Would you consider the DNSOne RTAS plugin? I've used it and I know other people rave about it. It really doesn't use much system resources and is fully automatable too. And of course you can have multiple instances of it. Plus if you're whole workflow has it then you could have it from your dial edit phase to the final mix without having to re-record anything.
Please note: When commenting on someone's question or answer, you will not receive notification on follow up comments unless you are mentioned with @Username. So for Andrew Spitz, you only need to use @Andrew
We use ours in a dialogue chain. All of the dialogue goes through it together (not separately per track). Also, using a Cedar is sort of like mixing. You often have to ride the faders on it to get what you want, and you often do it in real-time. You can get automation support for it as well, but they never use it here.
The thing works wonders. I would recommend it to anyone who can possibly get their hands on it.