Acoustics, perception, and production of legato articulation on a digital piano.

This study investigated the perception and production of legato ("connected") articulation in repeatedly ascending and descending tone sequences on a digital piano (Roland RD-250s). Initial measurements of the synthetic tones revealed substantial decay times following key release. High tones decayed faster than low tones, as they did prior to key release, and long tones decayed sooner than short tones because of their more extensive prerelease decay. Musically trained subjects (including pianists) were asked to adjust the key overlap times (KOTs) of successive piano tones so that they sounded optimally, minimally, or maximally legato. The results supported two predictions based on the acoustic measurements: KOTs for successive tones judged to be optimally or maximally legato were greater for high than for low tones, and greater for long than for short tones, so that auditory overlap presumably remained more nearly constant. For minimal legato adjustments the effect of tone duration was reversed, however. Adjusted KOTs were also longer for relatively consonant tones (three semitones separation) than for dissonant tones (one semitone separation). Subsequently, KOTs were measured in skilled pianists' legato productions of tone sequences similar to those in the perceptual experiment. KOTs clearly increased with tone duration, an effect that was probably motoric in origin. There was no effect of tone height, suggesting that the pianists did not immediately adjust to differences in acoustic overlap. KOTs were slightly shorter for dissonant than for consonant tones. They also varied with position in the ascending-descending tone sequences, indicating that the pianists exerted strategic control over KOT as a continuous expressive dimension.(ABSTRACT TRUNCATED AT 250 WORDS)