This sentence presents an interesting  peculiarity: although the structure of the items enumerated is the same:

                                            X (x') , Y (y') , Z (z') , ......n(n')      ,

the nature of x', y' and z' are different, fact that raises the problem of how to tag theses elements in brackets:

1) @APP
2) @N<PRED
3) @NPHR
4) different tags according to the nature of each n'
 

Neither of the analyses seem to satisfy completely th case. On the one hand, @APP, would not certainly be the case as there is no identification between n and n' (except probably in Luís Madureira (voz)). If the n' elements say something about n, it is not really the pure case of a postnominal nexus predicative (for instance, Olga Pratts is not a (kind of)piano) is the voice- so it would work perfectly in this case). On the other hand, the use of the tag @NPHR would not show the relation between n and n'. And finally, there could be different tags for each case / nature, but the evaluation of the nature would very much depend on a personal point of view and use of language:

(1) Os intérpretes são João Natividade (no movimento) @N<, Luís Madureira (voz) @APP, Olga Pratts (ao piano)@N<, Pedro Wallenstein (no contrabaixo) @N<, António de Sousa Dias (na percussão) @N< e Clemente Cuba (responsável pelo desenho de luzes) @N<PRED.

But this analysis would not be, at all, consistent.

Another possibility would be introducing a new tag, but for the moment that would have to be introduced fully manually and again it would still not be consistent or sistematic, as the parser would face, for instance, two nouns (like Olga Pratts (piano)) and it would be impossible for it to determine how the n' would semantically match n. For instance, there could be the case of having, instead of the proper noun Olga Pratts (piano), Hammon (piano) which in that case piano would be an @APP.

Therefore, these cases are for the moment being dealt as @N<PRED which was the closest and relatively satisfactory tag to apply to such cases.