New, here… been tearin’ it up entering on a good number of new releases as well as adding/editing info on existing releases.
It came up yesterday that I may have improperly uploaded artwork from a digital release and added it to a physical CD release.
I have been using iTunes’ high-resolution digital artwork in a number of instances where the artwork is the same as the physical release, for believing it a superior image to any scan or potential scan.
In review, another use called me out on it and said that I should not use digital artwork if the release is a physical one. Admittedly, I went toe-to-toe and debated that per any rule I am aware of, so far, there is no straight, clear-cut guideline precluding the use of digital artwork over scans for a physical release, and that it seemed to boil down to a fundamental philosophy or preference that a scan always be used over a digital artwork for physical releases.
When I go to the style guidelines CA (Cover Art = Artwork, with a slightly misleading name) isn’t on the menu.
When I search in Documentation I find https://musicbrainz.org/doc/Cover_Art Cover art, also known as “album art” or “album artwork”, is artwork that provides a visual representation of a release.
Normally it refers to the front of the release packaging, but the Cover Art Archive can store images of the back of the release packing, of the media itself, and of many other pieces — right down to the sticker on the shrinkwrap.
With a MB “Release” being defined in part by it’s cover art, it looks to me that the desired CA is Release specific.
This can leave some releases only having lo-fi CA.
CA is a problematic area. It is very easy for scanning/photographing of physical covers to produce artifacts.
However iTunes CA does sometimes appear to be a significantly Photoshopped (altered) version when compared to scans of the physical cover.
I think the aim is to preserve a fairly high level of information about each release and it’s specific CA.
As an example of the problems and the very subtle differences which can distinguish releases here are the covers of 2 releases from the same release group.
The colour differences of the actual CA may be zero, or they might be different. One release here only has CA from a sun damaged cover. And both covers were scanned by different scanners. The only difference that I am confident about is the width of the mountain peak in the background. If a easily found hi-res image was used in place of these problematic scans then there would only be one release on MusicBrainz as the only certain difference between the releases is that mountain peak’s width.
This is probably more than you wanted to know right now. I apologise.
We may not always think of it that way, but the artwork on the release is data. Just as we would never knowingly add the wrong catalog number or barcode to a release, we shouldn’t knowingly add the wrong artwork.
For this particular edit, I think there’s even more of a problem.
You say here you’re adding the digital artwork where it’s the same as the physical release, but that’s obviously not the case here. The font size on the physical release is much larger than on the digital release:
Agreed and understood. I do see the font size discrepancy. My background understanding of this particular album release is that it received numerous Cover Art variations from physical product to physical product, on the shelf, at the time of original retail sale. Some of the album covers were the silver/gray background, while others were white. In some instances the headshots of each band member were used individually as the entire front cover.
As such, I would propose that this is still a valid representation of the original release, even within markets and per barcode / catalog number.
If, however, consensus is against my proposal, I will be happy to remove my iTunes Artwork uploads from this and all other releases that I’ve contributed toward to date.
Please do not apologize! You are one who is bringing reason and sympathy to a conversation that is otherwise appearing to be somewhat charged without compromise (myself included)
Your post serves to illustrate my point perfectly! For the reasons you’ve already indicated and outlined I am persuaded that scans are not superior to digital artwork when the digital artwork is representative of the original image used on a physical release’s cover.
Question to all:
If someone has the physical artwork in hand and compares it critically to the scan on a good monitor and has the skills needed to identify possible artifacts and distortions in both images, and finds the images identical, then their seems to be no advantage (and possibly much to be lost) in insisting on some other image being used.???
I think we need to use a word other than “representative” here.
How about, “when the digital artwork is a highly accurate rendering of the original image on a physical release’s cover”.
So if the digital artwork is produced from the physical artwork with processing, filtering, photoshopping etc limited to that needed to produce a close replica of the original physical artwork.
I think this would be highly suitable cause that is the best that a scan of physical artwork can be.
I believe in this case the color of the artwork of the actual physical release (at least the one in my collection) is somewhere in between the color gradients you posted, @mmirG
Here’s a scan of my cover done just a moment ago, with no color alteration or correction.
I agree with what @mmirG and I would go even further to say that this is exactly what we don’t want. The cover art on the 1997 US CD release of this album should be the actual cover of the 1997 US CD release, not some piece of art that we think is the most representative of what the graphic designer presented to the marketing department. These variations between releases are what make this artwork correct or not. The fact that one release has the RCA logo on the front instead of the back, or this release has the catalog number in the lower left instead of the lower right, or this release has a pale yellow background instead of cream/off white (see Edit #34234849).
