Sometimes, digital releases, such as iTunes, also come with a digital booklet usually in PDF form. Does MB have a policy on uploading said booklet into cover art? Does that fall into release artwork, or does that cross into the copyrighted material that is covered under piracy law(s)?
My understanding is that this is not an MB issue.
Rather it is a Cover Art Archive (CAA) issue. Who are a separate entity related to archive.org.
And they seem to have no problem accepting pdfs from most sources.
I remember one label (classical) being said to have requested that their pdf’s not be uploaded and understand that their request is adhered to.
CAA accepts the images, stores them and distributes the images freely. MB is one of the users of their data.
Sounds good. I have uploaded a few in the past, and I just wanted to make sure it is appropriate. I know there is a policy of not using pirate sites as references, etc… and I respect/understand/agree with that, so I do not want to be posting infringing materials. I will continue as-is unless there is some indication to stop.
Your statement after reading made me do some thinking. So MusicBrainz is dependant on Cover Art Archive / Internet Archive / etc for all of its uploaded artwork. This seems like a risk point, at which I would ask… what if this third party organization shuts down this portion (or all of) their services?
I am sure I speak for others with my statement, but I have spent a fair amount of time looking through my release image and scrubbing the net for rare release images both in order to have images for all releases and also to provide a reference depicting the actual real release vs a web page simply telling us what the release is. I know there is the same risk for MusicBrainz itself, but if MB goes, either so does the artwork or the artwork becomes useless for this application. I know some will try and respond that this third party organization is stable and not going anywhere, but I will reply right away with the principals of risk management negating the response but keeping it as the likely status.
It might be possible to back up the CAA “privately” - it would be interesting to know its current size.
I would not be concerned that archive.org going away.
A few mb of images is nothing compared to an almost complete backup of the internet every few months.
I agree. In addition, I believe that MB is at least partially insulated from any potential copyright infringement issues by having the images stored on archive.org.
Sorry if my statement was misunderstood by anyone, I did not mean much by it. Let me explain where it came from…
Of all the issues I have had using the MB system, most of them are with cover art. Errors uploading, edits timing out, issues with thumb generation, etc. When I had asked on the cover art here, the response was tending the same as when the upload issues were reported… a “pass the blame” or “not us” type of thing. So my mind is thinking… this service is not us when it comes to performance problems, copyright issues or regulation on allowed content… from a business sense and the general web operator sense, that is risky. Especially since many releases use the cover art as a key portion of the release.
I was not suggesting that CAA or Archive.org is going away, but suggesting there is a good deal of reliance on them for a key portion. So I simply thought in the basics of risk management for a business, website owner/operator/etc… should my vendor disappear without notice, do I survive? Many things can happen aside from shutting down… change in strategic direction of a company, a “souring” of business relationships, financial issues, resource allocations, etc… I guess that is just a part of my thinking when a project has “outsourced” a portion and has no contingency plans.
As it relates to me personally as what I said above is business centric, this type of thinking is important… if I was relying on Microsoft OneDrive to store my data when it was 15GB, I would have a problem when they reduced it to 5GB and I failed to opt in for the grandfathering to keep the 15GB. I would need to find somewhere to go with my 10GB of data. So although most would say… Microsoft is not going anywhere, my 10GB of data is nothing compared to the data they store, they had a change in strategic direction of their product, which in turn causes me to change whether that is find a new provider, store locally or pay them to get more storage, it is not business as usual and I need to know my options.
Hope that clears up what I was thinking and explained the reasoning of my thinking. I did not intend anything negative if anyone took it that way, I just think in this way.
I didn’t interpret any of your comments to be overly negative. You raise some valid concerns, in that the MB / CAA stuff has been troublesome of late. I believe that steps are being taken to help the situation, but it’s not as solid as we’d like just yet.
Then we would have to drop cover art, probably. MB is not going to store cover art on their servers, because nobody (including the labels, in many cases) knows who owns the rights. The IA has a lot of experience with copyright issues and can take stuff down if appropriately, and by being a library it gets more leeway with these things than we might ourselves, plus by having a lot of money it can afford to fight law cases if someone is problematic about it.
Mind, we do not expect any court cases over cover art to happen - it’s just that with our income, in the US, we just couldn’t afford to fight one. Even if we’d win it anyway.
Thanks, there we have the answer. I can understand that, IP law is tricky and can get expensive, as you said, even if you win… you still need to represent and defend yourself. But I understand, the artwork is there currently and it is a very nice feature. There is no contingency plan however as the risks, although small, are still there and MusicBrainz/MetaBrainz is not willing, at least at this time, to take on such a risk.
@reosarevok - Do you happen to know or have an idea of the size of the cover art storage? Also, I know users have access to the artwork, do you know if there is a means of pulling the database, in whole or in parts? For example, lets say I would want to write a query to pull artwork from ___ artist(s), or all releases I have made an edit for or even just all artwork I submitted with my account? I can check over at IA as well for info, but I thought you might know.
Currently has 787,760 images.
Thanks, that answered all my questions via that link.
I’m calculating around a 1000 images per GB. And a million images.
Which is a thousand thousand images and 1TB.
So far (if I’m right) we can easily back-up the whole of CAA.
I found the same, it could be easily backed up. My whole point is just connecting the dots of logic. Very generally speaking, when it comes to physical media, what makes a release is the physical release. Second, we use other resources that index, list or otherwise give a reputable description of the release. So when the primary reference point is a physical product, and editors do not give their CD/vinyl/etc to MB to store in a vault, a picture is a very close substitute, especially in the fairly advanced digital age we are at…
This means that I can enter 100 releases with full scans of the release… front, back, spine, medium, booklet, etc… the whole deal. This would be considered almost a perfect release with a top quality reference. Now say we lose CAA (and thus end up no longer having images per speculative comments), those 100 quality release entries become unreferenced and lose their quality. So since in regard to a physical release, given how much emphasis in release differentiation is placed on the physical piece and packaging, this becomes a critical portion of the release. In all honesty in these cases, MB is really indexing release packaging, not music. This is even worse on digital media where in addition to not classifying the music itself, but the packaging (container and meta) are also disregarded as a norm.
For digital, which was the initial conversation media type, I do not see the artwork as so important. What really is important there and differentiates the release is the audio encoding and container it is placed in… generally speaking. With physical media, I personally have duplicate releases… some that were released with different color cases, different sticker sets, covers, etc. I do not do this on digital on the same criteria. I do not care about covers, barcodes, or anything like that. I care about the encoder used, the mastering process, the encoder settings, the container used, etc.
So… for digital, MB can currently store all the data it needs, it just opts not to. All the factors that differentiate a digital release can be stored in the MB databases and kept on the MB servers and backups, etc. For physical releases, the primary release differentiators are more on the disposable end, based on this conversation. Considered not critical data, but in practice is critical to the release itself. I hope all would agree… if I make a release and call it different because it has a different barcode, and I just make a dup of the release data, that is useless unless I can provide proof that this packaging with this different barcode does exist. Best way to do that… provide a picture.