We are on a mission
to transcribe all music
ever
recorded 🎼

πŸ•š 5 min read

The ecosystem for sheet music is stuck in the 19th century

Over the years we have become frustrated with the state of the sheet music ecosystem. Our guitar teachers had to transcribe the songs we wanted to learn themselves before we could start the actual lesson. Scores that we sought when practicing did not exist, were incomplete, missing the most crucial instruments and solos, or were outright incorrect. Some of the complete scores were paper scans stuck in PDF formats, depriving us of the possibility to play back the score and slow it down on the computer like we have gotten used to with Guitar Pro. Letting computers digitize these scores turned out to be surprisingly difficult. The available digital scores we did rely on disappeared because of website server failures, or because they were taken down. Even official store-bought sheet music was often not true to how the artist would play it.

When composing our own songs and in need for inspiration, there was no way to query existing scores for sections in a particular key, chord progression or any other contextual element. Then, when our bands pooled money together to get transcriptions produced to play covers, no open markets existed at all that allowed spreading the cost of transcribing over a large (unknown) group of people. A perfect transcription costs €900 for an average song. That's too much for individual musicians to afford on a continuous basis. Transcribing music is hard for humans, and even harder for computers. No wonder that qualitative scores are so hard to come by!

What we need...

We've had enough of this. The situation is not improving. After more than 13 years of suffering we have come to this conclusion:
πŸ’‘
What we need is a public internet-wide permanent digital archive of source code to all recorded music that can stand the test of time and is intended to be complete. A planet-owned system that incentivizes every musician on Earth to make the collection perfect and freely available. An immutable Wikipedia-meets-Github for the source code of music.
But it cannot exist yet. More than 91+ million songs exist in recordings. Yet only 0.1% of all music has ever been transcribed into its source code. Only a fraction of the sheet music is digital. And only a fraction of that is machine-readable. More importantly, no economic incentive system exists to fix this. At €900 per song, this is an €81+ billion opportunity, totally overlooked by technologists.

Imagine the possibilities...

Imagine what can be done with the complete dataset of digital scores from all music ever recorded. We could:
‍unambigiously determine a perfect classification of genres, fix cultural & copyright attribution, render the complete tree of music genealogy; make copyright dispute settlement fair and honest, have AI generate new music and new genres on the fly, create next-gen CAPTCHA's and provide developers with API's to the complete source code of music.
  • πŸ“
    mathematically Β perfect genre classifications
  • 🌿
    render the complete tree of music genealogy
  • πŸ’»
    let developers create apps that use the data
  • 🎀
    karaoke backing tracks for any song
  • 🦾
    train AI to generate new music on the fly
  • ©️
    fix cultural & copyright attribution
  • πŸ›
    make copyright settlement fair and honest
  • πŸ€–
    create
    next-gen CAPTCHA's

Our goals

We are visionary chief technologists building an infrastructure layer that can be used by all existing score-editing software and transcription service companies. A protocol to incentivize every musician and music education institution on the planet to blitzscale the work of transcribing the complete collection of Earth music, including your niche artist that no one made guitar tabs for, as quickly as possible, and having fun while doing it. This requires building 3 open-source products:
  • ο“„

    Web collaboration

    A real-time free collaborative digital score editor, community builder and project management system. Supports standard notation, guitar tabs, percussion, Klavarskribo, Chromatic Staff, Hummingbird, Dodeka and PDF printing. A foundational layer that can be used by Noteflight, Flat.io, Sibelius, Jellynote, Soundslice and UltimateGuitar.com to build on.

  • ο”ž

    Incentive layer

    An incentive protocol that aligns interest for our goal better and more profitable than the current publisher-oriented system does. A global reputation system for transcription and validation that economically rewards musicians for using their ears with claims on future platform fee revenue, relative to their contribution.

  • ο•‚

    Permanent immutable storage

    An immutable decentralized public permissionless blockchain that maintains the total content-addressable score graph & recording metadata, tracks all revisions, manages the incentive layer and is resistant against calamities, political unrest and regulatory madness.

What can we do with it?

The system we build revolves around a metadata index of all recordings known to mankind. Participants build reputation in communities around artists and genres. They are incentivized to administer, track and mark which scores, parts and solos are still incorrect or missing. This way, you can tell how far the collection for an artist is on its way to completion. Having this collaborative score editor and project management system doesn't constrain you to only work on scores for recognized records. You are free to compose your own songs, create new compositions with others, or base new work off of parts from other songs. Due to its branching version control system, attribution is perfect and you can always tell where any inspiration came from.

What's in it for me?

Like a library, access to scores is free and for public use. Use of the complete dataset for calculations, research or app development is available for a fee, which subsidizes the free use.
  • 

    for Musicians, Transcribers
    & Archivers

    Access the most accurate score of every song ever recorded, for free, forever. Get paid for every note you transcribe.

  • ο”˜

    for Developers, Academics, AI Researchers, Copyright lawyers, Historians, Music Education Institutes

    Query the complete dataset and all its interrelations using custom logic for a usage-based fee.

  • ο‡Ή

    for Rights Holders

    Earn more with your sheet music than you ever could by selling song books.

How does it work?

The platform works by bringing the future value of the complete dataset to the present using a financial instrument (like a bank). When contributors transcribe songs or missing parts, they are awarded Shares in those songs proportional to the size of the work. When the transcript is used in a search result, the contributor gets paid. It is effectively a collectivized lottery based on transcription submissions. Shares may be traded, giving them a floating value to speculate on. This increases the value of transcriptions beyond their current market value.

Imagine the fun we would have

Imagine we have launched. Suddenly, you can get rewarded for transcribing your favorite artist. Even for the severely unpopular songs you love! Gone are the days of sharing sheet music online and it being forgotten, or worse, you and your skills being forgotten. If you are an established artist earning nothing on the sale of your (non-existent?) sheet music through publishers, you are at an advantage being able to import your own work quickly. Organize yourself in communities around genres and artists to find friends-to-be that feel the same commitment and love for scores you like. You may find yourself being a reputable member that is asked to review scores with your expertise and sharp ears. If you are more financial-minded you can keep track of the share prices for different artists and songs. You may see an interesting opportunity and want to fund the creation of specific transcriptions for their associated shares, or get your friends to do the work and sell you the shares. What a goldrush!

Let's compose history together

To make our dream possible, we are welcoming a lot of new friends, and need a lot of helping hands. Are you excited? Please get in touch if you are:
On our mission, we welcome developers, architects and distributed systems engineers with affinity to music and experience in Rust, OCaml, Go and Elm; cryptographers acquainted with Zero-Knowledge Proofs; blockchain specialists; business developers with contacts to (international) libraries, conservatoria, archives and music education institutions; lawyers and (digital) copyright law specialists; token (economics) engineers; public-interest technologists; Information Scientists; OMR specialists; AI & ML engineers; IFPI insiders, transcription service representatives, and investors.
See you soon!
πŸ’– Get in touch
30/01/2020 - Institute for Partiture Securement Technology