle, controllable texture tied to the animation | your own cod

| Procedural (FFmpeg [`noise`](https://ffmpeg.org/ffmpeg-filters.html#noise), [`lut3d`](https://ffmpeg.org/ffmpeg-filters.html#lut3d), an original shader) | subtle, controllable texture tied to the animation | your own code; record its prov

| [Grainzilla](https://grainzilla.com/) | a small set of real grain scans | commercial use including client work and ads; no redistribution | 1080p pack $24.90; 4K pack $39.90 |

| [CineGrain](https://www.cinegrain.com/texture/35mm) | a deliberately film-derived treatment central to the look | Indie tier capped by production budget; Professional for larger productions, under its own licence | Indie 35mm $96; Profess

| [FilmConvert Nitrate](https://www.filmconvert.com/nitrate) | exposure-dependent grain and camera-profile film looks | software licence with three activations | from $119 |

| Source | Use it for | Licence notes | Price level, read 2026-10-28 | |---|---|---|---| | [Lucide](https://lucide.dev/license) | a consistent, code-native icon vocabulary | ISC licence; keep the notice | free |

|---|---|---|---| | [Lucide](https://lucide.dev/license) | a consistent, code-native icon vocabulary | ISC licence; keep the notice | free | | [Envato Elements](https://help.elements.envato.com/hc/en-us/articles/360000628966-Envato-Elements

| [Lucide](https://lucide.dev/license) | a consistent, code-native icon vocabulary | ISC licence; keep the notice | free |

| [Envato Elements](https://help.elements.envato.com/hc/en-us/articles/360000628966-Envato-Elements-License) | breadth in illustration, mockups and templates | per-project registration; embedded components may have their own licences | from

| [UI8](https://ui8.net/terms) | a coherent illustration set or design kit | marketing videos and client work permitted; no standalone redistribution of the assets | per item |

| [Apple Design Resources](https://developer.apple.com/design/resources/) | accurate presentation of a qualifying Apple app | separate conditions per resource | free downloads; terms per resource |

**Device and platform owners have their own rules.** Apple's App Store marketing guidance forbids modifying, spinning or animating its supplied product images and 3D simulations of its products in that context, and requires trademark credit

or animating its supplied product images and 3D simulations of its products in that context, and requires trademark credit lines in video ([guidelines](https://developer.apple.com/app-store/marketing/guidelines/)). Google forbids implying a

credit lines in video ([guidelines](https://developer.apple.com/app-store/marketing/guidelines/)). Google forbids

implying an endorsement that does not exist ([brand guidance](https://about.google/brand-resource-center/guidance/)).

| Problem | Case | What it shows | |---|---|---| | Proving a capability so the viewer believes it | Volvo Trucks, "The Epic Split" (2013; Forsman & Bodenfors; dir. Andreas Nilsson) | Jean-Claude Van Damme does the splits between two reversi

|---|---|---| | Proving a capability so the viewer believes it | Volvo Trucks, "The Epic Split" (2013; Forsman & Bodenfors; dir. Andreas Nilsson) | Jean-Claude Van Damme does the splits between two reversing trucks to demonstrate the stabil

| Proving a capability so the viewer believes it | Volvo Trucks, "The Epic Split" (2013; Forsman & Bodenfors; dir. Andreas Nilsson) | Jean-Claude Van Damme does the splits between two reversing trucks to demonstrate the stability and precis

| A chain of visible causes | Honda, "Cog" (2003; Wieden+Kennedy London; dir. Antoine Bardou-Jacquet) | A chain reaction of real car parts, each setting off the next, largely without computer graphics. [Wikipedia](https://en.wikipedia.org/w

| One rule of the world changes | Apple, AirPods "Bounce" (2019; TBWA\Media Arts Lab; dir. Oscar Hudson) | A commute in which the city becomes a trampoline; the film explores the consequences of one changed rule. [The Drum](https://www.thed

| Telling a human story through use of the product | Google, "Parisian Love" (2009 Search Stories film; aired during Super Bowl XLIV, 2010) | A love story told through one person's searches. | | A belief behind the product | Honda, "Grrr" (

| A belief behind the product | Honda, "Grrr" (2004; Wieden+Kennedy London; dirs. Adam Foulkes and Alan Smith) | Reportedly grew from a Honda engine designer's dislike of diesel engines ("Hate something, change something"). [Wikipedia](http

| A number the viewer must understand | Ed Hawkins, "climate spiral" (2016) | Monthly temperature anomalies drawn on fixed axes, month as angle and anomaly as radius, with colour encoding time. [Wikipedia](https://en.wikipedia.org/wiki/Clim

| Live data as short animations | IBM, "Datagrams" (2013; IBM with Ogilvy & Mather; animation by Buck) | Live match data from the US Open turned into short animations for each match. [PR Newswire](https://www.prnewswire.com/news-releases/li

| A values statement in few words | Apple, "Intention" (2013, Designed by Apple in California; TBWA\Media Arts Lab) | Short sentences given simple visual relationships. | | One silhouette through many materials | Nike, "Air Max '17: The Art

| One silhouette through many materials | Nike, "Air Max '17: The Art of Negative Space" (ManvsMachine) | The negative-space outline of the shoe recurs through scenes of different materials; the sound re-pitches footsteps into melody and rh

| Delaying recognition of a title | "Alien" (1979; R/Greenberg Associates, Richard Greenberg) | The title's letters assemble gradually, so viewers are not ready to read them as a word. [Art of the Title](https://www.artofthetitle.com/title/

| Motion that belongs to its period | "Stranger Things" (2016; Imaginary Forces, Michelle Dougherty) | An early "locking in" letter motion felt rigid and snappy, and the creators did not respond to it because it was not a familiar 1980s mot

| Unsettling the viewer on purpose | "Severance" main titles (2022; Oliver Latta / Extraweg; typography Teddy Blanks; music Theodore Shapiro) | A title sequence built to disorient rather than reassure. [It's Nice That](https://itsnicethat.c

| Problem | Case | What it shows | |---|---|---| | Scoring a technical subject without cliché | Antfood, score for Google's "Hacking Google" series | Tactile acoustic instruments rather than the dark synthetic textures of cyber tropes, alon

|---|---|---| | Scoring a technical subject without cliché | Antfood, score for Google's "Hacking Google" series | Tactile acoustic instruments rather than the dark synthetic textures of cyber tropes, alongside orchestral and synth material

| Scoring a technical subject without cliché | Antfood, score for Google's "Hacking Google" series | Tactile acoustic instruments rather than the dark synthetic textures of cyber tropes, alongside orchestral and synth material. [Antfood](ht

| Effects only, no music | Kia EV6, "Silence Is Powerful" (2023; David&Goliath / Innocean) | By its sound designer's account, picture and sound design only, with no voiceover or music. Watch it to confirm before citing. | | An ending that r

| An ending that refuses triumph | UNHCR, "The Journey" (Don't Panic; sound by String & Tins) | "The music is not triumphant at the end of the film"; the sound resists telling the listener how to feel. [D&AD](https://www.dandad.org/work/ann

| Music made from the product's own sounds | Apple, "Start Up" (2021 event opener; A.G. Cook) | A song built from the sounds of Apple devices, opening on a startup chime. | | Detailed sound on simple visuals | David Kamp (Studiokamp) | Choo

| Detailed sound on simple visuals | David Kamp (Studiokamp) | Chooses which visual elements need a sound, and uses sound to say what minimal visuals do not. [It's Nice That](https://www.itsnicethat.com/articles/david-kamp-interview/) |

| Provider | Output ownership and commercial use | Training on your inputs/outputs by default | Source | |---|---|---|---| | OpenAI API | rights in output assigned to the user (per terms; page not machine-readable at check time) | API: no, 

|---|---|---|---| | OpenAI API | rights in output assigned to the user (per terms; page not machine-readable at check time) | API: no, unless you opt in; consumer apps: may be used unless you opt out | [terms](https://openai.com/policies/te

| OpenAI API | rights in output assigned to the user (per terms; page not machine-readable at check time) | API: no, unless you opt in; consumer apps: may be used unless you opt out | [terms](https://openai.com/policies/terms-of-use/); [API

| Google Gemini API (incl. image and video models) | Google does not claim ownership; preview products can restrict commercial use | unpaid tier: yes; paid tier: no | [Gemini API terms](https://ai.google.dev/gemini-api/terms) |

| Runway | Runway does not claim ownership of inputs or outputs | yes (no opt-out described in the terms) | [terms](https://runway.com/terms-of-use) |

| Luma | customer owns output; commercial use only for output made during an active subscription | yes | [terms](https://lumalabs.ai/legal/terms-of-service) |

| Adobe Firefly | Adobe claims no ownership; marketed as commercially safe | no: Adobe states it does not train Firefly on user content | [Firefly approach](https://www.adobe.com/ai/overview/firefly/gen-ai-approach.html) |

| ElevenLabs | user retains rights; free plan non-commercial, paid plans commercial | yes, with an opt-out | [terms](https://elevenlabs.io/terms-of-use) |

| Stability AI | rights assigned to the user; community licence free up to USD 1M annual revenue | yes, with an opt-out | [terms](https://stability.ai/terms-of-service) |

| Transcripts and caption timing | **WhisperX** for word alignment | faster-whisper or whisper.cpp for speed on CPU; FFmpeg 8+ `whisper` filter where the build includes it | | Generated plates, voice, music, effects | the provider the perso

| Review with collaborators: watching a version, notes on moments, drawings, replies | **Flask** ([flask.do](https://flask.do)), through its MCP server ([review](review.md) §1) | the local Screening Room (`screen.py`) for private work on th

