Music Sample Provenance: 14 Real Copyright Cases — and What Documented Sourcing Would Have Changed

The Receipt — A Live Example You Can Verify Right Now

Open this URL in another tab: https://selektaudio.com/audit/454ff65b-b98b-4ad9-9dcd-09a698941d39. You will land on the public audit chain for a real sample in the Selekt catalog. Press the green "Verify Signature" button. The cryptographic check that runs in your browser — no network round-trip to Selekt, no trust required — tells you three things at once: this is the actual record Selekt published, the contents have not been altered since publish, and the signing key checks against the public key Selekt has published on its own server. The same Ed25519 math that secures HTTPS in your browser address bar is what is being used here.

That is the whole pitch in one click. Every sample in the Selekt catalog has one of these. Every download generates a license certificate that points back to one. The pattern is portable: any platform handling audio assets can do the same thing, and the math is the same math.

This post is about why that matters. It is organized around six audiences and fourteen documented, public copyright cases — cases where the affected creators had real legal exposure, real costs, or real distribution headaches because the music in their work did not come with a re-verifiable record. We are not arguing that documented sourcing would have made every dispute disappear; some of these were genuine clearance failures and would have failed any screen. The point is narrower and more important: in every single one of these cases, a tamper-proof, time-stamped, publicly verifiable provenance chain would have changed the timeline, the burden of proof, or the cost of resolution.

What You Are Actually Looking At When You Verify

The technical primitives are deliberately boring. The audit chain is a sequence of signed records, each one Ed25519-signed by Selekt's private key and hash-chained to the previous record so that altering any record retroactively would break every signature downstream of it. When your browser runs the verify, it pulls the public key from Selekt's server, checks each signature, and confirms the hash chain is intact.

The chain itself records substantive events: when the sample was added to the catalog, what its sourcing was (Library of Congress, Internet Archive, Citizen DJ, contributor, etc.), what license terms apply, when it was screened against the ACR Cloud fingerprint database — the industry tool that compares audio against an indexed corpus of commercial recordings to flag matches — what the screening result was, and any subsequent re-screenings or amendments.

Two design choices are load-bearing: the chain is append-only, and the verification key is published openly. Append-only means Selekt can never quietly amend history; the worst we can do is add a contradicting record after the fact, and that contradicting record itself becomes part of the chain. Public key means the people you are trying to prove this to — a label's clearance team, a distributor's compliance department, a sync supervisor, a platform reviewer, a rights-holder representative — do not have to ask Selekt for anything. They just verify.

This is not novel cryptography. It is the same Ed25519 signature scheme used in Tor, Signal, GitHub commit signing, and a substantial fraction of TLS deployments. What is novel — at least for the music sample catalog — is binding it to the per-asset clearance record so that "where did this sample come from" becomes a question with a re-verifiable answer instead of a question whose answer is "trust the platform."

For Music Producers — When Sample Clearance Becomes a Lawsuit

The producer's clearance problem is the oldest one in sampling. From the Vanilla Ice / "Under Pressure" matter resolved with songwriting credit and a reported eight-figure payment that ultimately transferred the publishing of "Ice Ice Baby" to Queen and Bowie's estate, to Nicki Minaj's $450,000 January 2021 settlement with Tracy Chapman over an unreleased interpolation that leaked from a studio session, every well-documented producer-side sampling lawsuit follows the same arc. The infringement happens in a studio. The release happens publicly. Discovery happens months or years later. The evidence is reconstructed from emails, witness statements, and forensic listening, because nothing about the original session was structured as a verifiable, machine-readable record.

A signed audit chain does not undo what was sampled. If you sample a copyrighted recording without clearance, your chain will document exactly that. What changes is the timeline. With a chain, the screen against the recorded-music fingerprint database — the same kind of comparison that powers YouTube's Content ID — happens before release, not after a lawsuit. Matches get flagged at the demo stage, when removing them is a creative decision worth a few hours of session work. After release, removing them is a lawsuit worth six figures.

