FLAC vs MP3 vs AAC: Can You Actually Hear the Difference?
3 de agosto de 2026
What lossless and lossy audio actually mean, where MP3 and AAC stop being distinguishable from FLAC, and the file size math behind the decision.
Most people can't reliably tell 256kbps AAC or a well-encoded MP3 apart from FLAC on a phone speaker, stock earbuds, or Bluetooth headphones. The gap opens up with a decent DAC, wired headphones, and a quiet room, and only past a certain bitrate. Below that floor, lossy encoding audibly cuts detail no listening setup can hide.
What "lossless" actually means
Lossless means the decoded audio is bit-for-bit identical to the original. FLAC (Free Lossless Audio Codec) doesn't throw away any audio data. It works more like a ZIP file for sound: it finds patterns and redundancy in the waveform and packs them tighter, then unpacks to the exact same samples on playback. Decode a FLAC file and a WAV of the same recording, and you get identical numbers back.
That's the entire pitch. FLAC isn't "better sounding" than the source, it's the source, just smaller on disk. Whatever's in the original master, good or flawed, comes through unchanged.
What perceptual coding throws away
MP3 and AAC are lossy. They use perceptual coding, which means they model human hearing and delete the parts of the signal a listener is statistically unlikely to notice. The main tool is masking: a loud sound at one frequency hides quieter sounds nearby in frequency and in time, the same way you can't hear someone whisper next to a running blender. The encoder finds those masked components and spends fewer bits on them, sometimes discarding them outright.
This is genuinely well-engineered psychoacoustics (a word which I didn't know up until this moment). It's not just "make the file smaller by cutting corners." At high enough bitrates the discarded information overlaps almost entirely with what masking theory predicts you can't hear anyway. At low bitrates, the encoder runs out of room and starts cutting into audible territory: high-frequency detail smears, cymbals and reverb tails get a swirly "underwater" quality, and transients lose their edge. That's what people mean by compression artifacts. If you haven't paid attention to them so far, there's no going back after you have.
AAC (Advanced Audio Coding) is the newer of the two, standardized after MP3, and it's generally more efficient at the same bitrate: better handling of stereo imaging, higher frequencies, and transients. It's why Apple Music, YouTube Music, and most modern streaming catalogues default to AAC rather than MP3.
Where each codec gets transparent
"Transparent" means a properly conducted ABX test (listener switches blind between the original and the encode, has to guess which is which, repeatedly) can't beat chance. Rough, widely cited thresholds:
- →MP3 at 320kbps CBR is transparent or close to it for nearly everyone, on nearly any gear. The rare person who can pick it out in a blind test usually needs a treated room, a resolving headphone or DAC, and a track chosen specifically to expose codec weaknesses (cymbal-heavy jazz, applause, certain synth pads).
- →MP3 at V0 VBR (variable bitrate, averaging around 245kbps) lands close to 320 CBR quality at a smaller size, because it spends more bits on complex passages and fewer on simple ones instead of a flat rate throughout.
- →AAC at 256kbps is the streaming-industry's own bar for "high quality," and it's a reasonable one. Most listeners, on most equipment, will not reliably distinguish it from lossless.
- →Below roughly 192kbps for either codec, artifacts start showing up on ordinary headphones for people who aren't even trying to listen for them. This is where a lot of pirated and radio-ripped files that circulated in the 2000s actually lived, which is a chunk of why lossy audio has a worse reputation than the modern codecs deserve.
None of this is a controversial claim in audio engineering. It's the whole reason streaming services standardized around 256-320kbps AAC or Ogg Vorbis instead of either the very low bitrates of early MP3 or full lossless for everyone: it's the point past which more bits stopped buying audible improvement for the overwhelming majority of listening conditions.
File size math: is FLAC worth the storage
Take a typical 4-minute track, 16-bit/44.1kHz stereo, the standard CD-quality format:
- →Uncompressed WAV runs about 1,411kbps, which works out to roughly 42MB.
- →FLAC compresses that losslessly, typically down to 50-60% of the WAV size depending on how dense the recording is. Call it 21-25MB for the same 4 minutes, decoding back to the identical waveform.
- →MP3 at 320kbps CBR comes in around 9.6MB.
- →AAC at 256kbps comes in around 7.7MB.
So FLAC costs roughly two to three times the storage of a well-encoded lossy file, for audio that's provably identical to the source rather than "probably indistinguishable." On a phone with 128GB or more, that difference is not the constraint it used to be. On a budget device still shipping 32GB or 64GB with half of it eaten by the OS and apps, it adds up fast across a few thousand tracks.
