How to decompress mp3 to uncompressed wav — browser tool
- Step 1Open the decompressor — Open the mp3-to-wav tool. 'Decompress to WAV' is exactly what MP3 to WAV does — decode the compressed bitstream to PCM.
- Step 2Drop the compressed MP3 — Add a single
.mp3. It's decoded locally by FFmpeg WASM; nothing uploads. The tool processes one file at a time. - Step 3Choose the sample rate — For analysis, keep the source rate (usually 44.1 kHz) so you don't introduce a resample. Use 48 kHz to match a video, 96 kHz for a high-rate session, or 16 kHz for speech.
- Step 4Choose stereo or mono — Keep Stereo to inspect channel differences; pick Mono to collapse a dual-mono file into a single trace for cleaner spectral reading.
- Step 5Decompress — Run it. FFmpeg decodes the MP3 frames to
pcm_s16leand writes the WAV. No quality is added — the artifacts baked into the MP3 are now visible in uncompressed form. - Step 6Load into your analysis/re-encode tool — Open the WAV in a spectral analyzer or sample editor, or feed it to a re-encoder. To inspect the spectrum here, use the spectrum-analyzer; to view the waveform, the waveform-generator.
What decompression does — and doesn't — do
Decompressing to WAV decodes the bitstream to PCM. It exposes artifacts; it never removes them.
| Aspect | Result of MP3 → WAV | Why |
|---|---|---|
| Bitstream → samples | Decoded to 16-bit PCM | WAV stores raw samples, not frequency coefficients |
| Compression artifacts | Made visible, not removed | Pre-echo, birdies, lowpass are encoded into the audio |
| High-frequency lowpass | Stays (e.g. brick wall ~16–20 kHz) | Most MP3 encoders cut highs; WAV preserves the cut spectrum |
| File size | Much larger | Uncompressed PCM vs compressed bitstream |
| Quality | Unchanged (no recovery) | Lossy data is gone before decompression |
Decompression options (real controls)
Sample rate and channels only. Bit depth is fixed at 16-bit PCM.
| Control | Choices | Default | FFmpeg arg |
|---|---|---|---|
| Sample rate | 44.1 / 48 / 96 / 16 kHz | 44.1 kHz | -ar <hz> |
| Channels | Stereo / Mono | Stereo | -ac <1|2> |
| Bit depth | 16-bit (fixed) | 16-bit | pcm_s16le |
Tier limits for decompression
Per-file size and duration caps. The cap is on the compressed MP3 input; the WAV output is larger.
| Tier | Max size | Max duration | Files |
|---|---|---|---|
| Free | 50 MB | 30 min | 1 |
| Pro | 200 MB | 120 min | 10 |
| Pro-media | 100 GB | Unlimited | 100 |
| Developer | 100 GB | Unlimited | Unlimited |
Cookbook
Decompression recipes for analysis and clean re-encoding. Output is uncompressed 16-bit PCM WAV.
Decompress for spectral inspection of the lowpass shelf
You suspect a file is a transcoded MP3 (re-encoded from another MP3). Decompress to WAV, then read the spectrum — a hard lowpass around 16 kHz is a tell.
Input: suspect.mp3 (128 kbps, 44.1 kHz, stereo)
Options: Sample rate = 44.1 kHz · Channels = Stereo
Output: suspect.wav (16-bit PCM)
Then: /audio-tools/spectrum-analyzer -> look for energy cutoff ~16 kHz
(classic 128 kbps MP3 brick-wall lowpass).Clean PCM intermediate before re-encoding
You must deliver an MP3 at a different bitrate. Decompress to WAV once, then encode from PCM — avoiding a double-decode you don't control inside a chained converter.
Step 1: mp3-to-wav -> master.wav (16-bit PCM, 44.1 kHz, stereo) Step 2: /audio-tools/wav-to-mp3 -> choose target bitrate (e.g. 192k) Why: one explicit decode + one explicit encode beats an opaque MP3->MP3 chain.
Forensic decompression, mono trace
A case recording is dual-mono MP3. Decompress to a single mono WAV trace for a cleaner waveform reading in a forensic editor.
Input: evidence.mp3 (64 kbps, 44.1 kHz, dual-mono) Options: Sample rate = 44.1 kHz · Channels = Mono (-ac 1) Output: evidence.wav (16-bit PCM, mono) Stays on the examiner's workstation — nothing uploaded.
Decompress and view the waveform
You need a waveform image of an MP3 for documentation. Decompress, then render the waveform.
Step 1: mp3-to-wav -> clip.wav (16-bit PCM) Step 2: /audio-tools/waveform-generator -> PNG of the waveform (The waveform tool reads decoded audio; PCM WAV is the cleanest input.)
Decompress to 16 kHz for speech artifact analysis
For a speech-only file where you only care about the voice band, decompress to 16 kHz mono — smaller, and focused on the relevant spectrum.