[Strobe](https://github.com/hridaya423/strobe) (Apache-2.0; v0.1.0, first released 2026-10-31, one author) renders an

| Model | Code and weight licence | Exact ONNX weights | Use | |---|---|---|---| | MODNet, Xenova conversion | [Code: Apache-2.0](https://github.com/ZHKKKe/MODNet/blob/master/LICENSE); [weight model card: Apache-2.0](https://huggingface.co/

|---|---|---|---| | MODNet, Xenova conversion | [Code: Apache-2.0](https://github.com/ZHKKKe/MODNet/blob/master/LICENSE); [weight model card: Apache-2.0](https://huggingface.co/Xenova/modnet) | [Pinned download](https://huggingface.co/Xenov

| MODNet, Xenova conversion | [Code: Apache-2.0](https://github.com/ZHKKKe/MODNet/blob/master/LICENSE); [weight model card: Apache-2.0](https://huggingface.co/Xenova/modnet) | [Pinned download](https://huggingface.co/Xenova/modnet/resolve/f

| BiRefNet Lite, Studio Ludens 512 conversion | [Code: MIT](https://github.com/ZhengPeng7/BiRefNet/blob/main/LICENSE); [converted weights and licence: MIT](https://huggingface.co/studioludens/birefnet-lite-512); [upstream weights: MIT](http

the spotted list and the palette) and `listen.py` (measure, transcript, sync, sheet and labels; [sound](sound.md)   §13). - **[earshot](https://github.com/Leonxlnx/earshot)** (Leon Lin, MIT; young) gives agents measurements and labels for

§13). - **[earshot](https://github.com/Leonxlnx/earshot)** (Leon Lin, MIT; young) gives agents measurements and labels for   sounds, cue sync sheets and code-made effects. The skill's `listen.py` follows its shape, takes its launch-film

- **[earshot](https://github.com/Leonxlnx/earshot)** (Leon Lin, MIT; young) gives agents measurements and labels for

| Platform | What the official page says | Source | |---|---|---| | YouTube | MP4 with the moov atom at the front; H.264 High, progressive, 4:2:0, BT.709 for SDR; AAC-LC 48 kHz; upload at the recorded frame rate; SDR guidance about 8 Mbps f

|---|---|---| | YouTube | MP4 with the moov atom at the front; H.264 High, progressive, 4:2:0, BT.709 for SDR; AAC-LC 48 kHz; upload at the recorded frame rate; SDR guidance about 8 Mbps for 1080p at 24-30 fps and 12 Mbps at 48-60 fps | [Yo

| YouTube | MP4 with the moov atom at the front; H.264 High, progressive, 4:2:0, BT.709 for SDR; AAC-LC 48 kHz; upload at the recorded frame rate; SDR guidance about 8 Mbps for 1080p at 24-30 fps and 12 Mbps at 48-60 fps | [YouTube help](ht

| LinkedIn (ads) | MP4; 75 KB-500 MB; 3 s-30 min (15-30 s recommended); 16:9, 1:1, 4:5 and 9:16; under 30 fps; H.264 or VP8 with AAC or MPEG-4 audio | [LinkedIn help](https://www.linkedin.com/help/lms/answer/a424737) |

| Instagram Reels (ads) | 9:16; 1440x2560 recommended; up to 15 min and 4 GB; H.264, square pixels, fixed frame rate, progressive; stereo AAC at 128 kbps or more; the safe zones in §3 | [Meta ads guide](https://www.facebook.com/business/ads

| TikTok (in-feed ads, non-Spark) | 9:16 recommended at 540x960 or more; up to 10 min and 500 MB; at least 516 kbps | [TikTok ads help](https://ads.tiktok.com/resources/help/article/tiktok-auction-in-feed-ads) |

| X | the help pages could not be read at check time, and third-party figures conflict | [X help](https://help.x.com) |

whole film in one clip. Keep text and logos out of the generated plate and composite them exactly afterwards.   Phrase positively, as Runway's Gen-4 guide recommends   ([guide](https://help.runwayml.com/hc/en-us/articles/39789879462419-Gen-

Phrase positively, as Runway's Gen-4 guide recommends   ([guide](https://help.runwayml.com/hc/en-us/articles/39789879462419-Gen-4-Video-Prompting-Guide)), and follow the   chosen model's own guide.

([guide](https://help.runwayml.com/hc/en-us/articles/39789879462419-Gen-4-Video-Prompting-Guide)), and follow the

structure in the edit. - **Voice:** save a designed voice through the vendor's persistent voice feature (ElevenLabs, for example, creates a   reusable voice ID from a generated preview: [create a voice](https://elevenlabs.io/docs/api-refere

- **Voice:** save a designed voice through the vendor's persistent voice feature (ElevenLabs, for example, creates a   reusable voice ID from a generated preview: [create a voice](https://elevenlabs.io/docs/api-reference/text-to-voice/creat

reusable voice ID from a generated preview: [create a voice](https://elevenlabs.io/docs/api-reference/text-to-voice/create))

the customer for the intended use; a clause making the customer indemnify the vendor is the opposite. - **Free tiers are often non-commercial,** and some outputs stay non-commercial after upgrading   ([World Labs terms](https://www.worldlab

- **Free tiers are often non-commercial,** and some outputs stay non-commercial after upgrading   ([World Labs terms](https://www.worldlabs.ai/terms-of-service)). - **Open weights are not unrestricted:** read each checkpoint's licence. Some

([World Labs terms](https://www.worldlabs.ai/terms-of-service)).

- **Open weights are not unrestricted:** read each checkpoint's licence. Some carry revenue thresholds, branding duties   or territorial limits (one 3D model's licence excludes the EU, the UK and South Korea:   [Hunyuan3D-2.1](https://githu

or territorial limits (one 3D model's licence excludes the EU, the UK and South Korea:   [Hunyuan3D-2.1](https://github.com/Tencent-Hunyuan/Hunyuan3D-2.1/blob/main/LICENSE)). Running locally does not make   a hosted wrapper private.

[Hunyuan3D-2.1](https://github.com/Tencent-Hunyuan/Hunyuan3D-2.1/blob/main/LICENSE)). Running locally does not make

logos in the scene"). - **Automated access:** some services forbid scripted or agent use of their apps   ([Midjourney terms](https://docs.midjourney.com/hc/en-us/articles/32083055291277-Terms-of-Service)). Use the

- **Automated access:** some services forbid scripted or agent use of their apps   ([Midjourney terms](https://docs.midjourney.com/hc/en-us/articles/32083055291277-Terms-of-Service)). Use the   documented API or have the person operate the 

([Midjourney terms](https://docs.midjourney.com/hc/en-us/articles/32083055291277-Terms-of-Service)). Use the

- **Disclose** realistic generated or altered people, products, events or behaviour where a viewer could take them as   evidence, and follow each platform's labelling rules   ([YouTube disclosure policy](https://support.google.com/youtube/a

([YouTube disclosure policy](https://support.google.com/youtube/answer/14328491?hl=en-GB)).

- **OpenAI GPT Image 2.5** (Sunburst for precise edits, Flare for throughput): custom sizes in multiples of 16, from   1:3 to 3:1, up to 3,840 px a side and 655,360 to 8,294,400 pixels in all, with anything above 2560x1440 marked   experime

1:3 to 3:1, up to 3,840 px a side and 655,360 to 8,294,400 pixels in all, with anything above 2560x1440 marked   experimental; transparent backgrounds ([guide](https://developers.openai.com/api/docs/guides/image-generation)). - **Google Nan

experimental; transparent backgrounds ([guide](https://developers.openai.com/api/docs/guides/image-generation)).

- **Google Nano Banana 2.1**: reference-led compositions and conversational revisions at 1K-4K   ([pricing](https://ai.google.dev/gemini-api/docs/pricing)). - **Black Forest Labs FLUX 3 Image and Klein**: layout control with bounding boxes;

([pricing](https://ai.google.dev/gemini-api/docs/pricing)).

- **Black Forest Labs FLUX 3 Image and Klein**: layout control with bounding boxes; Klein has open weights under its   own commercial terms ([FLUX 3](https://bfl.ai/models/flux-3-image)). - **Midjourney V8.2**: art direction and look develo

own commercial terms ([FLUX 3](https://bfl.ai/models/flux-3-image)).

- **Midjourney V8.2**: art direction and look development; no automated access under its terms. - **Ideogram 4.5 and Vector**: typography and genuine SVG output ([Vector](https://ideogram.ai/models/vector/)). - **Adobe Firefly Image 5** for

- **Ideogram 4.5 and Vector**: typography and genuine SVG output ([Vector](https://ideogram.ai/models/vector/)).

