to be analysed is created in . The abstract images (shapes) of the motives are calculated in . The important step corresponds to the identification of motives (from (1)) into gestalts (classes of motives) with respect to an action of a paradigmatic group. In the small gestalt relation between all pairs of gestalts (from (4)) is evaluated. Note that this relation on gestalts is the corresponding relation to the submotif relation on motives, and it is exactly after this step that all other calculations are done on gestalts instead of on motives, until we finally come back, at , to the level of motives.

Directly resulting from step the collection of generic gestalts, i.e. the maximal elements of the gestalt composition space for the score under the topological dominance relation, is created in , and the -matrix in . The -matrix of small gestalt relations corresponds to the -neighborhoods, , of all gestalts.

The distance function between all pairs of gestalts of the same cardinality is evaluated in , as well as the collection of all changing epsilons, i.e., those neighborhood radii where neighborhoods (they are all finite and only differ for specific ”jumping”radii) become larger while increasing the radii. Box is the unit where the collection of all neighborhood radii at which quantification functions, i.e. presence, content and weight functions, will be evaluated, is created.

The boxes from (10) to (14), and from (24) to comprise a cycle for the calculation of dimensions of function spaces and of stalk dimension of function spaces for content and presence. More precisely, in the presence/content loop variables are initialized. The running variable deals with the enumeration index of neighborhood radii from (9), and with the enumeration index of gestalts in S. In (11) the gestalt multiplicities as well as the linear combination gestalt coefficients are initialized as respectively the -Matrix and -vector. Steps - correspond to the presence/content loop first () on the whole motivic composition space, and then restricted on the -neighborhood of the th gestalt. In fact, linear operations in evaluate the presence and the content of each gestalt at a given neighborhood radius. It underlines the reciprocity property of these two quantification functions as we can see by comparing their respective linear operations. Step checks if all neighborhood radii from have been evaluated in , and in , the -matrix is modified with respect to the growth of the gestalt neighborhoods for a larger radius. The resulting evaluated presence and content functions at all radii, in the form of a matrix, are finally given in .

Weights of gestalts can then be calculated in , again in the form of a matrix. In - the dimension of each quantification function vector space is calculated by a linear algebra operation.

The qualitative function is evaluated in which then leads to the motivic evolution tree (MET) in . We recall that the MET of a score is a graphical representation of the overall motivic spectrum of with respect to the chosen parameters for the analysis.

At we go back to the motif level by assigning to each motif its respective weight. Weights of notes are calculated in .

Finally steps - (together with ,-) correspond to the presence/content sheaf loop. More precisely, presence and content function stalks for