Have you ever recorded your guitar, only to realize afterward that the signal was hitting the red on your DAW or software?
Or have you ever had to crank up the volume of your preamp to record whispers, making the background noise too audible?
If so, it’s probably because you’re not recording at the right level.
Indeed, there are “best practices” regarding setting the recording level. Not following them puts your mix quality at risk in a way.
Explanations…
If you prefer videos, you can find a complete explanation here:
Once Upon a Time in Analog Recording
Until at least the early 80s, most recordings were made using analog equipment, specifically on magnetic tape.
With these tapes, it was necessary to maximize the Signal-to-Noise ratio. Therefore, recordings were often made as loud as possible, close to “the red zone.”
This limit, concretely, corresponds to the 0 VU measurement that can be read on analog VU meters (like in the image above).
If the signal exceeded this value and went into the red zone, a rather warm saturation and, in any case, a progressive intensity appeared.
This undoubtedly contributed to giving what we now call a “vintage” quality to recordings of that time.
Clipping
Today, the situation is completely different as the operation of digital recording is quite different.
First of all, the digital signal is measured in dBFS, a unit that actually corresponds to the difference between the signal level and the highest signal that can exist.
It has nothing to do with the 0 VU measurements we mentioned earlier.
In other words:
- As long as your signal stays below 0 dBFS, there’s no problem and no coloration, except for that of your preamp / processing chain
- If the signal exceeds 0 dBFS, it will not be reproduced when listened to (for example, through headphones or speakers, once the digital-to-analog conversion has taken place).
So, what will happen?
Well, the sound will clip.
This means that any signal exceeding the 0 dBFS threshold will be simply truncated. The consequence is a strong distortion devoid of any musicality, unlike the distortion found on analog equipment.
So, if you record a sound whose signal exceeds this limit, it will be destroyed by distortion and generally unusable.
An example will clarify this — here’s an acoustic guitar recorded too loudly:
It’s horrible, isn’t it? 🙁
Moreover, we can clearly see that the shape of the signal we obtain on our DAW is truncated at the peaks due to the clipping phenomenon:
At What Level Should You Record Then?
Perhaps by trying to imitate analog recording, many people record their voice or guitar aiming for a level of 0 dBFS.
The problem is that doing so makes it very difficult to ensure that the signal won’t clip.
Indeed, even if you don’t hear it during tracking, it’s very likely that one or two peaks exceed the critical threshold of 0 dBFS. And when you notice it during mixing, it will be too late…
Moreover, by trying to reach this level, you risk overloading your preamps. If you’re using entry-level or mid-range equipment, this will likely degrade the sound quality.
Indeed, if you record in 24 bits (which you should do), the dynamic range available is wide.
No need to make high-level recordings.
On the contrary, it’s a much better practice to align with an average RMS level between, say, -12 and -24 dBFS.
As for the peaks, they will naturally be a bit higher — but as long as they don’t exceed the famous 0 dBFS, there’s no problem.
The consequence of all this: your recordings may seem less loud than before (if you were used to recording close to 0).
And that’s normal: once the mixing is done, it’s during mastering that you will raise the overall level of your mix to achieve, at will, the volume found on commercial CDs. And it’s only at that moment that the peak level will be close to 0 dBFS.
Good luck with your recordings ;)!