- **Adobe Firefly Image 5** for images, and Illustrator's Text to Vector Graphic for editable vectors   ([vectors](https://helpx.adobe.com/illustrator/desktop/use-generative-ai/generate-scenes-subjects-and-icons.html));   Adobe's indemnity 

([vectors](https://helpx.adobe.com/illustrator/desktop/use-generative-ai/generate-scenes-subjects-and-icons.html));

([conditions](https://helpx.adobe.com/legal/product-descriptions/adobe-firefly.html)).

- **Google Gemini Omni 1.1 Flash**: conversational generation and editing, extension and interpolation   ([model](https://ai.google.dev/gemini-api/docs/models/gemini-omni-flash)).

([model](https://ai.google.dev/gemini-api/docs/models/gemini-omni-flash)).

- **Google Veo 3.1**: first and last frames and content credentials, on Google's enterprise platform or a partner   route; its Gemini API preview endpoints shut down on 2026-10-22 ([video guide](https://ai.google.dev/gemini-api/docs/video))

route; its Gemini API preview endpoints shut down on 2026-10-22 ([video guide](https://ai.google.dev/gemini-api/docs/video)).

- **Runway Gen-4.5, Act-Two and Aleph 2.0**: Act-Two transfers a real performance to a character; Aleph edits and   relights existing shots ([API pricing](https://docs.dev.runwayml.com/guides/pricing/)). - **Luma Ray3.2**: up to sixteen key

relights existing shots ([API pricing](https://docs.dev.runwayml.com/guides/pricing/)).

- **Luma Ray3.2**: up to sixteen keyframes, HDR and 16-bit EXR output for compositing ([Ray](https://lumalabs.ai/ray)).

- **Seedance 2.5**: up to 30 s with audio and rich image, video and audio references   ([release](https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5)). - **Kling 4.0**: long takes with up to t

([release](https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5)).

- **Kling 4.0**: long takes with up to ten keyframes ([guide](https://kling.ai/blog/kling-4-ai-video-model)).

- **FLUX 3 Video**: up to ten keyframes, a cheap draft mode that continues into the final   ([overview](https://help.bfl.ai/articles/7342522296-what-is-flux-3)). - **HeyGen Video 1**: economical reference-driven scenes with native sound; fr

([overview](https://help.bfl.ai/articles/7342522296-what-is-flux-3)).

- **HeyGen Video 1**: economical reference-driven scenes with native sound; free output is non-commercial   ([API](https://www.heygen.com/ai-video-api)). - **Open weights:** LTX-2.5 ([model card](https://huggingface.co/Lightricks/LTX-2.5/bl

([API](https://www.heygen.com/ai-video-api)).

- **Open weights:** LTX-2.5 ([model card](https://huggingface.co/Lightricks/LTX-2.5/blob/main/README.md)) and MiniMax H3

([repository](https://github.com/MiniMax-AI/MiniMax-H3)); at full precision both exceed a typical laptop's GPU

revenue-threshold licences. - **The Sora 2 API is gone:** OpenAI shut down the Videos API and its Sora 2 models on 2026-09-24   ([deprecations](https://developers.openai.com/api/docs/deprecations)).

([deprecations](https://developers.openai.com/api/docs/deprecations)).

**Voice:** ElevenLabs v4 ([models](https://elevenlabs.io/docs/overview/models)), Cartesia Sonic

([pricing](https://www.cartesia.ai/pricing)), Google Gemini 3.8 Flash TTS for directed delivery

([model](https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash-tts)), and Chatterbox to run locally under the

MIT licence ([repository](https://github.com/resemble-ai/chatterbox)). Cloning conditions differ by route, so read them

**Music:** Suno v6 grants commercial rights on paid downloads only ([rights](https://help.suno.com/en/articles/9601665));

Eleven Music's self-serve rights exclude film, TV and radio, so check how its terms classify your use ([music terms](https://elevenlabs.io/music-terms)); Google Lyria 3.5 is cheap for exploration ([guide](https://ai.google.dev/gemini-api/do

([music terms](https://elevenlabs.io/music-terms)); Google

Lyria 3.5 is cheap for exploration ([guide](https://ai.google.dev/gemini-api/docs/music-generation)); Stable Audio 3

runs locally under a revenue-threshold licence ([licence](https://stability.ai/license)). Udio had disabled downloads

at the time of reading ([notice](https://help.udio.com/en/articles/12683565-changes-associated-with-the-universal-music-group-umg-partnership)).

**Sound effects:** ElevenLabs Sound Effects, 0.1-30 s with looping ([docs](https://elevenlabs.io/docs/overview/capabilities/sound-effects)); Adobe Firefly sound effects at 48 kHz; Stable

([docs](https://elevenlabs.io/docs/overview/capabilities/sound-effects)); Adobe Firefly sound effects at 48 kHz; Stable

- **Restoration and upscaling:** Topaz ([pricing](https://www.topazlabs.com/pricing)); DaVinci Resolve Studio's noise

reduction, SuperScale and Speed Warp ([Studio](https://www.blackmagicdesign.com/products/davinciresolve/studio));

Neat Video for camera noise; Real-ESRGAN locally ([repository](https://github.com/xinntao/Real-ESRGAN)).

- **Interpolation:** Topaz's models, Resolve Speed Warp, or RIFE locally ([repository](https://github.com/hzwer/ECCV2022-RIFE)).

- **Matting:** After Effects Object Matte and Roto Brush, Resolve Magic Mask, SAM 3 for segmentation masks to refine   into a soft matte, under its own licence ([repository](https://github.com/facebookresearch/sam3)), and Robust Video   Mat

into a soft matte, under its own licence ([repository](https://github.com/facebookresearch/sam3)), and Robust Video

**3D and scenes:** Meshy and Tripo for props to clean up afterwards; World Labs Marble for explorable environments ([docs](https://docs.worldlabs.ai/marble/models)). Use manufacturer CAD or measured modelling for a hero product.

([docs](https://docs.worldlabs.ai/marble/models)). Use manufacturer CAD or measured modelling for a hero product.

is scored under Accessibility. BBC subtitle guidance describes roughly 160-180 words per minute for many genres, with editorial adjustments; this is a review prompt, not a pass/fail score or proof for burned text. See [BBC subtitle guidance

for many genres, with editorial adjustments; this is a review prompt, not a pass/fail score or proof for burned text. See [BBC subtitle guidance](https://www.bbc.co.uk/accessibility/forproducts/guides/subtitles/) and [BBC research on subtit

See [BBC subtitle guidance](https://www.bbc.co.uk/accessibility/forproducts/guides/subtitles/) and

[BBC research on subtitle speed](https://downloads.bbc.co.uk/rd/pubs/whp/whp-pdf-files/WHP306.pdf). Cite and quote the

Signals: encoded shape and duration, first frame, captions and declared platform; these do not establish current platform limits, safe-zone clearance or muted comprehension. Cite the requirement and the affected frame. Platform overlays and

limits, safe-zone clearance or muted comprehension. Cite the requirement and the affected frame. Platform overlays and policies change; verify the current official guide, for example [YouTube upload guidance](https://support.google.com/yout

policies change; verify the current official guide, for example [YouTube upload guidance](https://support.google.com/youtube/answer/1722171).

red flashes and is not a photosensitivity test. Cite the check, standard and affected interval; a draft transcript cannot establish caption accuracy. For applicable web content, [WCAG 2.2](https://www.w3.org/TR/WCAG22/) requires captions fo

affected interval; a draft transcript cannot establish caption accuracy. For applicable web content, [WCAG 2.2](https://www.w3.org/TR/WCAG22/) requires captions for prerecorded synchronized audio except its stated exception, and limits flas

[WCAG 2.2](https://www.w3.org/TR/WCAG22/) requires captions for prerecorded synchronized audio except its stated

copy of this sidecar. `scopes/index.json` maps waveform and vectorscope pictures to the representative frame and time interval for each automatically detected candidate shot; long cuts cap the scope count and disclose that sampling. [QCTool

interval for each automatically detected candidate shot; long cuts cap the scope count and disclose that sampling. [QCTools filter descriptions](https://bavc.github.io/qctools/filter_descriptions.html) inspired these diagnostics; their mean

[QCTools filter descriptions](https://bavc.github.io/qctools/filter_descriptions.html) inspired these diagnostics;

count); dark bars below 160 cd/m² with at least 20 cd/m² between light and dark; and more than 40% of the screen   when still, or more than 25% when it changes direction, oscillates, flashes or reverses contrast. Smooth flow in one   direct

when still, or more than 25% when it changes direction, oscillates, flashes or reverses contrast. Smooth flow in one   direction is exempt ([Ofcom, Section Two guidance, Annex 1](https://www.ofcom.org.uk/siteassets/resources/documents/tv-ra

direction is exempt ([Ofcom, Section Two guidance, Annex 1](https://www.ofcom.org.uk/siteassets/resources/documents/tv-radio-and-on-demand/broadcast-guidance/programme-guidance/broadcast-code-guidance/section-2-guidance-notes.pdf)).

6. **Stills** of the hardest frames in every size, at full size and at feed size, judged as [picture](picture.md)    says: with the preview served,    `node "<SKILL_DIR>/scripts/film_capture.mjs" stills http://127.0.0.1:8000/?render --at 2.

says: with the preview served,    `node "<SKILL_DIR>/scripts/film_capture.mjs" stills http://127.0.0.1:8000/?render --at 2.5,b12 --size 1080x1920    --out qa/stills`, or a quick `make_film.py --preview` and `frame_sheet.py` on its renders.

