Also those record-label-bot channels will always have the channel name listed as "[artist name] - Topic".
I suppose I shouldn't be surprised that you of all people had figured out this trick as well.
Long ago, when Youtube still offered a stand-alone 256 kbit/s AAC-LC track option for every video (and even 192 kbit/s AAC-LC in their muxed format 22 legacy encodes), I considered lossless bot uploads the perfect backdoor means of mass-acquiring "first generation, keeper quality" lossy encodes of musical unobtanium. Alas, the Youtube forces that be removed the 256 kbit/s tracks eventually (and later changed even format 22 to only contain 128 kbit/s). The Opus tracks, when they were first introduced, ran at a tolerably tolerable 160 kbit/s too (~192 kbit/s in AAC-LC dog years). But they too got downgraded in the end to 128-130 kbit/s, and are still that way today. It appears with most people listening through iPhones, TV bezel speakers, and cheap earbuds these days, Youtube doesn't think anyone can hear better.
But the audio being provided directly by the copyright owner isn't always a guarantee that they'll use the highest-quality copy available.
Yeah, I should have said more about this. They aren't guaranteed sources of highest-quality-available audio. They simply
tend to be. I have even encountered some bot uploads now and then that were audibly sourced from lossy encodings. Even in cases where I don't clearly hear evidence of multi-generational lossy encoding, I always choose the Opus format and then inspect them using CoolEdit Pro's spectral view for insurance's sake. If I see the spectrum cutting off sharply at 16 kHz, that's a sure sign of the source having been an MP3. And since some MP3 encoders actually preserve audio above that limit, I also look for those characteristic MP3 encoder "spectral holes" -- little square-shaped (or even larger, gouge-sized) blocks of silence -- that many MP3 encoders' multiband gating functions tend to freckle throughout the highest frequencies during periods of bit starvation. The Opus codec doesn't leave those holes, so seeing them in an Opus encode is another dead giveaway that an MP3 was its source material. (Those holes are what's responsible for the "ringy" tonality of MP3 treble artifacting. They essentially consist of millisecond-length bandstops with skirts that are completely vertical and unsplined -- hence creating actual filter ringing that turns on and off at seemingly random frequencies with the speed and density of heavy rain plicking against a windshield. The designers of Opus ingeniously designed their encoder so that during any slices of time where it was starving for bits, it would leave meta instructions for the decoder to fill all the spectral holes it was creating with white noise matching the immediately adjacent spectrum's volume. This is why Opus spectral views look "solid to the top," like lossless audio spectral views -- and as a useful coincidence, it also means you can spot lossy MP3 source material, as with the Opus lossyness never adding a "layer" of its own spectral "holes," you always know that any that are visible are "showing through" from upstream.)
Especially if it's an album that was never officially released in digital form. You excitedly click on the video, thinking "Wow, they finally found the master tape and did a new digital transfer of it!" -- but no, it's just a needledrop from a scratchy copy of the vinyl record, played on a not-very-good turntable. They don't care how good or bad it sounds; they just want to add it into YouTube's Content ID system so they can get the copyright revenue from it.
Yep. Although I have personally experienced the "needle drop digital" phenomenon since long before Youtube hit the scene. I still remember buying CD re-issues of two early 1970s albums (
https://www.discogs.com/release/3914536 and
https://www.discogs.com/release/2222790) in 1990 when I was just a naive kid, and being completely fascinated and perplexed by why I was hearing clicks, pops, and vinyl surface noise in every track on both CDs.
Or, worse yet, it could be a re-recording! I've seen that quite often, especially from one-hit-wonder artists, and groups which had a revolving door of members, like the Ink Spots and the Drifters.
Finicky artists sometimes do the same thing simply "because." Didn't Enya famously try to memory hole the Gaelic version of "Book of Days" by covertly replacing it on all future album pressings with an english re-record? Also, Abby Travis' "Blythe," from her "Glitter Mouth" album, exists in a 5:29 version as well as in a 5:08 version. The 5:08 is the one you see everywhere today, on Youtube, at Bandcamp, and listed in Discogs. The 5:29 version, conversely, only showed up on Napster and on my local NPR station right before the album's official public release. Guess she decided to re-record it at the last second and scrub the original from existence -- or something.