Voiceover Studio Tech: What Do We Gain? (A bit about setting levels)

We’ve definitely wandered out into the weeds of 32 bit recording these past couple of posts. Hopefully, you’ve gained a clearer understanding of benefits in using this type of recording. 32 bit’s ability to change the input gain settings in your audio after you have recorded something provides an incredible buffer of safety.
But, is 32 bit essential? With input gain set conservatively, and an appropriate bit depth to maintain a decent amount of dynamic range, we’re likely covered for most recording situations. Input gain is something we can get “right” with a reasonable margin of error. It does not have to be precise. It’s easy to miss that idea.
Gain “Truth” – just don’t clip
In my booth, with my voice and go-to microphone, an input setting on the interface around 2 o’clock is a good starting place for most of the work I do. That will record at safe levels when I’m being conversational. If I’m going to be loud and energetic, the gain may be dropped down somewhere around 10 o’clock on the dial.
I’ve encountered other voice actors who have been told not to move their input settings. This leads to all kinds of tonal variance as they back away from their microphones when their performance has more energy. Others have adopted a highly controlled delivery volume which can cause things to sound highly robotic.
As listeners, a natural variance of volume keeps our attention. The dynamic nature of our instrument is why we’re interesting.
How to set input gain when recording voiceover
Manually setting input gain is a necessary skill, but not an unlearnable one. Besides, on any given day, we may have a noticeably different amount of vocal energy. Some days we’re simply louder or quieter.
There are many ways to describe the guidelines setting proper gain:
When recording…
…don’t go above -6 dB on their Peak…
…generally aim for Peaks in the minus twelve-ish range…
…stay mostly “in the yellow” on Twisted Wave…
…set a level that brings in the audio “a bit above -18 dB or so.”
Each of those terms are relatively vague. That is because ultimately there’s no “magic” number for input gain – all of those rules of thumb are conservative approaches which leave enough “headroom” in our input so that getting a bit louder will still not hit 0 dB.
And we are voice actors. We will get excited. That’s what we get paid to do.
Do you need a 32 bit recording solution?
Honestly, if you are reliably recording in 24 bit with peaks generally somewhere between -18 and -12 dB on the raw record, it’s OK to stick with that approach. Nothing about 32 bit recording makes any current system obsolete. If clipping your input is a consistent challenge for you, 32 bit could be a solution. For anyone doing a lot of animation and trying to record themselves at home with no engineer, 32 bit recording takes the pressure off of the technical end of things. It’s impossible to set the input level wrong.
There are other approaches. If you aren’t good at setting input, there are interfaces such as the EVO series from Audient which will set rational input levels based on an input sample. You talk into the mic for 10 or 15 seconds at the performance volume, and the EVO engages a conservative input setting which should cover the recording. As long as your sample and the performance match, this works well. Those types of tools simply automate what engineers have always been doing in the studio – setting the input levels for the loudest part of our performance.
In the voiceover booth – Learn your instrument
Like everyone who ever stepped behind a microphone, I’ve definitely ruined audition takes because I didn’t pay attention to my input gain. But after doing that a few times, I learned where to set things so it usually doesn’t happen. And I’ve developed much more of a feel for my “instrument” – the combination of my microphone and my voice inside of my recording booth – so that I know how far I can push things.
Want to receive these resources (and more!) the day they come out?
Each week I send out a new article to my email community focused on creatively using recording technology in your home voiceover studio, and how to balance those technical challenges with the need to be brilliant behind the microphone.
If you would like to join in to receive those emails the day they publish, please take a moment to share your contact information through this sign up form.
Thank you!
If this resource has been helpful to you, please consider sharing it with one of the buttons below!
I currently record at 16 bit, as per ACX standards. I do get my specs quite on point without problem 98% of the time and since it ain’t broke, probably shouldn’t try to fix it-
….BUT, would I be better off recording at 24 or 32?
What’s the negative trade off? Does it just create larger files?
Note that it is the _delivery_ spec is 16 bit MP3. You can actually record at 24 bit in your initial WAV file, which provides a better dynanmic range in the raw recording. I recommend the higher bit depth mostly because every other deliverable these days seems to be 24 bit. You are always better off downsampling (e.g. going from 24 bit to 16 as you save into MP3) than going the other way.
It’s hard to argue that there’s any noticeable difference between 16 bit raw and 24 bit raw recording. I see it more as future proofing things.
Recording at 32 bit is a slightly different animal. Most interfaces are only capable of 24 bit analog-to-digital (A/D) conversion, so if you record into a 32 bit file, you are just writing extra zeroes. So, yeah, no benefit there.
But if you have a 32 bit “floating point” input device – either the Rode NT1 Gen5 in USB mode or one of the very few interfaces that record in 32 bit floating – then you can actually change the input volume AFTER you record with no distortion. You cannot actually clip a 32 bit floating point recording. More about that yonder – https://justaskjimvo.studio/32-bit-recording/
That’s fascinating.
I’ll read the link you included for me about 32-bit floating, but can I use my Neumann u87?
If you have an interface which utilizes 32 bit floating point technology, sure. The limiting factor is the interface.
…annnnd there aren’t a lot of 32 bit float interfaces. Even the “32 bit” interfaces (like the new SSL) does not use floating point technology.