Comment by crazygringo | Hacker News Reader

Comment by crazygringo | Hacker News Reader

Parent Full thread

> Genius said it found more than 100 examples of songs on Google that came from its site.

> In a statement, Google said the lyrics on its site, which pop up in little search-result squares called “information panels,” are licensed from partners, not created by Google... In 2016, Google partnered with LyricFind Inc., a Canadian company that secures deals with music publishers allowing companies such as Google to publish lyrics online.

So Google is properly licensing lyrics from song publishers and uses a third party to provide the lyrics, and somehow that third party wound up with more than a "hundred" -- not a million but a hundred -- that were copy-pasted from Genius at some point.

Except for the apostrophes thing being clever, this sounds like a complete non-issue? That Genius's issue is not with Google but the company they source from... AND that they'd have to prove a wholesale systematic copying (closer to millions of songs) as opposed to it being e.g. some rando employee who got lazy and took a shortcut one week?

I think Google are being deceptive here too. I find a lot of lyrics on Google that are mis-attributed and seem very unlikely to be correctly licensed.

The best example is that Google claim that the copyright to the Star Spangled Banner, the National Anthem of the U.S.A., is owned by "Concord Music Publishing LLC". It's beyond parody.

Song-writers don't seem to have much of a say in their lyrics showing up on these sites, and are often not credited at all. This has been going on a long long time and getting it right doesn't seem to be a priority.

When I search for "Star Spangled Banner" I see "The Star Spangled Banner (Super Bowl XXV) lyrics © Concord Music Publishing LLC"

They're saying the lyrics come from the arrangement used for Super Bowl XXV, and that arrangement is copyright Concord Music Publishing. It would be better if they used the lyrics from the official public domain national anthem, but this isn't that weird.

Google are stating that the lyrics to the Star Spangled Banner are owned by Concord Music Publishing and are somehow derivative of the Super Bowl. That's just plain false, absolute misinformation.

It is obvious in this case that the information is wrong, but what of the many songs that Google misleads people into thinking are likewise locked up in rights? It's simply ridiculous on the face of it.

It is allowed for someone to copyright a performance of a piece of music, separate from the music itself. The text "The Star Spangled Banner (Super Bowl XXV) lyrics © Concord Music Publishing LLC" is saying the lyrics came from a performance of the Star Spangled Banner at Super Bowl XXV, and that Concord Music Publishing LLC owns the copyright of that performance.

But merely performing a song doesn't give you any kind of monopoly whatsoever on the lyrics. Even arranging a song wouldn't give you copyright on the lyrics unless you substantially changed them. It is simply false that Concord Music Publishing LLC has copyright on the lyrics to "The Star Spangled Banner," even if these public domain lyrics were used in an arrangement controlled by that company.

Edit: Getting into the weeds a bit here, but I don't think performances themselves can be copyrighted (in the US, that is). Recordings can be copyrighted, as can musical compositions, lyrics, and subsequent arrangements. It is my understanding that, in order for something to be eligible for copyright, it must be fixed in a "tangible medium of expression."

It's also worth mentioning that music publishing companies generally control compositions, not recordings. But even supposing Concord Music Publishing has copyright on the arrangement used on that particular occasion, that does not imply they have copyright on the lyrics.

Performances can be copyrighted. All of Woodstock 99 is a performance under copyright. I only know because I went and tried to use some footage later.

Parent is saying that the event of Woodstock cannot be copyrighted, but that the footage of the event can be

I initially thought that maybe the lyrics from the Super Bowl performance were different from the public domain ones, but comparing the one on Google with the one at https://amhistory.si.edu/starspangledbanner/the-lyrics.aspx all the differences are just typographical: https://www.jefftk.com/national-anthem-lyrics-comparision.pn...

While other aspects of the Super Bowl arrangement and performance are under copyright, the lyrics aren't.

> "hundred" -- not a million but a hundred -- that were copy-pasted from Genius at some point.

To be fair, they can probably only confirm incorrect lyrics. Correct ones would look the exact same. Presumably, the vast majority are correct

The article says it uses apostrophe fingerprinting... and a lot of songs have lots of apostrophes. Assuming just 4 apostrophes per song, there's only a 1/16 chance of it being an accidental match... but given that it's already quite unlikely for any text (including lyrics) to contain both types of apostrophes, practically speaking it's probably closer to a 1/100 or 1/1000 chance of false positives.

They should be able to accurately confirm whether hundreds of thousands of songs were taken from Genius, without any difficulty.

I'm surprised Google wasn't already normalizing apostrophes and other punctuation anyway. I'm sure they will be now.

> Correct ones would look the exact same.

Unlikely. Official lyrics (printed in CD inserts, for example) often have punctuation mistakes, or they have no punctuation at all.

Unofficial lyrics can have the same words while still differing in punctuation, line breaks, capitalization, or spelling (since most languages have multiple ways to spell the same word -- color/colour, for example).

All the tiny variations could easily create a distinct fingerprint for each source of lyrics.

Official ≠ correct.

See: Van Halen, R.E.M., and a million other bands.

As an aside, Converge’s official lyrics include the phrase “soul survivors”. While it is almost certainly an error, it’s felt poetic to me, as if implying that suffering is an inevitable consequence of consciousness.

Genius does not appear to give a number on how many of their lyric pages had the hidden apostrophe codes, so we have no idea what the actual total count of copied songs is. I think the bigger issue is that Google denied having done anything wrong until this came under public attention, and there is no way for smaller companies to effectively pursue legal action against tech giants. >“We take data quality and creator rights very seriously and hold our licensing partners accountable to the terms of our agreement,” Google said. *when it comes under public scrutiny, otherwise we continue to profit from it as long as we can.

The same pattern you can use for any big company. Make some shady proxy one-day living pocket company, sign contract, let'em steal others IP, you are clear, ..., PROFIT!

I don't think that company even should know how their content will be used. For example if they just offer web-scraping service and provide raw data? For example for statistical analizes. And then Google publish this data as is on SERP and make money on adv. How is responsible in this case?

Reply on news.ycombinator.com