Setting Proper Audio Levels

When you work with audio, you need to make sure you set proper levels at each stage of your production:

Setting Levels for Capture

When you capture digital audio, you usually cannot make level adjustments because an exact copy of the digital information is transferred to your hard disk. However, if you are capturing analog audio using a third-party audio interface, make sure you set each input channel so the meters in the Clip Settings tab of the Log and Capture window match the audio meters on your video or audio device. For more information, see Capturing Audio from Tape.

Detecting Audio Peaks

When you capture audio, clipping occurs if any part of the audio signal goes over 0 dBFS. Because 0 dBFS is the maximum digital level possible, all levels that would have been above 0 dBFS are set (clipped) at 0 dBFS. Because of the nature of digital audio recording, such clipped audio typically results in a crackly, brittle sound. Excessive peaks indicate that your audio was recorded at unsuitable levels.

If your program has peaks in the audio, you can either recapture the audio at a better level or edit the audio appropriately to avoid the peaks. You can use the Mark Audio Peaks command to identify audio peaks in your clips. It’s then up to you to decide whether to not use those sections of audio or rerecord them.

Note: Final Cut Pro considers your clips’ audio levels when analyzing levels. For example, if you set a clip’s audio level to +12 dB, audio peaks may be detected. However, if you reset the audio level to 0 dB, audio peaks may no longer be detected.

To find and mark audio peaks
  1. To detect peaks in a clip, do one of the following:

    • Select one or more clips in the Browser.

    • Open a sequence clip from the Timeline.

  2. Choose Mark > Audio Peaks > Mark.

    A status window appears with a progress bar showing how much of the process is complete. Markers are placed at each peak.

    • If you selected a clip in the Browser: Markers appear for the clip and are labeled “Audio Peak N,” where N starts at 1 and increases, depending on how many audio peaks are detected. These markers also appear in the Viewer when the clip is opened there.
      Figure. Browser window showing audio peak markers for a selected clip.
      • If you selected a clip in the Timeline: Markers appear in both the Timeline and Canvas.
        Figure. Timeline window showing audio peak markers for a sequence clip.

You can clear audio peak markers that were previously added, if you like.

To clear all audio peak markers in a clip
  1. Select one or more clips in the Browser or Timeline.

  2. Choose Mark > Audio Peaks > Clear.

Raising Audio Levels Using Audio Normalization and Gain

When you edit, your audio may come from a variety of sources, and the levels often vary. Final Cut Pro includes a Gain audio filter that allows you to amplify (or attenuate) the level of an audio clip far beyond the +12 dB gain available with audio level keyframes.

To amplify the audio level of clips whose levels are too low, you can manually apply the Gain filter. The only risk with manually applying gain adjustments to a clip is that you may amplify the audio too much, resulting in distorted audio. To guarantee that clips with low audio levels have the optimal gain, you can normalize your audio clips using the Apply Normalization Gain command.

How Normalization Gain Works in Final Cut Pro

Audio normalization works by scanning audio for the peak (loudest) sample level and then applying a Gain filter that brings the peak level to the level you request. By default this value is 0 dBFS, the highest level possible before clipping occurs. The Gain filter raises the overall audio level.

Figure. Diagrams showing an audio waveform with a peak, the same waveform with gain added so that the audio peak is set to 0 dBFS, and the waveform normalized with the peak level set at 0 dBFS.

In most applications, audio normalization is a destructive process because it permanently modifies audio files. Final Cut Pro applies normalization nondestructively by applying a Gain filter to a clip instead of affecting the clip’s audio file. You can disable or remove the Gain filter and hear the original, unmodified audio file.

To apply normalization gain to audio clip items in a sequence
  1. Select one or more audio clip items in a sequence.

  2. Choose Modify > Audio > Apply Normalization Gain.

    The Apply Normalization Gain dialog appears.

  3. In the “Normalize to” field, enter the value you want to raise each audio clip’s peak value to, then click OK.

    The dialog displays a progress bar and Final Cut Pro begins calculating the peak value for each clip. After processing, each selected clip has its own Gain filter applied with a gain adjustment appropriate for that clip’s normalization.

How Linked Mono and Stereo Clips Are Normalized

The Apply Normalization Gain command works differently depending on the type of clip items selected:

  • Single mono clip item or multiple linked mono clip items: A separate Gain filter is applied to each clip item, and peaks for each clip item are calculated independently.
  • Stereo clip items: A stereo Gain filter is applied to the stereo clip items, and the Gain setting is based on the peak value across both channels.

