Basic idea is to use two delay lines and and two (possibly) random LFOs to oscillate the different delay times between 0 and some very small upper bound.
This would not sum to mono well however. Possible synchronization of time in each line could balance out the phasing?
Also maybe think about a frequency domain based approach, which would sum to mono. That's probably for a different ticket.
Basic idea is to use two delay lines and and two (possibly) random LFOs to oscillate the different delay times between 0 and some very small upper bound.
This would not sum to mono well however. Possible synchronization of time in each line could balance out the phasing?
Also maybe think about a frequency domain based approach, which would sum to mono. That's probably for a different ticket.