rcjach/piano_textures

Collection of piano textures

PIANO TEXTURES

WHAT IS THIS

This repo is a collection of piano accompaniment textures found in classical or other genres of music, designed to be a pool of inspiration to steal learn from, written in Lilypond language.

WHAT IS IT GOOD FOR

There is scarcity of piano textures in the pop, rock, or even jazz context. Most of the materials contains only block chords, arpeggios, riffs, and at most some chorale lines, thus incorporating other textures from time to time can be helpful to sustain interest.

WHAT IS INCLUDED

The lilypond project and the rendered image is in the data folder of the source code, organized by genre, composer, opus - piece, then movement. Within each folder, there are three type of files:

1. A .ly file which is the main project.
2. A header.ily file containing information about the piece.
3. Many image files which is the rendered result of the project.

However it's really a hassle opening every image one by one, a better way of previewing is to have a searchable and tagged database which lists all images in one place. But that takes a lot of time. So until I decide to make one, please use the wiki for now. I will make one page for every uploaded piece, including motives and combinations of textures used in the movement, explained below.

MOTIVES

Motives page lists a set of basic texture patterns used in the piece, transposed to C major.

Example:

Each motif is given a short name with three parts:

1. Alphabet: The name of the super-category of the texture, in the case above is Arpeggio.
2. Number: The main variation of the texture in concept.
3. Apostrophes: The minor applied variation with the same concept. For example a1 is a 64th-note arpeggio with only C and G, and a1' differs by one note. Each variation may have different voicing, rhythm, or notes, but should share the core concept. For example a4 differs completely from a4' and a4'', but they are all based on the concept of arpeggio with multiple notes.

COMBINATIONS

Combinations are how motives are combined in the actual music, usually with an extra melody and/or a bass layer. However, since this is not music analysis, not all contents are covered, some combinations or even motives might be left out to simply and shorten the result.

Indicated in the upper-left corner, combinations are categorized by the number of voices within a measure, or part measures. Each combination has its motivic elements listed in a box above the staff, with "+" representing motives happening simultaneously, and "|" representing otherwise.

Different voices are colored differently, with red being the melody, blue the bass, and others the accompaniment voices.

Two hand playing the identical motif or variations are considered as two voices:

However, them contributing to a single motif are considered as one voice.

Sometimes, I may include some observations of things to consider as comments below the staff.

piano_textures's Issues

List of music pieces to analyze

This issue function as a list of potential music pieces to extract motives from.

I started with Franz Liszt's Transedente Etudes because it's hard, especially with interlocking motives and numerous layers, which qualifies for making a great test suite when designing the lilypond template.
But the problem is also that, it's hard.
There are too many variations and combinations in each movement, and even though I'm not doing a full analysis of the piece, it's still time consuming just to extract motives and decide what count as a separated textural motif.

Another time sucker is the process of finding pieces to work on, since I definitely haven't listened to enough music.

So feel free to give suggestions.

I'm considering:

• Ravel's Mirror
• Chopin's Etudes, especially no.1 in C major with only two layers.

Maybe I'll start with Bach's Well-Tempered Claviers.

FR: Create a searchable database

Currently there are two paths to get to the actual data:

1. Dive into the source code and traverse through layers of folders, opening files one by one.
2. Go to the wiki for a complete but unorganized list of uploaded pieces, with all images of the piece on its separated page.

The problem with 1 is that, well, it's time consuming.
Path 2 makes it better, but although fine for now, it will be bloated when this project grows into a larger scale.
Also, a simple list doesn't do justice for this project, a list of pieces to the actual sheet music on IMSLP would probably suffice.

The actual value in extracting motives from music, rather than presenting the music itself, is that we can more easily compare and contrast different variations and applications of a simple textual concept.
For that purpose, I think a tagged database would be more fitting, where one can use combination of tags to filter motivic results without knowing a thing about the actual source it came from.

I have added some tags to the header.ily file for each piece that I uploaded, I need a Static Site Generator (SSG) at this point.
I have little experience with Hugo, and I recall it being troublesome in reading external datafile.
My current bet is on Eleventy.
It's JavaScript based, which I have only elementary knowledge of, but it is definitely more flexible.

If that fails, I'll have to fall back to python based ones.

Anyway I just need to spend the time and do the work.

I'll leave this hanging until I get a considerable quantity of entries for the project.