Reapplying Normalization Gain

Final Cut Pro searches for peak audio only between a clip’s In and Out points, not for the entire duration of the clip’s media file. If you trim a clip’s In or Out point, new peaks may be introduced and the Gain adjustment may no longer be appropriate. In this case, you can easily reapply normalization gain to set an appropriate level.

Reapplying normalization gain is no different from applying normalization gain for the first time. The only difference is that no new Gain filters are added to clips that already have them. Instead, the values of the existing Gain filters are adjusted based on the current audio peaks of the clips.

Choosing Normalization Gain Versus Audio Level Keyframing

The Gain filter and the Apply Normalization Gain command are best used for broad audio level adjustments, such as when you have clips with fairly low audio levels. For subtle level adjustments and more complex mixing, you should use audio level keyframes in the Viewer or Timeline.

Troubleshooting Audio Normalization

There are a few issues to be aware of when you use the Gain filter and the Apply Normalization Gain command:

  • Applying gain raises the level of an audio signal, including the noise. Very quiet audio, when normalized, may be very noisy. When possible, the best solution is to rerecord the audio. If this is not possible, you may be able to minimize the noise using Soundtrack Pro.

  • Loud peaks in audio clips that otherwise contain low audio levels make audio normalization more difficult to use. For example, suppose you have a clip containing dialogue that was recorded too quietly. At the beginning of this clip, there is a brief peak when the slate was clapped together. When you attempt to normalize the audio of this clip, the sound of the slate is so loud that very little gain is applied. To apply more gain, simply trim the clip until the audio peak from the slate is gone, then use the Apply Normalization Gain command again.

What Reference Level Should You Use for Mixing and Output?

The dynamic range of your mix is dependent on the final viewing environment. For example, movie theaters have large, relatively expensive sound systems that can reproduce a large dynamic range. Television speakers are much smaller, and often the listening environment has more ambient noise, so very quiet sounds may not even be noticeable unless the overall signal is compressed and the level increased, reducing the dynamic range.

For example, television stations normally accommodate only 6 dB between the average loudness and the peaks. Dolby Digital feature film soundtracks, on the other hand, can accommodate up to 20 dB between average and peak levels. This is why loud sounds in a movie theater sound so loud: they are much louder than the average level.

Venue
Acceptable amount of dynamic range
Theatrical Dolby Digital
20 dB
Average videotape
12 dB
Television broadcast
6 dB

When you mix your final audio, you choose a consistent reference for the average level. When you choose the average reference level, you are actually choosing how much additional headroom you have before your signal distorts. The higher you set the average level, the less safety margin you have for peaks in the signal. This means that the loudest sounds in your mix cannot be much louder than the average levels, so the mix is less dynamic.

If you set the reference level of the Final Cut Pro floating audio meters to -20 dBFS, you have nearly 20 dB of headroom because 0 dBFS is the digital limit for the loudest sound. If you set the reference level in your sequence to -12 dBFS instead, you have less headroom. Even though the average level of your audio is higher, there won’t be as much dynamic range.

Figure. Diagram of audio meters showing the available headroom when the reference level is set to -12 dBFS.

How much dynamic range you allow in your audio mix depends on its ultimate destination. If you’re editing a program for TV broadcast, a reference level of -12 dBFS is fine, because you are only allowed 6 dB of dynamic range anyway. But if you’re working on a production to be shown in movie theaters, consider using a reference level closer to -18 or even -20 dBFS (both of these are frequently used standards).

Remember that the ultimate goal is to ensure that audio doesn’t peak over 0 dBFS in your mix (as displayed in the Final Cut Pro audio meters) and won’t peak over +3 dB or so on an analog meter.

Outputting Bars and Tone at the Head of Your Tape

When you output your program to a tape for duplication or delivery to a broadcast facility, you’ll typically include a 1 kHz reference tone at the beginning of the tape. The level of this tone is supposed to indicate what the average level of your audio mix is. For this tone to be meaningful, you must mix your audio so that the average level of your mix matches the level of the tone. Here’s why:

  • If you are duplicating the tape: Most tape duplication facilities use the reference tone at the beginning of the tape to set the audio recording levels when copying your master tape. If your average mix levels are too quiet or too loud relative to this tone, the audio on the copies will be either too low or distorted, respectively.
  • If you’re delivering your program for broadcast: Most broadcast facilities have very stringent requirements about what they’ll air. If your program’s audio levels are too hot (loud) or too soft, you might run into trouble with the broadcast engineer. In the worst cases, they’ll return your tape to you as unsuitable for broadcast and require you to send them a new one with proper levels.

Labeling Your Tapes

If you’re outputting to a digital format, make sure you note what level your 1 kHz tone is set to on the label of your tape. If you’re outputting to an analog format, like Betacam SP, you’ll always set your 1 kHz tone to 0 dB.