How to Trim Silence at the Start and End and Add Fades

Every microphone recording ends up with awkward silence: between pressing record and starting to talk, and between finishing and pressing stop. Leave it in and the file is longer than it needs to be, and joining several recordings leaves odd gaps. The Fade In/Out & Remove Silence tool does the cleanup in a few clicks. This guide covers how to set the detection threshold, how long fades should be, and normalizing.

1. Know what dBFS means

Digital audio level is usually measured in dBFS (decibels relative to full scale). 0 dBFS is the loudest sound digital audio can represent, and smaller numbers (-40 dBFS, -60 dBFS) are quieter. Setting the "threshold" in a silence trimmer means drawing a line: anything quieter than this number counts as silence and gets cut.

2. Get a feel for the threshold slider

Start low, for example -50 dBFS. Then slowly move the slider to the right (toward -30 dBFS) while watching the red detected areas on the waveform. If only truly silent parts turn red, the threshold is right. When parts of speech or music start turning red, step back to the previous, lower value.

3. Suggested thresholds by recording setup

A close microphone in a quiet room has little background noise, so strict thresholds like -45 to -55 dBFS work well. If a fan or street noise was in the room, the background itself can sit near -35 dBFS, and a threshold that's too low finds no silence at all. In that case, raise it to around -30 dBFS so only the quietest start and end are picked.

4. Always preview before applying

Moving the slider doesn't cut anything yet; it only shows the detected areas on the waveform. When the settings look right, use "Preview" to hear the actual result, then "Apply" or "Save". Following this order avoids cutting things you meant to keep.

5. Choose fade lengths

A fade makes sound grow and shrink smoothly instead of starting or stopping abruptly. For speech (interviews, lectures, calls), short fades of 0.2 to 0.5 seconds sound natural. For music or sound effects where the tail matters, 1 to 3 seconds lets it trail off gently. For ringtones, a 0.5 to 1 second fade out at the end removes the sudden stop.

6. Normalize to fix the level

Normalizing finds the loudest point (peak) in the file and scales the whole file so that peak lands at -1 dB. Recordings made far from the mic or with low input gain sound quiet overall, and normalizing brings them up to a listenable level in one step. Because it's based on the single loudest point, a stray loud sound (a cough or a bump) can keep the rest from getting much louder. In that case, also try LUFS matching in Increase Volume.

7. Processing order and saving

The tool processes in this order: ① trim silence → ② fade in/out → ③ normalize. Pressing "Save" applies all the settings you have turned on in one file. You can choose WAV (lossless) or MP3 (128 to 320 kbps), and everything happens in your browser, so the file isn't sent to a server.

Common mistakes

The most common mistake is jumping the threshold too high at once (for example -15 dBFS). Whispered words or quietly played passages are then treated as silence and cut. Always start low and raise it gradually while watching the red areas. Also, if nothing gets trimmed, check that the "Trim silence" checkbox is actually on. To remove long pauses and "um"s in the middle of a recording (not just the ends), use Silence & Filler Word Cutter; see this guide.

Sources and date checked

Checked: October 6, 2026

• Wikipedia, dBFS — https://en.wikipedia.org/wiki/DBFS
• Wikipedia, Audio normalization — https://en.wikipedia.org/wiki/Audio_normalization

This article is general information. Results depend on the recording environment and file.

Frequently asked questions