Skip to content

Spectral Editing

The spectral editing mode allows you to make selections in the time-frequency domain (see Time and Frequency Domains for more information).

Processing is applied to the selection in time-frequency so that you can easily edit out isolated problems visible in the spectrogram. This is also a very powerful tool for sound design. A specialized Retouch tool is available in the spectral editing mode that you can use to attenuate noise while using information from surrounding "clean" areas in the spectrogram.

You can activate the spectral editing mode by clicking the spectrogram button on the toolbar:

Waveform-Spectrogram Toggle

Use the waveform button (left) and spectral editing button (right) in the main toolbar to switch between modes.

Click the dropdown arrow beside the Spectrogram toggle button to change Channel Mode and Waveform Overlay Opacity.

Spectrogram drop down menu

Three channel modes are available:

  • Discrete Channels: displays each audio channel as its own spectrogram.
  • Channel Mean: takes the mean of all channels and displays a single spectrogram.
  • Panning as Hue: displays a single spectrogram that conveys panning via color hue, and lets you restrict the selection to a range of pan positions. See Stereo View.

To view both the Waveplot and Spectrogram at the same time, select an opacity greater than 0% from the Waveform Overlay Opacity options. Below we can see the 60% option:

Spectrogram drop down menu

Alternatively, you can choose View Spectrogram from the View menu or press Shift+S. The waveform view is changed to a spectrogram after activating the spectral editing mode:

Spectral Editing

To switch back, select View Waveplot from the View menu or press Shift+W.

Three further features of the spectral editing mode are described in their own chapters:

  • Stereo View draws the stereo image as colour, and lets you confine a selection to a range of pan positions.

  • Time and Frequency Feathering fades processing in and out at the edges of a selection, so that repairs blend into the surrounding audio.

  • Finding Similar Events searches the recording for further occurrences of the sound you have selected, so that a repeating noise can be repaired in one operation.