Apple Soundtrack Pro Tutorial

As with video, it's hard to capture audio perfectly. Fortunately, Apple Soundtrack Pro makes it easy to correct many common problems. In this article, I'll detail how to remove pops and clicks from your audio file, boost volume via normalization, and remove background noise.

Figure 1. Getting from Final Cut Pro to Soundtrack Pro.

Note that although Final Cut Pro allows you to perform some of these functions on the timeline, I recommend working in Soundtrack Pro because it provides better tools and a superior interface. To send an audio file from Final Cut Pro to Soundtrack Pro, click the audio file on the timeline, right-click, and choose Send To > Soundtrack Pro Audio File Project. Select a name for the file (or accept the default given) and click Save, and Soundtrack Pro will load the project file. Edit the file as detailed below, and once you save it and close the Soundtrack Pro project, Final Cut Pro will automatically update the audio file on the timeline with the edited file.

If you'd like, you can work along with me by dowloading the Rod_audio.wav file from here.

Let's jump in.

Figure 2. Soundtrack Pro's Clicks and Pops remover.

Pops and clicks have an annoying habit of appearing in many audio tracks, so it's no surprise that Soundtrack Pro has an automated function for removing them. Unfortunately, in my tests (and attempted uses), Soundtrack Pro doesn't identify all artifacts, produces lots of false positives, and doesn't resolve many of the problems that it does find. Perhaps your luck will be better than mine, so I'll detail how to use the tool and then move on to the technique that I actually use most of the time.

Related Links

Apple Logic Studio Review The release of Apple Logic Studio a couple of years ago heralded a complete graphic redesign of Logic; updated instrument and effects plug-ins; tons of new content; bundled applications MainStage, Soundtrack Pro and WaveBurner; and a variety of enhancements... Apple Final Cut Pro 7: First Look Review By now you know that Apple has launched an update to Final Cut Studio. Here's a look at the newest features.... Apple Final Cut Pro Color Correction Of all the skills necessary to successfully edit video, one of the most important is brightness and color correction

To access the Clicks and Pops tool, click the Analysis tab in the left pane, check the Clicks and Pops checkbox, and click Analyze. Soundtrack Pro will analyze the file and identify the clicks and pops. You can attempt to fix all the problems found by clicking the Fix All button on the bottom of the left pane, but you should preview before you fix the problem to make sure you don't erroneously repair a false positive.

To preview, double-click the problem in the pane, and Soundtrack Pro will isolate it in the waveform file. Click Play (or press the Space bar) to listen to the problem, and click the Fix button if you want Soundtrack Pro to attempt to remove the noise. If the results are inadequate, try the following.

Figure 3. Here I'm selecting an audio segment with no other content, and telling Soundtrack Pro that it's ambient noise.

Zoom in to the artifact in the .wav file as shown in Figure 3. Then select a portion of the audio file that contains no speech or music, and click Process > Ambient Noise > Set Ambient Noise Print. This tells Soundtrack Pro that the chosen segment is ambient noise. Then, highlight the pop or click, and choose Process > Ambient Noise > Replace with Ambient Noise (see Figure 4). Soundtrack Pro will replace the selected area with the ambient noise.

By replacing the artifact with adjacent ambient noise, the fix should be unnoticeable to your listeners. Had you just reduced the volume of the click or muted the area (Process > Insert > Silence), many of your listeners would likely have noticed the dead air.

Figure 4. Replacing the pop with the ambient noise.

Of course, this example is a simple one, because the problem is well isolated on the timeline and not located within speech or music. When pops and clicks are located within speech or music, you can still use this technique, but you have to be more precise: zooming in close, choosing small regions very proximate to the problem area, and applying the ambient noise only to the problem area. And previewing carefully, of course.

Figure 5. Anatomy of an audio file with inadequate volume.

One quick glance at Figure 5, and you instantly know that volume is inadequate. I'm sure that it's never happened with video that you've shot, but if someone hands you a file that looks like this, here's what you do.

Figure 6. Anatomy of an audio file with peaking.

Your first impulse would be to just boost the volume (Process > Adjust Amplitude, choose a level, and click OK). The problem is, you have to choose a level, and if you choose too high, you could introduce distortion into the audio file in the form of peaking, where the tops of the waveform are jammed against the max upper and lower levels. That's what you see in Figure 6. Obviously, if you choose a level that's too low, you haven't fixed the problem.

Figure 7. Soundtrack Pro's normalize dialog.

What you want is a tool that increases the audio file to the maximum level possible without creating distortion, which is what normalization does. To access the normalization tool, choose Process > Normalize, and you'll see the dialog in Figure 7.

Figure 8. Anatomy of a successfully normalized file. Looks great, no peaks!