The Minaj case is particularly instructive on the timing point. Minaj's team sought clearance to interpolate Chapman's song multiple times in July 2018; the requests were denied; the track was recorded anyway, ultimately left off the album Queen, and then leaked through a radio station. Once that leak happened, the boundary between "shelved session work" and "released material" was a question of trust and email screenshots. A signed log of when clearance was requested, when it was denied, and when the bounce was archived as "do not ship" would not have made the interpolation legal — but it would have made the timeline an evidentiary fact rather than a litigated narrative. For the 99% of producer sessions that never end in leaks, the same chain is the difference between answering a label's chain-of-title question in an afternoon and reconstructing it from old emails a year later.

The De La Soul case is the producer-side variant that does not involve a lawsuit at all, and may be the most instructive of the three. De La Soul's first six albums — released between 1989 and 2001 — were absent from streaming services for more than a decade. Not because the samples on them were infringing, but because the original 1989-era clearance paperwork covered physical and radio formats only. When Reservoir Media acquired the rights in 2021, DMG Clearances spent roughly a year manually re-identifying and re-licensing the samples so the catalog could land on Spotify, Apple Music, and Tidal in March 2023. The legal problem was always solvable. The provenance documentation was not there to solve it efficiently.

A signed manifest of which underlying recordings appear in which tracks, with timestamps and license terms, turns "what is actually in this album?" from a year-long forensic project into a database lookup. That is true whether you are the producer trying to release the album in 2023, the label trying to license it for film placement in 2030, or the artist trying to license the master back to your own catalog in 2045. The receipt outlasts the people, the labels, and the formats.

This is exactly what Selekt builds for producers. Every sample in the Selekt catalog is sourced from public-domain or CC0 material, pre-screened against the ACR Cloud commercial-music fingerprint database before it ever appears in the catalog, and bound to an Ed25519-signed audit chain you can hand to a label, a sync supervisor, or a distributor without needing to ask Selekt for anything. The cert URL is on the sample page. The verification button is on the audit page. The chain is built into every download. A producer using Selekt does not have an after-the-fact clearance project waiting for them in 2030 — the receipts are already filed.

For Composers — When Original Work Still Gets Sued

The composer's risk profile is different from the producer's. Composers rarely sample. Their copyright exposure usually comes from accusations of substantial similarity to existing works — a chord progression, a melodic phrase, a tonal palette someone else thinks they recognize. Sometimes those accusations are right. Often they are not. Either way, the cost of defending against one is substantial.

Hans Zimmer has been on both sides of this. In April 2006, the estate of Gustav Holst and music publisher G. Schirmer alleged that Zimmer's "The Battle" from the Gladiator score borrowed from "Mars, the Bringer of War" — the opening movement of Holst's The Planets — without permission. The matter was settled privately on undisclosed terms, with Zimmer's liner notes acknowledging shared "language" and "vocabulary." Nine years later, in 2015, composer Richard Friedman sued Zimmer alleging that the principal theme of 12 Years a Slave copied his 2004 piece "To Our Fallen." Friedman voluntarily dismissed his claims after evidence "resoundingly disproved" them, and issued a written apology.

The Friedman case is the cleaner illustration of why documented session history matters. The defending composer won, but only after months of expert testimony, deposition, and discovery aimed at proving independent creation. Independent creation is the standard legal defense against substantial-similarity claims, but proving it in court is expensive precisely because most composing happens in private, in a DAW session, with no third-party-verifiable record of what reference material was loaded, what was sketched, when, in what order.

A signed creation log changes this. Not a copyright deposit — those exist already and do not help here. A live, time-stamped audit chain that records, at session time: this is the project, here are the reference tracks that were loaded into the session, here are the sketches that came out of it, here is when the final mix was bounced. Years later, when a substantial-similarity claim shows up, the chain is something a third party can re-verify in their browser. It does not decide whether the resemblance rises to infringement — that is still for the court — but it converts "did the composer hear my piece?" from a litigated question with no good answer into a record. For the composer, the same chain is a professional-liability hedge made of math.

