Audio trimmer

Not when the output format matches the input. The file is cut and the audio data is copied across untouched, so what comes out is bit for bit what went in between those two points.

Quality is lost when the format changes, because that means decoding and re-encoding. Trimming an MP3 and saving as MP3 costs nothing. Trimming an MP3 and saving as AAC costs a generation.

The MP3 quirk

MP3 stores audio in frames of roughly 26 milliseconds. A cut can only land on a frame boundary, so asking to start at exactly 12.500 seconds may actually start a few milliseconds either side.

It's inaudible for speech and music. It matters if you're cutting to a beat grid or assembling something sample-accurate, and for that a lossless format is the right working choice.

WAV and FLAC have no such constraint: a cut lands exactly where you asked.

Which format to keep

Keep whatever you started with, unless there's a reason not to. That's the whole rule for preserving quality.

MP3 plays everywhere and is the safe choice for sending to somebody.

WAV is uncompressed, so it's large, perfect, and the right format if the audio is going into an editor afterwards.

M4A or AAC is better than MP3 at the same bitrate and plays almost everywhere now.

OGG is excellent and less universally supported outside browsers.

Converting between two lossy formats always costs something, even when the numbers look better. Going from a 128kbps MP3 to a 320kbps AAC does not recover anything: it just stores the existing damage more precisely, in a bigger file.

Picking the points

Listen before you cut, and give yourself a little room at both ends. A cut landing exactly on the first syllable clips it, and the recording sounds truncated even when the words are all there.

For speech, a beat of silence before and after is what makes it sound deliberate rather than snipped.

For music, cutting on a beat is worth more than cutting precisely. A cut half a beat early sounds like a mistake; the same length cut on the beat sounds intentional.

Fading

A hard cut in the middle of a sound produces a click, because the waveform jumps instantly from wherever it was to silence. It's most noticeable on sustained sounds: a held note, a hum, room tone.

A very short fade, ten or twenty milliseconds, removes it entirely and is inaudible as a fade.

If your trimmed file clicks at the start or end, that's what happened, and it's a property of where you cut rather than a fault in the file.

What trimming keeps

The audio, at its original quality, and usually the tags: title, artist, album art. Some conversions drop the artwork.

Very long files take longer, and the work happens on your own machine, so a two hour recording is limited by your memory rather than by an upload.

Nothing is uploaded

The audio stays on your machine. The converter itself is about 30MB and downloads the first time you use it, then stays cached, which is why the first trim is slower than the ones after it.

That matters for the common cases here: interview recordings, voice notes, lectures and legal recordings are exactly the things that shouldn't be handed to a server to save thirty seconds.