Mixing 101: Peak vs RMS

Hey, folks. Matthew Weiss here — weiss-sound.com, theproaudiofiles.com, and mixing101.co.

This is going to be a quick crash course in peak level versus RMS level. What they are, why we care, maybe why we don’t care, here we go.

[kick drum]

Alright, so we’ve got a kick drum, and it is peaking at I can’t see it, oh, minus 3.7 decibels full scale. So what does that mean? Well, that means that the absolute loudest signal — or part of the signal in our kick drum — is hitting minus 3.7dB Full Scale.

Why do we care about this? Well, no matter what kind of system we are going through, that system is going to have limitations and characteristics.
Now, in a digital system, we have what is described as a very linear system, and what that means is that if something is peaking at minus 20 dB Full Scale, it’s still roughly the same as something that’s peaking at minus 3.7dB Full Scale.

However, once we hit 0dB Full Scale, all of that changes. That’s what’s called the digital ceiling, and once our signal goes above that, “quish.” It gets hard clipped, and suddenly we have distortion, and it’s absolute square wave distortion at that.

So why do we care about the peak level in a digital system? Well, it’s because this is telling us we’re very close to clipping right now, and we probably need to back off the level, so if I want to say, I don’t know, add a bunch of top end to this pretty dark sounding kick drum, well, I’ll be able to do it without having to clip the signal or accidentally clip the signal.

Okay, now let’s talk about RMS.

[kick]

Now, my RMS is coming in at about negative 19 dB Full Scale. That is significantly quieter than the peak signal.

Well, that actually makes a lot of sense, right? Our peak signal is telling us what the absolute loudest point in the overall level is, whereas our RMS is actually telling us the average signal level over time, and if we think about that, well, that’s not going to just include the very top of the attack of the kick drum, which is the loudest point, but it’s also going to include some of the sustain and the decay, and it’s going to take all of those different levels and factor it all together and come out with an average.

So it makes sense that when we have our very top, top loud signal being negative 3.7, but maybe out bottom signal is down at like, negative 40 or something like that, because we’re getting into the release of the drum, then we’re going to come up with an average that’s much closer to the middle of that.

Now, here’s something that we need to note about RMS. RMS is note just a function of level, but it is also a function of time. Watch what happens if I speed up the response time of my RMS signal.

[kick]

My RMS level jumps five dB! Well, that makes sense, right? I’m taking basically a shorter window of time, which means that the kick is not decaying for as long as the signal is being measured.

This is giving us a higher RMS level. So this is one of the things about RMS. I said, you know, this might be a crash course on why we care, but also maybe why we don’t?

The thing is that when we’re looking at a limiter, or we’re looking at a meter, all of those meters and limiters and things can be calibrated differently, and they might have different response times, so if you were to say pull up the Slate FGX versus the FabFilter Pro-L, and you used the exact same amount of compression on both, well, we might not get the same RMS reading, because they are using different meters.

This is why it’s not really a perfect system. Oh, the other reason it’s not a perfect system is because RMS is effectively telling us a barometer for how loud something will be perceived.

If peak signals can happen over the course of half a millisecond, we might not even register that in our brain as actual loudness, but an RMS level, that’s something that takes place over a longer period of time, and will be more akin to how loud something is perceived, except it doesn’t factor in the fact that we are subject to the Fletcher-Munson curve, which means that different signals at different frequencies are perceived with different apparent loudnesses.

What does that mean without the jargon? Something that’s bassy and dark versus something that’s bright, even at the same signal level, the bright one is going to sound louder.

So RMS is not going to include that. But, it will give us an idea of say, how compressed something is, for example, because the more sustain that is happening in a signal, the larger the RMS value is going to be.

So if we start compressing everything and squishing it all towards the top, that means that the louder signals are sustaining for longer, and we’re getting a higher RMS value.

So it’s not maybe the most accurate measurement of perceived loudness — it might be a good indicator or a good indirect measurement of that, but it’s a fairly good, accurate measurement of how compressed something may be.

Alright, guys. Hope that you learned something. Until next time. Take care.

Matthew Weiss

Matthew Weiss

Matthew Weiss is a Grammy nominated and Spellemann Award winning audio engineer from Philadelphia. Matthew has mixed songs for Snoop, Sonny Digital, Gorilla Zoe, Uri Caine, Dizzee Rascal, Arrested Development, 9th Wonder, !llmind & more. Get in touch: Weiss-Sound.com.
Smiley face
Recommended