How audio conversion works
Every audio format is a different way of storing sound waves as numbers. Converting means decoding the file back to raw samples (tens of thousands of numbers per second for each channel) and encoding those samples in the new format. This tool does both steps inside your browser:
- MP3 is encoded with a WebAssembly build of LAME, the open-source MP3 encoder.
- M4A (AAC) and OGG (Opus) use the encoders built into your browser.
- WAV simply writes the raw 16-bit samples, so it needs no encoder at all.
When nothing needs to change, for example AAC audio going into an M4A file, the tool copies the compressed audio as it is. That is instant and lossless.
Lossy, lossless and uncompressed
| Format | Type | Typical bitrate | 3-minute song |
|---|---|---|---|
| WAV (16-bit, 44.1 kHz stereo) | Uncompressed | 1,411 kbps | about 31.8 MB |
| FLAC | Lossless (input only) | 700–1,000 kbps | about 16–23 MB |
| MP3 | Lossy | 128–320 kbps | 2.9–7.2 MB |
| M4A (AAC) | Lossy | 96–256 kbps | 2.2–5.8 MB |
| OGG (Opus) | Lossy | 64–160 kbps | 1.4–3.6 MB |
Lossy formats save space by removing sounds people are unlikely to hear. Each lossy conversion removes a little more, so convert from the best source you have (WAV or FLAC rather than an MP3) and avoid converting the same file back and forth.
The size formula
For MP3, M4A and OGG the size depends only on the bitrate and the length:
For WAV it depends on the sample rate, channels and 2 bytes per sample:
Worked examples
A 3-minute 20-second FLAC song (200 s) to MP3 at 192 kbps. 192,000 × 200 ÷ 8 = 4,800,000 bytes, about 4.8 MB. The FLAC was probably around 20 MB, so the MP3 is roughly a quarter of the size and sounds the same to most listeners.
The same song to WAV. 44,100 × 2 × 2 × 200 = 35,280,000 bytes, about 35 MB: bigger than the FLAC, with exactly the same sound.
A 30-minute voice memo to mono MP3 at 64 kbps. 64,000 × 1,800 ÷ 8 = 14,400,000 bytes, about 14 MB. Speech stays clear, and the file is easy to email.
Choosing settings
- Bitrate. 128 kbps is fine for speech and podcasts; 192 kbps is a safe default for music; 256–320 kbps is for critical listening. Opus sounds good at lower bitrates than MP3: 96–128 kbps is plenty for music.
- Sample rate. Keep the original unless a service asks for something specific. 44.1 kHz is the CD standard, 48 kHz is the video standard, and 22.05 kHz halves WAV size for speech. Opus always runs at 48 kHz.
- Mono. Halves WAV size and lets lossy formats spend their bits on one channel. Good for voice recordings, not for music.
Tips and edge cases
- Higher bitrate cannot add quality. A 128 kbps MP3 converted to 320 kbps is just a bigger file.
- Browser support varies for M4A and OGG. Firefox cannot encode AAC, and Safari only recently gained audio encoders, so those options are marked when unavailable. MP3 and WAV always work.
- Older browsers without the WebCodecs audio engine still convert files: the sound is decoded with the Web Audio API instead, which uses more memory for very long files.
- Batches are processed one file at a time, and you can download everything as a ZIP.
- Need only a part? Cut it with the Audio Trimmer, which also adds fades. Record something new with the Voice Recorder, or pull the sound out of a video with MP4 to MP3.