|
13 | 13 | "metadata": {},
|
14 | 14 | "source": [
|
15 | 15 | "## Overview\n",
|
| 16 | + "\n", |
16 | 17 | "The information in this sub-section is based on the [MUSB18 dataset page](https://sigsep.github.io/datasets/musdb.html). {cite}`musdb18,musdb18-hq` Here we have edited down the content to focus on the details relevant to this tutorial while keeping it concise. For more details about the datataset please consult the dataset page.\n",
|
17 | 18 | "\n",
|
18 | 19 | "MUSDB18 is a dataset of 150 full length music tracks (~10h total duration) of varying genres. For each track it provides:\n",
|
|
27 | 28 | "As its name suggests, the \"other\" stem contains all other sources in the mix that are not the drums, bass or vocals (labeled as \n",
|
28 | 29 | "\"accompaniment\" in the diagram below):\n",
|
29 | 30 | "\n",
|
30 |
| - ":::{figure,myclass} fig-musdb18\n", |
31 |
| - "<img src=\"../images/data/musdb18.png\" alt=\"musdb18\" class=\"bg-primary mb-1\" height=\"300px\">\n", |
32 |
| - "The MUSDB18 dataset.\n", |
33 |
| - ":::\n", |
| 31 | + "```{figure} ../images/data/musdb18.png\n", |
| 32 | + "---\n", |
| 33 | + "height: 400px\n", |
| 34 | + "name: fig-musdb18\n", |
| 35 | + "---\n", |
| 36 | + "Illustration of the stems comprising the mixture in a MUSDB18 track. Source: [https://sigsep.github.io/](https://sigsep.github.io/).\n", |
| 37 | + "```\n", |
34 | 38 | "\n",
|
35 | 39 | "All audio signals in the dataset are stereo and encoded at a sampling rate of 44.1 kHz. The mixture signal is identical to the sum of the stems.\n",
|
36 | 40 | "\n",
|
37 | 41 | "The data in MUSDB18 is compiled from multiple sources: \n",
|
38 |
| - "the [DSD100 dataset](dsd100.md), \n", |
| 42 | + "the [DSD100 dataset](https://sigsep.github.io/datasets/dsd100.html), \n", |
39 | 43 | "the [MedleyDB dataset](http://medleydb.weebly.com), \n",
|
40 | 44 | "the [Native Instruments stems pack](https://www.native-instruments.com/en/specials/stems-for-all/free-stems-tracks/),\n",
|
41 | 45 | "and the [The Easton Ellises - heise stems remix competition](https://www.heise.de/ct/artikel/c-t-Remix-Wettbewerb-The-Easton-Ellises-2542427.html#englisch).\n",
|
|
79 | 83 | "cell_type": "markdown",
|
80 | 84 | "metadata": {},
|
81 | 85 | "source": [
|
82 |
| - "## Downloading and inspecting MUSDB18 clips" |
83 |
| - ] |
84 |
| - }, |
85 |
| - { |
86 |
| - "cell_type": "markdown", |
87 |
| - "metadata": {}, |
88 |
| - "source": [ |
| 86 | + "## Downloading and inspecting MUSDB18 clips\n", |
| 87 | + "\n", |
89 | 88 | "### Downloading MUSDB18 clips with nussl\n",
|
90 | 89 | "\n",
|
91 | 90 | "We'll use `nussl`, the source separation library used in this tutorial, to download 7-second clips from MUSDB18. We'll get into `nussl` in greater detail in the next Chapter, but for we'll just use it to download and inspect the audio clips.\n",
|
|
0 commit comments