Audio Trimmer

How the waveform is made

When you drop a file, your browser's own Web Audio API decodes it — the same decoder it uses to play audio on any web page — into raw samples. The trimmer measures the peak level in about 1,600 slices of the file and draws one bar for each, so loud passages stand tall and pauses show as gaps. That is usually enough to spot where a sentence starts, where the chorus comes in, or where the silence at the end begins.

Formats your browser cannot decode itself — WMA, AMR and some others — are read with ffmpeg instead: it converts a small copy for the waveform, and the full file is still used for the export.

Selections are set to a tenth of a second, which is finer than you can place a cut by ear. Use Play selection to check the edges, and nudge them with the arrow keys until the cut lands exactly where you want it.

Controls

Drag anywhere on the waveform and the nearer handle follows the pointer, so you can grab an edge without aiming for it. Each handle is also a keyboard slider: tab to it, then use the arrow keys to move by 0.1 s, Shift with an arrow for 1 s, Page Up and Page Down for 5 s, and Home or End to jump.

The Start and End boxes take exact times as minutes and seconds (1:23.4) or plain seconds (83.4). This is the quickest way when you already know the timestamps, for example from a podcast's show notes.

Fade in and Fade out ramp the volume linearly over the number of seconds you set. A short fade of half a second to a second removes the click that an abrupt cut can make in the middle of a sound, and a longer fade-out is the classic way to end a music clip gracefully. If the two fades add up to more than the clip, both are shortened in proportion.

Choosing the export format

WAV is uncompressed 16-bit PCM at 44.1 kHz. It is written directly from the decoded samples in the page — no engine to download, finished the instant you press the button — and it is the right choice for editing software or anything that will be processed further. Expect about 10 MB per minute in stereo.

MP3 is the universal choice for sharing; 192 kbps is a good default for music and 128 kbps is ample for speech. M4A (AAC) suits Apple devices and sounds a little better than MP3 at the same bitrate. OGG Vorbis is an open format widely used on Android and in games. These three are encoded by ffmpeg, cutting straight from your original file with sample-accurate seeking.

Trimming an MP3 and saving it as MP3 re-encodes it, which costs a little quality — at 192 kbps or more it is very hard to hear. If you plan to trim the same file several times, export WAV while you work and convert to MP3 once at the end.

Long files and limits

Decoded audio is large: about 21 MB per minute in memory for stereo at 44.1 kHz. Files up to ten minutes are decoded at full resolution, so the native WAV export can work from the samples directly. Longer files — podcasts, lectures, hours of recording — are decoded at a lower rate just to draw the waveform, and every format, WAV included, is then exported by ffmpeg from the original file, which keeps memory use down.

The trimmer keeps one continuous section. To remove something from the middle, export the part before it and the part after it as two clips and join them in an editor. It does not change the volume of the clip or remove noise; those need an audio editor.

Private by design

Decoding, drawing, playback and WAV export all use your browser's built-in audio APIs, so for WAV nothing at all is downloaded beyond the page. For the other formats, ffmpeg is compiled to WebAssembly and runs in this tab after a one-time 32 MB download. Either way the recording never leaves your device — useful when it is an interview, a voice note or a meeting.