In this communication, we introduce a theoretical model for the analysis of musical texture and orchestration. Its computer implementation is also presented, together with a demonstration in the context of a contemporary orchestral score. Our methodology presupposes two complementary strategies, one focusing on the symbolic domain and the other tackling psychoacoustic features. The analysis starts by defining the universe of sonic possibilities, that is, a record of every instrumental resource prescribed by the composer. We call this record as the Sonic Resources Index (SRI). Through the SRI, the total number of resources available in the composition’s palette along with the potential maximum number of simultaneous resources are established. By tracking down every change in the use of SRI’s components, we can segment the musical work based almost exclusively on orchestral information. We call each of such segments as a Local Sonic Setup (LSS). Each LSS is then explored for its homogeneity attributes. That is, questions like the following might be put forward: when and how often the composer tends to blend groups of sonic resources by means of coordinated lines (voices)? How much independence is given for sonic resources of the same type? To answer them, we adapted principles of a theory known as the Theory of Partitions, which allows us to represent and manipulate structures, and calculate the level of agglomeration or dispersion of resources in a particular LSS (see Andrews 1984, Gentil-Nunes 2009) Our premise is that the more dispersed the resources, the more complex the orchestration (Berry 1987). Diversity and homogeneity measurements yield by those procedures are compared with other dimensions of compositional discourse. In a second step, the model calls for recording(s) of the work to evaluate, through the analysis of the audio signal, the actual result of the score’s prescriptions. Local Audio Units (LAU) are delineated in synchronicity to the formerly identified LSS’s. Audio features, psychoacoustic models and statistical measurements are applied to LAU’s. Audio and symbolic measurements are ultimately combined to form a relative-complexity index. In this talk, we concentrate on the symbolic aspect of the model, using, as a case study, the 4th movement from “Recovery Zones” (2013) for large orchestra and electronics, by the Spanish composer Núria Giménez Comas (1980). “Recovery Zones” constitutes a singular example of recent music where the orchestral writing, which meticulously employs a wide range of extended techniques, stays at the top of the hierarchy of structural components. We aim to shed light on the leading role of the orchestration in the score, which stands out for continuously weaving dispersed threads of sonic resources, as for its overlapping orchestral setups, and refined play between micro and macro-structural features.