Fade Audio Online: Fade In and Fade Out
Add a fade-in and fade-out to any audio file, with equal-power, linear or s-curve ramps, in your browser.
A fade is a gain ramp, and the shape of the ramp is what you hear. A linear ramp loses perceived loudness too quickly at the quiet end; an equal-power ramp keeps the perceived energy moving evenly, which is why it is the default here and in most editing software; an s-curve eases into and out of the ramp for a softer, more cinematic entry and exit. This tool applies both fades in one pass, lands the first and last sample on exact zero so the edges cannot click, and leaves everything between the fades untouched.
How to Use It
- Upload your audio file and wait for the local preview to load.
- Adjust the settings: Set the fade-in and fade-out lengths in milliseconds and pick a curve - equal-power, linear or s-curve.
- Download the result: Download the processed file directly to your device.
Controls and Options
Equal-power suits almost everything; try the s-curve when an ending should feel softer, and give fade-outs a little more time than fade-ins - endings breathe better long. If the two fades overlap on a short clip, they are shortened to meet once and the applied lengths are reported.
Formats and Output
Input: MP3, WAV, M4A, AAC, OGG, FLAC - anything your browser can decode. Output: WAV at the source sample rate, written without a further lossy pass.
Frequently Asked Questions
What is the difference between the three curves?
They are three shapes for the same gain ramp. Linear raises the gain in a straight line, which sounds like it spends too long quiet because hearing is roughly logarithmic. Equal-power follows a sine, so the perceived energy moves evenly across the whole ramp. S-curve is a smoothstep: it starts and ends gently and moves fastest in the middle, which reads as soft and cinematic on pads and ambience.
Why is equal-power the default?
Because it matches how loudness is perceived. The sine-in, cosine-out pair keeps the summed power constant through a crossfade, and on a single fade it distributes the audible change evenly over the ramp instead of bunching it at one end. It is the same default most editing software uses.
What happens when fade-in plus fade-out is longer than the file?
Both fades are shortened proportionally so they meet exactly once, and the result reports the lengths that were actually applied. No sample is ever faded twice, so the level dips to a single lowest point in the middle rather than collapsing.
Can I set one of the fades to zero?
Yes. A zero-length fade skips the audible ramp, but a 10-millisecond linear guard is still applied at that edge. Without it, a waveform that does not happen to pass through zero at the cut point produces a click; the guard is far too short to hear as a fade and removes that step.
Is my audio uploaded anywhere?
Supported audio content is not uploaded for processing. Decoding, fading, and export run locally in your browser; site analytics and the assistant are separate from the media path.
Why does the result come out as WAV?
The fade is computed on the decoded samples and written out without a further lossy pass. Re-encoding to MP3 would add a second generation of compression on top of whatever the source already had; WAV preserves the shaped audio exactly, at the cost of a larger file.