representation / Symbolic
Piano roll
A 2-D pitch×time grid — the canonical symbolic ML tensor. Each cell marks whether a pitch sounds at a time step; the spatial layout of a score / note-sequence laid out on a grid. A representation, not a new structural object: it renders existing symbolic music for models and editors.
How it works
This object can carry concrete values or a musical interpretation, depending on its notation.
This describes or analyzes music; it may not directly specify playable notes.
Written forms
Ways to write this concept
Each notation describes the same underlying object. Load a playable example to hear one MIDI realization in the Code Sequencer.
matrix
Dense pitch×time grid: rows are pitches, columns are time steps, 1 = sounding
[[0,1,0],[1,0,0],[0,0,1]]
events
Sparse on/off events as pitch:start:duration (time-step units)
60:0:2,64:0:2,67:2:2