It is the author's suspicion that the reason why some timbres did not benefit much from just intonation in his experiments is that they were produced from electronic, rather than physical sources. Tones from polyphonic electronic instruments do not interact: they superpose in a perfectly linear fashion. In contrast, it is possible that tones played on a real polyphonic instrument such as the violin interact with each other due to physical nonlinearities in the instrument body. These interactions could serve to change the character of the sound itself depending on whether the tones were harmonically or slightly inharmonically related, making just intonation ``matter'' more. An investigation into the possibility of existence of such nonlinearities, and their simulation in synthesis could be an interesting extension of this work.
Another future direction related to synthesis is the idea of sounds with tempered spectra. Much of tuning theory is concerned with bringing the frequencies of tones in alignment with the harmonics of others. Why not turn this problem on its head and concern ourselves with aligning the harmonics of tones with the set frequencies of a given tuning? Early experiments have been conducted by the author in which the partials of a tone are stretched so as to be slightly inharmonic and better aligned with the fundamentals of tones in 12TET. Unfortunately such stretching results in beating within the tone itself, due to the phenomenon of second-order beats. Although this beating is quite strong, it can be made much slower than the beating normally associated with an M3 in 12TET.