Measuring VO Volume: How Loud is “Loud”? – Tuesday Tech Tip
Following the discussion of Limiters a few weeks ago, I received a few questions about Loudness versus Peak.
When working in our home studios, it’s easy to get thrown off by the idea of “Loudness.” Part of the problem comes from the casual use of the word “loud” – I know there were times in my younger days when it was pointed out that my music was, in fact, “too loud” for other members of the household.
Before we began recording our own voices, we dealt only with playback volume. When we play audio back (or “monitor” it), we are controlling how comfortably the sound hits our hearing. If it’s too loud, we turn it down. If we need more detail, we turn it up.
However, when we are doing the recording, the key variable is the level of the input signal.
Controlling Recording Level
The level is controlled by an INPUT volume setting. Depending upon your recording chain, that might be a slider in software, a hardware dial on your USB-direct-connected microphone, or the control on your interface. It is independent from the idea of how loud it “sounds”.
I’ll typically recommend a value around -12 dB-ish as a good target on the input. However, that’s actually a bit of a shorthand. The value I’m talking about is the Peak measurement of the recording, a momentary value for the loudest single cycle of the audio wave form.
Our hearing, however, is not that reactive to “Peak” values of single waves. It tends to be comparatively slow to react (remember there are hundreds of those single wave cycles going into our ears every second for sound in the speech range, and thousands per second for higher frequencies). We use Peak values for controlling our input, but those numbers aren’t addressing the overall acoustic energy – the true “Loudness”. I’ll come back to that in a second.
Why “Peak” values work for input
There are three reasons we typically focus on the Peak value when recording VO tracks:
- Peak is very easy to measure.
- Aiming for roughly -12 dB for peaks tends to make the overall loudness fall into a usable place.
- If we hit -0 dB when recording, we have likely distorted the input by hitting the maximum input amount. Distortion is not desirable in VO recordings.
Loudness differs from Peak
When we measure Loudness, we are describing the overall energy of the audio over a specific period of time. For much of what we deliver – particularly audiobook recordings, for example – we are concerned with the overall loudness of the entire file. The most common measurement scale here is either RMS (Root Mean Squared) or LUFS/LKFS (Loudness Units Full Scale / Loudness K-weighted Full Scale). The important distinction is that those values measure the energy of audio over time, as opposed to a single waveform Peak.
For example, in the audio below, there’s a single Peak that hits -0 dB, but the overall Loudness is relatively low –
If we simply highlight and attenuate the Peak (using Amplify and giving a negative value) that fixes the single wave. We could then amplify to make everything a little louder as we see here. Even though the Peak is lower, the RMS has been increased.
Further resource: Tuesday Tech Tip – Why be Normal?
Using a dynamic control like a Limiter, we can force the RMS to be significantly higher, but control the maximum value for the Peaks. We’d end up with a “squashed” wave which is actually “louder” (the RMS is higher) even though the Peak height of the wave is lower.
If we were to listen to these three audio samples in the above order, each would sound a bit louder than the previous. That perceived Loudness is what RMS and LUFS/LKFS scales measure, by giving a numerical value to how we actually perceive sound. Loudness values – commonly specified using LUFS in most broadcast deliverables and RMS for the audiobook industry (which hopefully may catch up and change to LUFS) – are a more accurate measurement scale for what sounds “loud” to our ears.
As we can infer from the examples above, the “Normalize to -3 dB Peak” delivery approach listed on some online VO marketplace sites may not always be appropriate. A single waveform can throw that off measurement – it can sound “quieter” even though a single Peak is at a high value. It would be a welcome change if we could use a Loudness value – RMS or LUFS – as a deliverable target for auditions. Until that occurs, the question remains – what are the “right” settings for voiceover recordings?
It would be really nice if there was an actual loudness delivery spec for auditions. Though there has been a bit of interest in that idea recently, it unfortunately has not yet gained momentum. So it falls to us, as voice actors, to understand how these values can work together. Understanding Loudness measurements can help you make corrections to deliver your audio at an appropriate and competitive level.
More JustAskJimVO.studio Resources for Loudness, RMS, and Delivery Volume
Since this post originally appeared, I’ve covered a few more ideas about loudness and levels when recording in our home voiceover studios. Among the key thoughts is the idea of being “Competitively Loud” in your auditions.
Please check out these other posts –
How Loud Do They Want – understanding client’s specs –
https://justaskjimvo.studio/how-loud-do-they-want/
Commercial Audition Loudness Delivery Survey Results –
https://justaskjimvo.studio/fall-forward-and-audition-loudness-survey-results/
This information recently went out to my email community.
If you would like to join in to receive these the day they publish, please take a moment to share your contact information through this sign up form.
Thank you.
Previous Post
Next Post