Hip Hop/Jazz Analysis

This repository tracks my work for Georgia Tech's Robotic Musicianship VIP team for the Spring 2019 semester. This semester, I am working to create tools for automated rhyme scheme analysis.

Contents

lyric_scraping/ contains code for a python script that collects verses from given artists using the LyricsGenius python library UPDATE: My lyric scraping tool has been moved to its own GitHub repository. You can find it here.

nltk_rhyme.py and words_rhyme.py are small scripts I wrote to test out two python libraries, NLTK and Jellyfish, respectively.

RhymeSchemer/ contains code utilizing the Raplyzer project. I utilized Raplyzer to make rhyme_schemer.py. This python program can generate a visual representation of rhyme schemes to either HTML or the command line. Do note that due to limitations in the number of colors available at the command line, I recommend using the HTML route. Please note that the only code I personally wrote in here was rhyme_schemer.py. All credit for the Raplyzer code goes to the original authors of said project.

Example usage of rhyme_schemer.py:

from rhyme_schemer import RhymeSchemer
rs = RhymeSchemer(filename="aquemini_3k.txt")
rs.rhyme_schemes_to_html(filename="aquemini_rhyme_schemes")

This will generate an HTML file called "aquemini_rhyme_schemes.html".

Example output:

Rhyme Schemer utilizes common vowel sounds in words to detect assonant rhymes. The words are then grouped by their shared vowel sounds and assigned a color. That is why words such as "my" and "mind" or "blend" and "again" are given the same color despite not being perfect rhymes.

The tool is not without its issues, of course. First, the colors are picked randomly, so there is no guarantee for noticeable differences between colors let alone accessibility for color-blind individuals.

Second, words are naively assigned to rhyming groups such that a word is put in the first group that shares one of its vowel sounds. Put plainly, a word may have multiple syllables, each with their own vowel sound. These syllables can each belong to distinct rhyming groups. Ideally, colorations of words would be split up into individual syllables.

Third, this tool does not take into account the actual delivery of the verse, let alone the rhythm or stresses put on syllables. Here, Andre 3000 uses a repeated pattern of words and vowel sounds for some of the verse. "Warps and bends", "floats the wind", "count to ten", "meet the twin", and "Andre Ben" do not rhyme by convential standards. However, these phrases are delivered in such a way that they do end up rhyming to the listener. The shared vowel sounds between each respective word in a phrase are not perfectly alike, but sound similar enough to pass as a rhyme. This is especially true when factoring in Andre 3000's southern accent. The words "ten" and "twin" sound like they share the same vowel sound when listening to the verse. Lastly, a consistent meter, or rhythmic structure, ties these phrases together to an additional degree.

In summary, Rhyme Schemer needs more accessible color schemes, categorization rhymes on the syllable level, and consideration for lyrical rhythm and stresses to more accurately capture the true rhyme scheme of a given verse.