`node "<SKILL_DIR>/scripts/film_capture.mjs" stills http://127.0.0.1:8000/?render --at 2.5,b12 --size 1080x1920

`screen.py serve PROJECT --open` for live mode, where their notes and ticks reach the project folder directly and the page follows new versions. When the person reviews away from this machine or with others (a client, an editor, a team), [F

page follows new versions. When the person reviews away from this machine or with others (a client, an editor, a team), [Flask](https://flask.do) is the recommended collaboration and feedback layer: post each version there and bring every n

team), [Flask](https://flask.do) is the recommended collaboration and feedback layer: post each version there and

**Reviewing with others: Flask.** The Screening Room is local and private; when the person reviews away from this machine, or with a client, editor or team, use [Flask](https://flask.do), the recommended collaboration and feedback

machine, or with a client, editor or team, use [Flask](https://flask.do), the recommended collaboration and feedback

layer. Each version goes up as the next version of one asset, and reviewers watch it at speed and leave notes pinned to moments, with drawings on the frame and replies. An agent works it through Flask's MCP server (`https://api.flask.do/api

(`https://api.flask.do/api/mcp/mcp` when checked on 2026-10-31; the person signs in once):

You cannot hear, but you can listen by instrument. Measure, describe and compare the sound, and say what each measure cannot tell. `listen.py` follows [earshot](https://github.com/Leonxlnx/earshot)'s shape and uses its

measure cannot tell. `listen.py` follows [earshot](https://github.com/Leonxlnx/earshot)'s shape and uses its

flash thresholds; 0.006 steradians, about 25% of a 10° field; 341 x 256 px at 1024 x 768); 2.3.3 Animation from   Interactions (names parallax and non-essential motion); 1.4.1 Use of Color; 1.4.3 Contrast (4.5:1, 3:1 large   text); 1.4.11 N

Interactions (names parallax and non-essential motion); 1.4.1 Use of Color; 1.4.3 Contrast (4.5:1, 3:1 large   text); 1.4.11 Non-text Contrast (3:1); 1.2.2 Captions (Prerecorded). https://www.w3.org/TR/WCAG22/ - WCAG 2.2's note on saturated

text); 1.4.11 Non-text Contrast (3:1); 1.2.2 Captions (Prerecorded). https://www.w3.org/TR/WCAG22/

above 0.2. `flash_scan.py` implements the WCAG 2.0 working formula ((R−G−B)×320 > 20 where R/(R+G+B) ≥ 0.8), so   its red results are a screen, not a reading of the current definition. - Photosensitivity tools: Electronic Arts IRIS (https:/

its red results are a screen, not a reading of the current definition. - Photosensitivity tools: Electronic Arts IRIS (https://github.com/electronicarts/IRIS); Trace Center PEAT   (https://trace.umd.edu/peat/); Harding FPA (https://www.hard

- Photosensitivity tools: Electronic Arts IRIS (https://github.com/electronicarts/IRIS); Trace Center PEAT

(https://trace.umd.edu/peat/); Harding FPA (https://www.hardingfpa.com/); UK ASA on flashing images in ads

(https://www.asa.org.uk/news/flashing-images-in-advertising.html).

- [ITU-R BT.1359-1](https://www.itu.int/dms_pubrec/itu-r/rec/bt/R-REC-BT.1359-1-199811-I%21%21PDF-E.pdf), Relative

timing of sound and vision for broadcasting: detectability about +45 ms (sound early) to −125 ms (sound late),   acceptability about +90 ms to −185 ms. - [EBU R37-2007](https://tech.ebu.ch/publications/r037): 5 ms early to 15 ms late per st

acceptability about +90 ms to −185 ms. - [EBU R37-2007](https://tech.ebu.ch/publications/r037): 5 ms early to 15 ms late per stage; 40 ms early to 60 ms   late at emission.

- [EBU R37-2007](https://tech.ebu.ch/publications/r037): 5 ms early to 15 ms late per stage; 40 ms early to 60 ms

late at emission. - [ITU-R BS.1770-5](https://www.itu.int/rec/R-REC-BS.1770/en) (11/2023): loudness and true-peak measurement. - [EBU R128-2023](https://tech.ebu.ch/publications/r128) (−23.0 LUFS, true peak ≤ −1 dBTP) and

- [ITU-R BS.1770-5](https://www.itu.int/rec/R-REC-BS.1770/en) (11/2023): loudness and true-peak measurement.

- [EBU R128-2023](https://tech.ebu.ch/publications/r128) (−23.0 LUFS, true peak ≤ −1 dBTP) and   [R128 s1-2020](https://tech.ebu.ch/publications/r128s1) (short-form: −23 LUFS, maximum short-term −18 LUFS).   EBU's site may refuse automated 

[R128 s1-2020](https://tech.ebu.ch/publications/r128s1) (short-form: −23 LUFS, maximum short-term −18 LUFS).   EBU's site may refuse automated requests; open the links in a browser. - ATSC A/85:2026-07 (−24 LKFS; true peak below −2 dBTP; st

- ATSC A/85:2026-07 (−24 LKFS; true peak below −2 dBTP; streaming annex −23 to −27 LKFS), read from the ATSC PDF;   ATSC documents are listed at atsc.org. - [EBU Tech 3344](https://tech.ebu.ch/publications/tech3344): lossy codecs raise true

ATSC documents are listed at atsc.org. - [EBU Tech 3344](https://tech.ebu.ch/publications/tech3344): lossy codecs raise true peak, more at low bitrates. - [Spotify, Loudness normalization](https://support.spotify.com/us/artists/article/loud

- [EBU Tech 3344](https://tech.ebu.ch/publications/tech3344): lossy codecs raise true peak, more at low bitrates.

- [Spotify, Loudness normalization](https://support.spotify.com/us/artists/article/loudness-normalization/) (music):

- [Netflix, English (USA) Timed Text Style Guide](https://partnerhelp.netflixstudios.com/hc/en-us/articles/217350977-English-USA-Timed-Text-Style-Guide)

(revised 2025-12-19): 42 characters per line; up to 20 characters per second (adult) and 17 (children's).   [General requirements](https://partnerhelp.netflixstudios.com/hc/en-us/articles/215758617-Timed-Text-Style-Guide-General-Requirement

[General requirements](https://partnerhelp.netflixstudios.com/hc/en-us/articles/215758617-Timed-Text-Style-Guide-General-Requirements):

a minimum of 5/6 s (another Netflix page says 20 frames, about 0.8 s), a maximum of 7 s per event, two lines. - Brysbaert, M. (2019). How many words do we read per minute? A review and meta-analysis of reading rate. Journal   of Memory and 

of Memory and Language, 109, 104047. https://doi.org/10.1016/j.jml.2019.104047 (238 wpm silent non-fiction).

- YouTube recommended upload encoding settings. https://support.google.com/youtube/answer/1722171

- FFmpeg documentation (formats: `+faststart`; filters: `ebur128`, `loudnorm`, `psnr`, `libvmaf`, `subtitles`,   `xfade`, `whisper`). https://ffmpeg.org/documentation.html - Mux, "Your browser and my browser see different colors" (2020): un

`xfade`, `whisper`). https://ffmpeg.org/documentation.html

- Mux, "Your browser and my browser see different colors" (2020): untagged video renders differently by browser.   https://www.mux.com/blog/your-browser-and-my-browser-see-different-colors - Verified on FFmpeg 9.0.2 while writing this skill

https://www.mux.com/blog/your-browser-and-my-browser-see-different-colors

- LinkedIn video ads specs: https://www.linkedin.com/help/lms/answer/a424737

- Meta ads guide, Instagram Reels video ads: https://www.facebook.com/business/ads-guide/

- TikTok in-feed ads specifications: https://ads.tiktok.com/resources/help/article/tiktok-auction-in-feed-ads

HyperFrames (https://github.com/heygen-com/hyperframes), Remotion (https://www.remotion.dev/docs and its

LICENSE.md), Blender (https://www.blender.org), OpenTimelineIO (https://github.com/AcademySoftwareFoundation/OpenTimelineIO),

Playwright (https://playwright.dev/docs/videos), WhisperX (https://github.com/m-bain/whisperX), VMAF

(https://github.com/Netflix/vmaf), Demucs (https://github.com/adefossez/demucs, licence note of 2026-10-26).

film. Measure each motion reference's pace with `motion_scan.py` (still share, longest still stretch, cuts per    10 s; [kinetic-type](kinetic-type.md) §4) and record the numbers beside it. Name a look by its cause (halation,    bleach bypa

10 s; [kinetic-type](kinetic-type.md) §4) and record the numbers beside it. Name a look by its cause (halation,    bleach bypass, halftone, a dolly zoom): the [VFX Atlas](https://atlas.letz.ai/) describes 504 effects and looks    from 1820 

bleach bypass, halftone, a dolly zoom): the [VFX Atlas](https://atlas.letz.ai/) describes 504 effects and looks

Executable file

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

mask_bytes = base64.b64decode(encoded_mask, validate=True)