For example look at this cover here:
I’m sure nobody in the band really wanted that ugly “Best Buy” logo in the corner, but replacing the cover on this release with a cover lacking the logo would simply be wrong. This is what the 1982 RCA US Vinyl re-release looks like, and to put another cover in it’s place would be allowing inaccurate data into the database.
That is, I think, the root of the contention here. I’m not on the MB staff, and I don’t speak for MusicBrainz, but since I started here in 2009, the main thing I’ve learned is that the MusicBrainz community values accuracy above all. When a user comes to MB, whether it’s some guy pulling up the website to look up who played harmonica on a song he likes or the BBC downloading reams of data directly from the database, they know that what they’re seeing is the best, most accurate data we can give them. We may make mistakes, and there’s a lot of data in here that’s incomplete (and kudos to you for curating the U2 data), but what we don’t do is purposely allow wrong data into the database.
So when someone pulls up the 1997 US CD release of U2’s Pop, if they’re seeing the iTunes version of the vinyl release’s cover art, that’s just as bad as if we replaced the CD release’s catalog number with the vinyl release’s catalog number.
My take on this issue would be that if an iTunes cover/digital cover highly resembles (no discrepancies in font, color, size, graphics, etc) the actual cover on a physical release, then it should be accepted as a good cover art.
We must understand that by introducing scanning artifacts we are creating inaccurate representations of what the artist intended to be the cover. If the original cover file sent to the printer to be printed for the physical release (for modern albums at least) could be found and uploaded, I don’t see how the digital cover would be inferior, if not superior, to a scanned version.
The whole issue of cover art is fraught with controversy, partially because of differing opinions with respect to the intent of maintaining a cover art archive. I’ve tried to summarize my understanding in another thread, where I also suggest that the MusicBrainz brain trust provide some clarification with respect to intent.
If my understanding is correct, and the primary purpose of the CAA is to provide archival information (images) with respect to releases, then I believe that the images uploaded should be unaltered scans or pictures for releases on physical media. Obviously this does not apply to a digital release.
I also believe there is a place for high quality digital cover art suitable for tagging one’s music collection (such as one would find with a digital release). Unfortunately, there does not seem to be a place for this within the CAA portion of MusicBrainz.
I’ve been told by others that this digital cover art is more appropriately stored on sites such as fanart.tv, so I am no longer uploading it to the CAA. Until something changes or is clarified, I am now limiting my cover art uploads (for physical media releases) to unaltered scans showing information supporting my edits.
I have no issues with someone adding an artwork from the digital release IF it’s exactly the same as the CD release (meaning no color, font or any other differences). I assume people often search the CD cover via google images and pick the highest quality/resolution image, which could very well be a cover from the digital release. So you often cannot know which one is it if there are no actual differences on the cover art. And I say the higher quality/resolution image there is, the better.
One thing that we do weirdly in MusicBrainz is treat those as separate releases. (Well, ideally we do—our data is far from perfect). So all those different variations you describe should be different releases in MusicBrainz (linked together in the same release group, with the same recordings, etc.).
For any physical release, there will of course be variations—printing is not a perfect process, and when we’re scanning them many years later, different ones were subject to different storage and handling conditions. And different scanners digitize things slightly differently (sometimes more than slightly, even). So colors will vary a little, some will be more faded than others. Those I think aren’t justifications for a different release. But beyond that, different art → different release.
Yes, that is logical and intuitive and I believe I have a grasp of the concept. Admittedly, I’ve made a few errors over the past few days that I’ve since learned differently about so there may be existing edits that one could point to where I’m missing a concept or idea, but I don’t think I’ve ever been lost for the idea of uploading specific artwork to a specific release (as opposed to release group, etc.).
…appreciate the reminder… are you seeing that I’m in violation of this somewhere?
Ok, I got the impression that this was a “random high res digital artwork vs. scans” discussion. It’s imo ok to upload digital artwork to a CD release if it actually is identical to the printed version. The problem is that this rarely is the case. I would for example vote “no” if it’s a square (e.g. itunes) vs. a rectangle (e.g. digipak), but not if the only noticeable difference is that crop is a few pixels bigger or smaller.
