What would prevent this from being feasible 910nm-382nm-=528nm?????
Other wavelengths are produced by taking the sum or difference of the frequencies, or of 1/λ.
So using 910 nm and 382 nm:
1/382 - 1/910 = 0.001519
1/0.001519 = 658
So this would produce light at 658 nm. However, there must also be a nonlinear crystal that would work at these wavelengths, and I am not sure if that is the case.