How does CD Ripper scan UPC's from CD??

[Boffy]

Hello fellow metadata enthusiasts, and happy new year to my favorite internet forum.

I have trouble understanding how CD Ripper determines UPC's when reading CD's.

CD Ripper sometimes assigns an UPC value that differs from the one on the cover/disc.
Some of the UPC's doesn't give any results when looking them up on e.g. Discogs or barcodelookup.com.

As an example, I'm working with this specific release of Paul Simon - Graceland (Pictures attached)





As shown, the UPC printed on the cover is 0 7599-25447-2.
This UPC can be looked up both on Discogs and Barcodelookup.
10 characters long.

When loading the CD in CD Ripper, I can't get it to show this specific UPC, but I do get other results depending on the provider:

With only one provider enabled:
CD-UPC reads 7599254470200.
This UPC does not appear on either Discogs, Barcodelookup, or MusicBrainz.
13 characters long.

GD3 doesn't provide UPC.
MusicBrainz doesn't provide UPC.
AccurateRip Meta reads 081227480325.
This one can be found on discogs, however my version is manufactured in Germany, and the release on discogs is american. Also looks different on the backside.
12 characters long.

With all metadata providers enabled, it reads the same as CD-UPC only.

How does CD Ripper read the UPC and why does it differ from what is actually written on the cover/disc?
It is not an issue for most of the CD's I've tried, but I'm ripping a lot so it's good amount anyway.

Any help or insights are welcome.
Without knowing too much about it, I suspect I'm missing some kind of relationship between EAN and UPC?

[BrodyBoy]

It’s not as complicated as all that. CD Ripper is simply obtaining data from a mishmash of online databases, and the quality/accuracy of that data is only as good as what’s been input there.

There are essentially two main causes of inaccurate data in the databases (including UPCs). First, incorrect data was input…happens all the time since there’s really no “authoritative” source for metadata…garbage in, garbage out.) And second, many albums, especially older more popular ones like Graceland, have had LOTS of releases! Sometimes dozens or even hundreds. Might be the same set of tracks (though not always), but any & everything from format to cover art to label to UPC to country of release to mistakes in the credits can vary between them.

The chances of all that info being accurate and consistent for all albums and their releases across multiple online databases is slim-to-none. Again, there isn’t a single authoritative reference for online metadata.

Bottom line: If you’re going to be a stickler that all metadata be accurate for the disc you ripped, there’s no way around manually checking it yourself for inaccuracies & omissions and making corrections as needed. While I personally wouldn’t worry about UPCs (it’s data that virtually never impacts my use of the material), I do run every rip through a metadata editor to ensure that all relevent data is present and accurate. I’m very thorough, and the vast majority do in fact require some manual corrections or additions.

[Boffy]

Thanks BrodyBoy!

That's pretty much what I assumed, but it's reassuring to get it confirmed.

However, it still doesn't explain why CD Ripper gets the UPC wrong when only CD-UPC is enabled as a provider.
(Or maybe it does, please correct me if I'm wrong).
As I understand, CD-Text, CD-UPC, and CD-ISRC are all data embedded in the disc, right?
So in this case, no data from online databases are obtained, right?

Normally I wouldn't worry about UPC's either, but I'm ripping a huge collection for a broadcasting company.
Each release has a sequential unique ID assigned in the order of acquisition, so I figured the easiest way to map all the folders correctly was to get BatchRipper to name them after their UPC’s instead, as they are easy and quick to scan in our collection.
So far I have had success with the ones where the UPC's are correct, but to run the majority of rips through an editor would defeat the purpose of automating it.

[Spoon]

>As I understand, CD-Text, CD-UPC, and CD-ISRC are all data embedded in the disc, right?

Correct, if reading from CD-UPC then that code is on the disc, and the box is wrong.

CD-UPC would override the online database if present.

[Boffy]

Correct, if reading from CD-UPC then that code is on the disc, and the box is wrong.

Alright, that's great!

But then I'm curious why the CD-UPC reads 7599254470200, when it doesn't show up on either Barcodelookup.com or Discogs.
I got hold of a barcode scanner, and the specific CD in question reads 075992544726 on the cover.
This shows the results I need, but the one read from CD-UPC doesn't show any results anywhere.

Any idea why this is?
Could it have anything to do with the dash being read wrong?
I have no idea how the data is presented on the disc or if dash'es are even presented, so I'm sorry if it's a stupid question.