Cutting audio in your browser does not keep the original MP3 or AAC encoding — every download is a freshly encoded, uncompressed 16-bit PCM WAV file. This is a structural consequence of how browser audio works: the audio must be decoded into raw samples before any sample-level selection can occur, and a brand-new WAV container is then written around those samples. The source's MP3 or AAC frames, bitrate, encoder settings, and container fields are not preserved in the cut, even though the audible waveform inside the new file is identical to the selected range. The result is therefore larger, not smaller, than the same range extracted losslessly from a compressed source. Understanding this — that the encoding always becomes PCM16 WAV — is the key to planning how you trim, archive, and re-use cut audio, because the tool gives you sample-accurate cutting at the cost of switching to an uncompressed container. For most practical tasks such as clips for slides, voice notes, sound effects, and podcast excerpts that trade is exactly what is wanted, because WAV playback is universal and the file is ready to drag into another editor.
What format does the download actually use?
The download from Audio Cutter is always a RIFF/WAVE file. Inside the RIFF chunk sit a fmt sub-chunk and a data sub-chunk. The fmt sub-chunk declares format code 1 (uncompressed PCM), the channel count of the decoded buffer, the decoded sample rate in hertz, a byte rate, a block alignment of channels times two, and a sample size of 16 bits. The data sub-chunk contains interleaved, little-endian, signed 16-bit integer samples in the channel order of the decoded AudioBuffer. Each frame holds one sample per channel in original order, so a stereo file stays stereo and a file with up to eight channels retains all of its channels.
The two-byte-per-sample figure is the single most important number in the output. A 44.1 kHz stereo frame is therefore 4 bytes (2 channels times 2 bytes). A 48 kHz mono frame is 2 bytes. There is no compression, no predictor, and no bit reservoir — the data size is exactly what the header says it is, and the page shows the duration and file size before the download starts so you can confirm before saving.
Why a browser cutter cannot return your original MP3 or AAC
Browser audio does not let a script copy compressed frames from an MP3 or AAC file without first decoding them. The Web Audio API exposes decoded sample buffers through BaseAudioContext.decodeAudioData, which converts the entire encoded payload into a float-point AudioBuffer. Once audio is sitting in that buffer it is just an array of channel arrays of floats between -1 and 1 — the MP3 frames, the AAC frames, and the original encoder decisions are gone. Writing those floats back into a container is necessarily a re-encode, and the only container a browser script can reliably produce without shipping a custom encoder is RIFF/WAVE.
This is why the question of keeping the original encoding has a flat answer in this category of tools: the answer is no, by construction. Some desktop editors can cut at compressed-frame boundaries and emit a smaller MP3 or AAC because they have a local encoder library, but a browser tab does not. The cutter therefore commits to PCM16 WAV as its output format and stops there. Web Audio may also resample the source to the AudioContext sample rate while decoding, which means the sample rate written into the WAV header is the rate of the decoded working buffer rather than the encoded rate of the original file.
What else is lost when the codec changes
Because the codec is re-encoded from decoded samples, other fields carried by the original container also do not survive the cut. The downloaded WAV contains PCM data and a 44-byte header; nothing else.
| Original container field | Preserved in the cut WAV? |
|---|---|
| Compressed audio frames (MP3, AAC, Vorbis, Opus, FLAC) | No — re-encoded as PCM16 |
| ID3, iTunes, Vorbis comment, or Opus tags | No |
| Album art and embedded pictures | No |
| Cue points, chapters, lyrics | No |
| Loudness metadata (ReplayGain, EBU R128) | No |
| Encoder settings, bitrate, original encoded sample rate | No — the WAV reports the decoded rate |
| Audible waveform of the selected range | Yes, sample by sample |
| Channel count and channel order | Yes, up to the eight-channel limit |
If any of those metadata fields matter for the final use, plan to add them after the cut in a tagger or digital audio workstation. For a slide deck clip or a voice memo this rarely matters; for a music release it does, which is one reason to keep the original MP3 or AAC as the master file and treat the WAV from the cutter as a working intermediate.
How to cut audio and download the WAV
Open Audio Cutter in your browser tab. The whole job runs locally, so your file does not leave the device — for a deeper look at why that matters, see the article on uploads and browser-based trimming.
- Click the file picker and choose a browser-decodable audio file (MP3, WAV, M4A, AAC, Ogg, WebM, or FLAC) that is no larger than 25 MB.
