Join Paul Murphy for an in-depth discussion in this video Leveling dialog with compression, part of Adobe Audition: Mixing Music and Dialog.
- [Instructor] Earlier we looked at adjusting the overall loudness of each of these dialog clips, but now it's time to adjust the volume within the clip itself, and a great way to do that is by using compression. When you hear the word compression, you're probably thinking about file compression, which is used to reduce the size of a file, but in audio, compression also refers to reducing the peaks in a waveform to level out the volume. So I might just focus on this track here for now. So I'm just going to zoom in on this a little bit and I'll adjust my scroll bar as well so I can make it nice and high.
I probably don't need my essential sound panel here, so I'm just going to close that. Just give us a little bit more screen real estate. And just by looking at this file, you can already see that it's got quite a wide dynamic range, so we have some real high peaks in the dialog here, but a lot of the other speech is much softer around this level here, and this is not uncommon when you're recording people who are not professional voiceover artists. So why is this a problem? Well, firstly, it's going to limit how loud I can make the overall voice because I have to make room for these occasional peaks, but also, when it comes to combining this track with music, it means I'll have to turn the music down much lower so I can hear these softer moments in the voice.
Now what you might be tempted to do is keyframe your audio so go to your volume control here, and add a point here and add a point here and then just drag down this peak here, and while this is certainly going to solve my problem, the only issue is that this is a very manual process and it means I'm going to have to go through all of the audio clips in my session, visually identify those peaks, and then adjust them whereas with compression, it's an automated process. So it will automatically find those louder peaks for me and then turn them down.
And that's why keyframing is actually the last thing I do to dialog and only if it's something that a compressor can't fix. So I'll just get rid of these keyframes. Just Edit, Undo, Edit, Undo, Edit, Undo to get back to where we were before. Now before I jump in and start playing with compression, I like to get a little bit of information about the track that I'm working on. So I'm going to go to the amplitude statistics panel by choosing Window, Amplitude Statistics, and here it is over here. I'll just expand this so I can see it a little bit better, and this provides some incredibly useful information about clips.
The only probably is is that it's not available in the multitrack editor where I am now. I can only use this is the waveform editor, and that's why this scan button here is actually grayed out right now. So what I need to do is make all of these clips here on the track into a single clip that I can open in the waveform editor before I can get those statistics. So to do that, I'm just going to select my track, go up to the Multitrack menu, down to Bounce to New Track, and I'm going to choose Selected Track. And straightaway, it creates this new track here with a new clips on it that's a mix down of all of these clips here, including any effects that they had on them.
Now if I double-click that bounced file, it opens it up in the waveform editor, and if I go over to my amplitude statistics, you can see this scan button is no longer grayed out. I'll click this, and there's all the data on my clip. I'll just expand this across a little bit so I can see it better. There's even a handy section here for possibly clipped samples. So it will tell me if it thinks any of my waveforms have been clipped or distorted, but what I'm looking for is this number up at the very top, peak amplitude, and this will tell me what the highest peak in this clip is, and I can see that it's minus 4.63.
What I really like about this panel is if I want to go to where that peak amplitude is, I just click this little button here, and it will move my playhead to that point. So I can see that this is the hightest peak in all of these clips here. Now if I jump back into my multitrack session just by clicking the multitrack button up at the top, the playhead is actually moved to that peak so I no longer have to look for it. I don't actually need this bounced track anymore, so I'm just going to select the track, go up to my Multitrack menu, and choose Track, Delete Selected Track.
And now let's add a compression effect to our first track here, so I'll select that track, go over to my effects rack over here, and I'll just expand this a little bit so I can see it better. I'm going to make sure I'm on track effects. I don't want to add this to each clip. I just want to add it to the entire track, and I'll just go down to the next available slot down here, click the flyout menu, and go to Amplitude and Compression. This is where you can find all of the Compression effects inside of audition. And there are a lot of good compression effects that you can play with here like Dynamics or Dynamics Compression, but I really like to work with the Tube-modeled Compressor down here because it's got a very simple interface, and it has a nice sound to it.
I'm going to start with the default preset. This isn't actually making any change to my audio at the moment, and the first setting I'm going to work on is the threshold here. So think of the threshold as the loudest your audio can be before it's considered too loud. So everything that's above the threshold will be turned down, and everything below it will be untouched. So we know what our loudest peak is. That's this peak over here. How do we figure out what our threshold is? Well, one way you can come up with this is to go back to that bounced file.
If I go to my files panel over here and double-click my bounced file to open it up, over to my Amplitude Statistics, if I go to the maximum RMS amplitude. Let me just expand this a little bit so we can see that number better. Sometimes, this number here is a really good setting for your threshold, but I think I'm going to figure this one out myself. I'll go back to my multitrack session. So just looking at the waveform of my clips here, I would say that this section of dialog here is probably where the softer part is happening. So I'm just going to use my time selection tool and just select just that because I want to find out how loud this actually is, so I'll just make sure I've got a good selection of it, and then with the loop button turned on, I'm just going to loop this back and forth while looking at my audio meter down here to find out what the peak amplitude of this section is.
- [Man] Julie, Julie, Julie, Julie, Julie-- - [Instructor] Okay, so I can see it's about minus 18. So I'm going to use that as my threshold. I'll go over here to my effect and next to the threshold, I'm just going to punch in minus 18 decibels. So now, everything that's below minus 18 is going to be untouched, so this little selection here won't be compressed, but everything that's above that, everything that's louder than minus 18 is going to get turned down. The next question then is how much do I want those peaks that are above the threshold to be turned down? And that's where our ratio comes into it.
