Worked example. The experimentation begins from a fixed four-note melody (four equal notes C4 D4 E4 G4). The question: what does a sustained C3 below change? The trial is recorded twice (without, then with the added part); the melody and rhythm stay fixed while the texture moves from monophonic to two-part and the harmonic context shifts; the verdict names what moved and what stayed, with the balance dependent on timbre.
A second worked experiment tests accent: the same eight notes played twice, once with the accent on every second beat, once displaced onto every third. The question: does the metre the listener hears follow the notation or the accent? The evidence (two recordings, listeners' taps) shows the perceived grouping following the accent: the verdict feeds the creator's next piece, where a notated four is heard as three until the accent returns.