Cutting audio is the right fix when a recoverable section of a recording needs to be removed and the rest of the take is usable as-is. That one sentence is the decision rule behind the whole process, and it carries two checks at once: the problem has to sit inside a bounded region, and the audio around that region has to be worth keeping. When both conditions hold, a trim gives you a clean result without losing the rest of the file, and the tool that fits the case is Audio Cutter, which reads a browser-decodable file, lets you enter a start and end time, and exports the selected range as a new uncompressed PCM16 WAV. When only one condition holds, the right move is usually something else: a fresh take if the recording itself is bad, or a volume, fade, or noise fix from a related tool when the file is acceptable but the finishing work is missing. The rest of this article walks through the cues that separate those cases, describes exactly what Audio Cutter does to a file, and ends with the practical steps for cutting when the decision lands on a trim.
Signs Your Audio Needs Cutting, Not Rerecording
Cutting fixes problems that sit inside a bounded region. The audio around the problem has to be worth keeping: clean voice, in tune, correct level, no distortion, consistent room tone. If that condition holds, the cue to reach for a trim is usually one of the following:
- Dead air, breath noise, or handling sound at the very start before the first word
- A false start, sneeze, door slam, or background interruption that lasts a few seconds mid-take
- An intro or outro you want to drop from a podcast or voice memo
- A musical intro that runs longer than the platform you are uploading to allows
- A long pause inside a recording that compresses badly and sounds awkward when shared
- A spoken phrase that needs to be removed for length, clarity, or compliance reasons
The decisive test is whether the take as a whole is acceptable except for one section. If you would keep 90 percent of the file unchanged, cutting is almost always faster, safer, and more faithful than rerecording, and it avoids the small drift in tone, room, and energy that a fresh take tends to introduce.
Signs You Should Rerecord or Use a Different Tool
Cutting cannot rescue audio where the problem spans the entire file. The following cues mean a trim is the wrong tool and the take needs a different fix:
- A steady hum, hiss, or room tone is present in every section, not just one
- Clipping shows up at multiple peaks so the audio sounds distorted rather than just loud
- The recording level is far too low across the whole take and quiet passages disappear into noise
- The voice sounds boomy, boxy, or hollow throughout because of the room acoustics
- Multiple small errors are scattered randomly with no clean region large enough to keep
In these cases the better fix is either a fresh take or an effect that changes the whole file rather than a slice of it. Volume Changer raises or lowers the entire file; Audio Effects adds a fade, normalization, or simple echo; Audio Equalizer applies a tonal correction. None of those replaces a clean recording when the take itself is bad, but they help when the recording is acceptable and only needs finishing work after the cut.
| Cut it when | Rerecord or rework when |
|---|---|
| One region is bad and the rest of the take is clean | Noise, hum, or distortion runs through the whole file |
| The error has clear start and end times | The error has no clear boundary and bleeds into every section |
| You would keep roughly 80 percent or more of the original | You would keep less than half, or the keepers need heavy processing |
| The problem is content (a wrong word, an intro, a long pause) | The problem is capture (clipping, room tone, low level) |
What Audio Cutter Does to Your File (and What It Does Not)
Before you commit to cutting, it helps to know exactly what the output will contain. Audio Cutter reads the file in the current browser, lets Web Audio decode it, displays the decoded duration, sample rate, channel count, and frame count, then exports the selected range as a freshly encoded RIFF/WAVE file containing interleaved, little-endian, signed 16-bit PCM samples. Floating-point Web Audio samples are clamped to the PCM range before they are written, and multiple channels remain separate in their original order and are interleaved on export. That description is exact and worth comparing against what stays in many audio editors:
| Property | Cut WAV keeps it | Cut WAV strips it |
|---|---|---|
| Audio sample data in the selected range | Yes | — |
| Original channel order and count (up to 8 channels) | Yes | — |
| Decoded sample rate (may differ from source if Web Audio resampled) | Yes | — |
| MP3, AAC, Vorbis, Opus, or FLAC compressed encoding | — | No |
| Tags, album art, cue points, and chapters | — | No |
| Encoder settings, loudness metadata, and bitrate | — | No |
| Source filename and original MIME type | — | No |
Because the WAV is uncompressed, its data size is the selected frame count multiplied by the channel count and two bytes per sample, plus a 44-byte WAV header. A short cut from a compressed source can therefore produce a file noticeably larger than the original. The tool reports the output duration and file size before download so you can confirm the size before committing.
