The final selection of proto-melodies from the Alpha 1 version of my Magical Muzak Maker.
These proto-melodies demonstrate key signatures other than C major.
(Although as you can see they are still scored with the C major key signature. It would be easy enough to create the score with the correct key signature but I want to re-structure my code to make it more maintainable and extendable and will add a method to print the correct ket signature when I do this.)
D Major, 1 Octave Range, 4/4 Time, 4 Bars:
Magical Muzak Maker - Proto-melody 7
F Major, 1 Octave Range, 4/4 Time, 4 Bars:
Magical Muzak Maker - Proto-melody 8
The main point behind in generating these two proto-melodies was to check that sharps and flats were inserted properly for Lilypond, the scoring tool. In the case of F Major it took a few attempts before a melody containing the B flat was generated. This raises the question; how can you tell the difference between a melody generated in C Major and a melody generated in F Major containing no B flats … especially when both are scored using the C Major key signature?
Taking this further, how can you tell the difference between a melody in C Major and a melody in A Minor or the D Dorian mode? One of the, I think, more reasonable goals of this Magical Muzak Maker project is to come up with a system that might give an aspiring and learning musician an idea of what different keys and different modes sound like as an inspiration to their own musical endeavours. Not much good if you can’t tell the difference between them.
I am fairly sure that when I add some harmony generating features the nature of the different keys and modes may shine through. However I think I need to add some weighting to the note selection so that melodies tend to start on the root note of the scale, or maybe some other specific intervals.
In my more ambitious moments I start planning bringing together melodies generated by this Magical Muzak Maker into something actually resembling music. For this to happen I think I am going to have to guide the melody generation to try and force a resolution at the end (or at least have the option of trying to enforce this).
I do want to keep the process as random as possible but it is also interesting discovering how much you have to constrain the process before something worthwhile pops out.
If on the very rare chance either of these melody is not original I would be interested to find out. If either of these melody plagiarises your own music then apologies … but well, in that case your music could have been written by a very simple pseudo-random alogrithm.
