Audio in Rich Media

Fixing and Enhancing Audio with Soundtrack Pro

With these tips in mind, let’s make some structural changes to our interview audio in order to improve the finished piece. First, we want to remove the interviewer’s voice so we’re left with just the responses. Next, we want to remove some stutters and a few superfluous sounds like “um” and “you know.”

By cutting out these elements, we’ve reduced the total length of the audio file by more than 30 seconds, from 2:23 minutes to 1:52. Hear the results.

But we know the final project has to be less than two minutes long, including a narration, ambient sound, and music — so we’ll prune some more. We’ll look for nonessential phrases that might be cut to make a more compelling piece without changing the overall meaning of what the subject has to say. The resulting interview file clocks in at 1:19, which should work well.

Adjusting Levels

If the overall volume level of an audio file seems a bit low, you can increase it using Soundtrack Pro’s built-in tools — for example, the Normalize function. Rather than simply raising volume, Normalize analyzes the entire selection, locates the loudest peak, and calculates volume gain from that point, thus guarding against digital distortion.

Select the audio you’d like to make louder and choose Normalize under the Process menu. In the resulting pop-up window, move the Normalization Level slider to adjust the amount you want to increase gain. Before clicking OK, play the file and look at the audio level monitors at the right side of the interface (or in the Meters window in the upper left corner) while the audio is playing.

Soundtrack Pro defaults to a Normalize level of 0 dB — that is, it boosts the highest peak of the audio signal to the loudest possible level without distortion. However, you may get better-sounding results by reducing this setting to at least -2 dB, especially if you intend to incorporate this file into a mix with other audio sources. The level will still be plenty loud, and you’ll avoid possible distortion when this audio is combined with other sounds in the mix.

To change the gain level of one word, select it and choose Adjust Amplitude in the Process menu. The reduced gain level is reflected in the height of the waveform.

Sometimes just one word or syllable in a phrase is distractingly loud. One possible fix is to lower the level on just that syllable. As with cutting and pasting, this works best on isolated sounds that don’t flow directly into other words. Select the offending syllable and choose Adjust Amplitude under the Process menu. In the pop-up window that appears, move the amplitude level slider to the left to lower the volume of the selected sound. Modest level reductions — no more than a few dB — work best to retain a natural sound.

Decreasing Sibilance

Now it’s time to make this audio sound better. Our interview subject has a lot of sibilance (those prominent “s” sounds). We’ll try to reduce that hissing quality with a de-esser. This is a standard audio effect (basically a specialized type of EQ) that’s widely used in broadcast, music production, and other fields. It targets the specific high frequencies where “s” and “t” sounds are most noticeable and reduces them to a more manageable level.

To apply the de-esser effect, click on the Effects tab above the Browser window and select the Soundtrack Pro category. All the effects that come with Soundtrack Pro appear in a list. Double-click on DeEsser, then click on the Advanced button that appears in the Parameter column below. A small interface for this effect pops up. Clicking on the Show Presets button displays a number of options to experiment with. We’ll try the Vocal Female Soft DeEsser, which sounds pretty good — it takes the edge off the sibilance in the interview and makes the whole track sound more appealing.

A de-esser can decrease annoying sibilance in vocal tracks.

There’s one more step to make sure this de-essing effect is applied correctly. Under the Process menu, select Bounce Realtime to Action. The effect now appears in the Actions list, and plays back with the audio file.

The Soundtrack Pro library includes numerous other effects to enhance your audio. Try experimenting with effects like EQ, compression, and noise gates to see how they might improve the sound of your files.

Reducing Noise

Noise reduction is a double-edged and very sharp sword. There’s no such thing as “removing” noise — the best you can hope to do is make it less noticeable. Generally speaking, the more noise you take out, the more artificial your remaining audio sounds. Some types of noise are impossible to reduce by more than a small amount, if at all, without compromising the sound quality and intelligibility of your entire audio file. If you’re editing destructively, keep in mind that you’re permanently removing certain frequencies from your final sound — there’s no adding them back later.

Moderate noise reduction settings usually produce the best results.

Even though we recorded in a fairly quiet environment, there’s some background noise on this interview that we want to try to reduce. Step one is to select a few seconds of the room sound by itself, at the very end of the interview. With this bit of audio highlighted, go to the Process menu and choose Set Noise Print. Then select the entire audio file, return to the Process menu, and select Reduce Noise. In the dialog box that appears, we can set the level and amount to reduce the frequencies we initially highlighted.

We’ll keep the noise reduction settings at a modest level: a noise threshold of -50 dB with a reduction of 70%. Since this voice already has plenty of high-end emphasis, we’ll set the noise reduction to retain a bit more bass. (A good rule of thumb here is that enhancing bass adds warmth while enhancing treble adds intelligibility.)

In addition to the Set Noise Print/Reduce Noise settings under the Process menu, Soundtrack Pro provides a series of analysis tools that can help reduce noise, including scans for Clicks and Pops and Power Line Hum. To access these settings, click the Analysis button above the Actions list. Again, keep in mind that more processing often means less organic results. See if you can fix problems with less extreme settings first.

The Final Edit

The sequence in which different effects and processes are inserted can affect the final results. Fortunately, you can rearrange the order of individual actions in Soundtrack Pro’s Actions list. For example, we might decide that the de-esser sounds better when it’s added after the noise reduction, so we simply grab that action and drag it beneath the noise reduction action in the list. The ability to make adjustments like this is one of the benefits of nondestructive editing in Soundtrack Pro.

We’ve deleted the interviewer’s questions and other nonessential audio, removed “ums” and other space-fillers, cut out some phrases to make the interview shorter and more interesting, copied and pasted individual word sounds to improve flow, adjusted the level of a word or two, reduced sibilance and background noise, and checked for pops and clicks. Our edit is complete — listen to the final file.

The Clicks and Pops analysis screen shows areas that can be enhanced through processing.

These are just a few of the ways you can enhance and correct audio files in Soundtrack Pro. In addition, you can time-stretch or time-compress audio, use the Pencil tool to edit individual samples, set markers at specific points you want to work with, and instantly jump between them to speed your workflow. You can also move forward and backward between selections you’ve made on the audio waveform. And there’s a whole world of processing options in Soundtrack Pro’s built-in effects library. The application also comes with a huge library of instrument sounds, musical phrases, and sound effects that you can add to the final multitrack mix.

Now that we’ve sliced, shaped, and pruned our audio file into good shape, it’s time to move on to the next production stage: combining the interview with a narration, ambient sound, and music in Soundtrack Pro’s Multitrack Project view. (We’ll cover this in the next article [link to “Audio for Rich Media, Module Five” article] in this series.) This is where things get really fun. And thanks to all your hard work on this audio file, you won’t have too much to fix in the mix.