Fictional capstone: The Last Bus Home

All names, works, amounts, and events are fictional. Do not place real private data in a public submission.

Scenario

A writer-producer develops a three-song EP appearing across a three-part short series. A performer, beat maker, and editor may join. A community media center offers equipment access. No buyer, platform, or collaborator is attached.

Assignment

  1. Write a charter separating existing facts from proposals.
  2. Map each song version to an episode.
  3. Draft a $5,000 illustrative budget; label every number “exercise estimate.”
  4. Build survival, working, and expansion cases; use $0 for unsigned licensing, grants, and acquisition.
  5. Compare buying a $1,200 camera with rental and media-center access.
  6. Complete fictional contribution, split, rights-envelope, and delivery templates.
  7. Identify composition, master, performer, artwork, privacy, and AI questions.
  8. Create a public-safe page and private evidence index.
  9. Write an exit plan from a fictional video platform.
  10. List five professional-review questions and five producer-controlled decisions.

Label every claim: verified within the fictional packet, proposal, estimate, interpretation, or unresolved.

AI-assisted release decision

Use this practice in Module 6 before completing the delivery checklist. Everything below—including tools, terms, dates, recordings, and permissions—is synthetic. No audio or real account is needed. The written observation cards are the equivalent text-based path; they describe imagined listening, not a test of real systems.

Selected reading

Read Sections 11.2–11.4 (human evaluation, listening-test design, and reporting limits) and 12, especially its opening discussion and 12.2–12.3 (memorization, watermarking, and detection), in Modrzejewski, Kurkowska, and Herremans, A Tutorial on Generative Models for Music Audio, arXiv v1, submitted October 28, 2026. This is an arXiv preprint; see the source record.

Use Section 11 to ask what an evaluation actually measures and how listening conditions affect a comparison. Use Section 12 to separate training-source documentation, permission, reproduction, and output identification. An automated score, detector result, or watermark is not release permission. Broad training-source descriptions do not establish creator consent or compensation. This classroom comparison is not a controlled study or a ranking of tools.

For an actual release, check current first-party model documentation and the terms applicable to the exact tool version, account plan, inputs, and intended use. Record the source and date checked; this paper and the fictional packet cannot authorize a release.

Same brief, two candidates

Choose a 30-second instrumental bed for the series trailer, aimed at young adult viewers. It should feel warm but uneasy, leave dialogue intelligible, build gently, and end cleanly at 30 seconds. Compare audience/mood fit, dialogue space, structure/ending, and audible artifacts. Record specific observations rather than a single quality score.

Written observation card Candidate A: Night Route Candidate B: Window Seat
Audience and mood Warm keys over an uneasy pulse fit the late-night scene Bright, playful guitar feels less uneasy
Dialogue space Sparse middle leaves the supplied dialogue easy to follow Busy lead competes with dialogue between 12 and 20 seconds
Structure and ending Gentle build; clean stop at 30 seconds Gentle build; clean stop at 30 seconds
Artifacts No artifacts described in this synthetic observation A brief metallic flutter is described at 22 seconds

Synthetic evidence packet

All records below were checked on November 3, 2026 within this fictional exercise. Their identifiers are source references, not live documentation links. Neither candidate samples a real artist or claims a real person's endorsement.

Record Candidate A Candidate B
Tool and version A1: fictional Lantern Studio 2.1, paid Creator plan B1: fictional Cedar Audio 1.4, paid Producer plan
Training-source disclosure A1 model card, October 20, 2026: “partner collections and web-collected audio”; no itemized permissions, credits, or compensation records B1 model card, October 21, 2026: “licensed commissioned library”; no underlying agreements supplied
Output-use terms A2 terms, October 22, 2026: commercial output use permitted on this plan, subject to the user's authority over uploaded inputs; no third-party clearance provided B2 terms, October 23, 2026: commercial output use permitted on this plan, subject to input permissions; no third-party clearance provided
Input and rights record A3 contribution record, October 24, 2026: fictional performer Q's texture recording was uploaded. Signed permission covers private workshop testing only. No trailer/public-release authorization supplied B3 contribution and rights records, October 25, 2026: fictional team owns the composition and recording inputs; signed contributor permission includes commercial trailer use and processing with Cedar Audio 1.4
Credit and compensation A3: workshop fee recorded; public-use fee and credit remain unagreed B3: agreed credit and fixed fee recorded as paid for the stated use

Keep the model-card claims separate from output-use terms. Neither model card proves that every training contributor consented or was compensated. Candidate A's specific hold concerns its uploaded performer recording, independently of that uncertainty. Candidate B's packet supports a limited classroom decision, not a general legal guarantee about the tool.

Your submission

Add this copyable record to the existing rights envelope's Evidence section and link it from your delivery checklist. Use synthetic records only. There is no additional account, upload, or PIXIE schema required.

Done means you have made both judgments, referenced the records, and named what would resolve any hold. Passing the learning exercise does not clear a track for release. Use Check your work.