When the difference is real, and when it isn't
It matters when the whole chain supports it: a source file that's actually a good transfer (not a lossy file re-encoded to FLAC, which some sketchy sources do and which does nothing but waste space), a wired connection or a genuinely high-bitrate Bluetooth codec, headphones or speakers resolving enough to reveal the difference, and a listening environment quiet enough to hear it.
It stops mattering the moment any link in that chain is the actual bottleneck. Standard Bluetooth SBC caps out well below what a 256kbps AAC file already carries, so pairing FLAC with a cheap Bluetooth earbud accomplishes nothing except using more storage to deliver the same audible result. Commuting with traffic noise, phone speakers, and budget wired earbuds all cap what's perceivable long before codec choice does.
If you're archiving music you care about, rip or buy lossless, because storage is cheap and you only get one shot at a clean transfer. If you're building a library to actually carry around and listen to on a commute, a 256-320kbps lossy file is not a compromise, it's the correct bitrate for the equipment doing the listening.
Comparison table
| Format | Lossy or lossless | Typical bitrate | Approx. size, 4-min track | Practical verdict |
|---|---|---|---|---|
| WAV | Lossless, uncompressed | ~1,411 kbps | ~42 MB | Full source data, no compression at all; more space than almost anyone needs to spend |
| FLAC | Lossless, compressed | Variable, ~700-1,000 kbps | ~21-25 MB | Identical to the source on decode, roughly half the size of WAV |
| MP3 320 CBR | Lossy | 320 kbps | ~9.6 MB | Transparent or near it for nearly everyone, on nearly any gear |
| MP3 V0 VBR | Lossy | ~245 kbps average | ~7.4 MB | Close to 320 CBR quality, smaller file |
| AAC 256 | Lossy | 256 kbps | ~7.7 MB | The streaming industry's own "high quality" bar; transparent for most listeners |
| MP3 128 | Lossy | 128 kbps | ~3.8 MB | Audible artifacts on decent headphones; skip for anything you actually care about |
Key takeaways
- →FLAC is provably identical to the source on decode. Lossy formats are a bet on human hearing, a well-tested one, but still a bet.
- →256kbps AAC and 320kbps MP3 are transparent for nearly everyone, on nearly any playback chain that isn't specifically built to expose codec weaknesses.
- →FLAC costs roughly two to three times the storage of a well-encoded lossy file for a benefit that only shows up with resolving gear in a quiet room.
- →Bluetooth SBC caps below what a good lossy file already delivers, so lossless plus a cheap Bluetooth earbud buys nothing audible.
- →Archive in lossless if storage is not the constraint. Encode down to 256-320kbps for anything you're actually carrying around and listening to on the move.
FAQ
Can you actually hear the difference between FLAC and 320kbps MP3?
For the vast majority of listeners on the vast majority of equipment, no, not reliably in a blind test. The difference becomes audible only with resolving headphones or a dedicated DAC, a quiet room, and specific types of recordings that stress lossy codecs, like dense cymbal work or wide reverb tails.
Is AAC better than MP3 at the same bitrate?
Generally yes. AAC is a newer codec with more efficient encoding of stereo imaging, high frequencies, and transients, which is why most streaming services default to AAC rather than MP3 today.
What bitrate is considered CD quality?
CD audio itself is uncompressed 16-bit/44.1kHz, about 1,411kbps. FLAC at that same bit depth and sample rate is lossless and equivalent to CD quality, just compressed to a smaller file. No lossy bitrate is technically "CD quality," but 256-320kbps lossy is commonly accepted as audibly equivalent for most listening conditions.
Does Bluetooth ruin lossless audio?
Standard Bluetooth SBC caps out at a bitrate below what a well-encoded 256kbps AAC file already carries, so playing FLAC over a basic Bluetooth connection doesn't deliver any of FLAC's advantage. Higher-bitrate codecs like LDAC or aptX HD close that gap, but the receiving headphones or earbuds have to actually support them.
Should I convert my MP3 collection to FLAC?
No. Converting a lossy file to FLAC doesn't recover any of the data the original encode threw away, it just wraps a lossy file in a lossless container and wastes storage doing it. Only encode to FLAC from an actual lossless source: a CD rip, a lossless purchase, or a lossless master.
If you've settled on a bitrate and you're now picking what to play it back on, our comparison of the best offline Android music players covers FLAC support, EQ depth, and file size across eight apps, not just NeoMusic.
Whatever format you land on, you need a player that decodes it correctly instead of choking on FLAC containers the way some stock players do. NeoMusic runs FLAC and the common lossy formats through Media3/ExoPlayer with a built-in EQ, free, no account required.