Chopin Etude op.25 n.2 reloaded :-)

This time I put to hard test Synfire asking it to recognize the chords of the first 15 or 16 beats of the Chopin study.

I thought it was easier to start from the accompaniment (left hand), because it contains less notes extraneous to each chord respect to the fast paced right hand.

From the MIDI Import dialog I have therefore excluded the right hand channel ("Too many notes!") double clicking on its name.

As you can see from the animated GIF that follows, the result was not encouraging: Synfire introduced a myriad of chords extraneous to the simple harmonic progression used by Chopin.

I then deleted the Harmony parameter, then I proceeded to the recognition of the figures. For scruple, I clicked the checkbox Accidentals, although perhaps Synfire could do without it since the left hand plays only the notes of the chords. I then clicked also on Monophonic, being the accompaniment made of a single voice.

I went into the Harmonizer window, selected the track "Piano left" and set the correct global key (F minor). The only major change in the default settings of the Harmonizer was to assign a minimum duration of the chords corresponding to a whole measure (1/1), as the composer had thought them that way. I then clicked on Harmonize, and the result was surprising:

As you can see by comparing the harmonies proposed by Synfire with the original score of Chopin, Synfire has guessed everything, including diminished chords. For reading convenience I reported in red the chords elaborated by Synfire. The only note missed by the software was a Db (indicated by the red arrow), i.e. the sixth of the F minor scale. But it is a really negligible "error" and easy to correct manually.

It was a first experiment, others will follow.

A final annotation: I use classical scores because my musical training is classical. But I suppose the techniques I'm testing can be applied successfully to any style!

Please post yours, thank you ;-)

_Roberto