At the moment, my ratio is set to one to one, which means the loud audio is not being turned down at all. The more that I turn this number up, the more those loud peaks will be turned down. Now there is some math going on behind this number, and while it's not essential, it does help to understand what it's doing. So if I set this number to three, I'm going to set my ratio to three to one, then that means that for every three decibels something goes over the threshold, it will be reduced to one decibel over the threshold.
So if I encountered a peak that was minus 15 decibels, which is three decibels louder than the threshold, then after compression, it will only be one decibel over the threshold or minus 17 decibels. If the peak was minus 12 decibels, which is six decibels louder than minus 18, then it will be two decibels over, or minus 16 decibels. So there is a little bit of math involved here, and if you're trying to hit a specific peak level, it's good to understand that math behind the control, but otherwise, as with everything, it's find to just use your ears as the guide while also keeping your eye on the meter down here.
For something like dialog, you probably don't want to push the threshold too high. Somewhere between two and four should do a good job. Also, you'll notice if I click over this number here and drag over it to adjust it, I'm moving in whole numbers at the moment, but if you wanted much finer control, I'm going to hold down the command key that's control on Windows, and now this will actually allow me to move in decibels and let me get a much finer control over the ratio. So now to set my ratio, I'm actually going to change my selection to that area where we had that loudest peak before, and I'm going to keep my loop button on and while I loop through this, I'm going to gradually turn my ratio up.
And that peak was hitting around about minus five decibels before, so I'll be watching my meter as I turn the ratio up to see how much I can reduce that peak without it sounding overly compressed or distorted. So let's see how we go here. - [Man] And help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and help them, and-- - [Instructor] Okay, you can see why I usually ask producers to leave the room whenever I'm doing a sound mix because most of them find listening to something like that very annoying.
I'm always looping small sections and listening to fine details, but it looked like there that getting a ratio of about 3.1 was actually taking that high peak from minus five to minus nine, which is sounding much better to my ears. Now that we've set that ratio, let's have a listen to the entire track, but as I do that, I'm going to be paying attention to this meter here. This is the gain reduction meter, and this lights up red whenever the threshold is exceeded and shows me how much it's turning down the audio in decibels.
If this is red all the time, that means I've set my threshold too low and all my audio is being compressed. What I want is for it to light up only some of the time so that I know that the loud peaks are being turned down and the soft peaks aren't being touched at all. - [Man] There's so many veterans out there. If you took each veteran individually and helped them get a job, that process doesn't scale. - [Instructor] Okay, that sounds good to me, and everything looks good on the gain reduction meter. The next two settings that we have in our compressor effect is attack and release, and these are controlling how fast the compressor reacts to a loud peak, and how long before it stops compressing it, and I find that these default settings are actually quite good for dialog.
Finally, we have our output gain, and now that we've turned down parts of the track, we have the ability to turn up the entire track more. And one use for this is to hear the difference between your audio with and without compression. So that loudest peak that we turned down over here was minus five, now it's closer to minus nine decibels. So if I turn the output gain up four decibels, that gets that highest peak back to the way it was, and now if I play back my session while toggling the power state on and off, I should be able to hear the difference that the compression is making to the track.
- [Man] There are so many veterans out there. If you took each veteran individually and helped them get a job, that process doesn't scale. - [Instructor] Okay, so you can hear that it's definitely making the loudness of the dialog a lot more consistent now. It's also adding a bit more power to the voice as well. But we can also use this output gain setting to make sure that this track matches the loudness of the rest of the tracks, and there are a few different ways of figuring out how to do this, but I find that the best way is to bounce the track again. So I'm just going to set my output gain to zero, and I'll just close my compressor for now.
I'll make sure I have the track selected, and then I'll choose Multitrack, Bounce to New Track, Selected Track, and now it's created a rendered version of my track that has the compressor actually rendered into it. Now, in fact, if I double-click this, I can open it in the waveform, and I can see how much of a difference that compressor is making to the waveform. It's a much smaller dynamic range that I have now. So I'll just back into my multitrack session and what I'm going to do to match this to a specific setting is to go up to my Clip menu, and I'm going to choose Match Clip Loudness, and I'm going to set this to a target loudness of minus 23 loudness units full scale, and this is a good standard setting to work with for dialog.
So I'll click okay on that, and now if I go down here, I can see that it's actually increased the track by 1.84 decibels in order to get it to minus 23 loudness. So now if I select my dialog track back here, go over to my compressor in the effects rack, I'll double-click this to open it up, now if I set my output gain to 1.84, that means I know that this track is now minus 23 LUFS. So I'll just close that. I don't need my bounce track anymore, so I'll select it and choose Multitrack, Track, Delete Selected Track.
Important to keep in mind if you're bouncing a lot of tracks, I have deleted that track, but it doesn't delete the actual file. It stays over in the files panel. So if you are building up a lot of these and you need to clear up some hard drive space, you can just right-click on them and choose reveal in Finder or reveal in Explorer if you're on Windows, and just go over to your Bounced Files folder. This just creates this automatically, and you can delete the bounced files that you're not using or just delete the entire folder if you don't need that that all, and then I'll return back to Audition.
So that's my workflow for compressing the peaks of my audio clips and reducing their dynamic range. Now my audio will be a lot easier to listen to and a lot easier to mix with music, and if I find that compression is still not leveling out my dialog enough, only then will I start adding keyframes to my clips and do it manually.
- Fixing recording issues such as clipped audio and background noise
- Enhancing dialog with compression and EQ
- Automating music levels
- Setting the right levels for venue
- Exporting your mix back to your editing project