Audio Converter

Which audio format should you actually use?

Audio formats are one of those subjects where the technically correct answer and the practically useful answer are rarely the same. The best-sounding format is easy to name. The one that will actually play on the thing you need it to play on is a different question, and usually the more important one.

Lossy and lossless, in plain terms

Every audio format falls into one of two camps. Lossless formats — WAV, FLAC, AIFF — keep every single sample the recording captured. Nothing is thrown away, so the file is an exact representation of the original. Lossy formats — MP3, M4A, OGG, Opus — permanently discard the parts of the sound that human hearing is least likely to notice, and get dramatically smaller files as the reward.

The discarding is cleverer than it sounds. A loud sound masks quieter ones near it in pitch, and a lossy encoder simply does not store what you were never going to hear. At a high enough bitrate the result is genuinely indistinguishable from the original for almost everyone, on almost any equipment.

The formats worth knowing

Picking a bitrate

For lossy formats, bitrate is the single dial that decides quality. It is measured in kilobits per second, and more of them means more of the original sound survives.

• 128 kbps — audibly compressed on music, perfectly fine for speech
• 192 kbps — the sensible default; most people cannot reliably tell it from the original
• 256–320 kbps — for music you care about, or listening on good headphones

One rule is worth internalising: converting from one lossy format to another loses quality every time. An MP3 turned into an M4A is not restored to its original fidelity — it is a lossy copy of a lossy copy. If you have the lossless original, always convert from that. If you only have the MP3, convert at a bitrate at least as high as the source and accept the small loss.

Speech is a different problem

Everything above assumes music. Voice recordings behave differently: the frequency range is narrower, there is usually one source rather than a full mix, and a listener is far more forgiving of artefacts than they are of a missing cymbal.

For an interview, a lecture, a voice memo or a podcast, 96–128 kbps mono is more than enough, and the file will be a fraction of the size. If you are transcribing it later, quality matters even less — speech recognition works from a heavily simplified version of the audio anyway.

Why this converter does not upload anything

Most free audio converters work by taking your file to a server, converting it there, and giving you a download link. That is a reasonable engineering choice, and it means your recording sits on a machine you do not control, for a period nobody tells you, in a jurisdiction you did not pick.

For a song you downloaded, that may not matter at all. For a recorded meeting, a medical consultation, a legal interview, a voice memo about something private — it should.

This converter runs ffmpeg, the same engine most of those servers use, compiled to run inside your browser instead. Your file is read from disk, converted in your device’s memory, and written back to your downloads folder. There is no server in the path, which is why there is no account, no queue, no daily cap and no size limit beyond what your own machine can hold.

The engine is about 32 MB and is fetched the first time you convert something. After that your browser caches it, and subsequent conversions start immediately — including with no internet connection at all, which is the easiest way to prove nothing is being uploaded.