{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "\n", "\n", "In this notebook, we show how one may synthetically generate self-similarity matrices from structure annotations. In particular, based on the notions introduced in Section 4.2.1 of [Müller, FMP, Springer 2015], we consider SSMs with path and block structures. \n", "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Annotations\n", "\n", "In previous notebooks, we have looked at the [Hungarian Dance No. 5 by Johannes Brahms](../C4/C4S1_MusicStructureGeneral.html), which has the musical structure $A_1A_2B_1B_2CA_3B_3B_4D$. Starting with an audio recording of the piece, the goal is not only to derive a **symbolic** description of the musical structure, but also to determine the exact time positions when the structural parts start and end. Let us have a look at the Ormandy recording of the Brahms piece. \n", "\n", "\n", "\n", "\n", " | start | \n", "end | \n", "label | \n", "
---|---|---|---|
0 | \n", "0.00 | \n", "1.01 | \n", "\n", " |
1 | \n", "1.01 | \n", "22.11 | \n", "A1 | \n", "
2 | \n", "22.11 | \n", "43.06 | \n", "A2 | \n", "
3 | \n", "43.06 | \n", "69.42 | \n", "B1 | \n", "
4 | \n", "69.42 | \n", "89.57 | \n", "B2 | \n", "
5 | \n", "89.57 | \n", "131.64 | \n", "C | \n", "
6 | \n", "131.64 | \n", "150.84 | \n", "A3 | \n", "
7 | \n", "150.84 | \n", "176.96 | \n", "B3 | \n", "
8 | \n", "176.96 | \n", "196.90 | \n", "B4 | \n", "
9 | \n", "196.90 | \n", "199.64 | \n", "\n", " |
\n", " | \n", " | \n", " | \n", " | \n", " | \n", " | \n", " | \n", " | \n", " |