Every audio tool eventually asks you the same question: which format do you want? MP3, AAC, WAV, FLAC, OGG, Opus, M4A, WMA. Most guides answer it with a wall of specifications that leaves you no better off than before.
The honest answer is that there are only three questions that matter, and once you have answered them the format picks itself. This guide walks through those three questions, then gives you the numbers and a straight recommendation for each common situation.
The one distinction that explains everything else
Audio formats split into two families, and almost every practical difference follows from which family a format belongs to.
Lossless formats keep every sample the recorder captured. WAV and AIFF store them raw; FLAC and ALAC compress them the way a ZIP file compresses a document — smaller, but you get back exactly what went in. Decode a FLAC and you have the original bit for bit.
Lossy formats — MP3, AAC, Opus, Vorbis, WMA — throw information away permanently. They use a model of human hearing to discard what you are unlikely to notice: sounds masked by louder sounds, frequencies past the edge of your hearing. The result is dramatically smaller, and for listening it is usually indistinguishable. But the discarded data is gone, and no later conversion brings it back.
This is why “convert my MP3 to FLAC to improve the quality” does not work. You get a large file that faithfully preserves the damage the MP3 already did.
The formats, briefly and honestly
- MP3 — the oldest lossy format still in wide use, and technically the weakest. It survives because it plays on absolutely everything: car stereos, gym equipment, DJ gear, cheap MP3 players, every phone and computer ever made. Compatibility is its entire argument, and it is a very strong argument.
- AAC (usually in an .m4a file) — the successor to MP3 and meaningfully better at the same bitrate, especially below 128 kbps. It is what Apple Music, YouTube and iPhone recordings use. Support is excellent nearly everywhere, with older or cheaper hardware the exception.
- WAV — raw, uncompressed, universally readable, and enormous. The standard interchange format for editing and mastering.
- AIFF — Apple’s equivalent of WAV. Same idea, same size, slightly less universal.
- FLAC — lossless compression at roughly half the size of WAV. The right choice for archiving. Its weakness is playback support: it works in most modern players but not reliably in car systems or older devices.
- ALAC — Apple’s lossless format. Essentially FLAC’s equivalent, preferred inside the Apple ecosystem.
- Opus — the best-sounding lossy codec at low bitrates by a clear margin, and the reason voice chat sounds as good as it does. It is what Discord, WhatsApp voice notes and modern video calls run on. Its weakness is that it is a web and messaging codec: many editors and consumer devices still will not open a bare
.opusfile. - OGG Vorbis — an older open-source lossy format, largely superseded by Opus. You mostly meet it in game audio and older downloads.
- WMA — a Microsoft format from the Windows Media era. There is no reason to choose it today; you only encounter it in old libraries and need to convert away from it.
The numbers side by side
Sizes below are for five minutes of stereo audio at CD quality (44.1 kHz, 16-bit). They are arithmetic, not estimates: uncompressed stereo CD audio runs at 1,411 kbps, so five minutes is 1411 × 300 ÷ 8 ≈ 53,000 KB.
| Format | Type | Typical bitrate | 5 min stereo | Plays on |
|---|---|---|---|---|
| WAV / AIFF | Uncompressed | 1,411 kbps | ~53 MB | Everything |
| FLAC / ALAC | Lossless | ~700–900 kbps | ~28 MB | Most modern players |
| MP3 (320) | Lossy | 320 kbps | 12 MB | Everything |
| MP3 (192) | Lossy | 192 kbps | 7.2 MB | Everything |
| MP3 (128) | Lossy | 128 kbps | 4.8 MB | Everything |
| AAC (128) | Lossy | 128 kbps | 4.8 MB | Nearly everything |
| Opus (96) | Lossy | 96 kbps | 3.6 MB | Browsers, apps |
| Opus (64) | Lossy | 64 kbps | 2.4 MB | Browsers, apps |
The important comparison in that table is not top to bottom, it is AAC 128 against MP3 128. Identical size, audibly different result — the newer codec simply spends its bits better. Wherever compatibility allows it, AAC beats MP3 at the same size, and Opus beats both.
So which one should you use?
| Your situation | Use | Why |
|---|---|---|
| Sharing music or a podcast with anyone, anywhere | MP3, 192–256 kbps | Nothing on earth fails to play it, and the quality is beyond most listening conditions |
| Spoken word — lectures, interviews, meetings | MP3, 96–128 kbps mono | Speech needs a fraction of music’s bitrate; stereo buys you nothing |
| Archiving a master recording | FLAC | Lossless, but half the size of WAV |
| Handing audio to an editor or studio | WAV | The universal interchange format; no decode step, no ambiguity |
| Audio for a website or app you control | Opus, or AAC as a fallback | Smallest files at listenable quality; browsers all support it |
| Inside the Apple ecosystem | AAC/M4A, or ALAC for lossless | Native everywhere Apple made it |
| You have WMA, OGG or bare AAC files | Convert to MP3 | These are legacy or niche; MP3 ends the compatibility problem permanently |
If you want the single-sentence version: MP3 when the file is going to other people, FLAC when it is going into storage, WAV when it is going into an editor. That covers the large majority of real decisions.
