Understanding camera levels in audio post-production
 

On a recent documentary shoot I took Steve Mullen's suggestions and matched the input-trim function of my V1U with the mics I would be using (i.e. using an input trim value of -8db to match the ECM77B sensitivity of -52db with the camcorders input sensitivity that Steve measured at -60db.) and now that I am bringing this footage into my NLE (Premiere Pro CS3) I'm noticing my levels are around -12db. Is this because I used the input trim function, or would my camcorder have captured it at -12db regardless of my input trim settings.

I haven't bought Jay Rose's book on "Audio Post Production for Digital Video" yet, but I have recently read his book "Producing Great Sound for Digital Video". He may have brushed on it in the book I just read, but I'm still a little confused about it.

When I'm finished my editing, do I keep all the footage at -12db or will I want to have Premiere Pro (or a DAW, or whatever) raise the gain slightly until around -3db?

I assume a bit of headroom is good so that it doesn't peak 0db, but is 12db too much headroom? Is that a low signal to put onto a DVD?

TV stations have very strict guidelines for audio for broadcast material. For a couple of documentariesI made I had to make sure my dialog was from -20 to -12 dBFS level, maximum levels at -6 dBFS (all timecodes where audio hit -6 dBFS had to be listed on a separte raport) and the absolute maximum peak allowed was -3 dBFS.

I compressed the audio ever so slightly with knee at -25 dBFS and using 1:1.5 ratio above that. After doing that I ran the audio thru "analyze" function in Audition and lift the levels a bit to confirm with the requirements.

I think DVDs are not so different from broadcast, and the above rules are quite universal.

Peter is correct, in part, but note that he is writing from Europe. EBU and SMPTE standards are quite different, plus North American networks make a distinction in delivery setups for normal and HD programming. A good example of broadcast requirements can be found at PBS's technical requirements site - they're arguably the most stringent and programs that meet their technical specs should be acceptable almost anywhere. Take a look at the Producers Red Book technical delivery specs at http://www.pbs.org/producers/redbook...on_8_20_07.pdf

With digital audio a lot of headroom is very important. -9dBFS (digital, "full scale") equals 0dB analog. With analog you have a lot of headroom above 0dB, with digital, 0dBFS is the absolute end of the line. Everything above 0dBFS clips and destroys your recording. Here in Europe it is common to use -9dBFS as the maximum level. The common rule is the audio should touch -9dBFS as often as possible, but peak over it as seldom as possible (and not by much!)

You do need to be careful as to what the signal source and meter types are when making those measurments. Voice versus music versus a steady sinewave tone all with the same average levels will meter differently. The EBU standards call for 0VU to equal 0dBu average signal level and a steady sinewave tone at reading 0VU on an analog meter should read -18dBFS on a digital recorder's peak reading meter. That setting would place a voice signal at that meters the same 0VU average level analog to indicate peaks at about -9 or -8dBFS when viewed on the same digtal peak reading meter.