[pmlnr]

EVERYONE can tell 128kbp vs FLAC if they listen even a bit. 128kbps was always bad, especially if it's an old upload.

320kbps (which is ~Spotify 'extreme') vs FLAC is a different story, you do need some reasonable (read at least a $30 soundmagic E10 in-ear, not some wunderbar looking designer latest fashion trend on-ear for $200+) equipment and some training to spot that.

EDIT Also doesn't tackle the problem that most people hear music, not see it, so 480p vs 1080p for MUSIC isn't a deal breaker now, is it?

[nonbirithm]

I've always wondered if 320 is supposed to be distinguishable from lossless.

Like, wasn't MP3 encoded at 320kbps designed to sound like a CD, while still being compressed? If it's still possible to tell 320 from FLAC in enough circumstances, then is that a failing of 320? If so, why stop at 320? Why wouldn't there be a bitrate somewhere between 320 and uncompressed that makes it impossible for practical purposes to tell the difference between lossy and lossless?

[pmlnr]

Does a 1% compression jpeg _supposed_ to be distinguishable from the orginal? Probably not, until it goes into a 4m x 3m print.

MP3 does the same to sound that jpg does to colour: those similar enough become one. And it also cuts off the "inaudible" sounds. If you open a spectogram, 320 still cuts at 20khz, whereas most CDs have until 21-22khz - because there are people out there who hear those sounds. (There are some "ultrasonic" cat & mammal scare away devices, which I hear, and they are literally painful to me.)

But answering your question: no, the goal was to have a good enough, but massively smaller file. The goal was never not to be possible to tell them apart.

FLAC compresses a CD to roughly half of it's size - that's already a big win.