Evidence-Based Instrument Making: Robust Experimental Design Using Finite Element Modelling

This paper examines how evidence-based approaches applied to experimental design can enhance the epistemic robustness of research in musical acoustics.Using guitar soundboard bracing pattern as a case study, this paper combines finite element modelling for sample-size planning, vibratory measurements and blind perceptual tests. Two batches of guitars with scalloped or regular bracing patterns of their soundboard are compared, both experimentally and numerically. Treating finite element based methods as an experimental-planning tool, the approach helps determine what counts as a measurable and decision-relevant effect before committing resources to experiments, by quantifying the number of paired samples required for the experimental campaign. Additionally, brace-induced differences, while statistically significant in structural dynamics, are not easily perceptible in informal playing tests. The approach outlines a general workflow for reducing bias, avoiding under-powered studies, and clarifying the burden of proof in applied experimental science.