Input: call.mp3 (48 kbps, 44.1 kHz, mono) Options: Sample rate = 16 kHz · Channels = Mono Output: call.wav (16-bit PCM, 16 kHz, mono) Nyquist = 8 kHz — appropriate for telephone-band speech analysis.
Edge cases and what actually happens
Expecting decompression to remove MP3 artifacts
Not removedDecompressing decodes the bitstream but the artifacts (pre-echo, birdies, lowpass) are part of the audio. WAV makes them visible/measurable, not gone. For perceived cleanup, a denoiser addresses noise — not lossy artifacts.
High-frequency content is missing after decompression
ExpectedMany MP3 encoders apply a lowpass (e.g. ~16 kHz at 128 kbps). The WAV faithfully preserves the already-cut spectrum. Missing highs were removed at encode time, not by the conversion.
WAV vastly larger than the MP3
By designUncompressed PCM is far bigger than a compressed bitstream — ~10.1 MB/min at 16-bit/44.1 kHz/stereo. That's the cost of an analysis-ready uncompressed file.
Re-encoding the WAV reintroduces compression
ExpectedGoing WAV → MP3 with wav-to-mp3 is lossy again. The benefit of decompressing first is a controlled single decode/encode, not a lossless round-trip.
Need 24-bit for a forensic standard
16-bit onlyOutput is 16-bit PCM. The MP3 holds no >16-bit information, but if a procedure mandates 24-bit WAV, this tool can't produce it.
Transcoded (MP3-of-MP3) source
DetectableDecompress and check the spectrum: stacked lowpass shelves or unusually low cutoff suggest the file was re-encoded from another lossy source. The WAV is your evidence; use spectrum-analyzer.
Sample rate change during decompression
AvoidableIf you pick a rate other than the source rate, FFmpeg resamples during decode. For pristine artifact analysis, keep the source rate so the only change is bitstream→PCM.
Input exceeds the tier cap
rejectedLarge/long MP3s can exceed the size or duration cap (Free 50 MB / 30 min). Split with audio-trimmer or use a higher tier.
Multiple case files at once
single fileDecompress one file at a time. For a batch, run files individually — the tool has no in-UI queue, and everything stays in-browser.
Encoder-delay padding shifts the start
PreservedMP3 padding sits at the file head. The decompressed WAV reflects the audible audio; for frame-exact timing, trim the lead with audio-trimmer.
Frequently asked questions
What does 'decompress MP3 to WAV' actually mean?
It decodes the compressed MP3 bitstream back into raw samples and stores them as uncompressed 16-bit PCM WAV. That's the format analysis and re-encode tools expect as input.
Does decompressing remove MP3 artifacts?
No. Pre-echo, high-frequency 'birdies', and the encoder's lowpass are baked into the audio. Decompressing makes them visible and measurable in uncompressed form — it doesn't undo them.
Why are the highs gone after decompression?
Most MP3 encoders apply a lowpass (often ~16 kHz at 128 kbps). The WAV preserves the already-cut spectrum. The high frequencies were removed when the MP3 was made, not by the conversion.
What output format do I get?
16-bit PCM WAV (pcm_s16le) at your chosen sample rate and channel count. No bit-depth control — always 16-bit.
Why decompress before re-encoding?
Re-encoding from a clean PCM intermediate gives you a controlled single decode + single encode, instead of an opaque MP3-to-MP3 chain. Use wav-to-mp3 from the WAV for the new bitrate.
Can I tell if an MP3 was transcoded from another MP3?
Often yes — decompress to WAV and inspect the spectrum with the spectrum-analyzer. Stacked lowpass shelves or an unusually low cutoff point to a re-encoded lossy source.
Is this safe for forensic/evidence audio?
The conversion runs in-browser with no upload, so the file never leaves your workstation. Your own chain-of-custody and procedural requirements still apply to the files you hold.
Will the WAV be much bigger than the MP3?
Yes — uncompressed PCM is far larger, about 10.1 MB/min at 16-bit/44.1 kHz/stereo. That's the trade for an analysis-ready file.
Should I keep the source sample rate?
For artifact analysis, yes — keep the source rate so the only transformation is bitstream→PCM. Resampling adds a second processing step that can blur what you're measuring.
Can I view the spectrum or waveform here?
Yes. After decompressing, use the spectrum-analyzer for frequency content or the waveform-generator for an amplitude image.
Does it run offline / without upload?
FFmpeg 8.1 runs as WASM in your browser. The MP3 is decoded locally and never uploaded; after the page loads, the decode works offline.
Can I decompress several files at once?
The tool is single-file. Decompress each individually. If you then need to combine them, use the audio-merger after converting each to WAV.
Privacy first
Every JAD Audio tool runs entirely in your browser via FFmpeg (WebAssembly) and RNNoise. Your audio files never leave your device — verified by zero outbound network requests during processing.