Ramp the level up at the beginning of an audio file and down at the end. Fade lengths run from 0 to 30 seconds each, and setting both to 0 leaves the samples untouched.
Curve shapes
The curve decides how the gain travels between silence and full level over the length you set.
- Linear moves the amplitude at a constant rate. Because hearing is logarithmic, a linear fade-out seems to drop quickly at first and then linger near the bottom.
- Equal power follows a sine curve. It holds the level higher through the middle of the fade, which keeps perceived loudness steadier and is the standard shape where two pieces of audio overlap.
- Exponential squares the position. It stays near full level and then falls away sharply, which suits an ending that should clear out rather than trail.
Where the fade is anchored
Both fades are positional. The fade-in is measured from the first frame of the file and the fade-out from the last, and the render always covers the whole file.
Marking a range on the waveform does not move either anchor. The range is there to audition a boundary with looped playback, so fading a clip out of the middle of a longer recording means trimming it first, then fading the result.
Overlapping fades
Nothing stops the two fades from covering the same ground. When they overlap, the lower gain wins at every point, so a long fade-in and long fade-out on a short file produce a smooth rise and fall that never reaches full level. It is a workable way to shape a short sting, though the peak level ends up lower than the source.
Sources without a reported duration
A fade-out needs to know where the end is. A container that reports no duration in its metadata leaves nothing to anchor against, so the audio passes through unchanged rather than fading at an arbitrary point. Re-exporting the source to WAV or FLAC usually restores a readable duration.