About audio levels (mic sensibilty) to be recorded and edited.

Hi,

I have made some tests on speech record from interviews. I am using a recorder + lav microphone. The recorder provides two audio levels (mic sensibilty).


a) I am not confident about setting the recorder to either low (1) or medium (2). With 1 I catch less noise/ambient sound than with level 2, but the volume is lower. I am thinking on recording in level 1 to avoid the noise (the interviews will happen in offices, then external voices and phone rings, even air conditioner or car noise from street would be minimized). Then I would normalize/increase the volume through post audio edition (some electronic hiss appeared after normalize but by using a filter I removed the hiss and the sound is clear). Does it make sense? (Attached there are the wave forms before and after treatment=> mic_level.png).

b) On the other hand, the low level (1) of course provides a low wave form than level (2). Is it supposed that the level 1 provides less audio information than level 2? I mean, theoretically, should it be better to record the audio louder than record lower+normalize?

Thanks in advance,
Cadu

[Carlos Silva] "a) I am not confident about setting the recorder to either low (1) or medium (2). With 1 I catch less noise/ambient sound than with level 2, but the volume is lower. I am thinking on recording in level 1 to avoid the noise (the interviews will happen in offices, then external voices and phone rings, even air conditioner or car noise from street would be minimized)."

Your best recording possible will be maximizing the digital levels... you want your peaks to be close to 0dB (eg, full scale), but never over. If you have a real 24-bit recorder and intend to deliver 16-bit audio, it's easy and safe to leave a few dB of guard at the top of the scale.

For "run and gun" type video shooting (usually on-camera mics), I use a single mic going to a stereo channel, with the gain up on one channel and down on the other. This pretty much guarantees I'll have good levels on at least one track, no matter what happens.

Recording at a lower level really won't help you at all. You'll have a higher level on background noise, sure, but it'll track with the subject. Lower the volume in Vegas, and you'll have exactly the thing you would have had with the lower recording volume.. only, with less system noise.

Also, with the higher levels, you have a higher quality signal that may hold together through a bit more manipulation. Background noise can be removed or at least minimized in various ways: noise reduction tools like Sony's plug-ins, EQ, etc. My rule for recording -- get as much useful information when recording, as full a level as you can on each track (I do 16-track recordings from time to time), and you have done as well as you can during the performance. Fail at this, and you'll inevitably have more work to do in editing, and likely a lower quality result.

-Dave

I always record at 96kHz/24-bit with my Zoom or my Tascam. There's little reason not to... memory is cheap and plentiful. Sure, it's not always necessary, but it doesn't hurt.

And like many one-man production crews, I'm usually doing multiple things at once. If I'm shooting a video will all guns blazing, that means three camcorders (well, one is an HDSLR) and I'm running at least one field recorder, like the Zoom H4n, if not several, or the multi-track. I probably only need 16-bit, the end result is likely going to DVD + Blu-ray or online or CD or all of the above. But you can think of your 24-bit channel as a 16-bit channel with an extra 48dB of overhead, if you're not really planning for a 24-bit output. If you can't monitor audio the whole time, this can make a huge difference in the end result.

-Dave