MP3 vs WAV vs FLAC: Which One to Send, and Which to Keep

One is lossy, two are not, and that split matters more than the name

WAV and FLAC both store every bit of the original recording. The only difference between them is that FLAC compresses those bits losslessly, typically to half the size, while WAV doesn't compress at all. MP3 is a different kind of thing. It throws away the parts of the sound a human ear is least likely to notice, and once that is done, the discarded detail cannot be recovered by any means, including converting back to WAV afterwards. A WAV made from an MP3 with MP3 to WAV is not a restored original. It is the same lossy audio taking up ten times the space.

That is the one fact worth remembering when choosing between them. WAV and FLAC are the same audio, one just packed smaller. MP3 is a different, smaller audio signal. The choice isn't really MP3 versus WAV versus FLAC. It is lossy versus lossless, and then a size decision within whichever side you land on.

When lossless earns its size

Keeping a master recording in WAV or FLAC matters when the file might be edited again later, whether trimmed, mixed or layered with other tracks, because every one of those steps re-encodes lossy audio and stacks fresh quality loss on top of whatever the format already threw away. FLAC is the better lossless choice for archiving in almost every case: the same audio, roughly half the disk space, and still reconstructible back to the original WAV whenever a program needs one. The only real reason to keep plain WAV around is a piece of hardware or software old enough that it doesn't understand FLAC at all.

Bitrate is the real decision inside MP3, not the format itself

Once lossy is the right call, for sending a track, background music or a phone recording, the bitrate matters more than anything else. 320 kbps is the format's ceiling and the right pick when the source was already lossless, the usual case when FLAC to MP3 is the tool in hand, and file size barely matters. 192 kbps is where most people stop hearing a difference on ordinary speakers or headphones, which is why it is the sensible default for everyday music. Speech needs far less. 96 or even 64 kbps is enough for an interview or a lecture, because human speech carries almost nothing in the frequency range those lower bitrates cut.

One detail that surprises people: converting an already-compressed MP3 or Opus file up to a higher bitrate doesn't recover anything. The detail that a low bitrate threw away the first time is gone, and requesting a bigger number just stores the same reduced sound in a bigger file.

What survives a conversion, and what doesn't

Artist, title, album and the rest of a track's tags travel across a format change without trouble, along with embedded cover art where the source container carried one. Cover art is common in FLAC and M4A files and essentially never present in plain WAV. None of that is affected by the lossy-versus-lossless question. It is a separate, much simpler copying step that happens alongside whichever encoding decision you made.

A simple rule that covers most situations

Keep the master as FLAC or WAV if there is any chance the audio gets edited, mixed or re-exported later. The point of lossless is to protect against redoing that work from a degraded copy. Send MP3 for anything that is just meant to be listened to: an email attachment, a voice note, a podcast episode, a track shared with a friend. Audio to MP3 makes one from any other format, and Compress Audio shrinks a file that only needs to be smaller. Within MP3, default to 192 kbps for music and drop to 96 or 64 for speech-only recordings. And don't try to fix a low-bitrate file by re-encoding it at a higher number. The only way back to better quality is the original source, not the compressed copy.

What the numbers mean: one minute in each format

Bitrate is how many bits each second of sound gets, so file size is plain arithmetic: bits per second times seconds, divided by eight for bytes. A CD-quality WAV (44.1 kHz, 16 bits, stereo) runs at about 1,411 kbps, which is roughly 10 MB per minute: a four-minute song is around 42 MB and an hour of audio around 635 MB. FLAC packs the same signal into roughly half of that, around 5 MB per minute. MP3 at 320 kbps takes 2.4 MB per minute, at 192 kbps 1.4 MB, at 128 kbps just under 1 MB, and at 64 kbps about 0.5 MB.

Those figures explain the choices people make. A twenty-minute interview recorded as WAV is over 200 MB, far past the 25 MB that Gmail allows in one message; at 64 kbps the same twenty minutes is under 10 MB and goes as an attachment with room to spare, even though mail encoding inflates a file by about a third. An hour is another matter: 635 MB in WAV, about 29 MB at 64 kbps, and around 22 MB with the Voice preset in Compress Audio, 48 kbps and mono, which after that encoding is still too big to attach, so an hour of speech is better shared as a link. A four-minute song at 192 kbps is about 5.8 MB; 192 is also the default in Audio to MP3. Variable bitrate shaves a further ten to twenty per cent off at the same perceived quality. And an album in FLAC takes half the room of the same album in WAV, which is the whole argument for FLAC as the archive format.