For composers scoring film, TV, advertising, games, or library catalogs, Selekt is what makes the underlying material safe to load into the session in the first place. Every reference cue, every found-sound texture, every public-domain stem you pull from Selekt arrives with a signed provenance chain. The cue sheet you hand the music supervisor can carry that chain. If a substantial-similarity claim ever lands, your "I sourced this from a public-domain catalog with an Ed25519-signed audit record" is no longer a verbal claim — it is a URL anyone can click and verify. That is the artifact Friedman's case took months to construct from scratch.

For Game Developers — When Music Licenses Expire and Games Get Delisted

The game industry has a music licensing problem that producers and composers do not have, and it is the opposite shape. The music in a game is almost always licensed correctly at ship time. What kills game soundtracks is what happens to those licenses years after release.

Alan Wake — Remedy's 2010 atmospheric thriller — was pulled from Steam, Xbox Live, and other digital storefronts on May 16, 2017, when its seven-year music licenses expired. The game stayed delisted until Microsoft renegotiated the licenses and the title returned to sale in October 2018. In September 2024, a further update removed David Bowie's "Space Oddity" from the game entirely, replacing it with an original cue by composer Petri Alanko. The game has now shipped with three different soundtracks over fourteen years — not because the developers wanted that, but because the licensing structure underneath each release demanded it.

The Grand Theft Auto series is a parallel case at scale. The 2014 Steam update for GTA: San Andreas removed 17 licensed songs whose deals had expired — including tracks by 2Pac, Rage Against the Machine, and Tom Petty. The 2021 GTA Trilogy: The Definitive Edition retained those absences and added new ones. Anyone who bought the original CDs in 2004 and the Definitive Edition in 2021 is playing two different games, sonically, because the asset-to-license mapping was not structured as durable, queryable metadata bound to each track.

Forza Horizon 4 — Playground Games' 2018 racer — was delisted from sale on Steam and the Microsoft Store on December 15, 2024 for the same reason. Existing owners can still play, but the title is no longer purchasable. A six-year-old game came off the shelf because the licensing clock ran out.

A signed audit chain does not make a Tom Petty sync cheaper. It does not make a Bowie estate negotiation go away. What it does is make the licensing structure of a soundtrack legible. Every audio asset in the build is bound, at ship time, to its license terms and expiry date. Six months before a license lapses, the system surfaces it. Renewal decisions become per-track P&L exercises, not all-or-nothing delisting threats. For the substantial portion of a game soundtrack that comes from indie composers, original score, library music, or public-domain sourced material — assets with no expiry — the chain proves it. Those tracks never end up in the "do we re-license or cut?" pile. The game stays on sale with a trimmed playlist instead of getting pulled.

Selekt is the upstream provider for that "no expiry" portion of a game soundtrack. Every Selekt sample is sourced from public-domain or CC0 material with a signed audit chain documenting the source, the screening result, and the license terms. Use a Selekt-sourced loop in a level cue, ship the game in 2026, and the asset still has the same provenance chain in 2046 — verifiable in a browser, against the same published public key. For studios planning around remasters, re-releases, and decade-long ports, that is the part of the soundtrack that does not need a renewal conversation. Ever.

For App Developers — When the Platform Music Deal Disappears

The app developer's music problem is the cliff edge. You build a feature that depends on music — a fitness platform that schedules tracks to workout intensity, a karaoke app, a creator tool that lets users add background audio. The music in that feature is licensed through a platform-side deal with major labels or publishers. The deal expires, renegotiates, or collapses. The next morning, you have a feature with no music.

The TikTok / Universal Music Group standoff in early 2024 is the textbook example. UMG's licensing agreement with TikTok lapsed on January 31, 2024. The next day, roughly three million UMG-controlled recordings became unlicensed on the platform. The publishing catalog followed on February 27. Videos by Taylor Swift, Olivia Rodrigo, Billie Eilish — and millions of smaller creators whose videos happened to include any of the affected songs — were muted overnight. A new deal was struck in early May 2024 and audio was restored over the following weeks. The platform handled the dispute. The creators experienced the dispute.