You have two options: Peak, which I'll use for this file, and RMS, which I'll cover in a moment. With either option, you set the target level for the adjusted volume, or 0.0000dB in the figure, which is essentially maximum volume and the target that I typically use. If you'll be passing the audio on to someone else, use -.05dB, which takes the volume to about 5 percent from the max and gives them some headroom. I used 0.0000dB to produce the file shown in Figure 8.

Figure 9. How do you normalize a file with varying volumes?

As you can see, you have one peak on the lower right barely touching the 0dB volume level, while the rest of the clip looks a good deal more healthy than Figure 5 and without the clipping in Figure 6. As mentioned, that's because when using the Peak volume, normalization boosts the audio volume of all regions in the file by an equal decibel level, but only enough to increase the audio volume of the loudest segment to the selected target.

Figure 10. Normalizing to a peak value doesn't work well when your source file has segments with varying volumes.

That works well in the first clip because the volume is uniform, but what happens when you have a clip like that shown in Figure 9, where you have different sections with different volumes. To set the scene, this is from a project involving multiple speakers, the PA announcer, and several dancers on stage. The PA announcer is at the start and end of the waveform, and the microphone's volume is much higher than that of any of the other speakers. If you normalize to peak values, you'll get a waveform that looks like Figure 10, where the PA announcer is boosted to adequate levels, while the volume for the speakers on stage is still too low.

Figure 11. Normalizing using the root mean square option.

In cases such as these, consider normalizing using RMS (root mean square), which adjusts the average volume of all regions in the audio file.

Figure 12. Using the RMS values produced some peaking.

Be careful with this adjustment, however, as it can introduce clipping in some instances, as you can see in Figure 12 when I used the default values. You can undo and try to find a value that boosts lower levels adequately without causing peaking at the loud end, or consider splitting the audio file into the relevant segments in Final Cut Pro and normalizing them separately.

Figure 13. Normalizing different segments in Final Cut Pro.

That's what I'm showing in Figure 13. Use the Razor tool to split the audio file as necessary, then select each audio clip and choose Normalize > Audio > Apply Normalization Gain to see the control shown in Figure 13, which works pretty much the same as Soundtrack Pro's peak normalization adjustment.

The only bummer is that Final Cut Pro doesn't adjust the waveform to reflect the volume increase, so you're flying a bit blind, but if you listen to the audio file before and after, you'll definitely hear the difference.

So that's normalization. Now let's move onto our final function, noise reduction.

Figure 14. Identifying background noise for Soundtrack Pro.

Soundtrack Pro's noise-reduction function attempts to remove background noise such as microphone hum and other similar noises from your audio track. I say attempts because it doesn't work with all types of background noises. In addition, unlike normalization and our ambient sound switcheroo trick for pops and clicks, both of which are fairly mechanical and introduce distortion into your clip only if you screw something up, noise reduction is subjective. Careless application will produce a clip that sounds worse than your original clip.

For these reasons, you should never count on fixing background noise in post. In addition, you should do your best to capture adequate levels because large volume adjustments also boost background noise. You can see this in Figure 14, a closeup of the file that I normalized back in Figure 8. When you zoom in, you can see the bushy area around the waveform centerline, which typically represents background noise.

Soundtrack Pro has a two-step noise reduction workflow. First, you identify the noise for Soundtrack Pro, and then you remove it. To identify the noise, select a region in the file that contains only background noise, like the selected region in Figure 14, and then choose Process > Noise Reduction > Set Noise Print.

Figure 15. Soundtrack Pro's Reduce Noise dialog.

Then, double-click the waveform to select the entire file, and choose Process > Noise Reduction > Reduce Noise. Soundtrack Pro opens the dialog shown in Figure 15.

My first step is to listen to the sound being removed, because that's the best indicator of whether the adjustment will produce distortion. To do this, click the Noise Only checkbox, and then click Play. While previewing, I adjust the Noise Threshold and Reduction parameters until I hear only the noise that I want to remove and none of the audio that I'm trying to preserve. In this case, if I heard Rod's voice during preview, I'd back off either the Threshhold or Reduction until I can't hear him at all.

Then I'd uncheck the Noise Only checkbox, and play the audio file again, toggling the Preview Effect Bypass button on and off to hear the before and after file. If the audio doesn't sound metallic or otherwise distorted, I click Apply and move onto my next edit.

As a side note, there are multiple schools of thought as to when you should apply noise reduction. Specifically, some producers apply noise reduction before normalization, some after, and some both before and after. I generally normalize first, then apply noise reduction, but you should see what works best with your source audio.

So that's the audio side of the equation. As with brightness adjustments and color correction, these three audio techniques can go a long way toward improving the quality of your work in a pretty short time.