Principally you are right, an opto CAN help a lot.
But: Up to now we don't know what Rob is all having on his PCB, means where the noise actually develops. So, after a final analysis there might be the need for introducing an opto, but so far...

So far we only know, that he has a simple and a rather slowly transistor inverter and a rather slowly 4049 powered by 12V. These parts are so softly working, that I cannot imagine that an opto is needed to overcome any noise problems caused by their use. What seems to be lacking is just an adequate supply voltage decoupling measure, at least in my eyes. Or maybe he followed a poor ground management, not using star point grounding, etc.

So, at this point just following good design practices is needed with his application, means introducing of proper decoupling measures and star ground routing, but not so much the introduce of an opto.

By the way, people are often a bit disappointed, when they finally use optos in a design which suffered from problems and notice that situation is only slightly improved or that new problems arise. Often overlooked is, that optos show a certain input-to-output capacitance, which can act like a short circuit for fast glitches!! In really fast designs not much is gained by introducing an opto. In such cases additional common mode filtering by the help of common mode soft ferrite chokes is a must.
There are still some other disadvantages of optos but that's a different story.

Kai