LazyScales is an experiment with musical scales, (guitar) fretboards and Neo4j, the graph database. LazyScales is all about basing everything on intervals, and representing these as relationships in the database. I wanted to see how this line of thought works out in practice, and also, if it can be leveraged for scale and chord analysis.
To start with, this is how scales are represented by relationships in the database:
So a scale is just a series of intervals. And as the scale doesn’t really end after one octave, it’s connected back to itself with the last interval. This way, the scale can be represented by a relationship pointing to it’s starting point within a circular chain of intervals.
One cool thing about this is that scales relate to each other by something called modes. This means that different scales use the same sequence of intervals, just with a different starting point. The image above depicts some modes of the common Ionian scale (that’s just a plain major scale).
To keep track of all the scales I like to put them into families (and subfamilies). Here’s an example with a minimal set of data:
The Chromatic scale is a special case; there’s a shortcut to it for internal use in the application.
Notes are represented by nodes in the database:
The database is populated with all possible relationships between the notes — the image is just a small excerpt of that data.
So having a note, finding related notes by interval is really easy. What’s the note a major second above the C note? That translates into "follow the outgoing MAJOR_SECOND relationship from the C node" on the database level. Which in turn is already implemented in the Neo4j API for us to use.
Of course we want to combine a scale with a starting note to make things usable. That’s basically done by traversing the scale and notes in parallell. The algorithm is the following:
fetch the next interval/relationship from the scale
use that relationship to pick the next note from the notes
repeat until done
Among musical instrument, I prefer playing guitar. And you might not be aware of it, but people tune their guitars in a lot different ways. In my case I go for something called E-flat tuning, meaning that I tune all strings down by a half step compared to the standard guitar tuning. Let’s have a look at how to represent this in the database.
Again, we use intervals as the basis. The intervals will also implicitly tell us how many strings there are on the instrument!
The tuning itself is represented by a node. From that node, relationships point to the starting node of the intervals as well as to the root note/node of the tuning. I this case we have the standard tuning on the left, pointing to the E root note/node, while the E-flat tuning on the right points to E-flat. Both tunings share the same sequence of intervals.
At the top of the image, you can see that there’s a structure in place to categorize the tunings, just like with the scales.
So what’s the deal with this exercise, delving into the world of intervals? Well, I tend to get confused when it comes to projecting all the scales onto the guitar fretboard. To remedy that, I wrote up a GUI which visualizes the notes on a fretboard as well.
This is the C major scale projected on a standard tuned guitar:
One feature I specifically wanted is the ability to combine two scales on the fretboard:
In this case there’s a B minor pentatonic scale added to the C major scale. These scales have a lot of notes in common, but F and G are only there in the C major scale, while F# is only present in the B minor penta. If you’re into jazz like me you might have an idea as to how this feature is useful!
Building / Downloads
To build the application from source, use Apache Maven. It works at least with Maven version 3. Even the Windows exectuable is built using Maven.
Java version 7 or newer is required to run the application. Download current snapshot builds from here.
The GUI was created by using the WindowBuilder editor in Eclipse IDE. The looks come from a fork of the excellent Substance library. The fonts used are all from the DejaVu family.