Bumping up audio in post- bad for quality of audio?

S

soursteezle14

Member
I recently recorded a shoot and the db levels were lower than I expected. I added some volume effects (in premiere) and boosted the volume. Just wanted to know how much this kills the quality of the audio. Is it better to record lower to be sure it doesn't peak, and then just boost it in post?
 
Depends on what your recorded levels are, and how the sound was recorded. At 24 bits, sound that was recorded at little low can be brought up without artifacting. At 16 bits, the artifacts will come in sooner because you aren't really using all 16 bits... so levels below -12dB are really only using the first 6 or 8 bits, which is pretty low resolution when it comes to post processing. In 24-bit, even if you're below -12dB, you're likely still using the first 10-12 bits.

That is, of course, assuming that levels weren't drastically low. Again, what levels are we talking about?

Another thing that can happen depends on the self-noise of the sound gear and whether the signal chain was properly gain-staged. If all components are properly aligned, bringing up lower levels won't be as much of an issue... but if not, the self-noise wasn't minimized and bringing the levels up also brings up noise (hiss). And on the topic of noise, were the levels low just because of the levels set on mixer/recorder/camera, or were they low because the mic wasn't close enough or was aimed improperly? If it was the latter, then you've recorded a higher ratio of background noise to dialog so bringing up levels brings up everything that was recorded.

There are lots of variables here, so the short answer to your questions is, "Maybe, maybe not."

As for ideal recording levels, it's best to set your sound so that it records at the right level with the right headroom... and if peaks are an issue, you may have to ride gain or set a safety track or use a limiter.
 
I did set a safety track but should have cranked the original track up a bit i guess. It seemed to peak very easily. I'm using a DR40 with a Rode ntg1. It was 24 bits, it'll work fine but i'm just not good at telling if it's good audio or not. I don't quite have the ear for it yet.

If someone wants to tell me what they think of this quality, that'd be great. Also, this was a backup mic so it wasn't placed as close as the main one.

Here's a link,
http://youtu.be/_DS-2aLpfGE


Thanks
 
Recording levels too low and then boosting them is typically bad form, but sometimes you can get away with it and/or disguise it. There is a "Goldilocks Zone": not too hot and not too cold.

If you record at sub-optimal levels, you are reducing your signal-to-noise ratio (SNR) and probably imposing avoidable (and difficult-to-impossible to remove) NOISE into your recording. When you boost the signal at a later stage in the process (whether analog or digital), you boost the noise along with it. And the "hiss" type of noise typical of all electronic equipment is one of the most difficult to remove. Many times people will just declare the recording unusable and record it over again (if possible). You REALLY want to avoid doing this.

And, of course, if you set the recording levels "too hot" you run the risk of clipping on peaks. And in the digital domain, this is particularly noxious and equally to be avoided.
 
With good quality gear and clean audio it is possible to add gain even 30 dB without a noticeable effect. I had to do this once with a SD722 recorded 24 bit file and nobody noticed. With 16 bit original or lesser quality recorder (where 24 bit files have no more DR than a 16 bit file due to hissy analog front end and so-so converter) hiss might have started to show. Still, for many a decade even a 60 dB dynamic range, 10 bits in modern talk, was considered good, as there was nothing better. It is the final product which determines the real quality, if one component of the audio mix is a bit hissy, it might well be hidden and not noticed by the casual listener using home reproduction gear, even if the sound man is painfully aware of the less than stellar quality of the original file.
 
Generally changing the levels in post is "mixing". Ideally you don't want to have to boost levels in post, mostly because your boosting everything else and if it takes a lot of boost the sound will start to fall apart. Actually the sound was always bad you just couldn't hear it because it was too low. But as long as it's not a lot it's not usually a problem.

But I would do the gain when your mixing. What you don't want to do is boost it all because you think it's too low and then lower it when you mix. If you need more gain do it in the mix.
 
You must log in or register to reply here.
Back
Top