Audio Cutter

The practical question is not only where a clip should begin and end. A good cut also avoids clipped words, audible clicks and an output setting that is unnecessarily large or too compressed for the next destination. This page keeps those decisions beside the waveform and reports what it actually created.

How to use this Audio Cutter

1. Choose one audio file. Add a file up to 100 MB and 30 minutes long. MP3 and WAV are the clearest starting points. M4A, AAC, OGG, FLAC or WebM audio may work when the current browser can decode the specific file and codec.
2. Select the part to keep. Drag the coral region across the waveform, resize either edge or enter a start and end time such as 0:12.500 and 0:28.200.
3. Preview the boundaries. Play the selection and listen first to its opening and ending. Turn on Loop selection preview when repeated comparison makes the timing easier to judge.
4. Choose boundary and fade settings. Keep zero-crossing adjustment on for a small search around each cut point, and use short fades when a hard edge sounds abrupt.
5. Choose the output. Select MP3 or WAV, then choose MP3 bitrate and stereo or mono output where relevant.
6. Cut and inspect. The result panel provides playback, actual start and end points, duration, bytes, sample rate, channels, peak level, fades, metadata status and a one-click download.

Changing a selection or output setting clears the previous result. That behavior prevents an older file from being downloaded after the visible controls have changed.

Choose the exact section

The waveform is useful for finding obvious silence, speech, beats and transitions. It is not a transcript and does not identify words automatically. Zoom in when a short sound needs a more exact boundary; zoom out to understand the position of the selection within the whole recording.

The start and end fields accept seconds, M:SS.mmm or H:MM:SS.mmm. Milliseconds help when the difference between two edits is too small to place comfortably by dragging. The minimum selection is 0.05 seconds, which prevents an empty or effectively unusable export.

Three playback actions serve different checks:

• Play starts normal playback from the current cursor.
• Play selection plays only the selected region and stops at its end.
• Loop selection preview repeats that selected region so you can compare the opening and closing points without restarting it manually.

For spoken audio, leave a small amount of natural room before the first consonant and after the last syllable. For music, listen in context rather than cutting only by the visual peak. A waveform can show amplitude, but it cannot tell whether the selected musical phrase feels complete.

Prevent clicks at the cut points

A hard cut can click when it begins or ends while the waveform is away from its center line. The sound jumps suddenly from silence to a non-zero sample, or from a non-zero sample to silence.

Snap near zero crossings searches up to about 10 milliseconds around each requested boundary for a point with lower combined channel amplitude. This is a small corrective search, not AI analysis and not a guarantee that every click will disappear. The result panel shows whether the actual start or end moved and by how many milliseconds.

Turn zero-crossing adjustment off when an exact legal, editorial or synchronization timecode must be preserved. In that case, use a short fade if the boundary is audible. The fade controls on this page affect the generated file, not only preview playback.

A 0.05–0.15 second fade is a practical starting point for many abrupt clip edges. Longer fades can be useful for music, but they may soften consonants or remove the impact of a transient. The tool limits each fade to half of the selected duration so the fade-in and fade-out cannot extend beyond the clip.

These controls directly connect the editing principle to the tool: use the reported boundary adjustment to confirm what zero-crossing search changed, and use result playback to decide whether the fade solved the audible problem without masking wanted sound.

Choose high-quality MP3 or WAV output

“High quality” depends on the source and destination. An output cannot recover detail that is absent from the original file, and a larger bitrate is not automatically useful when the destination recompresses audio again.

MP3 for a smaller, widely accepted file

MP3 output is newly encoded. Choose among 128, 192, 256 and 320 kbps:

These are starting points, not universal quality grades. Listen to the generated result with the actual source. Re-encoding an existing MP3 introduces another lossy generation even when the selected bitrate is high.

WAV for uncompressed 16-bit PCM output

WAV output stores the selected samples as 16-bit PCM and keeps the source sample rate. The file is usually much larger than MP3. It is useful when the next editing step accepts WAV or when you want to avoid adding MP3 compression at this stage.

WAV does not make a compressed source lossless again. It also does not retain the source file's bit depth as a promise: this tool's WAV encoder specifically produces 16-bit PCM.

Stereo or mono

Keep stereo when left-right information matters. Choose mono when the recording is primarily one centered voice or the destination specifically needs one channel. Mono mixes the available source channels together; it does not isolate speech from music or remove background sound.

