Convert Audio to WAV

Features

• MP3 / M4A / AAC / WAV / FLAC / OGG / Opus — anything the browser decodes
• Sample rate: keep, 48k, 44.1k, 22.05k, 16k or 8k
• Channels: keep, mono (averaging left and right) or stereo
• Bit depth: 16-bit integer (universal) or 32-bit float (for further processing)
• Peak normalisation to −1 dBFS, reporting exactly how many dB it raised
• Anti-aliased resampling (through the browser's offline audio context, not hand-written interpolation)
• Preview before processing, so you can confirm the right file
• Nothing is uploaded and no server is involved; it works offline

How to use

1. Drop an audio file in (MP3, M4A, WAV, FLAC all work)
2. Pick a sample rate and channel layout — 16k mono for speech
3. Tick normalise if the recording came out quiet
4. Click Start, then download the .wav

FAQ

Why can I not convert to MP3?

Because browsers ship no MP3 encoder. The MP3 algorithm was patented for years, and browsers implemented decoding (playing) but not encoding (producing). Encoding would mean loading several megabytes of JavaScript encoder, which would make this page slower than every other tool here. So the export is WAV, which is lossless; convert it with any player if you need MP3, and that step costs nothing in quality because the WAV already holds the full waveform.

Why is the exported WAV so large?

Because WAV is uncompressed: every sample is written out as-is, about 10 MB per minute for 44.1 kHz stereo 16-bit (44100 × 2 channels × 2 bytes ≈ 176 KB per second). That is exactly why it is lossless — compression and quality trade off against each other. Two ways to shrink it: drop the sample rate (22.05 kHz halves it) and downmix to mono (halves it again), so both together give a quarter. Speech at 16 kHz mono comes out about a ninth of the original.

Does downsampling degrade the sound?

It loses high frequencies, but it does not add noise — two different things. 44.1 kHz captures up to 22 kHz (the ear tops out near 20 kHz); 16 kHz captures up to 8 kHz, so the result sounds duller while speech stays clear. What genuinely ruins quality is downsampling without an anti-aliasing filter: high frequencies fold back into the audible range as metallic buzzing, and plenty of online converters do exactly that. This tool uses the browser's offline audio context, which filters properly, matching the quality of a media player.

Does normalisation change the tone?

No. Normalisation is a single multiplier applied to the whole clip, scaling every frequency equally, so the timbre (the relative strength of frequencies) is untouched — unlike a compressor or limiter, which alter dynamics. The one risk is raising something already clipped: if the recording peaked at the input, normalising lifts the distortion along with everything else. That is why the target is −1 dBFS rather than 0, and why the result reports how many dB were added.

Is the file uploaded?

No. Decoding uses the browser's `decodeAudioData`, the maths happens in memory, and the export writes a WAV header in a few lines. There are no network requests and it works offline. Recordings and meeting audio should not pass through someone else's server.