Commons:Featured picture candidates/File:Upernavik cemetary 2007-08-06 original stitch.jpg

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

File:Upernavik cemetary 2007-08-06 original stitch.jpg, featured[edit]

Voting period is over. Please don't add any new votes.Voting period ends on 24 Dec 2011 at 08:21:56 (UTC)
Visit the nomination page to add or modify image notes.

Upernavik cemetery, North-West Greenland
  •  Info Created, stitched and uploaded by Slaunger - edited by Paolo Costa - nominated by Slaunger. This is a follow-up on a previous nomination I withdrew recently based on review comments suggesting some improvements. More foreground, and quality issues addressed. Please consider, that no downsampling has been used in this 35 Mpixel edit, and the stitch is based on single frames taken in 2007 with a compact camera. To see how it looks in, e.g., 10 Mpixel resolution see the discussion page of this nomination. Thanks to benh and Paolo for a fruitful dialog regarding the possibilities for image improvement. Independent of the outcome, it has been a fun process. -- Slaunger (talk) 08:21, 15 December 2011 (UTC)[reply]
  •  Support -- Slaunger (talk) 08:21, 15 December 2011 (UTC)[reply]
  •  Support great group job.. I followed the discussion and progress here User_talk:Benh#A_preview_before_I_proceed Ggia (talk) 10:07, 15 December 2011 (UTC)[reply]
  •  Support -- Nice picture Jean-Pol GRANDMONT (talk) 11:25, 15 December 2011 (UTC)[reply]
  •  Oppose wonderful view, nice colors but sadly nearly at all parts of the picture very noisy. --Wladyslaw (talk) 12:04, 15 December 2011 (UTC)[reply]
  •  Oppose -- Sorry guys, but quality is just not there: the image is unsharp, noisy and lacks detail. I also find the composition unbalanced, due to the shape of the hill, and wonder whether a different shooting positiom wold have helped. Alvesgaspar (talk) 12:19, 15 December 2011 (UTC)[reply]
  • Weak  Support. I know there's no weak support over here, but this is what I mean. Superb view of not mainstream place and nice composition à la Boringher ;) but lighting a bit flat on the right side and overall quite soft. But we have to take into account the size of the picture... It certainly would render better than many of our sharp FP (including some of mine...) on a given fixed size "canvas". And very nice fixes from Paolo !! - Benh (talk) 12:31, 15 December 2011 (UTC)[reply]
  •  Comment (ec) Regarding noise. Yes, if you open the file in full resolution and inspect a section of this 35 Mpixel image at a time, it is noisy, and also a little soft in focus. However, in my opinion, the pixel-by-pixel noise and softness is not noticable in any practical real-world application of the image. Be it as a full screen view on a monitor or as a large printout. Noise can be decreased a lot by downsampling, and I have done that previously here and here for instance. But I am done with this trick as the downsampling also results in a loss of information. The more I work with images, the more I feel that postprocessing like downsampling, sharpening and denoising ruins the photo - giving it either an artificial look or lower information value. There are exceptions though. like it can make sense to denoise a large homogenous area with basically no information. Like blue sky. That could be done here, but really, in the end, it would not matter in any real-world application. --Slaunger (talk) 12:51, 15 December 2011 (UTC)[reply]
  •  Comment regarding noise.. if the image will be downsample to 2megapixels, the noise will not be visible. If some of you have knowdelge of engineering.. you probably know the Signal to noise ratio. This ratio describes how strong is the signal concerning the noise. If the signal is very strong, the noise can be ignored. In electrical engineering (ie. telecommunications) you don't try to see who much actual noise you have but you are interested to have a strong signal. I believe that the community has to justify image in a similar way. When you have a 35mpixel image you have to see the ratio of image size (image information) / noise (lets call it INR: Image to Noise Ratio): Ggia (talk) 13:22, 15 December 2011 (UTC)[reply]
  •  Comment -- As any other regular, I'm perfectly aware that some allowance should be given when examining a digital photo in full size. "Full size" means here the size of the image as captured by the sensor and reproduced on a computer screen with the usual enlargement factor. Knowing that a panorama is just a mosaic of several individual pictures, I don't see why additional allowance for unsharpness & etc should be given in that case. The argument of Benh concerning the practical real-world application of the image is invalid in my opinion. Either the picture is supposed to be used in full size (on screen or as a printed copy) or it should be downsampled because the extra pixels are useless. I've always been against downsampling but, on the other hand, consider that big photos should be assessed exactly the same way as the small ones. A quick browsing through our FP panoramas (including Boringher's ones) shows that their sharpness, focus and detail are in general much better. Alvesgaspar (talk) 19:17, 15 December 2011 (UTC)[reply]
    • PS - I didn't nominate this recent pano of mine, whose size is similar to Slaunger's, because I consider that image quality is not good enough (the reason being poor lighting conditions). Alvesgaspar (talk) 19:24, 15 December 2011 (UTC)[reply]