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env python3

def default_fetch(url):     request = urllib.request.Request(url, headers={'User-Agent': 'legible-film cue_list freeze'})

request = urllib.request.Request(url, headers={'User-Agent': 'legible-film cue_list freeze'})

with urllib.request.urlopen(request, timeout=30) as response:

continue         try:             data = fetch(url)

try:             data = fetch(url)         except Exception as error:   # any fetch failure leaves the reference live and the binding unverified

data = fetch(url)

[Credential-related source omitted]

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env node

node film_capture.mjs info http://127.0.0.1:8000/?render

node film_capture.mjs analyse --url http://127.0.0.1:8000/?render

}     const url=new URL(a.url||process.env.LF_URL||'');     if(url.protocol!=='http:'||url.hostname!=='127.0.0.1')throw Error('page must be http://127.0.0.1:<port>/');

const url=new URL(a.url||process.env.LF_URL||'');     if(url.protocol!=='http:'||url.hostname!=='127.0.0.1')throw Error('page must be http://127.0.0.1:<port>/');     if(requestedSize) {url.searchParams.set('w',String(requestedSize.width));u

if(url.protocol!=='http:'||url.hostname!=='127.0.0.1')throw Error('page must be http://127.0.0.1:<port>/');

node film_capture.mjs cues film/ --out film/cues.json --source timelines --verify   node film_capture.mjs cues film/ --out cues.json --source declared --ready __cues_ready --global __cues --fps 60   node film_capture.mjs cues http://127.0.0

node film_capture.mjs cues http://127.0.0.1:8000/index.html?render --root film/ --out cues.json --source lf

const server = http.createServer((req, res) => {     let rel;     try { rel = decodeURIComponent(new URL(req.url, 'http://127.0.0.1').pathname); } catch { res.writeHead(400); res.end(); return; }

let rel;     try { rel = decodeURIComponent(new URL(req.url, 'http://127.0.0.1').pathname); } catch { res.writeHead(400); res.end(); return; }     if (rel.endsWith('/')) rel += 'index.html';

try { rel = decodeURIComponent(new URL(req.url, 'http://127.0.0.1').pathname); } catch { res.writeHead(400); res.end(); return; }

});   await new Promise((resolve, reject) => { server.once('error', reject); server.listen(0, '127.0.0.1', resolve); });   return {server, origin: `http://127.0.0.1:${server.address().port}`, seen};

await new Promise((resolve, reject) => { server.once('error', reject); server.listen(0, '127.0.0.1', resolve); });   return {server, origin: `http://127.0.0.1:${server.address().port}`, seen}; }

return {server, origin: `http://127.0.0.1:${server.address().port}`, seen};

host.setAttribute('data-lf-mounted', '');         const src = host.getAttribute('data-composition-src');         const r = await fetch(src);

const src = host.getAttribute('data-composition-src');         const r = await fetch(src);         if (!r.ok) throw Error('scene file ' + src + ' could not be read (' + r.status + ')');

const r = await fetch(src);

#!/usr/bin/env python3

(r'(?<![\w-])animation(-name)?\s*:', 'CSS animation'),     (r'@keyframes', 'CSS @keyframes'),     (r'\bfetch\s*\(', 'fetch (network)'),

(r'@keyframes', 'CSS @keyframes'),     (r'\bfetch\s*\(', 'fetch (network)'),     (r'XMLHttpRequest|\bnew\s+WebSocket\b|\bEventSource\b', 'network request'),

(r'\bfetch\s*\(', 'fetch (network)'),

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

raise RuntimeError('the cue export needs Node and film_capture.mjs beside this script')     target = os.path.join(folder, 'cues.json')     url = f'http://127.0.0.1/index.html?render&w={w}&h={h}' + (f'&res={a.res}' if a.res else '')

target = os.path.join(folder, 'cues.json')     url = f'http://127.0.0.1/index.html?render&w={w}&h={h}' + (f'&res={a.res}' if a.res else '')     cmd = [node, capture, 'cues', url, '--root', folder, '--out', target, '--source', 'lf', '--size'

url = f'http://127.0.0.1/index.html?render&w={w}&h={h}' + (f'&res={a.res}' if a.res else '')

httpd = None if file_only else serve(folder)     url = (Path(folder, 'index.html').as_uri() if file_only else            f'http://127.0.0.1:{httpd.server_address[1]}/index.html')

url = (Path(folder, 'index.html').as_uri() if file_only else            f'http://127.0.0.1:{httpd.server_address[1]}/index.html')     url += f'?render&w={w}&h={h}' + (f'&res={a.res}' if a.res else '')

f'http://127.0.0.1:{httpd.server_address[1]}/index.html')

[Credential-related source omitted]

else:                 httpd = serve(kept)                 url = f'http://127.0.0.1:{httpd.server_address[1]}/index.html'

httpd = serve(kept)                 url = f'http://127.0.0.1:{httpd.server_address[1]}/index.html'             url += f'?render&w={w}&h={h}' + (f'&res={a.res}' if a.res else '')

url = f'http://127.0.0.1:{httpd.server_address[1]}/index.html'

[Credential-related source omitted]

#!/usr/bin/env python3

raise ValueError('--page must stay inside the film folder')             page = a.page + ('&' if '?' in a.page else '?') + 'render'             url = f'http://127.0.0.1:{server.server_port}/{page}'

page = a.page + ('&' if '?' in a.page else '?') + 'render'             url = f'http://127.0.0.1:{server.server_port}/{page}'             try:

url = f'http://127.0.0.1:{server.server_port}/{page}'

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

payload = base64.b64decode(row.pop('png'))

#!/usr/bin/env python3

thread = threading.Thread(target=server.serve_forever, daemon=True)         thread.start()         url = f"http://127.0.0.1:{server.server_port}/?render"

url = f"http://127.0.0.1:{server.server_port}/?render"

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

'ffmpeg': pick('winget install Gyan.FFmpeg', 'brew install ffmpeg', 'sudo apt install ffmpeg'),

'node': pick('winget install OpenJS.NodeJS.LTS', 'brew install node', 'sudo apt install nodejs npm'),

[Credential-related source omitted]

try:             node = find_tool('node', self.a.node)             url = f'http://127.0.0.1:{server.server_port}/index.html?render'

node = find_tool('node', self.a.node)             url = f'http://127.0.0.1:{server.server_port}/index.html?render'             code, text = self.run('info', [node, HERE / 'film_capture.mjs', 'info', url, '--size', label])

url = f'http://127.0.0.1:{server.server_port}/index.html?render'

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3

MODEL_NAME = 'modnet.onnx' MODEL_SHA256 = '07c308cf0fc7e6e8b2065a12ed7fc07e1de8febb7dc7839d7b7f15dd66584df9' MODEL_URL = ('https://huggingface.co/Xenova/modnet/resolve/'

MODEL_SHA256 = '07c308cf0fc7e6e8b2065a12ed7fc07e1de8febb7dc7839d7b7f15dd66584df9' MODEL_URL = ('https://huggingface.co/Xenova/modnet/resolve/'              'fa2fa546052fba4c08921230a26cc69a333fca12/onnx/model.onnx')

MODEL_URL = ('https://huggingface.co/Xenova/modnet/resolve/'

'fa2fa546052fba4c08921230a26cc69a333fca12/onnx/model.onnx') MODEL_CARD = 'https://huggingface.co/Xenova/modnet' CODE_LICENCE = 'https://github.com/ZHKKKe/MODNet/blob/master/LICENSE'

MODEL_CARD = 'https://huggingface.co/Xenova/modnet'

CODE_LICENCE = 'https://github.com/ZHKKKe/MODNet/blob/master/LICENSE'

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

serve = project.python_command(['-m', 'http.server', '--bind', '127.0.0.1', '--directory', out, '8000'])     print(f'  3. Preview: {serve}')     print('     then open http://127.0.0.1:8000/')

print(f'  3. Preview: {serve}')     print('     then open http://127.0.0.1:8000/')     print('     (keys: space play, , . frame, arrows beat, 0-9 jump, S safe area, R reading states, L luminance).')

print('     then open http://127.0.0.1:8000/')

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

preload = Path(__file__).with_name('render_block.cjs').resolve().as_posix()     env['NODE_OPTIONS'] = f'--require "{preload}"' + (' ' + env['NODE_OPTIONS'] if env.get('NODE_OPTIONS') else '')     url = f'http://127.0.0.1:{sink.server_addres

env['NODE_OPTIONS'] = f'--require "{preload}"' + (' ' + env['NODE_OPTIONS'] if env.get('NODE_OPTIONS') else '')     url = f'http://127.0.0.1:{sink.server_address[1]}'     for key in ('HTTP_PROXY', 'HTTPS_PROXY', 'ALL_PROXY', 'http_proxy', '

url = f'http://127.0.0.1:{sink.server_address[1]}'

