One observation, FWIW: if the singer's not very good, a particularly fine microphone and preamp aren't going to fix that, and could actually make him or her sound worse, rather than better.
As already mentioned, what you need to do depends on what's wrong with the singer's performance. In some cases, the tool is kind of obvious: too sibilant -> de-esser; level all over the place -> compression, etc.
A few other random thoughts:
- Notwithstanding the above, reverb tends to make pretty much anyone (who's not quite good to start with) sound better.
- General pitch and timbre issues: fully double-track the vocal, i.e. have the singer sing it twice and use both in the mix. This tends to "smooth" problems. Of course, there's a cost, as it also smooths away expressive nuance that you might like ... but if the singer's not good, there's probably more badness than euphonious expressiveness.
- Inarticulate: might try recording a "whisper" track. I've never done that, but I've heard of it.
- Pitch: obviously there's autotune and similar software.
- Timbre: Autotune-like software can also be used to fool around with formants and the like. You may wind up making it worse, unless you're trying for something that doesn't sound like a real human being.
- Timing, phrasing: You can edit in your DAW, either by cutting up words and phrases, or with a tool that's specifically designed for the purpose. This can quickly turn into a lot of work that turns out to be counterproductive when you sit back and listen to it, though.