File talk:Dürer's Rhinoceros, 1515.jpg

Stitching defects

The image was copied from the site of the British museum, which uses a plugin to show scans a fragment at a time. To obtain the full scan, the user had to scroll around in the image viewer, capturing individual pieces with PrintScreen. In combining them he made at least two errors which I have marked by image annotations. I first wanted to rectify it by repeating the tedious work, but then I found an all-in-all better scan and suggest that this faulty image be deleted together with its .png duplicate and the black-and-white version, which I have recreated from the better scan. Although it is normal to keep several versions of the same image on Wikipedia, it should not apply to defects introduced by the uploader. Ant 222 (留言) 11:59, 3 October 2015 (UTC)

 

@Yann: Why is the request "bordering on vandalism"? The image is corrupted trough the uploader's fault. A better version without these defects is available. The situation seems to fall under the Redundant/Bad quality paragraph of the Commons:Deletion policy.Ant 222 (留言) 16:21, 3 October 2015 (UTC)

I don't see any serious defect, and certainly none which warrants the deletion of a such heavily used file. Yann (留言) 17:30, 3 October 2015 (UTC)

@Yann: Here are examples in 2:1 scale for your convenience: 1, 2, 3, 4, 5. See the misalignment? Ant 222 (留言) 19:44, 3 October 2015 (UTC)

I don't see anything, except advertisements on these links. Anyway, your reasons for deletion are not valid. Yann (留言) 23:27, 3 October 2015 (UTC)

I beg your pardon. Image hosting changed. Ant 222 (留言) 10:52, 4 October 2015 (UTC)

I see the issues, but they are entirely fixable (and are being fixed as I type this). And how is it the "uploader's fault"? The way I get these images, if there are any stitching errors, it was definitely server-side (I don't do what Ant seems to think I do; I have a much better system). Furthermore, as clearly noted on the file page, this is the Christie's scan, not the British Museum's. — Chris Woodrich (talk) 01:15, 4 October 2015 (UTC)

Here is a screenshot from Christie's showing the absence of the defect on sample 2. If it was a server-side problem, then it is fixed now and you can re-capture the scan using your "system" and upload it as a new version of the same image.Ant 222 (留言) 11:03, 4 October 2015 (UTC)

Oh, that's what you have done already. Thanks. Ant 222 (留言) 11:10, 4 October 2015 (UTC)