Promotional campaign example: a story reference is not lyric clearance

Optional discussion for Module 6; this does not add points or change the existing rubric.

Scenario. A creator promotes a release through a short browser story, as in the review work for Sewers & Shadows: A Night in Derry. The creator supplies a lyric transcription, video references and merchandise. Violet is a separate character cosplaying Pennywise, not the same character. These facts do not establish rights clearance, artist participation or endorsement.

Compare three proposed uses:

Proposed campaign use Evidence to record Defensible next step
Link to a video, or offer a click-to-load provider embed without autoplay Source URL/uploader, date checked, embedding availability, applicable terms, third-party data/accessibility implications Verify the destination and embedding conditions; retain a source-link alternative. Embedding does not clear a separate lyric adaptation.
Quote a short passage to discuss that specific passage Exact purpose, necessary excerpt, amount and significance, attribution/source, applicable law and effect on the original market Assess the particular use; reduce the excerpt or seek permission/professional advice when unresolved. There is no universally safe word count.
Turn lyrics into scene atmosphere, character dialogue or repeated game responses Proposed adaptation, composition/lyric rightsholders, exact permission scope; recording rights if audio is used Hold that adaptation until the necessary evidence is reviewed; use original prose meanwhile. A public demo is still a public use.

Worked decision. Proceed with original story prose and responses. Keep lyric transcription out of the build. Offer source links and only conditionally load embeds after a deliberate player action, subject to verified provider conditions. Hold lyric-derived dialogue pending permission or qualified review. Do not label the whole campaign “fair use” simply because it is promotional, attributed, noncommercial, or “demo only.”

For a U.S. fair-use analysis, examine purpose/character, nature of the work, amount/substantiality, and market effect. Other jurisdictions may apply different exceptions. These are assessment questions, not a legal opinion or a clearance certificate. See the U.S. Copyright Office fair-use overview and record the actual law and advice relevant to the release.

Check your work. Can you distinguish linking/embedding from copying lyrics and from using a recording? Identify what is documented, what remains unknown, and the evidence that would resolve the hold. A visible printed quotation in a product photograph is not blanket permission to turn those words into story dialogue. Buying merchandise does not grant music or adaptation rights. Purchases, sign-in and shop visits must stay outside story progression.

Disclosure example. “This independent story uses original drafted prose and scripted game responses. Optional source videos load only on request and are subject to the host's terms and data practices. No song lyrics or downloaded music are included in the story. Product photographs may display printed artwork; no artist affiliation, endorsement or adaptation permission is implied.” Use this only when it accurately describes the actual campaign build.

Evidence limit, 4 November 2026. This is an implementation/review example, not a legally cleared or independently accepted release. Owner-supplied video references are eBwec9gtuCI and H1wA4kTFC3g; official uploader status, playback, captions and embedding availability remain unverified in this review. The supplied transcript is not reproduced here; advertisement material is excluded. Distinguish intended embed behavior, automated local checks and a verified public experience.

Extend the Module 5/6 rights work with Who Owns the Continuation?, a proposed AI-music use with real platform-source checks and unresolved rights. This separate case is not part of the synthetic candidate packet above. Complete its eight evidence records in the existing rights envelope and delivery package. Use its worked HOLD example and the unchanged 24-point rubric; no audio, upload or artist approval is required or asserted.

Optional stewardship and enterprise records

Extend the same rights envelope with BRAND-001 or ENTERPRISE-001. Separate legal ownership, proposed economics, program reporting and evidence provenance. Use the portable synthetic sample to inspect a PDF/JSON/CSV package without submitting private data. The existing rubric and stop conditions remain unchanged; no personal benefit determination is made.

Applied production-budget example: Mayhem

This real-project discussion is separate from the fictional packet above. Read the Mayhem production-cost comparison, an owner-review draft based on a supplied summary rather than independently inspected source files. The link becomes available publicly only after owner-approved publication.

For your fictional EP’s artwork, merchandise and physical edition, separate three budget lines: quote preparation, production-art labor, and physical proofing. Mayhem’s founder reports an AI scope in under 15 minutes; its 24 / 48 / 60-hour production options are planning estimates, not commissioned work or confirmed delivery dates. Vendor intake and refund correspondence does not establish an equivalent itemized assessment.

Record what each stage gives you, cash already paid versus proposed costs, the evidence still needed, and the likely correction work. Keep USD and GBP separate. Do not assign the separate Veiled Dominion OpenAI spend to Mayhem or calculate savings from unknown founder hours.

Decide what you would pay for next and why: a specialist’s source-file inspection, specified artwork variants, embroidery digitizing, or a physical proof. Explain what would change that decision. Test the mark at card-index size and on the intended material; concept and placement feedback alone is not print approval. Keep private evidence in your private index and use aggregate figures in the review copy.

Download Markdown source · Presentation does not indicate legal review or clearance. See the document's stated status.

PIXIE