When you are recording an interview, do you ride the levels? For example if someone has a dynamic voice do you ride the levels. Do you anticipate riding the levels if someone is a loud speaker as they start a sentence? How much of this goes on if any or should you keep the knobs in one spot. This is for an office interview with no outside noise except the AC.
I mostly keep it in one spot. I just make sure to have enough headroom. :)
An exception would be when the speaker generally has gotten louder (maybe got more excited?) but on a per sentence basis, it's much more harder to anticipate.
I try not to. It makes things more difficult in post. The closest I'll get is a slight tweak or two, between sentences or while the opposite person is talking, because I find that people speak differently when they actually start the interview than they do when checking the mic. Those are very rare occasions.
I make sure to pay attention to what they sound like speaking before I approach them about putting on a lav (I always use one as backup). I then listen for differences during level checks, and try to predict if they'll be louder or softer once they've settled in.
Record at 24 bit, with reasonable levels, and you should be fine.
I never ever ride faders during session, though I do sometimes tweak it down if it shows to be way hotter than expected to begin with and will give a more sustained dist on several occasions during the take. In that case, often due to very spontaneous actors (which may be ever worse for your ears...), the take is ruined anyway so I might as well make the best possible of whats left of it.
In the Dark Ages (IE when DAT was something positive), it was often a real challenge to keep the voices above noisefloor but below hard clipping on dynamic actors, often including heavy compression which I've never really liked during recording anyway, but as we do have sweet sweet 24bit nowadays, it's MUCH better to keep it lower rather than higher and keep everything absolutely clean! Everything one can do during recording, can be made in post, with the difference that you can undo it there. Except keeping the material in focus, or anything else related to micking techniques, of course :-)
Only to avoid distortion, otherwise I try and keep everything the same.
I rather not touch the gain. it will increase the cups of tea in post. Leave plenty headroom -12dBFS? and use 24bit depth if possible (permits a wider dynamic range capture than 16 bit)
I would pay attention to mic placement. Probably go for a dynamic mic if the acoustics of the room are not good, also they not respond to fast transients which may cause clipping.
You could also try to get the interviewee on you side by telling him cool stuff that they may not understand about acoustics and microphone characteristics, and ask him to control his/shes dynamics. Unfortunately they tend to forget everything you told them as soon as they start talking.:)
I agree with the above - gain adjustments should definitely be kept to a minimum. However, keep in mind your environment.
You mentioned that the location was indoors with little to no ambient noise? For that, I'd set it and forget it because normalizing the signal will be fairly easy without bringing up the unwanted noise. In noisier environments it becomes much more necessary to optimize the gain on the way in.