Automatically Improve Your Audio Levels

Whether you use Adobe Premiere Pro CC or Apple Final Cut Pro X, one of the challenges we all face is getting our audio levels to sound nice, loud and consistent – without spending all our lives drawing keyframes and adjusting levels manually.

BACKGROUND

There are three audio rules that are true for all media applications:

1. Audio levels are additive, the more audio clips playing at the same time, the louder the total mix will be.
2. Audio levels must not exceed 0 dB for any project that you export. To do so will cause the audio to distort and distorted audio sounds terrible.
3. For every 6 dB drop in level, the perceived audio volume is cut in half. This means that we want to get our audio levels as close to 0 dB as we can, without ever going over.

Adding to this challenge is the fact that most professional audio engineers mix their projects based upon average audio levels (measured in LKFS), while the audio meters in both Premiere and Final Cut display audio levels as peaks (measured in dBFS).

So, while an audio engineer will talk about audio levels between -16 and -24 (LKFS), video editors talk about levels from 0 to – 6 (dB). The fact that these two numbers don’t match – but describe the same volume levels – causes all kinds of confusion!

To keep things simple, if you are mixing projects for the web, try to have your audio peaks bounce between -3 and -6 dB, as measured on the audio meters in your NLE. This equates to an average level around -16 LKFS. This is both loud enough to be clearly heard, yet not run the risk of distorting during output.

NOTE: Mixing for broadcast, cable or digital theater has more stringent levels. Professional audio help is encouraged for these projects.

THE CHALLENGE

How can we get our levels to meet these specs, when we don’t have a lot of time to mix the audio and without risking distortion?

There are two audio tools (filters), one in Premiere and the other in Final Cut, that do exactly that. I use one or the other of these in virtually every project I create.

Both of these filters use a form of “compression.” Unlike data compression, which makes files smaller, audio compression decreases the amount of dynamic range in an audio clip. Dynamic range is the distance between the loudest and softest portion of the audio.

Both of these filters raise the softest passages, without raising the loudest passages. This is similar to a color corrector that raises the shadows without changing highlights.

In video, if all the gray scale values are the same, you’ll see a very foggy image. However, with audio, having a consistent volume level means that all your speakers are equally easy to hear, even for speakers recorded at significantly different levels.

NOTE: While there are good creative reasons for variations in volume, most often in drama, this article assumes that you want the audience to easily hear what is being said.

While each of these filters can be used stand-alone or as part of a bus (Premiere) or compound clip (Final Cut), in this article I’ll illustrate how to use each of these filters by applying it to an individual clip. The operation is the same when used in a bus/compound clip, but creating the mix is a bit more complex.

ADOBE PREMIERE PRO CC – MULTI-BAND COMPRESSOR

Whether you mix in Premiere or Audition, the operation of this filter is the same, though the default settings are not.

In the Effects panel, apply Audio Effects > Multiband Compressor effect to the clip. This effect can be applied equally to stereo or mono clips.

Select the clip, then, in Effect Controls, click the Edit button inside the Multiband Compressor settings.

Next, don’t panic!!

This is audio playing in the Multiband Compressor and it looks REALLY scary!

While you can mess with every setting, here’s what I recommend to make most conversational audio sound great.

In the Presets menu at the top, select Broadcast. The rest of the presets are designed for music, which is an entirely different art.

Unfortunately, the Premiere version of this filter defaults to doing absolutely nothing. So, we need to change some settings.

NOTE: During this exercise, you’ll create your own custom preset, which you’ll use for all future applications of this effect.

First, change the Margin setting to -3 dB. This means that the filter will not allow any level in that clip to exceed -3dB.

Next, uncheck the Brickwall Limiter. When this is on, the filter will absolutely, positively clamp all audio levels in the clip so that they don’t exceed the Margin setting. When this is off, the audio levels may gracefully exceed the margin from time to time. While you can leave this setting checked, I prefer the sound of audio with this limiter unchecked.

Next, use this screen shot to match your numbers to mine. The ones you need to change include:

• All the Crossover settings just below the graph
• All the Thresh settings
• Gain for sections 1 and 4
• Ratio for sections 1 and 4
• All Attack settings

Then, when all settings are entered, go back to the Effect Controls panel, right-click the name of the filter and save this as a preset, using a name you’ll remember in the future.

Saved presets are found in Effects > Presets. Use this custom preset whenever you want to apply this filter to a clip.

NOTE: This filter is SO useful, that going thru these steps will be very helpful. It will be even MORE helpful when Adobe brings the Audition presets into Premiere.

WHAT THIS DOES

What this filter does is divide all the audio in a clip into four regions (from left to right):

• Frequencies below normal human speech
• Low frequencies in human speech (think vowels)
• Higher frequencies in human speech (think consonants)
• Frequencies above normal human speech.

The regions related to speech are boosted more than other frequencies and by different amounts.

Play the clip and turn this filter on and off and you’ll instantly see why this is such a help in doing mixes. Most of the time, I never need to set keyframes for levels.

APPLE FINAL CUT PRO X – LIMITER

Final Cut has a similar filter: the Limiter, though it is much less intimidating.

Select the clip(s) to which you want to apply the filter. Then, in the Effects Browser, scroll down to the Audio filters section, then apply Levels > Limiter.

NOTE: There are two versions of this filter, choose the one in the Logic section, not the Final Cut Limiter filter.

Open the Inspector, click the Audio icon at the top, then scroll down until you see the Limiter filter.

Click the almost-impossible-to-see icon on the right side of the filter.

This opens the Limiter interface.

Here are the settings you need to change:

• Set Output Level to -3 dB
• Set Release to any number around 500 ms
• Leave Softknee turned on (default)
• Leave Lookahead to 2ms (default)

Then, play your clip and adjust the Gain slider until you see the blue bar at the top bouncing around 1 – 4 during the loudest sections of the audio.

You are now set. You can save this as a preset if you wish, but, since the amount of the Gain slider changes for each clip, I’ve found it easier to set this manually for each clip.

NOTE: For projects with lots of clips, it is much better to apply this filter to a compound clip containing all your dialog. Here’s an article that explains more.

WHAT THIS DOES

What this filter does is raise the level of all the frequencies in a clip such that the maximum level does not exceed the level you specify in Output Level.

In this example, we are “limiting” the loudest portions of the audio to -3 dB. The filter will remain “on” for 470 ms after the last sound ends; this prevents “pumping” where the filter turns on and off quickly, making background noise levels jump. Larger numbers are better here.

The Lookahead determines how quickly the filter will respond when new audio appears. 2 ms is pretty darn quick.

Finally, the Gain setting tells the filter how much to increase the playback level of the clip; here, I’m raising it 11 dB.

The Gain Reduction bar at the top – which is just magical! – indicates when the instant-by-instant gain of the clip exceeds the limit of the Output Level and by how much.

By adjusting the Gain slider so that we are “throwing away” some of this excess volume, we can make our clip nice and loud without introducing any compression artifacts.

SUMMARY

Even without understanding how these filters work, if you use my settings, your audio will sound better, louder and you’ll reduce the amount of time you spend setting volume keyframes on dialog.

In other words, these can save you a TON of time on each of your projects.

---

Bookmark the permalink.