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

Executable file

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

import hashlib import html import http.client

import html import http.client import http.cookies

import http.client

raw = urllib.parse.unquote(urllib.parse.urlsplit(self.path).path)             if raw == '/favicon.ico':                 icon = (b'<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64">'

if raw == '/favicon.ico':                 icon = (b'<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64">'                         b'<rect width="64" height="64" rx="10" fill="#3b5cf6"/>'

icon = (b'<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64">'

origin = self.headers.get('Origin')             if origin and origin not in {f'http://127.0.0.1:{self.server.server_port}',

if origin and origin not in {f'http://127.0.0.1:{self.server.server_port}',

f'http://localhost:{self.server.server_port}'}:

return False             path = '/screening/index.html'         connection = http.client.HTTPConnection('127.0.0.1', port, timeout=1)

path = '/screening/index.html'         connection = http.client.HTTPConnection('127.0.0.1', port, timeout=1)         try:

connection = http.client.HTTPConnection('127.0.0.1', port, timeout=1)

finally:             connection.close()     except (OSError, ValueError, TypeError, http.client.HTTPException):

connection.close()     except (OSError, ValueError, TypeError, http.client.HTTPException):         return False

except (OSError, ValueError, TypeError, http.client.HTTPException):

server.server_close()         raise     url = 'http:' + '//127.0.0.1:' + str(server.server_port) + '/screening/index.html?t=' + token

raise     url = 'http:' + '//127.0.0.1:' + str(server.server_port) + '/screening/index.html?t=' + token     print(url, flush=True)

url = 'http:' + '//127.0.0.1:' + str(server.server_port) + '/screening/index.html?t=' + token

function expired(){let banner=$('connection');banner.hidden=false;   banner.textContent="This link has expired; reopen the room from the agent's link."} async function post(path,body){let response=await fetch('../api/'+path,{method:'POST',

banner.textContent="This link has expired; reopen the room from the agent's link."} async function post(path,body){let response=await fetch('../api/'+path,{method:'POST',   headers:{'Content-Type':'application/json'},body:JSON.stringify(bod

async function post(path,body){let response=await fetch('../api/'+path,{method:'POST',

}; let snapshot=JSON.stringify(data);async function poll(){try{   let r=await fetch('../api/status',{cache:'no-store'});if(r.status===403){expired();return}if(!r.ok){

let snapshot=JSON.stringify(data);async function poll(){try{   let r=await fetch('../api/status',{cache:'no-store'});if(r.status===403){expired();return}if(!r.ok){     $('connection').hidden=false;

let r=await fetch('../api/status',{cache:'no-store'});if(r.status===403){expired();return}if(!r.ok){

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

Executable file

#!/usr/bin/env python3

Executable file

Executable file

Executable file

[Credential-related source omitted]

text[-200:])     runtime = page_copy(root / 'runtime-fetch', html={"window.__timelines = window.__timelines || {};\n(function () {":         "window.__timelines = window.__timelines || {};\nfetch('https://' + 'example.com/lf-fixture.json').

runtime = page_copy(root / 'runtime-fetch', html={"window.__timelines = window.__timelines || {};\n(function () {":         "window.__timelines = window.__timelines || {};\nfetch('https://' + 'example.com/lf-fixture.json').catch(() => {});\

"window.__timelines = window.__timelines || {};\nfetch('https://' + 'example.com/lf-fixture.json').catch(() => {});\n(function () {"})

readme = '<!doctype html><title>notes</title><script src="http://example.com/tracker.js"></script>\n' \

'<link rel="stylesheet" href="http://example.com/theme.css">\n'

"s.src = 'late/' + 'later.js'; document.head.appendChild(s); })();\n(function () {"})     (late / 'late').mkdir()     (late / 'late' / 'later.js').write_text('if (false) import("https://example.com/unpinned/later.js");\n', encoding='utf-8')

(late / 'late').mkdir()     (late / 'late' / 'later.js').write_text('if (false) import("https://example.com/unpinned/later.js");\n', encoding='utf-8')     code, text, doc = export(late)

(late / 'late' / 'later.js').write_text('if (false) import("https://example.com/unpinned/later.js");\n', encoding='utf-8')

[Credential-related source omitted]

Executable file

Executable file

Executable file

[Credential-related source omitted]

Executable file

Executable file

#!/usr/bin/env python3

Executable file

Executable file

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

Executable file

#!/usr/bin/env python3

Executable file

Executable file

net = root / 'freeze-net' / 'page'     net.mkdir(parents=True)     pinned = 'https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js'

net.mkdir(parents=True)     pinned = 'https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js'     (net / 'index.html').write_text(f'<!doctype html><script src="{pinned}"></script>\n'

pinned = 'https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js'

(net / 'index.html').write_text(f'<!doctype html><script src="{pinned}"></script>\n'                                     '<a href="https://example.com/about">about</a>\n', encoding='utf-8')     fetched = []

'<a href="https://example.com/about">about</a>\n', encoding='utf-8')

def fetch(url):

loose.mkdir(parents=True)     (loose.parent / 'shared.js').write_text('// shared\n', encoding='utf-8')     (loose / 'index.html').write_text('<!doctype html><script src="https://cdn.jsdelivr.net/npm/gsap@3/dist/gsap.js">'

(loose.parent / 'shared.js').write_text('// shared\n', encoding='utf-8')     (loose / 'index.html').write_text('<!doctype html><script src="https://cdn.jsdelivr.net/npm/gsap@3/dist/gsap.js">'                                       '</script>

(loose / 'index.html').write_text('<!doctype html><script src="https://cdn.jsdelivr.net/npm/gsap@3/dist/gsap.js">'                                       '</script><script src="../shared.js"></script>\n', encoding='utf-8')     record_loose =

data = root / 'data-urls'     data.mkdir()     (data / 'index.html').write_bytes(b'<!doctype html><p data-href="https://example.com/p">x</p>'

data.mkdir()     (data / 'index.html').write_bytes(b'<!doctype html><p data-href="https://example.com/p">x</p>'                                       b'<script src="app.js"></script>\n')

(data / 'index.html').write_bytes(b'<!doctype html><p data-href="https://example.com/p">x</p>'

b'<script src="app.js"></script>\n')     (data / 'app.js').write_bytes(b'const sources = ["https://example.com/a.json", \'https://cdn.example.org/x.png\'];\n'                                   b'const link = `see https://example.net/docs`;\

(data / 'app.js').write_bytes(b'const sources = ["https://example.com/a.json", \'https://cdn.example.org/x.png\'];\n'

b'const link = `see https://example.net/docs`;\n')     (data / 'facts.json').write_bytes(b'{"site": "https://example.com", "logo": "https://example.com/logo.svg"}\n')     plain = cl.freeze(data, fetch=lambda url: b'')

(data / 'facts.json').write_bytes(b'{"site": "https://example.com", "logo": "https://example.com/logo.svg"}\n')     plain = cl.freeze(data, fetch=lambda url: b'')     record('cue_list: URLs in data strings (a JS array, a template string, a 

(loads / 'index.html').write_bytes(         b'<!doctype html><link rel="stylesheet" href="style.css">'         b'<img srcset="https://a.example/1.png 1x, https://a.example/2.png 2x" src="local.png">'

b'<!doctype html><link rel="stylesheet" href="style.css">'         b'<img srcset="https://a.example/1.png 1x, https://a.example/2.png 2x" src="local.png">'         b'<video poster="https://a.example/p.jpg"></video>'

b'<img srcset="https://a.example/1.png 1x, https://a.example/2.png 2x" src="local.png">'

b'<video poster="https://a.example/p.jpg"></video>'

b'<style>.x{background:url("https://a.example/inline.png")}</style>'

b'<script type="module" src="main.js"></script>\n')     (loads / 'style.css').write_bytes(b'@import "https://fonts.example/x.css";\n.b{background:url(https://a.example/bg.png)}\n')     (loads / 'main.js').write_bytes(b"import x from 'https:

(loads / 'style.css').write_bytes(b'@import "https://fonts.example/x.css";\n.b{background:url(https://a.example/bg.png)}\n')

(loads / 'main.js').write_bytes(b"import x from 'https://esm.example/m.js';\n"

b'export { y } from "https://esm.example/n.js";\n'

b"const later = () => import('https://esm.example/dyn.js');\n")     (loads / 'local.png').write_bytes(b'png')     positions = cl.freeze(loads, fetch=lambda url: b'')

(loads / 'local.png').write_bytes(b'png')     positions = cl.freeze(loads, fetch=lambda url: b'')     wanted = {'https://a.example/1.png', 'https://a.example/2.png', 'https://a.example/p.jpg',

positions = cl.freeze(loads, fetch=lambda url: b'')     wanted = {'https://a.example/1.png', 'https://a.example/2.png', 'https://a.example/p.jpg',               'https://a.example/inline.png', 'https://fonts.example/x.css', 'https://a.examp

wanted = {'https://a.example/1.png', 'https://a.example/2.png', 'https://a.example/p.jpg',

'https://a.example/inline.png', 'https://fonts.example/x.css', 'https://a.example/bg.png',

'https://esm.example/m.js', 'https://esm.example/n.js', 'https://esm.example/dyn.js'}

#!/usr/bin/env python3

if '/v1/audio/' in url:             return json.dumps({'results': [dict(id=c, license=c, license_version='1.0' if c == 'cc0' else '4.0',                   duration=550, url='https://example.org/' + c, title='Fixture', creator='Fixture autho

return json.dumps({'results': [dict(id=c, license=c, license_version='1.0' if c == 'cc0' else '4.0',                   duration=550, url='https://example.org/' + c, title='Fixture', creator='Fixture author',                   foreign_landin

duration=550, url='https://example.org/' + c, title='Fixture', creator='Fixture author',                   foreign_landing_url='https://example.org/source') for c in ('by-nc', 'by-nd', 'by-sa', 'cc0', 'by')]}).encode()         return (root 

foreign_landing_url='https://example.org/source') for c in ('by-nc', 'by-nd', 'by-sa', 'cc0', 'by')]}).encode()