    • By evaluating the per pixel sensor quality instead of the end result, which is the image in its entirety, you get some very odd drivers for promotion. For instance, this is a stitch of 9 photos. To get as much detail and information as possible I used zoom on the camera knowing well that the per pixel noise will increase, because I get less light on the sensor. I could as well have chosen to take just three photos at its widest angle, as the pixel noise will be at its minimum for the sensor. In that way I could have produced a 10 Mpixel stitch of the same scenary with a significantly lower pixel noise and better sharpness per pixel. But that image would not contain nearly as much information. You say, that you should see an image at full size. I do not think that makes sense. If on a screen, you simply do not have a screen 12000 pixels wide (at least I don't). If you click on an image to see it in its entirety on a monitor, the viewing software makes the downsampling for you to make it fit. and the the per pixel noise viewed decreases, and the per pixel shapness increases on screen. Likewise, if you use a printed media, I would claim that you do not chose a paper format based on the number of pixels, but rather scale the pixels to the format you want, tolerating a low or high dpi in the final result. I think it is an oddity to include the method used in the review. That just because it is a panorama, the detail level shall be much better than if you too the same scenary with a wide angle lens in a single shot. It is the end result, which matters. --Slaunger (talk) 21:16, 15 December 2011 (UTC)[reply]
  •  Neutral - it's a nice panorama, but there is visible noise for me at fullscreen, especially around the cross. --Claritas (talk) 21:45, 15 December 2011 (UTC)[reply]
  •  Info Last remark about the noise. The discussion above is very theoretical perhaps. To make it easier to understand I have collected some examples of the image sampled at different resolutions starting from HD resolution, to minimum FP resolution all the way up to full resolution. And finally, the apparently crisp HD upsampled to full resolution, where the information loss is evident. --Slaunger (talk) 21:51, 15 December 2011 (UTC)[reply]
  •  Comment opposing votes will turn photographers in the future to upload a non-full resolution of images. One can say.. Why to upload an image @35megapixels and receive oppose votes? I will upload @12 megapixels and everybody will be happy.. (no oppose votes due to noise). It is clear in the examples in the discussion-page that this image in lower resolutions has no noise. This image has high information and very low noise (concerning the size-resolution). Ggia (talk) 22:21, 15 December 2011 (UTC)[reply]
  •  Support As editor, and because, as I said from the beginning, I think it is important to have some panos from all around the world. The other files from Upernavik did not have as great light conditions as this file. Sky is nice, the sea is beautiful, houses are colorful, and those nice icebergs give an idea of how cold the place must be. This place even gives you an idea how a cemetary looks like. I always wonder what kind of cities and towns exist around the world, in such isolated places like Greenland, Tibet, Iceland, Far Oer islands, etc. This is a great way to help users like me get that picture. That's why I really do not care about quality in this case (besides, we are not even talkin about a disaster picture here, noise is not that bad imo). Finally, this was an example of a nice voluntary-group-job. Things like this should be supported on the Wiki projects. --Paolo Costa (talk) 04:42, 16 December 2011 (UTC)[reply]
  •  Neutral very interesting composition, but too unsharp Cathy Richards (talk) 21:17, 16 December 2011 (UTC)[reply]
  •  Support I'm convinced by Paolo Costa's arguments. Very nice collaborative job. I think Ggia's last comment is right too.--Jebulon (talk) 21:41, 16 December 2011 (UTC)[reply]
  •  Comment I've added another downsizing example to the talk page. I don't understand the comments about noise -- this image doesn't suffer from noise much (there's a little in the sky). What it suffers from is a lack of detail. Of the many reasons I listed on the talk page, the likely culprit is aggressive in-camera noise reduction. Like many pictures taken with a compact camera, the official resolution is an order of magnitude greater than that actually recorded onto JPG. Add to that the processing necessary to make a panorama, and we've got an image that can comfortably be reduced 50% with no loss of information. For panoramas especially, we should not be uploading huge images for no good reason. This image should, IMO, be downsized 50% from 35MP to 9MP. There are valid reasons to oppose downsizing but this watered-down image is not one of them. Colin (talk) 18:40, 17 December 2011 (UTC)[reply]
  •  Support This kind of panoramas are important. See Paolo Costa's arguments. Achird (talk) 00:20, 18 December 2011 (UTC)[reply]
  •  Support Very nice! Something which also makes me happy is that it's not among the most common locations of the earth photographed for Commons. We definitely need good pictures from all over the world. Novadays, especially from the more unexpected sites of this globe. --Ximonic (talk) 03:28, 18 December 2011 (UTC)[reply]
  •  Support Trongphu (talk) 23:01, 21 December 2011 (UTC)[reply]
Confirmed results:
Result: 9 support, 2 oppose, 2 neutral → featured. /--Claritas (talk) 11:33, 25 December 2011 (UTC)[reply]
This image will be added to the FP gallery: Places
Retrieved from "https://commons.wikimedia.org/w/index.php?title=Commons:Featured_picture_candidates/File:Upernavik_cemetary_2007-08-06_original_stitch.jpg&oldid=70824245"