Hey, so I’ve been improving my audio quality. Of course, when you record voiceover, they recommend your audio be recorded in a range of -18 to -12DBFS, with peaks no louder than -6.
I’m a very dynamic person when I speak, so I very easily hit peaks of -4 or even -3 without trying. I was told you could use a safety limiter or compressor when tracking.
Question is: if I set one up during tracking, would it be best to use the brickwall limiter or Standard Limiter?
When I hear someone say “safety limiter” I think that “safety” refers to preventing clipping and distortion. I have that on some of my outputs and groups for example, and back in the day when I used to do VO recording I had a compressor in the analog domain to prevent clipping the converter.
If you never go over 0dBFS and thereby clip your converter then you don’t need a “safety” limiter, in my opinion.
If you want to limit the dynamic range then you can do that in the analog domain or after you record. The benefits of doing it after is that it can be undone in case you go too far, and you can leave it to the mix engineer that receives your VO recording. If you are your own mix engineer then whatever sounds good and whatever is efficient works.
The idea is that you would use a separate analogue device before the input of the A/D so it never gets too hot a signal. If you clip the input the damage is already done, and a digital limiter won’t fix it.
Yes, that’s a guideline, assuming you’re working at 24-bit.
That would be fine too, as long as it doesn’t clip. Can you not just reduce the signal from the mic, or even just move it away a bit, and use a pop filter? All cheaper and easier than setting up an analogue limiter.
Whats your audio interface.
The best current audio interfaces (RME, Apogee, Prism, MOTU UltraLite mk5 -the one we use, even some SSL units) have ridiculously low EIN (Equivalent Input Noise) and huge dynamic range—often in the 118–124 dB range—which means you can safely record with peaks sitting at –12 dBFS or lower. The preamps are so clean and linear that the concept of a “sweet spot” from the analog days doesn’t really apply anymore.
Motu UltraLite mk5 has a 32 bit pipeline and the built in DSP can be setup to act as a unclipable compressor as shown live recording here
Previously we used Sound Devices 7-series recorders, which came with “unclip-able” analog limiters and were brilliant for unpredictable field work, but in a controlled VO or narration environment, modern 32-bit float converters (like those in the MixPre-II, RME ADI-2, or Zoom F-series) make them almost redundant. You get all the safety of limiter-like headroom without the coloration or intervention, and they are affordable—unlike any decent analog limiters.
For field recording, though, analog protection and redundancy still win: humidity, RF noise, mic movement, or shock pops can ruin takes that 32-bit float can’t always predict.
I use the Motu M2!
There is no such thing as a 32-bit float A/D converter. Some devices just parallel multiple converters with fixed attenuators and reconstruct the signal in the digital domain, making it appear as if you’re getting higher bit resolution, which is of course convenient for field reporters etc., but it’s not real.
I agree with the others: With how good ADCs and preamps are these days, the best idea is to just have a good ADC/preamp and record at a lower level. I recently did vocals for our Halloween show (my GF REALLY likes Halloween) and I used a Rode NT1 Gen 5 which has a built in “32-bit float” ADC. As others have pointed out, the 32-bit FP thing is kinda BS but what it actually has is just a very quiet mic pre and ADC. I didn’t adjust the gain at all and just recorded.
I was doing two sorts of speech, one at my normal speaking level, actually a bit less because I was speaking in a lower register, and one where I was literally shouting. Both came out well below 0dBFS, and both had no audible noise when boosted up to 0dBFS peak. The speech was -31dBFS peak, yet still, no audible noise. Amplifying the quiet parts (-70dB or lower) all the noise I can hear is me, the soft movement of my shirt as a move around, a slight pop of my lips, etc.
There just isn’t any meaningful amount of equipment noise, be it mic, preamp or converter, that I would be concerned about getting rid of with any sort of gain staging.
If you are just doing VO, something like the NT1 is actually an all-in-one option. There are a number of mics like it that have a preamp, DSP, and ADC inside, so it does everything in box, no need for an interface. They often have compressors, EQs, etc onboard so you can do your effects on them directly. The NT1 has a compressor, high pass filter, noise gate, and exciter onboard. There are other options out there, it was just the one I chose.
When it comes to achieving true dynamic range and making worthwhile investments, the difference is tangible. In our setup, we use both the older Sound Devices units and the Zoom F6. Personally, I find the F6 more flexible and modern in its workflow, but most clients still prefer Sound Devices — and are often willing to pay a premium for it. Both recorders, however, deliver exceptional results for dialogue and voiceover capture.
From a technical standpoint, there’s little reason to invest in an external safety limiter or compressor unless you’re deliberately pursuing a vintage or colored tone. In documentary and film work, analog coloration is generally a strict no-go; transparency and headroom are far more valuable. If you do decide to invest, a used Sound Devices is still one of the best options out there — built like a tank and engineered to last. Ours has been running flawlessly for over 15 years, and it’s safe to say it’ll probably last another 15.