Convert FLAC to MP3

Most conversions into a lossy format on this site come with a warning attached, because the source was already lossy and the second encode compounds the first. This one does not. A FLAC holds every sample of the original recording, so an MP3 made from it is a first-generation encode: whatever the encoder discards, it discards once, working from complete information.

That is the best position an MP3 can be made from, and it is why converting a library from FLAC produces noticeably better results than converting the same albums from someone else’s MP3s. The decision is not whether to lose something — it is whether the destination is worth the one loss, and a car stereo that will not play FLAC settles that question by itself.

A minute of FLAC runs about 5 to 6 MB for typical music, so a 60-minute album lands near 350 MB. The same album at 192 kbps is about 85 MB and at 320 kbps about 140 MB. A 200-album FLAC collection of roughly 70 GB becomes 17 GB at 192 kbps.

Those figures are what make the conversion worth doing rather than streaming instead. A phone, a memory card in a car, an old player with 32 GB in it — none of them will hold the FLAC library and all of them will hold the MP3 one, with room left over. The saving is not marginal and it is the entire reason this page exists.

A studio-master FLAC at 96 kHz or 192 kHz cannot be carried into MP3 at that rate: MP3 defines 32,000, 44,100 and 48,000 Hz and nothing above, so a high-resolution source is resampled on the way in. That step is not a defect of this converter — it is the format ceiling, and every MP3 encoder ever written does the same thing.

It is also, for practical purposes, not the loss anyone should be worried about. The audible damage in a 192 kbps MP3 comes from the encoder discarding detail inside the audible band, not from the ultrasonic content that the resampling removes. If the high sample rate is the point of the file, it was never going to survive as an MP3 and the FLAC is the copy to keep.

The quality control offers three bands rather than a kilobit figure, and for this pair the middle one is the answer for most libraries. MP3 at around 192 kbps is transparent enough that the difference from the FLAC is inaudible on the equipment most listening actually happens on — a car, a phone, earbuds, a kitchen speaker.

The high band earns itself in two situations. The first is material the encoder finds hard: live recordings with applause, orchestral music, anything with cymbals and long reverb tails. The second is when the MP3 will be handed on and possibly re-encoded by somebody else, because every later encode starts from what this one produced. If the destination has room, taking the high band costs disk and settles the question.

MP3 encodes audio in fixed-size blocks, and a track almost never ends on a block boundary, so the encoder pads the end and the start with a short run of silence — typically a few dozen milliseconds each. Encoders record how much they added so that a well-behaved player can trim it, but the information sits outside the audio and not every player reads it.

On separate songs nobody notices. On a live album, an opera, a mixed DJ set or any record where one track runs into the next, a listener hears a tick or a stutter at every boundary. There is no encoder setting that removes it. For those specific albums the right answer is to keep the FLAC and convert the rest, which is a much smaller compromise than converting nothing.

This is one of the few audio conversions where metadata has a destination on both sides. FLAC stores named text fields plus a dedicated block for cover art; MP3 stores ID3 frames that cover the same ground. The concepts line up well enough that a well-tagged FLAC library maps onto a well-tagged MP3 one.

The mapping is not guaranteed to be complete, and the sensible move is to convert one album, open it in whatever will play it, and look. Composer, disc number, replay gain and custom fields are where mismatches show up first. Fixing them in a tag editor across a finished batch takes minutes; discovering them after converting 200 albums is a different afternoon.

Treat the MP3s as a derived set that can be regenerated at any time and thrown away without a thought. That framing decides everything else: it is why the bitrate choice is low-stakes, why it does not matter if the tags need touching up, and why you can redo the whole thing differently in a year.

It also settles the question people actually mean when they ask whether to keep the FLACs. Deleting them converts a library that can produce any format into one that can only produce worse MP3s. The disk that holds a FLAC collection is cheap; the collection is not.

MP3 is the most requested audio target on the internet and the one format WebCodecs uniformly refuses to encode — no browser ships an MP3 encoder. This page works around it by downloading a WebAssembly build of LAME, the reference encoder, the first time you convert to MP3 in a session.

That download happens once, is loaded only for MP3 targets, and is the entire reason the conversion can stay on your machine. It also explains a small delay before the first track starts and none before the second. Someone converting a FLAC to WAV on this site downloads none of it.

The batch limit is a hundred files, which is most of a box set, and each comes back separately inside one ZIP. The per-file ceiling is 100 MB, and because FLAC compresses to roughly half of uncompressed audio, that works out at something like seventeen to twenty minutes per track — comfortable for songs, tight for a single-file concert recording.

Nothing is uploaded, so the constraint on a large library is processor time rather than a queue or an allowance. Two or three drops of a hundred files each will move most collections, and the tab can be left to work while you do something else.