Credits & Attribution
CABAL Crimean Campaign · Last updated: 15 August 2026
CABAL is built on open data and open-source software. This page records the third parties whose work makes the Crimean campaign possible.
Creator
CABAL was created by Nell Watson (github.com/NellInc).
With thanks for instrumental research and engineering assistance from Claude (Anthropic) and Codex (OpenAI), to Filip Alimpić for playtesting and product management, and to the Engineer and Logistic Staff Corps RE (formerly the Engineer and Railway Staff Corps) for inspiration.
Terrain — Copernicus DEM
The battlefield terrain is derived from the Copernicus GLO-30 Global Digital Elevation Model (COP30), cropped to the Sebastopol–Balaklava area (44.44°N–44.82°N, 33.34°E–33.86°E) at native 1 arc-second resolution. Redistribution of the derivative heightmap is permitted under the Copernicus DEM Licence.
© DLR e.V. 2010–2014 and © Airbus Defence and Space GmbH 2014–2018, provided under COPERNICUS by the European Union and ESA; all rights reserved.
The modern DEM reflects post-1855 terrain. Some features (for example the Chornaya reservoir, built after the Crimean War, and minor modern cuts and fills near Sebastopol and Balaklava) are noted rather than corrected.
Original media
The Campaign Briefing film was produced for CABAL by Nell Watson using Fable. Its English captions were authored for the film and are available as an optional track.
The Crimean campaign’s telegraph, cabinet, railway, writing, and clock cues are original procedural sound designs generated for CABAL. They evoke period machinery and working environments; they are compositions rather than archival recordings.
Terrain movement, artillery handling, railway machinery, harbour work, weather, camp, and medical details combine permissively sourced recordings with deterministic sound design. Complete drum and field-call performances come from public-domain United States Navy, Army, and Marine Corps recordings documented in the project source ledger. These functional game mappings do not claim that every Crimean belligerent used the exact United States calls.
Original soundtrack
The Lesson Came Before is a ten-song, non-diegetic soundtrack developed for CABAL. The album carries writing and arrangement credits to Nell Watson, Sol 5.6, and Suno 5.5. The selected performances and their original cover artwork were rendered with Suno v5.5. The songs are original compositions rather than archival music.
The album looks back on the Crimean conflict through the later Staff Corps lesson. The Staff Corps formed in 1865 and is not represented as an organisation active in the 1854 to 1855 campaign. Canonical lyric sheets always retain Corps. Generation-only lyric fields use Core where needed to secure the correct sung pronunciation.
Tracks 2 to 7 and 9 to 10 use their stored Suno word timing. The replacement Track 1 uses local machine alignment bound to its selected master. Track 8 combines its stored Suno timing with a disclosed local timing patch for three restored words. CABAL maps every timing source onto the canonical lyric sheets, records interpolated words explicitly, and never substitutes generation text into the displayed lyrics.
The soundtrack manifest records source identifiers, payload hashes, technical measurements, and the applicable terms snapshot. On 1 September 2026, Nell Watson authorized this exact project-owned soundtrack candidate for repository publication and public noncommercial CABAL deployment under the project’s CC BY-NC-SA 4.0 licence. Public builds retain the campaign score, Staff Corps masters, artwork, timed lyrics, and player, then verify them against a hash-bound authorization receipt.
AI-generated voices
The multilingual unit, nursing, engineering, logistics, and Staff Corps voices are AI-generated performances, rather than recordings of human voice actors. Their transcripts and English translations were authored for CABAL, then rendered as fixed and reviewable audio files. Captions remain available independently of campaign sound.
The current 322-clip technical candidate was rendered with Gemini 3.1 Flash TTS. CABAL records the provider, model, voice, transcript, source receipt, and asset hash for every clip. All 322 files in Candidate B.3 are Gemini renders. The authoring system can use OpenAI text to speech as an explicit fallback, with that provenance recorded per clip, but no fallback is present in the current candidate.
Public builds derive a 226-clip accepted baseline from the 322-clip technical candidate. The other 96 clips remain in the private/source review candidate and are excluded from the public catalogue and audio payload until their listening, language, context, rights, and release fields are accepted. A deterministic public receipt binds the published catalogue and asset hashes to the reviewed source candidate without publishing reviewer identities or notes.
