Logo

MonoCalc

/

Audio Trimmer

Media

or drop it here · MP3, WAV, M4A, FLAC, Ogg and anything else your browser decodes, up to 300 MB. The audio stays on your device.

Display gain

0:00.000 / 0:00.000

About This Tool

Audio Trimmer – Cut, Fade and Export Audio Without Uploading

This audio trimmer cuts a part out of any recording on your device. Open an MP3, WAV, M4A, FLAC or Ogg file, drag the two handles on the waveform, listen to the cut points and download the result as MP3 or WAV. Decoding, editing and encoding all happen in the browser tab, so the audio is never uploaded. Use Keep selection for the usual trim, or Remove selection to cut a cough, an ad break or a false start out of the middle and join what is left.

Samples, sample rate and frames

Digital audio is a list of samples: numbers that give the height of the sound wave at evenly spaced moments. A CD-quality file stores 44,100 samples per second per channel, so one second of stereo is 88,200 numbers. The trimmer stores the start and end of the selection as whole sample positions, not as rounded seconds, so moving a handle back and forth never drifts. Zoom in far enough and the waveform turns into individual dots, one per sample.

How an MP3 can be cut without re-encoding

An MP3 file is a run of short frames. Each Layer III frame holds 1,152 samples, which is about 26.1 ms at 44.1 kHz and 24 ms at 48 kHz. Every frame starts with a 4-byte header giving its bitrate and sample rate, so a program can step through the file frame by frame. Copying a run of whole frames gives a valid, shorter MP3 that is byte-for-byte identical to the original audio, with no quality loss and no waiting. That is what Same as source does, which is why the handles snap to frame boundaries in that mode. The tool also writes a fresh Xing/Info header so players show the right duration; files that used the rarer VBRI header get none, so some players may show an estimated length.

There are two small costs. Layer III frames can borrow spare bits from earlier frames (the bit reservoir), so the first frame after a cut may point back into audio that was removed, and many decoders briefly mute it. Encoders also add about 1,100 samples of delay at the start, which this tool ignores, so the waveform can sit up to about one frame away from what a player outputs.

Lossless WAV cutting

WAV files usually store plain PCM samples, so a cut is exact to the sample. With Same as source, untouched audio is copied straight from the file and only the fade and join regions are rewritten, in the file's own bit depth. Metadata chunks such as LIST, bext and cue points are dropped, because their times no longer match the shorter file.

Why cuts click, and how fades fix it

A cut in the middle of a waveform makes the signal jump from some value straight to silence. The ear hears that jump as a click. The trimmer applies a tiny 5 ms linear fade at every hard cut by default (220 samples at 44.1 kHz), which is too short to notice as a fade but long enough to remove the click. In Remove mode the two sides of the join overlap with an equal-power crossfade, where the outgoing side follows cos(p·π/2) and the incoming side sin(p·π/2), so the combined power stays constant through the join.

Fade shapes and loudness

We hear loudness roughly on a logarithmic scale measured in decibels, so a fade that is straight on paper does not sound straight. A linear fade changes the level steadily. An equal-power fade (sin curve) stays near full volume longer and drops at the end, which suits music. A cubic fade (p³) falls quickly at first and then tails off, which often sounds the most natural for a fade-out. The fade curve is drawn over the waveform and the preview plays exactly what the export will contain.

Generation loss
MP3, AAC and Ogg are lossy formats: each encode throws away detail the encoder thinks you won't miss. Decoding and re-encoding adds a second round of loss. Prefer Same as source for MP3s, and WAV when you need fades and want to keep full quality for further editing. LAME also adds about 25 ms of silence at the start of a new MP3, which is normal for the format.

Tips for common jobs

  • Ringtones and notification sounds: set the start, tap a 15 s or 30 s length chip, add a one-second fade-out and export as MP3.
  • Podcasts and voice memos: use Trim silence to drop dead air from both ends, and raise the display gain to see quiet speech.
  • Checking a cut: Hear start, Hear end and Hear the join play just the few seconds around each cut, which is where problems show up.

Credits

MP3 encoding uses LAME through @breezystack/lamejs, a JavaScript port licensed under the LGPL-3.0. It is loaded only when you export as MP3 (re-encode).

Frequently Asked Questions

Is the Audio Trimmer free?

Yes, Audio Trimmer is totally free :)

Can I use the Audio Trimmer offline?

Yes, you can install the webapp as PWA.

Is it safe to use Audio Trimmer?

Yes, any data related to Audio Trimmer only stored in your browser (if storage required). You can simply clear browser cache to clear all the stored data. We do not store any data on server.

Is my audio uploaded anywhere?

No. The file is read, decoded, edited and encoded inside this browser tab, and nothing is sent to a server. There is no URL input either, because a web page isn't allowed to read audio from another site without that site's permission, and fetching it through a server would mean uploading.

Does trimming lower the quality?

Not with "Same as source": MP3 frames and WAV samples are copied exactly, with no re-encoding. WAV output is lossless from the decoded audio. MP3 (re-encode) adds another generation of lossy compression, so use it only when you need fades on an MP3 or the source wasn't MP3 or WAV.

Why do my MP3 cut points move slightly?

A lossless MP3 cut can only happen between frames, and each frame holds 1,152 samples (about 26 ms at 44.1 kHz), so the handles snap to the nearest frame boundary. Because MP3 frames can borrow bits from earlier frames (the bit reservoir), some players briefly mute the first frame after a cut.

Why is there a length limit?

Editing needs the whole decoded recording in memory as 32-bit floating-point samples, about 21 MB per minute of stereo at 44.1 kHz. The tool allows up to 100 million samples: roughly 18 minutes of stereo or 37 minutes of mono at 44.1 kHz, and 17 or 34 minutes at 48 kHz. Files over 300 MB are refused before reading.

Can I make a ringtone with it?

Yes for Android, which accepts MP3: use the 15 s or 30 s length chips, add a short fade-out and export as MP3. iPhone ringtones must be .m4r (AAC) files, which this tool doesn't create, so you'd need to convert the trimmed WAV or MP3 afterwards.

Which audio formats can I open?

Whatever your browser can decode. MP3, WAV, AAC/M4A and FLAC open in current Chrome, Firefox and Safari. Ogg Vorbis, Opus and WebM depend on the browser (Safari is the usual exception), and WMA, AMR and other legacy formats don't open in any browser. MP4 and WebM videos work too; only the soundtrack is kept.