Peloton's 2019 lawsuit by NMPA-member publishers — initially $150 million in claimed damages, raised to $300 million after discovery — settled in February 2020 with undisclosed terms and a "joint collaboration agreement" on licensing systems. The damages claim grew during discovery because Peloton's catalog of music uses turned out to be larger than even the publishers initially knew. Triller's run-in with the major labels in 2022-2023 had the same shape: a platform with thousands of audio assets, no per-asset license manifest, and a major-label deal that deteriorated.

The audit-chain pattern does not replace platform-side licensing deals. What it does is give every individual creator on the platform — and the platform itself — a portable, per-asset record of which audio came from where. When a major-label deal collapses, the audio with documented non-major-label provenance is the audio that should never have been muted in the first place. With a signed chain, telling those two pools apart is a database query, not a discovery exercise. For the platform, it is a way to honor the deal and minimize creator harm at the same time. For the creator, it is the difference between losing a year of content overnight and losing only the small subset of content that genuinely depended on the lapsed catalog.

This is exactly the role Selekt plays for app developers. Build your music feature on a Selekt-backed catalog and every track your users hear arrives with a signed audit chain proving it is public-domain or CC0 sourced — not major-label catalog. When the next platform-vs-publisher standoff happens (and there will be a next one), the audio in your app is on the right side of the line, with the evidence already attached. For the app developer, that means your feature does not go dark at midnight on the day a deal expires. For your users, it means their workouts, karaoke runs, and creator-tool posts do not get muted in someone else's dispute.

For Podcasters — When the Music Licensing Gap Catches Up

Podcasting has a music problem that most podcasters do not know they have until it is too late. The blanket licenses from ASCAP, BMI, and SESAC that cover public performance for radio, restaurants, and live venues do not cover podcast use. Sync rights — the right to combine music with spoken-word audio in a recorded program — are a separate, per-use license that has to be cleared directly with the rightsholders. Most podcasters discover this when a takedown lands.

In February 2025, the National Music Publishers' Association launched what it called an "extensive takedown action" against Spotify podcasts using unlicensed music. The initial action covered over 2,500 detected uses of NMPA-member works. Spotify dismissed it publicly as a "press stunt"; the takedowns proceeded anyway. The same trade group has been pursuing similar actions against other podcast platforms intermittently for years. The first major podcast-music suit was filed in 2018 against a podcast operator that had been approached repeatedly by a major-label publisher about music use and continued using the music regardless.

A signed audit chain is the lowest-cost intervention in podcasting. Before publish, every cue in the episode gets screened against the recorded-music fingerprint database. The chain records which cues passed, which cues were swapped out, and the underlying sourcing of every retained piece of audio. When a takedown arrives — and at the trade-group volumes recently announced, takedowns will arrive — the response is a re-verifiable record rather than a frantic catalog audit.

For podcasters who do license correctly, the chain is the receipt the licensor probably has not asked for yet but will eventually want. For podcasters who source music from cleared catalogs to begin with, the chain is what makes that sourcing portable across hosting platforms, distribution networks, ad insertions, and any future format the show ends up syndicated to. The audio outlives the platform; the receipt has to outlive it too.

For podcasters, Selekt is built to be that source. Every theme cue, bumper, bed, transition, and sound effect from the Selekt catalog ships with its own audit chain — sourced from public-domain or CC0 material, pre-screened against the recorded-music fingerprint database, signed and timestamped on the way in. When an NMPA-style takedown notice lands on a Spotify-hosted episode in 2026 or 2027, the response is not a panicked audit through a year of show files; it is a list of URLs, each one already re-verifiable by the rightsholder. The cost of preparing that response with Selekt is essentially zero — the receipts were filed at download time.

