Audio files come in a confusing alphabet of formats: MP3, AAC, M4A, FLAC, WAV, OGG, Opus, WMA and more. Underneath, they differ in only a few ways, and once you know them the choice becomes easy. This guide explains the vocabulary first and then recommends a format for each common job.
The basics: sampling, bit depth and bitrate
Digital audio measures a sound wave many times per second. The sample rate is how often (44,100 times per second for CD audio, 48,000 for video). The bit depth is how precisely each measurement is stored (16 bits for CDs, 24 bits in studios). Uncompressed audio at CD settings needs about 10 MB per minute. That is what WAV and AIFF files contain.
Bitrate is the number of bits used per second of audio after compression, expressed in kbps. A 128 kbps MP3 is about one megabyte per minute, a tenth of the uncompressed size. Higher bitrate generally means better quality and bigger files, with diminishing returns after a point.
Lossy and lossless
Lossless formats such as FLAC, ALAC and WAV preserve every sample. FLAC and ALAC compress to roughly half the size of WAV, and decode back to identical audio. Lossy formats such as MP3, AAC, Opus and OGG Vorbis use a model of human hearing to remove detail you are least likely to notice, reaching sizes five to ten times smaller than lossless.
The formats, one by one
MP3
The format everyone knows. Plays on every phone, car stereo, computer and smart speaker made in the last two decades. At 192 to 256 kbps most listeners cannot distinguish it from the original in normal listening; at 128 kbps, artefacts are audible on good equipment. Choose MP3 when compatibility matters most.
AAC and M4A
A newer codec designed as MP3's successor, delivering better quality at the same bitrate. Files usually carry the .m4a extension. It is the native format for Apple devices and streaming services. A 128 kbps AAC file often sounds closer to the original than a 128 kbps MP3.
Opus
The most efficient lossy codec in common use. It excels at speech, sounding clear at very low bitrates such as 24 to 32 kbps, and also handles music well. It powers many voice-chat and messaging services. Older players and some car systems cannot play it, so it suits files you control the playback of.
OGG Vorbis
An open, patent-free alternative to MP3, popular in games and open-source software. Quality is comparable to AAC. Hardware support is more limited.
FLAC and ALAC
Lossless compressed formats for music collections. FLAC is open and widely supported on Android and desktop players; ALAC is Apple's equivalent. Use them to archive CDs or buy high-resolution downloads, then make smaller lossy copies for your phone.
WAV and AIFF
Uncompressed. They are ideal for editing, mastering and handing audio to a studio, because there is no decoding step and no loss. They are a poor choice for sharing, because the files are huge.
WMA and AMR
WMA is Microsoft's older format and plays reliably mainly on Windows. AMR is a low-bitrate speech format used by older phones for voice recordings. If you have files in these formats, converting them to MP3 or M4A makes them easier to use elsewhere.
Which format should I use?
- Music for everyday listening: MP3 at 192 to 256 kbps, or M4A (AAC) at 160 to 256 kbps.
- Podcasts and audiobooks: MP3 at 64 to 96 kbps mono, or Opus at 32 to 48 kbps, or M4A at 64 kbps. Speech needs far less bitrate than music.
- Voice notes and meeting recordings: M4A or Opus at a low bitrate keeps files tiny and clear.
- Archiving or editing: FLAC or WAV. Keep the lossless master and make lossy copies as needed.
- Ringtones: M4R for iPhone, MP3 or OGG for Android, trimmed to under 30 seconds.
Things that surprise people
Converting a low-quality MP3 to WAV or FLAC does not improve it. The detail was removed when the MP3 was made. You only get a bigger file. Likewise, converting MP3 to a higher-bitrate MP3 cannot restore anything.
Each lossy conversion costs a little. Going MP3 to M4A to MP3 adds two rounds of loss. Convert once, from the best source you have.
Mono is fine for speech. Converting a mono voice recording to stereo doubles nothing useful. Choosing mono at the same quality setting halves the size.
Loudness differs from quality. A file that sounds quiet is not low quality; normalisation is a separate step done in an editor.
Extracting audio from video
Lectures, interviews and screen recordings are often easier to use as audio. Video files such as MP4, MOV and MKV contain an audio track that can be saved on its own. Our Audio Converter accepts video files and saves the soundtrack as MP3, M4A, FLAC and many other formats.
A simple rule of thumb
If you do not know who will play the file, choose MP3. If you are keeping it for yourself and want the best sound, keep it lossless. If you are recording speech and want the smallest file, choose Opus or low-bitrate M4A. Everything else is a refinement of those three decisions.
Bitrate and sample-rate cheat sheet
| Content | MP3 | AAC / M4A | Opus |
|---|---|---|---|
| Speech, podcast (mono) | 48 - 64 kbps | 40 - 64 kbps | 24 - 32 kbps |
| Casual music | 128 - 160 kbps | 96 - 128 kbps | 80 - 96 kbps |
| Good-quality music | 192 - 256 kbps | 160 - 192 kbps | 112 - 128 kbps |
| Near-transparent | 320 kbps | 256 kbps | 160 kbps |
Sample rate rarely needs attention. 44.1 kHz suits music and 48 kHz is standard for video. Speech recordings sound fine at 22 or even 16 kHz, which is one reason phone voice notes are so small. Converting to a higher sample rate than the source adds nothing; converting down is a legitimate way to save space for speech.
Metadata and album art
Tags such as title, artist and cover art travel with the file in most formats. Conversions normally keep them, but formats differ in how well they store them: MP3 uses ID3 tags, M4A uses MP4 atoms and FLAC uses Vorbis comments. If tags go missing after a conversion, a tag editor can restore them without touching the sound.