- Wait for Web Audio to finish decoding. The page displays the decoded duration, sample rate, channel count, and total frame count once decoding succeeds; if decoding fails, an error message names the reason and no buffer is kept.
- Enter a non-negative start time in seconds and a later end time in moments. End must be greater than start, and end cannot exceed the exact decoded duration shown on the page.
- Select Cut and export WAV. The tool rounds the start and end to the nearest sample frame at the decoded sample rate, includes the start frame and excludes the end frame, and rejects selections that round to less than one frame rather than producing an empty file.
- Preview the resulting PCM16 audio in the page player. Confirm that a spoken word or musical transient falls correctly at both boundaries before downloading.
- Click Download to save the freshly encoded WAV. The page reports the output duration and file size before the download starts.
Because nothing is uploaded, you can repeat the cut, change the times, and re-download without leaving the tab. Changing either time field clears the previous WAV, and choosing a new file clears the entire buffer, players, errors, and prior result. In-flight reads and decodes are guarded so an older job cannot overwrite a newer selection.
Why the cut WAV is bigger than the source
A 30-second section of a 128 kbps MP3 is roughly 480 kilobytes. The same 30 seconds cut as PCM16 WAV at 44.1 kHz stereo is far larger, because PCM stores every sample uncompressed. The exact data size of the cut is selected frames times channels times 2 bytes per sample, and a 44-byte RIFF/WAVE header sits on top of that data.
Worked example: a 30-second selection at the decoded rate of 44,100 Hz, stereo, is 30 times 44,100 = 1,323,000 frames, times 2 channels = 2,646,000 samples, times 2 bytes = 5,292,000 bytes of audio data. Adding the 44-byte header gives 5,292,044 bytes total, or about 5.05 MiB — roughly ten times the size of an MP3 of the same selection.
Two consequences follow. First, large selections of long files approach the tool's decoded-sample ceiling faster than the duration limit alone suggests, because each second at a high rate costs more samples than the same second at a lower rate. Second, if the downstream consumer really needs MP3 or AAC, the WAV from the cutter is the right input for a second encoding pass — an uncompressed source keeps a re-encode honest, while re-encoding an already lossy file at the same bitrate compounds artifacts.
Which input formats work, and which do not
The file picker recognizes common MP3, WAV, M4A, AAC, Ogg, WebM, and FLAC file names and MIME types, but recognition does not guarantee decode. Actual decoding depends on the codec support inside the browser's Web Audio decoder, which differs between Safari, Chrome, Firefox, operating systems, and browser versions. Container and codec support is the kind of detail that varies quietly across releases, so a file that opens on one machine may fail on another.
| Container | Common codecs inside | Typical browser support |
|---|---|---|
| MP3 | MP3 | Widely supported |
| M4A | AAC, ALAC | Widely supported; ALAC less so |
| WAV | PCM, sometimes float | Widely supported |
| Ogg | Vorbis, Opus | Mostly supported; Opus on most modern browsers |
| WebM | Opus, Vorbis | Mostly supported in modern browsers |
| FLAC | FLAC | Supported in most recent browsers |
An unsupported, damaged, mislabeled, encrypted, or incomplete file produces a clear decode error rather than a silent fallback. The tool does not install a codec, send the file to a conversion server, or substitute a different source — if decoding fails, the page says so and stops.
Limits to check before you cut
Five concrete limits decide whether the file will cut at all, and they are enforced explicitly rather than silently. None of them shorten a file automatically; an over-limit file is rejected with a message.
| Limit | Value | What it means in practice |
|---|---|---|
| Encoded input file size | 25 MB | Larger MP3 or AAC files are rejected by the picker |
| Decoded duration | 15 minutes | Audio longer than this after decode is rejected |
| Channel count | 8 channels | Multichannel files above this are rejected |
| Decoded sample rate | 192 kHz | Higher-rate decoded buffers are rejected |
| Total decoded samples | 30,000,000 across all channels | Long high-rate or multichannel files hit this first |
Because browser decoding necessarily happens before duration and sample count can be inspected, a highly compressed long file can still consume temporary memory during the decode attempt. If the result approaches a limit, choose a shorter selection or a lower-rate source file, or trim the file in a dedicated editor that streams rather than fully decodes.
For dependable results, preview the cut WAV, confirm the spoken word or musical transient at both boundaries, keep the original MP3 or AAC as the master, and use a dedicated editor when you need waveform-level fades, lossless compressed-frame cutting, metadata preservation, or production mastering.