For Library and Production Music Composers — When Submission Becomes Rejection

Production music libraries operate on a simple risk model: the library cannot resell a track unless it can prove to its sync customers that the underlying ownership is one hundred percent clean. Master rights, publishing rights, every sample, every loop, every breath of audio in the track. When a composer submits a track and the library cannot verify ownership of every element, the safe default is rejection. Most library composers learn this through the rejection slip.

The plagiarism risk runs in the other direction too. Mood Music's 1967 objection to De Wolfe Music over the composition "Girl In The Dark" is one of the oldest documented library-on-library plagiarism disputes — two production music libraries arguing about which one came up with the same musical phrase first. Without a shared, neutral, time-stamped record of when each composition was written, the question became an evidentiary one. Sixty years later, the structural problem is the same.

Epidemic Sound's litigation against Meta — a $142 million suit filed in 2022 over Facebook and Instagram's use of its catalog, with a second suit filed in December 2025 covering an additional thousand works — illustrates the platform-scale version. When platform-side audio inventories are not built on per-asset provenance records, library compositions get duplicated, reuploaded, and re-served without the license chain attaching. The platform does not know which audio is the licensed copy and which is the reupload. The library does not know which uses it is being paid for and which it is not. Discovery has to construct the answer from scratch.

A signed creation manifest, generated at session time and re-verifiable for the life of the work, addresses all three of these. For composer-submission rejections, the library can verify the composer's ownership claim in seconds instead of taking it on faith. For library-on-library disputes, the timestamp settles the priority question. For platform-scale licensing, the license chain travels with the audio, on any platform, for any reuse. The same primitive solves all three problems because all three are the same problem — provenance without a third-party-verifiable record.

Library and production music composers using Selekt as their source material get exactly that primitive, for free, on every download. Build your library cue from Selekt-sourced public-domain stems and the submission package you hand the sync agency includes the signed provenance chain for every underlying element. The library does not have to take your ownership claim on faith — they can verify it in seconds. Your cue clears their internal screening on the first pass instead of getting bounced back as "unverifiable." That is the difference between getting accepted on submission day and getting rejected on suspicion.

The Pattern Across All Fourteen Cases

Fourteen cases. Six audiences. The common thread is not that any of them would have been "cleared" by a signed audit chain — that word means something specific and we are being careful not to overstate. The common thread is what changes about the timeline and the burden of proof.

For the cases that were genuine clearance failures — where uncleared material was knowingly used — the audit chain moves the discovery moment from years after release back to before release. The same fingerprint check, run at the demo stage, would have flagged the match. The choice that follows the flag is creative, not legal. Most producers, given that flag at the demo stage, would have picked a different sample.

For the cases that were not genuine clearance failures — the composer who was sued and won, the library composer who was rejected on suspicion rather than evidence, the creator caught in a platform-side dispute they had nothing to do with — the audit chain moves the burden of proof from "construct the timeline from scratch" to "re-verify the existing record." That changes a months-long discovery process into a minutes-long lookup.

For the cases that were licensing-clock problems — Alan Wake's expiring syncs, GTA's deleted songs, Forza Horizon 4's delisting, TikTok's UMG standoff, Triller's major-label suits — the audit chain converts an overnight cliff into a known, scheduled event with months of lead time. The legal structure of expiring licenses does not change. What changes is whether the engineering team, the legal team, and the platform partner are looking at the same data.

Selekt's Solution, Audience by Audience

Across every audience in this post, the underlying offer from Selekt is the same: a catalog of 100,000+ public-domain and CC0 samples, each one pre-screened against the ACR Cloud commercial-music fingerprint database, each one bound to an Ed25519-signed audit chain you can re-verify in any browser. What changes per audience is how that core gets used in practice.

For producers: drop a Selekt sample into your beat, hand the cert URL to your label or distributor, and the chain-of-title question is answered before they ask it. No clearance phone calls. No reconstructing emails years after release. The screening, the signature, and the receipt were already there at download time.

