From MIDI to Rich Tablatures: an Automatic Generative System incorporating Lead Guitarists' Fingering and Stylistic choices

TL;DR

The paper addresses automatic generation of lead guitar tablatures enriched with fingering and expressive articulations from monophonic MIDI melodies. It introduces a constrained, multi-attribute optimization that jointly minimizes time- and biomechanical-related costs while tracking fretting-hand position, solved via IBM ILOG CPLEX, with $PC$, $SC$, and $HS$ guiding the objective. Building on this, it inserts statistically grounded articulations and expressive techniques drawn from the mySongBook corpus using a rule-based framework, and outputs MusicXML for visualization and integration. The work enables applications in teaching, assisted composition, and computational expressive models, and points to future extensions to multi-note passages and production workflows.

Abstract

Although the automatic identification of the optimal fingering for the performance of melodies on fretted string instruments has already been addressed (at least partially) in the literature, the specific case regarding lead electric guitar requires a dedicated approach. We propose a system that can generate, from simple MIDI melodies, tablatures enriched by fingerings, articulations, and expressive techniques. The basic fingering is derived by solving a constrained and multi-attribute optimization problem, which derives the best position of the fretting hand, not just the finger used at each moment.Then, by analyzing statistical data from the mySongBook corpus, the most common clich{é}s and biomechanical feasibility, articulations, and expressive techniques are introduced. Finally, the obtained output is converted into MusicXML format, which allows for easy visualization and use. The quality of the tablatures derived and the high configurability of the proposed approach can have several impacts, in particular in the fields of instrumental teaching, assisted composition and arranging, and computational expressive music performance models.

Figures (5)

• Figure 1: To obtain this solution, the least penalized feature was the string change (SC). The hand spread (HS) was penalized twice as much as SC, while the position change (PC) was penalized twice as much as HS.
• Figure 2: To obtain this solution, the least penalized feature was HS. PC was penalized twice as much as HS. SC was penalized twice as much as PC.
• Figure 3: Ratio of the use of the considered techniques to the total number of notes for each string. Strings are labeled according to guitar standard tuning, the 6-th string being the low "E" and the 1-st string the high "e".
• Figure 4: Ratio of the notes played on the different strings to the total number of notes in the test dataset (applying our optimized fingering) and the entire mySongBook corpus.
• Figure 5: Rich tablature and standard notation of a sample lead part automatically generated by our system.