Search

I approach timbre here primarily from the perspective of the music creator/producer, and the processes and considerations they apply in order to make recordings. Over the course of this thesis, I build a new synthesiser, whose sound manipulations (FX) are modelled after what would be typically used in a compositional or audio mixing context. These FX are, themselves, treated as the timbre descriptors. In the first experiment I try to establish units for these FX as measured in steps of smallest perceivable change of said parameter; the perception threshold (PT). This unit makes timbral changes of any kind comparable on a perceptual foundation and also relates back to manipulation features that are directly accessible to composers/mixers/musicians. A noteworthy change in this research, in comparison with the more established just-noticeable difference (JND), is that the PT is measured for change in continuous sound, rather than for sets of separate, isolated sounds as used for JNDs. This again mirrors the applicability to musical processes, where sound change is often evoked gradually over periods of time, rather than always being distinct to isolated pieces of sound. With this specific attribute of continuous sound, I test the impact on PT for changing FX gradually over a set amount of time, rather than abruptly, the time-frame for the change is here denoted transition time (TT). Following that, I explore how these FX are used by inexpert participants when they are given the task of mixing musical example pieces using the established FX. Also measured is how they rate, post-task, their own efforts when compared to both standardised and random results of the same task, or an expert audio engineer’s resulting mix. Lastly I compare the impact of both the FX and more standard audio descriptors (see sections 2.2 – 2.3) on participants’ preferences for different mixes, as well as their relations to each other. Finally I use the insight gained in the previous experiments, to create a piece of music. The creative process is primarily informed by applying the established PTs to a musical context, which in itself is created almost exclusively by the synthesiser used in the experiments. Potential research questions that could arise from these efforts are briefly discussed as well.