For composers: source your reference material and your bed-track stems from Selekt. The audit chain becomes a documented session-side record of what you loaded, what you used, and what was public-domain underneath it. When a substantial-similarity claim lands — even a meritless one — you respond with URLs instead of months of discovery.

For game developers: ship Selekt-sourced cues for the portion of your soundtrack you do not want to renegotiate every seven years. Selekt assets travel with their provenance for the life of the work. Re-releases, ports, and remasters do not lose audio, because the chain has no expiry — only the major-label sync deals do.

For app developers: build the audio inventory in your platform on Selekt-backed material and the next major-label / platform standoff stops being your problem. Your users' content stays unmuted because the audio is provably non-major-label, with a re-verifiable record attached to every track.

For podcasters: cue music, beds, transitions, and stings from the Selekt catalog. Every episode you publish carries a signed manifest of every musical element in it. When the next trade-group takedown action sweeps through your hosting platform, you have URLs to show, not boxes of files to audit.

For library and production music composers: build your submissions on Selekt-sourced underlying material and the libraries you submit to can verify your ownership claim in seconds. Acceptance rates rise because the rejection-by-default risk model has nothing to push back on — the receipts are already in the file.

In every one of these cases, the difference is the same: Selekt is not a "trust us" catalog with a license PDF stapled to the end. Selekt is a catalog where every download is a piece of cryptographic evidence anyone in your professional chain can re-verify, today, in their own browser, without ever having to come back to Selekt to do it. That is the part of this that is genuinely new — and it is the part this entire post is about.

What Documented Sourcing Is Not

We have been careful all the way through this post not to say a Selekt audit chain "clears" anything, "guarantees" anything, or makes copyright disputes go away. None of those things are true, and we will not pretend they are.

The fingerprint database Selekt screens against — ACR Cloud — is the same kind of system that powers Content ID on YouTube. It is excellent at flagging matches against indexed commercial recordings. It is not one hundred percent comprehensive. There are recordings it does not index. There are derivative works it may miss. The honest figure on coverage is in the high nineties; the honest characterization is "the screening every serious reviewer wants to see," not "the screening that catches everything."

The audit chain is also not a legal opinion. It documents what was done — when, how, with what result. It does not constitute advice that any particular use is non-infringing. A use of a public-domain recording is overwhelmingly likely to be safe under U.S. copyright law, and the chain documents that the sourcing is from the public domain. Whether the specific use crosses into a derivative-work problem is a question for a lawyer, not a fingerprint scan.

What the chain is, precisely: a tamper-proof, time-stamped, publicly re-verifiable record of how a particular piece of audio was sourced and screened. That record is evidence — the kind of evidence that distributors, sync supervisors, platform reviewers, and rights-holder representatives actually want to see when they are trying to decide whether to take you on faith.

How Every Sample on Selekt Has One of These

Every sample in the Selekt catalog goes through the same pipeline. The audio is ingested from a public-domain or CC0 source — Library of Congress, Internet Archive, Citizen DJ, Freesound, Musopen, and similar — with its source URL and license terms recorded. It is then screened against the ACR Cloud fingerprint database to confirm no major-label catalog match. The screening result, the source URL, the license terms, and a content hash of the audio file get assembled into a signed audit record. The record is hashed onto the chain.

Every sample page on Selekt has a "Verifiable cert link" button that copies the public URL anyone can open. Every download generates a PDF certificate that includes the chain URL. Every cert URL is indexable, shareable, and re-verifiable by anyone with a browser.

The chain you can verify right now at https://selektaudio.com/audit/454ff65b-b98b-4ad9-9dcd-09a698941d39 is one of those. It is the audit chain for a real sample in the Selekt catalog. Click the green "Verify Signature" button on that page, and your browser runs the Ed25519 check entirely client-side, against Selekt's published public key, with no round-trip to Selekt's servers during the check. The math is the same math that secures HTTPS in your browser's address bar.

For producers, composers, game and app developers, podcasters, and library composers, this is what changes: you do not have to ask anyone to take your sourcing on faith. The receipt is built in.