The Web Audio specification defines the browser audio graph and decoded audio buffers used by this workflow. Browser codec support can still differ, so “browser-supported audio” is more accurate than promising every container and codec. See the Web Audio API specification for the underlying platform model.

Read the audio result

Listen before downloading. Start with the first and final second because those points reveal clipped speech, delayed starts, abrupt endings and fades that are too long. Then inspect the report:

• Start and end: the actual sample positions used for encoding, including any reported zero-crossing adjustment.
• Duration: the generated clip length and its percentage of the source duration.
• Format and bitrate: MP3 with the selected bitrate, or 16-bit PCM WAV.
• File size: the actual Blob size created in the browser, not an estimate.
• Sample rate: the output rate. WAV keeps the decoded source rate; MP3 uses the closest rate supported by this encoder when resampling is needed.
• Channels: stereo or mono after the selected channel setting.
• Peak level: the highest absolute sample after fades, expressed as dBFS. A value at 0.0 dBFS means the encoded sample data reached full scale; this report is not a loudness measurement such as LUFS.
• Metadata: “Not copied” means tags, artwork, chapters and other source metadata are not transferred.

Compare the output file size with the source only as context. A short WAV cut can still be larger than a longer compressed source, while a newly encoded MP3 may be smaller because its duration or bitrate changed.

Illustrative audio cutting examples

Keep one spoken answer

Assume a 6-minute stereo MP3 contains a useful answer from 2:14.300 to 2:48.900. Enter those times, preview the selection, keep a 0.10-second fade at both edges and export at 192 kbps. The expected selected duration is about 34.6 seconds before any small zero-crossing adjustment.

Prepare a short music excerpt

Assume a WAV source contains a phrase from 0:42.000 to 1:05.500. Zoom into both transitions, loop the selection and choose WAV when another editor will receive the result. Keep stereo if the source has meaningful left-right placement.

Illustrative examples. Exact cut points, file size, peak level and browser compatibility depend on the selected source and settings. The generated result panel is the source of truth for the actual file.

Scope and alternatives

This base97 tool focuses on one continuous section of one file. Its useful distinctions are the combination of draggable and typed timing, selection looping, optional zero-crossing search, fades that are applied to the output and a result report that explains the generated file. Encoding runs in a Web Worker so the four-stage processing display can remain responsive while the output is built.

Several established tools cover related workflows:

• 123apps Audio Cutter provides a waveform-based online cutter with fades and a broader suite of connected audio tools.
• AudioTrimmer offers a focused browser page for choosing a section and applying fade controls.
• Clideo Audio Cutter supports selecting or deleting a section and is part of a larger cloud media workflow.
• VEED and Kapwing place trimming within broader timeline editors that support continued video or content editing.

Those products may be a better fit when you need cloud projects, multiple clips, timeline composition, collaboration or continued video editing. The base97 page is narrower: it keeps one range, shows the exact generated audio properties and performs the current workflow in the browser without an account.

This version does not support multiple files, batch cutting, keeping several disconnected ranges, joining clips, deleting only the selected middle, multi-track editing, AI silence detection, transcription, noise removal, effects, recording, direct cloud imports, URL imports, APK installation or looper output. It also does not preserve source metadata.

Troubleshoot audio cutting

The file will not open

Confirm it is under 100 MB and no longer than 30 minutes. A familiar extension does not guarantee that the browser supports the codec inside the file. Try MP3 or WAV, or convert the source in a trusted editor first.

The beginning or ending clicks

Turn on zero-crossing adjustment and generate the result again. If the exact timecode must remain unchanged, turn snapping off and add a short fade instead. Always judge the generated playback, not only the source preview.

A word or beat is clipped

Zoom in, move the relevant edge slightly outward and loop the selection. Waveform peaks can help locate the sound, but listening determines whether enough context remains.

The MP3 sounds worse than the source

Raise the bitrate and avoid repeated MP3 save cycles. Try WAV when another editing step follows. Neither option can restore quality already lost in the source.

The WAV file is unexpectedly large

WAV output on this page is uncompressed 16-bit PCM. Choose MP3 when a smaller delivery file is more important than avoiding another lossy encoding stage.

Processing fails or the tab becomes slow

Long, high-sample-rate, multi-channel files require more memory because the browser decodes the complete source before cutting. Try a shorter source, choose a shorter selection, close memory-heavy tabs or use a desktop editor for large production recordings.

Audio Cutter FAQ