Executable file

[Credential-related source omitted]

Executable file

edited = look.replace('- The feeling in three words:', '- The feeling in three words: calm, exact, warm')     (root / 'look.md').write_text(edited, encoding='utf-8')     code, output = tool('look.py', 'add', root, '--link', 'https://example

(root / 'look.md').write_text(edited, encoding='utf-8')     code, output = tool('look.py', 'add', root, '--link', 'https://example.com/a-film', '--note', 'liked its pace')     code2, output2 = tool('look.py', 'set', root, 'R01', '--mode', '

code, output = tool('look.py', 'add', root, '--link', 'https://example.com/a-film', '--note', 'liked its pace')

code == code2 == 0 and 'calm, exact, warm' in after and '| L01 |' in after            and 'recreate closely' in after and not list((root / 'refs').glob('L01*'))            and next(row for row in index['refs'] if row['id'] == 'L01').get('ur

and 'recreate closely' in after and not list((root / 'refs').glob('L01*'))            and next(row for row in index['refs'] if row['id'] == 'L01').get('url') == 'https://example.com/a-film',            output + output2 + after[-500:])

and next(row for row in index['refs'] if row['id'] == 'L01').get('url') == 'https://example.com/a-film',

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

const value = f => flags.includes(f) ? flags[flags.indexOf(f) + 1] : null; if (value('--edit-live')) fs.appendFileSync(value('--edit-live'), '\n<!-- edited during the render -->\n'); if (value('--fetch')) { try { await fetch(value('--fetch'

if (value('--edit-live')) fs.appendFileSync(value('--edit-live'), '\n<!-- edited during the render -->\n'); if (value('--fetch')) { try { await fetch(value('--fetch')); } catch (e) { console.error('fetch refused: ' + (e.cause?.code || e.mes

if (value('--fetch')) { try { await fetch(value('--fetch')); } catch (e) { console.error('fetch refused: ' + (e.cause?.code || e.message)); } }

[Credential-related source omitted]

[Credential-related source omitted]

check_page(hf_fetch)         code, text = add(root / 'project-fetch', '--cues', hf_fetch / 'cues.json', '--file', root / 'renders' / 'fetch.mp4',                          '--render', stub_cmd('--fetch', 'https://example.com/lf-during-render

code, text = add(root / 'project-fetch', '--cues', hf_fetch / 'cues.json', '--file', root / 'renders' / 'fetch.mp4',                          '--render', stub_cmd('--fetch', 'https://example.com/lf-during-render.json'))         record('regi

'--render', stub_cmd('--fetch', 'https://example.com/lf-during-render.json'))

and 'the network (blocked): example.com:443' in text, text[-300:])         code, text = add(root / 'project-fetch', '--cues', hf_fetch / 'cues.json', '--file', root / 'renders' / 'fetch-b.mp4',                          '--render', stub_cmd(

code, text = add(root / 'project-fetch', '--cues', hf_fetch / 'cues.json', '--file', root / 'renders' / 'fetch-b.mp4',                          '--render', stub_cmd('--browser', browser['chrome'], '--goto', 'https://example.com/lf-browser')

'--render', stub_cmd('--browser', browser['chrome'], '--goto', 'https://example.com/lf-browser'))

gsap = make_page(root / 'page-gsap', head='<script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script>\n')

[Credential-related source omitted]

#!/usr/bin/env python3

Executable file

Executable file

Executable file

Executable file

stale_path = root / 'screening/server.json'     stale_path.write_text(json.dumps({'pid': os.getpid(),                                      'url': 'http://127.0.0.1:1/screening/index.html?t=stale'}), encoding='utf-8')

stale_path.write_text(json.dumps({'pid': os.getpid(),                                      'url': 'http://127.0.0.1:1/screening/index.html?t=stale'}), encoding='utf-8')     try:

'url': 'http://127.0.0.1:1/screening/index.html?t=stale'}), encoding='utf-8')

info = root / 'screening/server.json'     info.write_text(json.dumps({'pid': os.getpid(), 'port': 12345,                                 'url': 'http://127.0.0.1:12345/screening/index.html?t=stale-secret-XYZ'}),

info.write_text(json.dumps({'pid': os.getpid(), 'port': 12345,                                 'url': 'http://127.0.0.1:12345/screening/index.html?t=stale-secret-XYZ'}),                     encoding='utf-8')

'url': 'http://127.0.0.1:12345/screening/index.html?t=stale-secret-XYZ'}),

connection.getresponse.return_value = response     with patch.object(screen, 'pid_alive', return_value=True), \             patch.object(screen.http.client, 'HTTPConnection', return_value=connection):

with patch.object(screen, 'pid_alive', return_value=True),              patch.object(screen.http.client, 'HTTPConnection', return_value=connection):

patch.object(screen.http.client, 'HTTPConnection', return_value=connection):

def request(path, auth=True, headers=None, body=None):             target = base + path + (('?t=' + token) if auth else '')             req = urllib.request.Request(target, data=json.dumps(body).encode() if body is not None else None,

target = base + path + (('?t=' + token) if auth else '')             req = urllib.request.Request(target, data=json.dumps(body).encode() if body is not None else None,                                          headers=headers or {})

req = urllib.request.Request(target, data=json.dumps(body).encode() if body is not None else None,

headers=headers or {})             try:                 with urllib.request.urlopen(req, timeout=5) as response:

try:                 with urllib.request.urlopen(req, timeout=5) as response:                     return response.status, response.read()

with urllib.request.urlopen(req, timeout=5) as response:

except urllib.error.HTTPError as error:                 return error.code, error.read()         with urllib.request.urlopen(base + '/screening/index.html?t=' + token, timeout=5) as response:

return error.code, error.read()         with urllib.request.urlopen(base + '/screening/index.html?t=' + token, timeout=5) as response:             cookie_header = response.headers.get('Set-Cookie', '')

with urllib.request.urlopen(base + '/screening/index.html?t=' + token, timeout=5) as response:

depth_code, depth_body = request('/api/notes', headers={'Content-Type': 'application/json'},                                          body={'text': 'x', 't': .5, 'nested': nested})         deep_request = urllib.request.Request(base + '/api/

body={'text': 'x', 't': .5, 'nested': nested})         deep_request = urllib.request.Request(base + '/api/notes?t=' + token,                                               data=b'{"x":' + b'[' * 3000 + b'0' + b']' * 3000 + b'}',

deep_request = urllib.request.Request(base + '/api/notes?t=' + token,

headers={'Content-Type': 'application/json'})         try:             urllib.request.urlopen(deep_request, timeout=5)

try:             urllib.request.urlopen(deep_request, timeout=5)             deep_code, deep_body = 200, b''

urllib.request.urlopen(deep_request, timeout=5)

record('screen: server rejects requests without token', code == 403, str(code))         try:             urllib.request.urlopen(base + '/screening/index.html', timeout=5)

try:             urllib.request.urlopen(base + '/screening/index.html', timeout=5)             no_cookie = False

urllib.request.urlopen(base + '/screening/index.html', timeout=5)

record('screen launch room: the room serves only its own font files', code == 200 and body == shipped                and other == 404, str((code, other)))         req = urllib.request.Request(base + '/api/approve?t=' + token, data=b'{}',

and other == 404, str((code, other)))         req = urllib.request.Request(base + '/api/approve?t=' + token, data=b'{}',                                      headers={'Origin': 'http://attacker.example', 'Content-Type': 'application/json'})

req = urllib.request.Request(base + '/api/approve?t=' + token, data=b'{}',

headers={'Origin': 'http://attacker.example', 'Content-Type': 'application/json'})         try:             urllib.request.urlopen(req, timeout=5)

try:             urllib.request.urlopen(req, timeout=5)             no_cookie = False

urllib.request.urlopen(req, timeout=5)

const context=await browser.newContext();const a=await context.newPage(),b=await context.newPage();  await a.goto(process.argv[4]);await b.goto(process.argv[5]);  for(const page of [a,b])assert.equal(await page.evaluate(()=>fetch('/api/stat

await a.goto(process.argv[4]);await b.goto(process.argv[5]);  for(const page of [a,b])assert.equal(await page.evaluate(()=>fetch('/api/status').then(r=>r.status)),200);  const error=new Promise(resolve=>b.once('dialog',async dialog=>{

for(const page of [a,b])assert.equal(await page.evaluate(()=>fetch('/api/status').then(r=>r.status)),200);

await a.evaluate(()=>poll());  assert((await a.locator('#connection').innerText()).includes('link has expired'));  assert.equal(await b.evaluate(()=>fetch('/api/status').then(r=>r.status)),200);

assert((await a.locator('#connection').innerText()).includes('link has expired'));  assert.equal(await b.evaluate(()=>fetch('/api/status').then(r=>r.status)),200);  await browser.close();

assert.equal(await b.evaluate(()=>fetch('/api/status').then(r=>r.status)),200);

host, port = '127.0.0.1', info['port']         for _ in range(4):             client = socket.create_connection((host, port), timeout=5)

for _ in range(4):             client = socket.create_connection((host, port), timeout=5)             client.setsockopt(socket.SOL_SOCKET, socket.SO_LINGER, struct.pack('HH' if os.name == 'nt' else 'ii', 1, 0))

client = socket.create_connection((host, port), timeout=5)