How to Cut Audio in Your Browser with Audio Cutter
Once the decision lands on a trim, the practical work is short. Audio Cutter operates entirely inside the current browser tab, so the original file never leaves your device. Follow these steps in order:
- Open Audio Cutter in a desktop browser and click the file picker. Choose a browser-decodable audio file within the displayed file and decoded-audio limits (file no larger than 25 MB; decoded audio no longer than 15 minutes, using no more than eight channels or 192 kHz, and containing no more than 30 million samples across all channels).
- Wait for Web Audio to decode the file. The page shows the decoded duration, sample rate, channel count, and frame count before cutting. Confirm those numbers match what you expect from the source.
- Enter a non-negative start time in seconds. Enter an end time value that is later than the start and does not exceed the exact decoded duration. Both values are rounded to the nearest sample frame at the decoded rate, so the start frame is included and the end frame is excluded.
- Select Cut and export WAV. Preview the PCM16 result using the on-page player and listen for the spoken word or musical transient at both boundaries.
- Download the WAV when the preview matches what you want. Changing either time field clears the previous WAV, and cutting again replaces and releases the prior download URL. Choosing another file clears the old decoded buffer, players, errors, and result.
Two timing rules are enforced rather than silently fixed: end must be greater than start, and a selection that rounds to less than one frame at the decoded rate is rejected instead of producing an empty file.
Reading the Numbers Before You Press Cut
The page shows four numbers after decoding, and each one changes how the output behaves. The decoded duration is the working length of the buffer; your end time cannot exceed it. The sample rate describes the decoded working buffer and the exported WAV, not necessarily the encoded rate of the source file, because Web Audio may resample the source to the AudioContext sample rate while decoding, as documented in the MDN BaseAudioContext decodeAudioData reference and the MDN AudioBuffer reference. The channel count is what the WAV will keep in its interleaved layout; the tool supports up to eight channels and retains the original order. The frame count is the total number of sample frames in the buffer.
These four numbers let you predict the output size before you download. The output WAV data is:
frames in selection × channel count × 2 bytes per sample + 44-byte header
For a stereo (2-channel) selection of 60 seconds at a 44.1 kHz decoded rate: 60 × 44100 = 2,646,000 frames per channel. Total samples across both channels = 2 × 2,646,000 = 5,292,000. Bytes for sample data = 5,292,000 × 2 = 10,584,000 bytes. Add the 44-byte header and the final WAV is 10,584,044 bytes (about 10.09 MB). The tool reports the output duration and file size before download, so you can confirm this estimate against the displayed value.
Preview, Confirm, and Decide Again
Cutting is rarely a one-shot decision, so the preview is where the second decision happens. Listen for the first voiced consonant or first musical transient at the start boundary, and confirm the last syllable or downbeat at the end boundary is not clipped. If a breath, a reverb tail, or a sustained note is sitting on a boundary, move the time by a small amount and cut again. Each new cut replaces the previous WAV, so only the last download is the result you keep. Keep the original file so you can redo the cut at a different boundary without rerecording. The verification workflow in more detail is covered in How to Check the Result After You Cut Audio, which complements the steps above.
Two cautions follow from how the tool works: it is exact at the decoded sample rate, but it is not a full digital audio workstation. Reach for a dedicated editor when you need waveform-level fades, lossless compressed-frame cutting without re-encoding, metadata preservation, or production mastering.
Other Tools to Reach For When Cutting Is Not Enough
The decision to cut is often the first in a chain. If after the cut the result still has level, tonal, or transition problems, the rest of the Lizely audio category covers the common follow-ups, all of which share the same local-only model of decoding in the browser and exporting a new PCM16 WAV.
| If the problem is | Use this tool | Why it fits |
|---|---|---|
| Whole-file level is too quiet or too loud | Volume Changer | Applies uniform gain or mute before download |
| Tonal balance is off across the take | Audio Equalizer | Three-band peaking EQ with a single PCM-clipping safety scale |
| Click, hiss, or hum survived the cut | Audio Effects | Deterministic echo, reverb, reverse, fade, or normalization |
| A spoken line needs to play backwards | Reverse Audio | Reverses every decoded sample locally |
| Multiple clips need to sit in one file | Audio Joiner | Joins 2–10 decoded files in a chosen order |
| Key of a musical clip is wrong | Audio Pitch Changer | Resamples within a one-octave range up or down |
| Tempo needs to change without a re-record | Audio Speed Changer | Speeds up or slows down the decoded file |
When the next step is "fix the level," "fix the tone," or "join two clips," the path stays in the same place, with the same rule that applied to cutting: confirm the boundary you are changing is actually the problem, then apply the smallest fix that produces the result you want.
If you're weighing options, Document the Steps You Use to Cut Audio in a Browser covers this in detail.