Salutations Colanoscopy,
Lets suppose the spectrum were segmented into a couple dozen LEDs with overlaping bands each corresponding to specific peaks 15 ~ 20 nm wide, you’d still need to figure out which bands are actually part of a multi-stage vegetal process, as in some 2-stage ionic pump model evoked below:
If i’m interpreting this correctly then that means providing 680 nm alone or 700 nm alone ain’t good enough: they need to be paired together for optimal performance. So how many permutations out of 24 bands are there to try before ever falling on this 680 + 700 duo? IMO it might be necessary to start with a series of full spectrums each having only 1 single band removed to begin with, testing dark bands one after the other, trying to determine which lesser contributions can be safely rejected. Then i’d expect having to remove 2 simultaneously, being careful not to defeat beneficial processes, etc. Too bad my reflexion is purely fictive, but maybe that’s sufficient to inspire a hint what’s to search for next.
Best of lucks!
Good day, have fun!! 