Executable file

#!/usr/bin/env python3

Executable file

Executable file

Executable file

Executable file

Executable file

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

Executable file

Executable file

def eval(self): return self

[Credential-related source omitted]

record('setup ear: public CPU installs and model downloads are explicit and unauthenticated',                len(calls) == 3 and '--isolated' in calls[0] and '--isolated' in calls[1]                and 'https://download.pytorch.org/whl/cpu'

len(calls) == 3 and '--isolated' in calls[0] and '--isolated' in calls[1]                and 'https://download.pytorch.org/whl/cpu' in calls[0]                and 'https://pypi.org/simple' in calls[1]

and 'https://download.pytorch.org/whl/cpu' in calls[0]

and 'https://pypi.org/simple' in calls[1]

[Credential-related source omitted]

#!/usr/bin/env python3

SPATIAL_MODEL = 'modnet.onnx' SPATIAL_SHA256 = '07c308cf0fc7e6e8b2065a12ed7fc07e1de8febb7dc7839d7b7f15dd66584df9' SPATIAL_URL = ('https://huggingface.co/Xenova/modnet/resolve/'

SPATIAL_SHA256 = '07c308cf0fc7e6e8b2065a12ed7fc07e1de8febb7dc7839d7b7f15dd66584df9' SPATIAL_URL = ('https://huggingface.co/Xenova/modnet/resolve/'                'fa2fa546052fba4c08921230a26cc69a333fca12/onnx/model.onnx')

SPATIAL_URL = ('https://huggingface.co/Xenova/modnet/resolve/'

'Linux': ('sudo apt update && sudo apt install ffmpeg', 'sudo apt update && sudo apt install nodejs npm'),

}     if system not in commands:         return 'Install a supported build for your operating system; see https://ffmpeg.org/download.html'

if system not in commands:         return 'Install a supported build for your operating system; see https://ffmpeg.org/download.html'     command = commands[system][0 if kind == 'ffmpeg' else 1]

return 'Install a supported build for your operating system; see https://ffmpeg.org/download.html'

[Credential-related source omitted]

[Credential-related source omitted]

'(sudo apt install python3-venv), then run: ' + fix)

'python3-venv (sudo apt install python3-venv), then run this command again.') from error

'(sudo apt install python3-venv), then run this command again.')

[Credential-related source omitted]

[Credential-related source omitted]

return ('Ubuntu 22.04 commonly supplies FFmpeg 4.4; Debian 11 supplies 4.3. '                 'Upgrade to a release with FFmpeg 5.1 or later, or install a current build from '                 'https://ffmpeg.org/download.html and put it fir

'Upgrade to a release with FFmpeg 5.1 or later, or install a current build from '                 'https://ffmpeg.org/download.html and put it first on PATH.')     if system == 'Darwin':

'https://ffmpeg.org/download.html and put it first on PATH.')

if system == 'Windows':         return command_fix('ffmpeg', system) + '\n    FFmpeg 5.1 or later is required.'     return 'Install FFmpeg 5.1 or later from https://ffmpeg.org/download.html.'

return 'Install FFmpeg 5.1 or later from https://ffmpeg.org/download.html.'

[Credential-related source omitted]

[Credential-related source omitted]

url = 'https://api.github.com/repos/JamesDimachkie/legible-film/releases/latest'     request = urllib.request.Request(url, headers={'Accept': 'application/vnd.github+json',

request = urllib.request.Request(url, headers={'Accept': 'application/vnd.github+json',

'User-Agent': 'legible-film-setup'})     try:         with urllib.request.urlopen(request, timeout=4) as response:

try:         with urllib.request.urlopen(request, timeout=4) as response:             release = json.load(response)

with urllib.request.urlopen(request, timeout=4) as response:

[Credential-related source omitted]

[Credential-related source omitted]

newer, detail = latest_release(version)         add('Skill', 'limited' if newer is None or newer else 'ready', 'Updates', detail,             'Review the latest release at https://github.com/JamesDimachkie/legible-film/releases/latest')

add('Skill', 'limited' if newer is None or newer else 'ready', 'Updates', detail,             'Review the latest release at https://github.com/JamesDimachkie/legible-film/releases/latest')     return groups

'Review the latest release at https://github.com/JamesDimachkie/legible-film/releases/latest')

path = Path(temporary.name)     try:         urllib.request.urlretrieve(SPATIAL_URL, str(path))

try:         urllib.request.urlretrieve(SPATIAL_URL, str(path))         if project.file_sha256(path) != SPATIAL_SHA256:

urllib.request.urlretrieve(SPATIAL_URL, str(path))

libraries = (f" On Linux it also needs system libraries: sudo {npx} --prefix {package_dir} playwright-core "

python = str(python or sys.executable)     subprocess.run([python, '-I', '-m', 'pip', '--isolated', 'install', 'torch', '--index-url',                     'https://download.pytorch.org/whl/cpu'], check=True, stdout=sys.stderr)

subprocess.run([python, '-I', '-m', 'pip', '--isolated', 'install', 'torch', '--index-url',                     'https://download.pytorch.org/whl/cpu'], check=True, stdout=sys.stderr)     subprocess.run([python, '-I', '-m', 'pip', '--isolat

'https://download.pytorch.org/whl/cpu'], check=True, stdout=sys.stderr)

subprocess.run([python, '-I', '-m', 'pip', '--isolated', 'install', 'transformers<5', 'huggingface-hub',                     'safetensors', 'librosa>=0.10', 'soundfile>=0.12', 'soxr>=0.3',                     '--index-url', 'https://pypi.or

'safetensors', 'librosa>=0.10', 'soundfile>=0.12', 'soxr>=0.3',                     '--index-url', 'https://pypi.org/simple'],                    check=True, stdout=sys.stderr)

'--index-url', 'https://pypi.org/simple'],

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

Search uses GET https://api.openverse.org/v1/audio/ with q, license, page_size,

in the user's cache (or --rate-state). HTTP 429 stops with its Retry-After reason; no retries. See Openverse's API documentation before use: https://api.openverse.org/v1/ and

Retry-After reason; no retries. See Openverse's API documentation before use: https://api.openverse.org/v1/ and https://docs.openverse.org/api/reference/authentication_and_throttling.html

https://api.openverse.org/v1/ and

https://docs.openverse.org/api/reference/authentication_and_throttling.html

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

def fetch(url, limit):

public_url(url)     request = urllib.request.Request(url, headers={'User-Agent': 'LegibleFilm/1.2 sound-library (no authentication)'})     with urllib.request.build_opener(PublicRedirect()).open(request, timeout=FETCH_TIMEOUT) as response:

request = urllib.request.Request(url, headers={'User-Agent': 'LegibleFilm/1.2 sound-library (no authentication)'})

'page_size': a.max, 'duration': 'short'})     try:         payload = json.loads(fetch('https://api.openverse.org/v1/audio/?' + query, 4 * 1024 * 1024))

try:         payload = json.loads(fetch('https://api.openverse.org/v1/audio/?' + query, 4 * 1024 * 1024))     except (OSError, ValueError) as exc:

payload = json.loads(fetch('https://api.openverse.org/v1/audio/?' + query, 4 * 1024 * 1024))

raise ValueError('not a small effect (duration must be <=20 seconds)')             url = item.get('url')             data = fetch(url, 16 * 1024 * 1024)

url = item.get('url')             data = fetch(url, 16 * 1024 * 1024)             identity = hashlib.sha256(str(item.get('id', url)).encode()).hexdigest()[:16]

data = fetch(url, 16 * 1024 * 1024)

path = folder / ('audio.' + extension)                 path.write_bytes(data)             licence_url = ('https://creativecommons.org/publicdomain/zero/1.0/' if code == 'cc0' else

path.write_bytes(data)             licence_url = ('https://creativecommons.org/publicdomain/zero/1.0/' if code == 'cc0' else                            f'https://creativecommons.org/licenses/by/{version}/')

licence_url = ('https://creativecommons.org/publicdomain/zero/1.0/' if code == 'cc0' else

f'https://creativecommons.org/licenses/by/{version}/')

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env python3

Executable file

<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 10 10"><circle cx="5" cy="5" r="4" fill="#3080c0"/></svg>

Executable file

Copyright 2016 The Inter Project Authors (https://github.com/rsms/inter) Inter-Italic[opsz,wght].ttf: Copyright 2016 The Inter Project Authors (https://github.com/rsms/inter)

This Font Software is licensed under the SIL Open Font License, Version 1.1. This license is copied below, and is also available with a FAQ at: http://scripts.sil.org/OFL

http://scripts.sil.org/OFL

#inner { position: absolute; inset: 0; } </style> <script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script>

</style> <script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script> </head>

<script src="https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js"></script>

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

#!/usr/bin/env python3

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]