These days the CALM Act mandates that TV programs and commercials all adhere to a standard specification for average integrated program loudness. That spec is -24 dB LUFS, where LUFS stands for Loudness Units Full Scale. It can also be referred to as LKFS. If you’d like to dive deep into the details of this, here you go.
But suffice it to say, we have to meet this spec with the audio that we deliver. And frankly, it’s a pretty good spec; it retains a healthy amount of dynamic range potential for “brief excursions” above and below the spec, and also keeps things focused around dialog level in a pleasing way.
The most important means of meeting the spec is simply to mix to the spec. However, mixes often end up a little low or (more commonly) a little hot. So let’s look at a quick and simple method to push the mix back into spec using a component of Avid’s excellent new Pro Limiter plugin.
The Pro Limiter from Avid is a great new plugin that limits your audio in the “true peak” domain, which is another aspect of proper CALM act compliance. True peak is of course a subject unto itself, but suffice it to say this kind of a limiter is a Good Thing™. So how does a limiter help us with average program loudness levels? It’s not the limiter itself, it’s a separate component of the plugin called the Pro Limiter Loudness Analyzer.
Let’s say we have a composite mix stem for our show that needs to comply with the network’s LKFS spec. All we need to see if it’s in spec or not is the Pro Limiter Loudness Analyzer plugin, located at AudioSuite –> Other –> Pro Limiter Loudness Analyzer.
Then just select your composite mix stem (5.1 or stereo) and hit analyze. The plugin will return your integrated loudness value and true peak value.
Now all you have to do is trim your mix up or down by the difference between the measured value and -24dB. So if your mix measures at -21.9 dB LUFS, then trim it down by 2.1dB. If it reads at -25.3dB LUFS, just trim it up by 1.3 dB. The easiest way to achieve this is with clip gain.
Now, this measured value is an integrated loudness reading over time. It’s not like normalization that simply looks for the one highest peak and returns that value. It weighs all the mix that it sees, for however long you select. The point is, the length of time you analyze directly impacts the resulting measured LUFS value. The above example, for instance, is a reading of the entire episode (in this case about 23 minutes).
But sometimes one act is louder than another, so a full-length analysis like this isn’t ideal. There may still be sections that are outside of spec even after we adjust the overall level. A preferred method is to break your episode down into shorter lengths of time, and analyze them each separately.
The shortest length of time that any broadcast network should consider when checking your mix for compliance is a single act. So break your stem at each commercial break, and measure the individual acts. Then adjust them up or down as necessary to meet the required loudness spec, and you’re good to go!
This is probably the fastest and easiest way to guarantee that your mix complies with the new loudness rules that all broadcast networks must adhere to according to the CALM Act. Happy mixing!
What a handy tool. I’m surprised this hasn’t been developed earlier. I’ve always used durrough by Waves, and just eyeballed the rms while comparing it to my reference files. This is a huge step for pro tools and avid as a whole. Thank you for this blog. It’s articles like this that keep me coming back!
So, in the second act, with the fight scene/car chase/concert performance/whatever pushing the average LUFS up (say +2db), you’d use clip gain to just turn that entire act down? As the dialogue mixer levelled the dialogue consistently throughout all acts in the first instance, act 2 now has a dialogue level lower (a very noticable -2db) than the others.
So, although this is one method of creating broadcast legal CALM compliant mixes, it’s a rather crude broad brush stroke and can result in an unsatisfactory mix between acts. Much better to monitor LU throughout the mix sessions with regular checks across larger portions of material.
The downside I see with this plugin – and in fact AudioSuite and all PT implementation of LU analysis in general – is the inability to NRT analyse through the mixer processing prior to bounce. You have to render to check in non real time – which is a bit of a drag.
So, why not just monitor your LU throughout and do lots of listening of your work beginning to end? That’s the only way to be sure for me.