Name of Project:
Pentagrom
Proposal in one sentence:
Open source a music ML model using a 3x7 matrix for representing any music note.
Description of the project and what problem it is solving:
Representing any written music in a visual way is not an easy task. To be able to study and understand music, we need a way which enables us to see, listen, write and play music in a simple and orderly way.
What great musicians visualize when they think about music?
A musician visualize how the sounds on his instrument are represented when he thinks about music. A pianist visualizes the piano keys, a trumpeter the three key positions, etc… but the big problem is that this visualization is not directly connected to the musical script system: the stave.
This project could solve this problem by unifying the visualization in to a 3 x 7 array which might open a new way for reading, writing, learning and teaching music. At the end, we are combining the instrument with the stave - so that, for the first time, sight, hearing and touch go together.
Check this video to get an idea of the music controller that Jaime Iglesias (the inventor and member of this team) patented at that time:
[Transcripción de "Billie`s Bounce" (solo de Wes Montgomery) - YouTube](https://Youtube - “Billie`s Bounce” transcription)
Grant Deliverables:
- To have a proper music dataset for training (~3000-5000 midi files of non complex music).
- A way to convert midi files to 3x7 matrix (or a way around to have the dataset in the proper format).
- Have one or more base models that could fit this use case and get a first approach to them.
- Desirable: To have a test prototype which could bring us an idea of the product (model working).
Spread the Love
This project will be a team of 5 people, check the work on Github.
Squad
Squad Lead: AgustĂn GĂłmez
- @localaization (Personal @TuWebO)
- localaization#5156
Squad members:
-
Jaime Iglesias (Inventor and jazz musician and teacher)
-
Jaime#0140
-
Chad Kensington#9564 (ML)
-
EmeraldOdin#1991 (ML)
-
Dutschida (ML)