Typically these tasks can be performed in a manner that is non-linear. Audio editors may process the audio data non-destructively in real-time, or destructively as an "off-line" process, or a hybrid with some real-time effects and some off-line effects.
Comparison of destructive and real-time editing
Destructive editing modifies the data of the original audio file, as opposed to just editing its playback parameters. Destructive editors are also known as "sample editors". Destructive editing applies edits and processing directly to the audio data, changing the data immediately. If, for example, part of a track is deleted, the "deleted" audio data is immediately removed from that part of the track. Real-time editing does not apply changes immediately, but applies edits and processing on the fly during playback. If, for example, part of a track is deleted, the "deleted" audio data is not actually removed from the track, but is hidden and will be skipped on playback.
Advantages of destructive editing
In graphical editors, every change to the audio is usually visible immediately as the visible waveform is updated to match the audio data.
The number of effects that may be applied is virtually unlimited.
Effects may be applied to a precisely specified selected region.
Mixing down or exporting the edited audio is usually relatively quick as little additional processing is required.
Limitations of destructive editing
Once an effect has been applied, it cannot usually be changed. This is usually mitigated by the ability to "undo" the last performed action. Typically a destructive audio editor will maintain many levels of "undo history" so that multiple actions may be undone in the reverse order that they were applied.
Edits can only be undone in the reverse order that they were applied.
Advantages of real-time editing
Effects can usually be adjusted during playback, or at any other time.
Edits may be undone or adjusted at any time in any order.
Multiple effects and edits may be 'stacked' so that they are applied to the audio as an effect chain.
A stack of effects may be changed so that effects are applied in a different order, or effects inserted or removed from the chain.
Some real-time editors support effect automation so that changes to effect parameters may be programmed to occur at specified times during audio playback.
Limitations of real-time editing
The waveform does not usually show the effect of processing until the audio has been mixed-down or "bounced" to another track.
The number of effects that may be applied is limited by the available processing power of the computer or editing hardware. In some editors this may be mitigated by "freezing" the track.
It is not usually possible to have an effect only on part of a track. To apply a real-time effect to part of a track usually required that the effect is set to turn on at one point and turn off at another.
In multi-track editors, if audio is copied or moved from one track to another, the audio in the new track may sound different from how it sounded in the original track as there may be different real-time effects in each track.
In some applications, mixing down or exporting the edited audio may be slow as all effects and processing needs to be applied.
For use with speech
Editors designed for use in speech research add the ability to make measurements and perform acoustic analyses such as extracting and displaying a fundamental frequency contour or spectrogram. They typically lacks most or all of the effects that interest musicians.
