Commons:Featured picture candidates/File:Castillo de Zafra, Campillo de Dueñas, Guadalajara, España, 2017-01-04, DD 41-46 PAN.jpg

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

File:Castillo de Zafra, Campillo de Dueñas, Guadalajara, España, 2017-01-04, DD 41-46 PAN.jpg, featured[edit]

Voting period is over. Please don't add any new votes.Voting period ends on 27 Jan 2017 at 19:24:05 (UTC)
Visit the nomination page to add or modify image notes.

View of the Castle of Zafra, Campillo de Dueñas, Guadalajara, Spain.
  • I don't think it is that big of an issue on such a large picture though and probably that most wouldn't see even on a moderate large print. The blurred area doesn't cover parts of much interest. - Benh (talk) 11:21, 19 January 2017 (UTC)[reply]
  • And depending on what kind of shots you do, you really needn't be jealous of the 5DS. As it's been discussed, If you are a macro, or still object guy, it won't bring you any much advantage over any other FF or APS-C given sensor of same generation. - Benh (talk) 11:24, 19 January 2017 (UTC)[reply]
@Benh: 1. I would love to support this image, but the different levels of sharpness and visible seamline because of it are bothering me. Would this image be of less resolution then it might no be visible even at 100%, but given the resolution, these shortfalls are visible. 2. I haven't followed these discussions. I'd love to have more resolution available for my studio shots, just for the sake of seeing more details, so I'm jealous. --Lucasbosch 12:28, 19 January 2017 (UTC)[reply]
Lucasbosch, I agree a high-resolution sensor has an advantage for single-shot photos such as your studio. We try to avoid penalising photographers for uploading full-resolution photos or huge stitches like this. Try the 50% downsize I link below. That's still 28MP and very sharp. If you'd support that then there is no reason to not support this. Opposing over 100%-size pixel peeping of a >100MP image just encourages folk to downsize prior to upload, and then we lose detail that can never be retrieved. As I'm fond of saying, if your monitor is a standard 100DPI, then this image is over 4 metres wide, and you'd probably view it from a couple of metres distance at least, rather than normal monitor distance. -- Colin (talk) 12:58, 19 January 2017 (UTC)[reply]
@Colin: Thanks for your thorough explanation, I strike my oppose. I would welcome allowing/encouraging photographers to upload downsampled images instead of full sized ones like this, to hide flaws better and avoid pixel peepers like me. In fact I would have downsampled this in secret if I were in the same situation, as the resolution is plenty even downsampled. But I understand that having sharp parts of the image is considered more desirable for the Commons project than having less pixels but with the whole frame being perfect. --Lucasbosch 13:13, 19 January 2017 (UTC)[reply]
Lucasbosch, if the software can downsize on-the-fly then why downsize on upload? Actually, for huge stitches most photographers already downsize a bit to ensure the detail is sharp, which it is for most of this image. I don't see the point in uploading full size if it is soft/blurry all over, and for big stitches there is no value in making people download a big file that is not sharp. When we get folk uploading 6MP landscapes that pass FP, it isn't really fair to to penalise others who don't downsize. -- Colin (talk) 13:50, 19 January 2017 (UTC)[reply]
@Colin: You say yourself that some people already downsize so others don't have do download a panorma that isn't sharp corner to corner. I don't think this image is any different, albeit being much more resolution. My opinion is that such a panorama should have equal sharpness over the full frame, and not rely on downsampling to achieve this. If there is a soft part, downsample until everything is constistent and thus the seamlines become invisible. I find a 100MP image which is soft on some spots and sharp on others kind of more wonky than a pristine 6MP image. I see your point, too, that you throw away detail on parts of the image in the process. So I wouldn't want to penalize him for not downscaling, but the different levels of sharpness which reveal seamlines, and I don't want to see seamlines. I believe in a pixel perfect uploaded file not reliant on downsampling to achieve even sharpness. Even if this requires downsamling before upload. Agree to disagree ;) --Lucasbosch 14:57, 19 January 2017 (UTC)[reply]