Recorded sound and faction camp repertoires
CABAL’s complete recording ledger, source hashes, derivative method, and publication
boundary are documented in docs/CRIMEAN_SPATIAL_AUDIO_SOURCES.md and
docs/CRIMEAN_AUDIO_EXPANSION_SOURCES.md. Field calls and camp music remain
continuous performances. The builders use no MIDI, extracted notes, or sequenced
reconstruction.
The French Réveil comes from LaSonothèque sound 3263 by Joseph Sardin. Its source page marks it as a real, mono, indoor, 24-bit/48 kHz recording released under CC0. The reviewed game excerpt is included in the public faction-signal bank.
The selected British and French camp banks include credited performances of Last Post and The British Grenadiers. Their exact source and attribution records, along with the Canadian Forces Rouse and the Dordogne conservatoire performances, remain in the project ledger. Nell Watson authorized the ten reviewed field-music and faction-signal derivatives for this noncommercial public CABAL release. CABAL preserves each external performance credit and makes no ownership claim over those third-party recordings.
These routes represent faction camp repertoire. Some recordings are later performances of period compositions. CABAL does not claim that every excerpt was a tactical signal or that the recorded ensemble performed in the Crimean campaign.
Feature positions
Feature coordinates (bastions, redoubts, camps, and the like) derive from a mix of Wikidata P625 points, Wikipedia infobox coordinates, OSM/Mapcarta, and inference from battle narratives. Each feature’s positional precision is tagged in the source as high (<50 m), medium (~50–200 m), or low (200–500 m inference).
Historical cartography
Historical map layers and traced siege works use resized or cropped derivatives of these collection images:
- Charles Alexandre Fay, Map of port of Balaklava and route to Sevastopol. Wikimedia Commons records the image as public domain and free of known copyright restrictions.
- Eugene James Vaughan and James Wyld, Sevastopol, sketch of the Mamelon and new Russian defences, 1855. Credit: Library of Congress, Geography and Map Division.
- W. & A.K. Johnston, Johnstons’ Plan of the Siege of Sevastopol, 1855. Credit: David Rumsey Map Collection, David Rumsey Map Center, Stanford University Libraries.
These period sheets are interpreted sources. Their imagery is warped onto modern elevation data, and their disagreements are preserved as uncertainty rather than presented as exact survey truth.
Historical portraits
Cabinet likenesses use nineteenth-century photographs and prints. Roger Fenton’s 1855 Crimean portraits of John Fox Burgoyne, William Filder, Lord Raglan, James Yorke Scarlett, and James Simpson are public-domain works sourced through the Library of Congress, the Imperial War Museums, or their Wikimedia Commons records. The interface crops these images for medallion framing.
The portrait of Sir John Hall comes from the Wellcome Collection, licensed under CC BY 4.0. The Richard Airey and Richard Dacres likenesses are public-domain nineteenth-century portraits recorded by Wikimedia Commons and the National Portrait Gallery.
Open-source software
CABAL directly uses the following production packages, each under its own declared licence. Exact versions come from the release lockfile:
- @google/generative-ai@0.24.1 — Apache-2.0
- @mlc-ai/web-llm@0.2.84 — Apache-2.0
- @mlc-ai/web-tokenizers@0.1.6 — Apache-2.0
- @react-three/drei@10.7.8 — MIT
- @react-three/fiber@9.7.0 — MIT
- @react-three/postprocessing@3.0.5 — MIT
- d3-array@3.2.4 — ISC
- d3-scale@4.0.2 — ISC
- express@5.2.1 — MIT
- express-rate-limit@8.6.2 — MIT
- helmet@8.3.0 — MIT
- lucide-react@1.31.0 — ISC
- miniplex@2.0.0 — MIT
- pino@10.3.1 — MIT
- pino-http@11.0.0 — MIT
- postprocessing@6.39.4 — Zlib
- react@19.2.8 — MIT
- react-dom@19.2.8 — MIT
- three@0.185.1 — MIT
- three-stdlib@2.36.1 — MIT
- troika-three-text@0.52.5 — MIT
- typewriter-effect@2.22.0 — MIT
- uuid@14.0.1 — MIT
- ws@8.21.3 — MIT
- yuka@0.7.8 — MIT
- zod@4.4.3 — MIT
- zustand@5.0.15 — MIT
A generated dependency and declared-licence inventory ships with the project source. Package metadata is machine evidence rather than legal advice. CABAL itself is released under CC BY-NC-SA 4.0; see the Terms of Service.
← Return to the campaign