DJ, the whole trick on something like this is going to be figuring out the weakest link in the chain, fixing that, then finding the next weakest, yada, yada. Otherwise, you try an experiment on a stronger link, and maybe you improved it, but you can't even tell because the weak link is letting you down.
Some of these things are measurable. The runout Pete talks about is a good example. You can take a DTI and your tool in the spindle and just straight up see what the runout is. Post that up hear and you'll get an idea back of what to expact. That finish is showing quite a bit of variation. Looks a lot like what chatter looks like, BTW.
Hard to believe it is due to tool deflection from a 3/4" endmill that was likely not being run all that hard. You got a known bearing problem, though fixed. But do you have enough preload? Do the bearings run a little warm? You want them to run a little warm, but be stable. If they don't get warm at all, you can up the preload.
How are you feeding the cut? By hand? I remember my early hand feeds produced TERRIBLE surface finish seemed like no matter what I did. What a revelation power feed was. Then CNC trumped that by a lot.
Is something, anything loose? Is the machine leveled and firm on its stand? Is the tool tight in the taper? What are you using for a tool holder? A cheap set screw holder has a ton of runout versus a collet chuck. An R8 collet is somewhere in the middle. BTW, if its an R8 collet, is it properly tightened? Tormach has a great white paper on collet pull out. Scary how easy it is to miss a trick and make it easy for your tool to be slipping in an R8 collet.
You improve one thing, then you go around looking for more to improve. Then you come back to what you already improved and improve it some more. Weak link by weak link.
So on, and so forth. It's a methodical trial and error sort of thing. It's harder if you don't have the experience of someone looking over your shoulder to tell you what "right" is. Does that cut sound "right"? Heck, I dunno, just got this darned mill. Sure is noisy!