Enter a sequence of pitches in C Major, then click "Harmonize" to see various barbershop chord progressions for that melody. Brought to you by recursive backtracking.
- Down Our Way:
E B A G A E D
- Mary Had a Little Lamb:
E D C D E E E
- Chromatic scale:
C C# D D# E F F# G
- Santa Fe:
G(G7) F E(C) G F E(Am7)
- Forced chords. You can force a particular chord for a particular melody note. Just put the chord name in parentheses after the pitch: F#(D7). Make sure the forced chord actually contains the pitch—you won't see any validation messages or error messages if not.
- Progression permissivity. You can decide whether to allow only basic Circle of Fifths motion (e.g. II7 → V7 → I) or additional progressions as well (e.g. I#7 → I7, tritone substitution). Permitting all progressions will give many more results.
- Melody length. For performance reasons, melodies can be at most 15 pitches long. Computating anything longer would likely cook your browser.
- Letter casing. Case matters only for chord names (e.g. because AM7 and Am7 are distinct chords). Case does not matter for pitch names.
- Playing chords. You can play through progressions using Keyano, a browser-based piano, if you want.
While arranging barbershop music, I've often wished for a tool that could quickly show me all legal-ish chord progressions that are possible for a certain sequence of melody notes. This is a scrappy, two-day attempt at building such a tool.
This is a decidedly rules-based harmonization engine. That means that there is an explicit, curated list of rules dictating how each chord is allowed to progress. For instance, G7 may always go to C, while C can go anywhere, since it's the tonic. The rules are stitched together to build a graph of all possible chord progressions, then for a given melody, recursive backtracking is used to find all harmonic paths "out of the maze." This system isn't particularly robust—it's mostly just a quick proof of concept. But it does spit out some interesting results!
This is intentionally not powered by a Machine Learning model. In the past, I built an ML model for harmonizing melodies using a Hidden Markov Model trained on Bach chorales. That approach may yield a more statistically defensible result for any given melody, but it would have required more training data than I had time to create, and it also tends to give boring, least-common-denominator harmonizations. In contrast, the rules-based engine here just returns all reasonable-ish progressions for you to peruse. The result is a much greater appreciation for the complexities of barbershop arranging. :)
To learn more about barbershop, visit the Barbershop Harmony Society's website. To learn more about the chorus I direct, visit the Fog City Singers' website.
To deploy to Github Pages, run the following:
yarn run deploy
Then go to https://cmslewis.github.io/shop-it.
This project was bootstrapped with Create React App.
In the project directory, you can run:
Runs the app in the development mode.
Open http://localhost:3000 to view it in the browser.
The page will reload if you make edits.
You will also see any lint errors in the console.
Launches the test runner in the interactive watch mode.
See the section about running tests for more information.
Builds the app for production to the build
folder.
It correctly bundles React in production mode and optimizes the build for the best performance.
The build is minified and the filenames include the hashes.
Your app is ready to be deployed!
See the section about deployment for more information.