Planning the steps to cut audio means deciding which range of a file you want to keep, confirming your source can be decoded by the browser, and choosing exact start and end times in seconds before you export. A clean plan answers four questions in order: which file are you cutting, where does the useful audio start, where does it end, and what output do you need. The Audio Cutter follows that order: it waits for Web Audio to decode the file, displays the decoded duration, sample rate, channel count, and frame count, then trims one continuous range at sample-frame boundaries and writes an uncompressed PCM16 WAV for download. Nothing is uploaded; the file, decode, range selection, preview, and export all happen in the current browser tab. Planning ahead keeps you from clicking Cut on a file the browser cannot decode, from picking a start time that rounds below one frame, or from choosing an end time past the decoded duration. A short pre-flight check replaces trial-and-error with one deliberate pass, and the rest of this article walks through that pre-flight check in the same order the tool expects.
What planning the steps for an audio cut actually involves
Cutting audio sounds like a single click, but the planning is the part that decides whether the first export is useful or wasted. The work breaks into three pre-cut decisions and one verification step. The pre-cut decisions are: pick a file your browser can actually decode, decide where the useful audio starts, and decide where it ends. The verification step is previewing the resulting WAV at both boundaries before you commit the download.
Each decision depends on the previous one. You cannot pick boundaries until you know the file decoded successfully and you know its decoded duration. You cannot preview until you have cut, and you cannot cut until the start and end values are valid. Treating the workflow as a chain of dependencies is what makes a planned cut predictable. The tool itself enforces the chain by reading the file, reporting the decoded duration, then converting the time fields to frame indices at the decoded AudioBuffer sample rate before exporting. If you respect that order in your plan, the cut runs once.
Confirm the source file can be decoded first
The first item on the plan is a sanity check on the source. The file picker recognises common MP3, WAV, M4A, AAC, Ogg, WebM, and FLAC file names and MIME types, but recognition is not a guarantee that decoding will succeed. Actual decoding is handled by your browser's Web Audio decoder, and container and codec support differs between Safari, Chrome, Firefox, operating systems, and browser versions. An unsupported, damaged, mislabeled, encrypted, or incomplete file produces a clear decode error rather than a silent failure. There is no codec installation, no conversion server, and no fallback source.
The source also has a hard size limit: the encoded file must be no larger than 25 MB. After decoding, the audio is rejected if it is longer than 15 minutes, has more than eight channels, exceeds 192 kHz, or contains more than 30 million samples across all channels. The decoded-sample limit is the one that surprises people: a long, high-rate, multichannel file can hit the boundary before the duration limit does, and the message tells you which constraint triggered. Planning means knowing these limits up front so you are not surprised at the decode step.
Map cut boundaries in moments, not as guesses
Once decoding succeeds, the page displays the decoded duration, sample rate, channel count, and frame count. Those four numbers are the basis for planning your boundaries. Start and end times are entered almost instantly and converted to the nearest decoded sample-frame boundary using the AudioBuffer sample rate. Web Audio may resample the source to the AudioContext sample rate during decoding, which means the displayed sample rate describes the working buffer and the exported WAV, not necessarily the source file's encoded rate. If you try to base boundaries on the source file's nominal rate, your numbers will be slightly off in the exported file.
Three rules govern the time fields. Start cannot be negative and cannot sit at the end of the audio. End must be greater than start. End cannot exceed the exact decoded duration. A selection that rounds to less than one frame is rejected rather than producing an empty file. The start frame is included in the cut and the end frame is excluded, so a 10.000-second to 20.000-second selection on a 44.1 kHz decoded buffer covers exactly 441,000 frames. Planning the boundaries in moments that align to whole frames at the displayed sample rate keeps the selection unambiguous.
Run the planned cut with Audio Cutter
- Choose a browser-decodable audio file within the displayed file and decoded-audio limits.
- Enter a non-negative start time and a later end time that does not exceed the decoded duration.
- Select Cut and export WAV, preview the PCM16 result, then download it.
Each step in the list matches a button or field on the page. After you choose the file, wait until the decoded duration, sample rate, channel count, and frame count appear before you fill in the time fields. Type the start time almost instantly, type the end time in moments, then trigger the cut. Changing either time field clears the previous WAV, and cutting again replaces and releases the prior download URL, so you can iterate without piling up result files in the page.
Verify the WAV before you download
The export writes a newly encoded RIFF/WAVE file containing interleaved, little-endian, signed 16-bit PCM samples. The page reports the output duration and file size before download, which is your chance to sanity-check the math. For a stereo selection from 10.000 to 20.000 seconds at a decoded sample rate of 44,100 Hz, the frame count is (20 - 10) × 44,100 = 441,000 frames. The data size is 441,000 × 2 channels × 2 bytes = 1,764,000 bytes, and the full WAV is 1,764,000 + 44 = 1,764,044 bytes, or roughly 1.68 MB. If the reported size does not match that calculation within rounding, redo the cut before downloading.
Preview the WAV and confirm the spoken word or musical transient at both boundaries. If the start lands mid-word or the end cuts a sustain tail, go back, adjust the seconds, and cut again. Multiple channels remain separate and are interleaved in their original channel order; a stereo source stays stereo, and a file with up to eight channels is retained as long as playback software supports the layout. Floating-point Web Audio samples are clamped to the PCM range from -1 through 1, with negative full scale becoming -32768, positive full scale becoming 32767, and non-finite values written as silence. None of that changes the boundary, but it explains why very hot recordings occasionally need a follow-up normalize step in a dedicated editor.
Limits to plan around so the cut does not fail
Putting every guardrail in one place makes planning faster. The table below lists each limit the tool enforces, what it caps, and what happens at the boundary.
| What is checked | Maximum allowed | What happens at the limit |
|---|---|---|
| Encoded source file size | 25 MB | Rejected with a clear message before decoding |
| Decoded duration | 15 minutes | Rejected after the decode attempt |
| Channel count | 8 channels | Rejected after the decode attempt |
| Sample rate | 192 kHz | Rejected after the decode attempt |
| Total samples across all channels | 30,000,000 | Rejected after the decode attempt |
Browser decoding happens before the decoded duration and sample count can be inspected, so a highly compressed long file can still consume temporary memory during the decode attempt. Nothing is shortened automatically; an over-limit file is rejected rather than truncated. In-flight reads and decodes are guarded so an older job cannot overwrite a newer selection, which keeps repeated cuts consistent. Audio contexts are closed after decoding and during cleanup, and temporary local URLs are released when a file or result is replaced or when the page closes.
When planning reveals you need a different tool
The plan above is enough for trimming one continuous range and exporting a clean uncompressed WAV. It is not enough for fades, lossless compressed-frame cutting, metadata preservation, or production mastering. If you need waveform-level fades, change pitch, normalize loudness, remix, denoise, or keep the original MP3 or AAC encoding with tags, use a dedicated editor. The result does not retain the source's MP3, AAC, Vorbis, Opus, FLAC, or other compressed encoding, and it does not preserve tags, album art, cue points, chapters, loudness metadata, or encoder settings. Treating the cut as a planning step in a wider workflow keeps the right tool for the right job. For a broader view of the failure modes that show up when planning is skipped, the avoid mistakes when you cut audio in your browser guide walks through the planning gaps that cause those failures.
For a deeper look, see Online Metronome on Windows: Set Tempo and Start Clicks.