Project Title: AI Based Arpeggio Plugin

Description: 

Questions We Want to Explore:
How precisely can we tailor training material for machine learning and artificial neural networks? And is it precise enough that we can develop concrete, creative, and professional composition tools on the other end—tools that benefit composers and bedroom producers?
The Project:
Our initial goal is to conduct a focused research project that will provide us with the necessary experience to co-develop products together in the future.
Most composers, film composers, and producers use VST instruments and plugins when creating music. VST stands for Virtual Studio Technology instruments—computer programs that emulate real instruments, which can be used for sketching, demos, compositions, and background music.
For this initial research, we have chosen to focus on solving a specific problem with machine learning — one that no virtual instrument has been able to address using the conventional sample-based
Programs available:
Arpeggios Across Two or More Strings on a String Instrument
An arpeggio is a musical term meaning that a chord is broken up—i.e., its notes are played in quick succession, back and forth. This technique is over 500 years old and has been widely used in music. For example, string arpeggios are the core element of Guldimund’s hit “Det’ kun vigtigt, hvad det er”. There are several reasons why this is difficult to solve with traditional samples (small audio recordings of individual notes and playing techniques). For instance:
• The notes in a broken arpeggio are played in a single bow stroke, making it challenging to capture in samples.
• To provide full compositional flexibility with arpeggios, recording all possible arpeggio combinations would require an impractical number of studio hours.
• If a musician is not formally trained, playing arpeggios precisely and at high speed can be difficult.

We believe AI can help give composers and producers access to arpeggios within their virtual instrument arsenal, making this technique more accessible and flexible.
The main reason we want to collaborate in this constellation is that together we have the precise set of diverse skills that is required to develop new AI models for use in compositional VST instrument plugins. The long term goal is to collab on a series of software products exploiting these new methods,
which can be sold for composer on a global market. We initially want to do a limited research project with the aim of giving us the necessary experience to be able to co-develop products together in the future.

Duration: 

Start: 31 March 2025 

End:  30 June 2025