Music streaming tiers differ in three documented ways: the codec and bitrate each tier uses, whether playback is lossless or lossy, and whether the service applies loudness normalization — the level-matching decision that arguably shapes what you hear more than any bitrate jump. The numbers are public: major services run their standard tiers near 96–160 kbps on free plans and 256–320 kbps on paid ones, using codecs like AAC, Ogg Vorbis, or MP3, while hi-fi tiers add lossless FLAC streams at CD quality (16-bit, 44.1 kHz) and sometimes above. What the numbers mean for ears is the part worth unpacking.
AA Digital Sound publishes information, not purchasing advice — this is a factual breakdown of documented codec and bitrate behavior, not a recommendation between services.
What is a codec doing to the music?
A lossy codec compresses audio by discarding signal its psychoacoustic model predicts you will not miss — a science built on masking, the documented phenomenon where a loud sound hides quieter ones nearby in frequency and time. At high bitrates the discards fall below audibility for most listeners and most material; at low bitrates the model starts cutting into audible territory, and the classic artifacts appear: a dulling of cymbals («smearing»), a hollow watery shimmer in reverb tails, and collapsed stereo space. The MP3 era's worst artifacts came from 128 kbps encodes a generation ago; today's default encoders at 256 kbps and up are documented in blind tests to be largely transparent — indistinguishable from the source — for most listeners.
How do the tiers actually compare?
Across the major services, the structure is consistent even where brands and numbers differ.
| Tier type | Typical quality | Codec | What changes audibly |
|---|---|---|---|
| Free/ad-supported | ~96–160 kbps | AAC / Ogg Vorbis / MP3 | Fine artifacts on dense mixes; cymbal smearing; limited on some platforms |
| Standard paid | ~256–320 kbps | AAC / Ogg Vorbis | Documented as transparent for most listeners in blind tests |
| Lossless / Hi-fi | CD quality 16/44.1 lossless | FLAC / ALAC | Bit-exact to the master; differences from 320 kbps rarely identified blind |
| Hi-res tiers | 24-bit / up to 192 kHz | FLAC | Ultrasonic content beyond human hearing; see hi-res analysis elsewhere on this site |
Related stories: Is hi-res audio worth it? What the measurements say · The CD revival, and what collectors are finding.
What is loudness normalization, and why does it matter?
Because it can change the loudness of every track you play more than the codec does. To even out shuffle playback, major services normalize tracks toward a common loudness target — a practice standardized by broadcast loudness recommendations (EBU R128 in Europe, with streaming targets commonly cited near −14 LUFS). Tracks mastered louder than the target are turned down; quieter ones may be lifted or left alone depending on the service. The audible consequence: the loudness war's crushed masters no longer win on streaming — they get pulled down and sound smaller, while dynamic masters hold up. One documented quirk: some services apply limiting after normalization for quiet tracks, another small sonic fingerprint most subscribers never notice.
Does gapless and crossfade behavior differ?
Yes, and on some albums it matters more than bitrate. Live albums and through-composed records assume continuous audio; early streaming pipelines re-buffered between tracks, clipping applause and transitions. Gapless playback — now standard on paid tiers of major services — closes that gap. Some platforms also default to a crossfade that overlaps track endings, which quietly alters the album format itself. Listeners who care about hearing records as sequenced should check both settings; they are documented features, and they are free to fix.
What about download quality and data use?
The same tier numbers apply offline, and the arithmetic is public: a 320 kbps stream uses roughly 2.4 MB per minute — around 1.4 GB for ten hours — while CD-quality FLAC runs near 10 MB per minute, several times more. Free tiers commonly cap download quality or deny downloads entirely. None of this changes the sound; it changes the settings worth choosing before a flight.
Are you actually getting the tier you pay for?
A documented gap worth checking: quality settings are per-device, and many listeners discover their phone has been playing a lower bitrate than their subscription allows — default settings on mobile data are often conservative, and desktop apps default differently from phones. Connectivity matters too: a stream that buffers downshifts temporarily rather than stopping. Audiophile communities treat «check the settings screen first» as the default first step before any claim that a service sounds worse than another, because most such comparisons are really configuration comparisons. The fix is free and takes a minute on every device you own.
Which tier is «good enough»?
The documented answer: a standard paid tier at 256–320 kbps with normalization on is, in blind tests, transparent for most listeners on most equipment. Lossless is a guarantee rather than a revelation — it removes the question instead of adding an audible layer, which is exactly why archival listeners and those with good quiet-listening setups value it. The genuinely audible upgrades in streaming's history were not bitrate jumps but the switch from the 128 kbps era to the 256+ era, and the arrival of loudness normalization that stopped crushed masters from dominating queues. Both are already in every paid tier.