Bitrate matters more than format
People agonise over MP3 versus AAC and then export at whatever the default happens to be. Bitrate is the bigger lever. A 320 kbps MP3 sounds better than a 96 kbps AAC, despite AAC being the superior codec, because it was given three times the data to work with.
Useful anchors for MP3:
- 320 kbps — transparent for practical purposes. Use when storage is irrelevant.
- 192–256 kbps — the sweet spot for music. Most listeners cannot reliably distinguish 256 from the original on normal equipment.
- 128 kbps — fine for casual listening, audibly thin on good headphones.
- 96 kbps or below — acceptable for speech, poor for music.
Doubling the bitrate doubles the file. Going from 128 to 256 kbps is a real, hearable improvement on music; going from 256 to 320 is mostly insurance.
The mistake that actually costs you quality
Format choice is a small effect. Repeated lossy re-encoding is a large one.
Every time a lossy file is decoded and encoded again — MP3 to AAC, AAC back to MP3, or even MP3 to MP3 at a different bitrate — the new encoder discards data based on what it hears, and what it hears is already a lossy approximation. The damage compounds. This is generation loss, and it is the one audio mistake that is genuinely irreversible.
Worked example. You record a 40-minute interview. The recorder gives you a 400 MB WAV. Two sensible paths:
- Keep the WAV as your master (or convert it once to FLAC, roughly 200 MB, to halve the storage without losing anything).
- Export a 128 kbps mono MP3 for distribution — 40 × 60 × 128 ÷ 8 ÷ 1024 ≈ 37 MB. Small enough to email, and speech at 128 kbps mono is comfortably clear.
What you should not do is edit the 37 MB MP3, export it as MP3 again, then send that to someone who converts it once more. Every export from your lossless master is a first-generation copy; every export from an MP3 is another step down. Keep the master, generate distribution copies from it, and never chain conversions.
Converting between formats without uploading anything
Audio files are frequently the most private things people convert — voice memos, therapy notes, interviews with sources, confidential meetings, unreleased music. Most online converters upload all of it to a server and process it there.
Our audio converter runs the encoder inside your browser tab with WebAssembly, so the recording never leaves your device. There is no upload step and no queue — conversion starts the moment you pick the file. It reads MP3, M4A, AAC, WAV, FLAC, OGG, Opus and WMA, plus the audio track of any common video file.
Direct converters for the most common jobs: WAV to MP3, M4A to MP3, FLAC to MP3, WMA to MP3, Opus to MP3 and MP3 to WAV. If you only need the audio out of a video, use video to MP3.
Frequently asked questions
Does converting MP3 to FLAC improve the quality?
No. FLAC will perfectly preserve whatever it is given, but the data the MP3 encoder discarded is permanently gone. You end up with a file several times larger that sounds exactly like the MP3. Convert to FLAC only when your source is lossless to begin with.
Can people actually hear the difference between MP3 and lossless?
At 256–320 kbps, in blind tests and on ordinary equipment, most listeners cannot reliably tell. It becomes easier on studio monitors or good headphones, with certain material — cymbals, applause, reverb tails. The practical reason to keep lossless is not daily listening, it is that a lossless master survives future editing and re-encoding without compounding damage.
What is the difference between M4A and AAC?
M4A is a container; AAC is the codec inside it. Almost every .m4a holds AAC audio (Apple Lossless is the exception). A bare .aac file is the same audio without the container, which is why .m4a is better behaved — the container carries duration, chapters and artwork.
Which format is best for a podcast?
Record and edit in WAV, publish as MP3. Most podcast hosts require or strongly prefer MP3, and it is the only format guaranteed to work across every podcast app and car system. 96–128 kbps mono for pure speech, 128–192 kbps if you have music beds.
The short version
Decide lossy or lossless first — that follows from whether the file is for listening or for keeping. Then let compatibility break the tie: MP3 for anything going to other people, FLAC for archives, WAV for editors, AAC or Opus when you control the playback. Pick a sensible bitrate, keep your lossless master, and never convert a lossy file into another lossy file.
Need to change a file’s format right now? The RedPandaCompress audio converter is free, has no signup or watermark, and runs entirely in your browser — nothing is uploaded. For a closer look at the lossless-versus-lossy trade-off, see WAV vs MP3, and if you are curious how in-browser conversion works at all, we took the pipeline apart here.