I appreciate you don't want to see the seamlines, and I wish the stitching was better. But you have a choice to examine the image at this level of detail. Nobody forces you to download the full size image and then display it on your screen at 100% while examining it closely. A good review test might be to judge the whole image fullscreen on your monitor, and then to examine for flaws at some intermediate resolution. If you'd support this at 12MP, say, or 24MP, say, then any extra resolution is simply a bonus. I think that unless the image was huge and very soft/noisy all over, then I'd be reluctant to complain about the size being too large. We have a culture here of pixel peeping that harms people's generosity in uploading/creating high-resolution images (vs Flickr where many images don't even fill one's screen). The result is some photographers really do upload 6MP landscapes from their 36Mp cameras and get and expect to get FP. Of course, minor errors only visible at 100% on a large image may be worth pointing out to see if they can be fixed. If you have a high DPI screen, then much of this pixel peeping concern simply disappears. Our standard 100DPI monitors are the equivalent of taking a magnifying glass to a print. -- Colin (talk) 15:54, 19 January 2017 (UTC)[reply]
  •  Support That bottom right is an issue, but given where it is and the image resolution, I think it can get away with it. -- KTC (talk) 12:45, 19 January 2017 (UTC)[reply]
  •  Support There is a flawed frame and perhaps the stitching could be improved (IIRC you just use Lightroom - have you tried PtGui with SmartBlend as the blend tool?). But the resolution of 111MP makes this visible at 100%. A reduction (see this link to a 28MP 50% downsize) hides such sharpness problems and the whole image is very sharp indeed. -- Colin (talk) 12:58, 19 January 2017 (UTC)[reply]
  •  Comment I've uploaded a new version where I cropped the right side a bit to get rid of the area that somehow wasn't as sharp as others. If the result is not satisfying I can offer also this other version with a far wider view. Thank you Benh and Colin for making understandable that images with more resolution are not always comparable with lower ones and users of a camera like 5DS shouldn't be punished for that. It is indeed not as easy as it was with the 5D Mark II to get all images sharp. Poco2 18:09, 19 January 2017 (UTC)[reply]
  • Colin did most of the lobbying :) And, pardon me if you already knew, but If your pipeline gives you access to the seam mask, it's very easy to edit it and soften the transition from sharp to blurred area. This would give a better result I think, but this take for granted the "common" area between the source pictures is large enough. - Benh (talk) 18:43, 19 January 2017 (UTC)[reply]
The new crop doesn't just chop off a bit on the right, but introduces some to the left. I like the road on the left, but this is a different enough picture that I think you should ping all who have voted so far. It isn't like you just removed a dust spot. As for the other one you link, it is far too wide and also has quality issues. There is still an issue with a seam (to the right of the rocks) that could be handled better if, like Benh says, you took control over the join there (or used Smartblend, which I find is often better are placing seams and not crearting blurred seams). -- Colin (talk) 19:57, 19 January 2017 (UTC)[reply]
@The Photographer, Alchemist-hp, Ikan Kekek, and King of Hearts: @Johann Jaritz, Tomascastelazo, INeverCry, and Martin Falbisoner: @Code, Cayambe, Michielverbeek, Gnosis, Pudelek, and Benh: @Lucasbosch, KTC, and Colin: Dear all and sorry for the disturbance, I just wanted to let you know that I've cropped the image (mainly on the right) to get rid of the blurred are. I'm informing you just in the case that this change would affect your already emitted vote. Poco2 20:10, 19 January 2017 (UTC)[reply]
  • Poco, I like the second crop the most. Now the main subject is no longer centered, and this is a bit weird IMO (but this still has my support). I don't garantee anything, but just in case, I offer assistance to implement the above mentioned solution. - Benh (talk) 20:34, 19 January 2017 (UTC)[reply]
Thank you for your offer Benh. Will look into it this weekend and probably come back to you then. Poco2 20:36, 19 January 2017 (UTC)[reply]
I prefer the crop from the time of original nomination. There's still a visible seam in the new crop, which wouldn't go away unless the you crop much closer to the rock. While now the rock the castle is sitting on is actually centered, the castle itself is now a bit off to much to the side to me. Then again, if the new crop was the only version offered, I would still had supported so I'm certainly not going to object now. -- KTC (talk) 21:40, 19 January 2017 (UTC)[reply]
Please go back to v2 and ignore the naysayers. It was The Perfect Composition. KennyOMG (talk) 02:21, 20 January 2017 (UTC)[reply]
I think that this one is "better", however, both are perfect to me, IMHO --The Photographer 11:05, 20 January 2017 (UTC)[reply]
Confirmed results:
Result: 28 support, 0 oppose, 0 neutral → featured. /George Chernilevsky talk 06:14, 24 January 2017 (UTC)[reply]
This image will be added to the FP gallery: Places/Architecture/Castles and fortifications
Retrieved from "https://commons.wikimedia.org/w/index.php?title=Commons:Featured_picture_candidates/File:Castillo_de_Zafra,_Campillo_de_Dueñas,_Guadalajara,_España,_2017-01-04,_DD_41-46_PAN.jpg&oldid=230815996"