I have a long way to go when it comes to PixelMath. Thankfully, helpful folks like David Ault (Cloudy Nights) have provided some great examples for me to reverse engineer and try to understand. Here is my second attempt at processing the original Rosette subs into something similar to a bicolor SHO palette. This was accomplished by generating a synthetic green channel with the following PixelMath:

0.5*(0.5*min(Ha,OIII)+0.5*max(Ha,OIII))+0.3*Ha+0.2*OIII