Holy crap - this discussion is all over the place but I'll bite... (got time here at the in-laws farm on dialup
)
Preamp or not with JB3If your mics have enough output (i.e. you don't need additional gain for your peaks to approach 0 dB), then you're good running line in. Battery boxes with the popular stealth mics usually just provide bias to the mic, not gain.
If you end up with mics that require 48v phantom, then you can choose between a pre with phantom or just straight phantom (again, only if the mic output is high enough).
JB3 mic-in vs. line-inLine-in is the only option for music and the 12 dB gain is true analog and a better alternative to increasing amplitude digitally in post. See this for more:
http://ca.geocities.com/dkleined@rogers.com/audio/JB3analog/JB3_Analog_recording.htmThe a>d on the JB3 isn't very goodI think it's actually quite alright. I have a sample of a recent show where 2 of us took the same board feed - the JB3 and an M-1. I used +6dB of JB3 gain. I gotta listen again but I did some ABX testing and couldn't differentiate. The bigger issue with JB3 analog recording is the preamp section.
JB3 Analog preNo phantom and it's definitely not the quietest but in my opinion, the larger issue is electrical interference from the hard drive. If you crank the gain on the pre, you will pick up noise as the hard drive writes - a high pitch buzz that lasts a half a second or so and comes in at alternating 5 second/10 second intervals. It's not real loud unless you boost amplitude again in post. I just stealthed a show with CSBs and the JB3 at +12 dB. Signal was so low I had to add another 20 dB in post and then, during the really quiet parts between songs, you can hear the hard drive noise.
I've run a higher sensitivity mic into the JB3 analog without the same problem - it's a combo of low sensitivity mic, and a quiet show that required lots of boost after the fact.
levels and bitsYou get 6 dB per bit so if you're peaking at -12 dB on your M-1, you're really only creating a 14 bit recording
This describes the resolution (or detail). Instead of making use of any one of 65,536 possible values for each sample, you're reducing it to one of 16,384 values.
You can boost it later to occupy the higher values but you're basically just shifting the whole thing up, not adding detail. It will be most noticeable on quiet passages and decaying sounds.
levels and the a>dI would think that all a>d chips perform best when provided with full range signal. The only thing I could come up with is perhaps how nasty it gets if it clips, but that's probably the analog front end getting used to tame the overs anyways...