Actors are too much to afford, non-actors can't act

I want to use recorded conversations for my project but, unfortunately, ever person I've considered cannot act (I should have taken theatre). Maybe I can use natural sounds -- or Stephen Hawkings voice synthesizer -- to make my clips. I was looking forward to script dialog and record natural conversations but, for the subjects(people) I have chosen, they sound like they're trying to act in both situations -- I know we're suppose to blur the boundaries but it sounds horrible. Maybe I shouldn't use my friends (they're both computer nerds) but I don't want to ask strangers to perform for me -- it's weird and creepy. I think I'll have to work with them, hoan their skills. Maybe I can use recordings of other people talking and mash it up. I guess that's a good idea... maybe.

Sound recorded, sound manipulated

Hurm, sound is an odd effect. It's probable the most frustrating thing to capture and edit. I've dealt with it in the past, while filming a few documentary and short films, but never has it been the focus of my attention. Unfortunately, when it comes to film, sound is a vital component that shouldn't be ignored -- it's also the most annoying sensory stimulus to edit. Without the proper equipment, you could really clutter your recording with too much sound or you could neglect important ambient noise that makes your recording less realistic. I think I'll use a lav mic to record singular sound (people's voices, conversations, breathing) and use a directional mic to capture ambient noise (environmental sounds like wind, echoes, people, distant chattering, animals, water, etc). I think it would be best to capture sound separately for the constructed portion of the project. That way I can layer the sounds I think are vital to the overall "realistic" qualities that it imbues and delete the sound that makes it sound artificial. For the captured piece, I have little to no idea what to do; however, I am sure that I will use the directional mic -- without ambient noise, it'll sound fake. For both sequences, I want to have similar attributes so I'm thinking about using conversations for both pieces. I am unsure whether I will script it or not -- seeing that the whole project is ambiguous on how I should approach both segments. I'm leaning towards having everything captured will be unscripted -- or what has been termed as "reality" programing -- so I'll have to attempt to capture people in their natural disposition. It'll be hard but I think I can find some locations today to record -- that is, if I don't become bored of this project first and give up, which I probably will.