Wikipedia:Featured picture candidates/Trevi Fountain, Rome

The Trevi Fountain of Rome

 * Reason:Again, a reprocessing of an old image that I wasn't 100% happy with. This is a very detailed rectilinear-stitched mosaic image of the Trevi Fountain in Rome. It is derived from 25 individual frames. I needed to use this many due to the extreme nature of the rectilinear stitching (the edges are very stretched/distorted, but due to downsampling of the image from 15000x10000 to 3500x2300, softness is not as visible). I've never seen a rectilinear view of the Trevi Fountain before. Typical views only include part of the scene (This for example), are taken from a long way back and obscured by the crowds, or are distorted by the upward angle. This image avoids all of these issues and is a pleasing view. This is perhaps despite the slightly ordinary sky, which I think I've managed to make look 'atmospheric' rather than dull, and does contribute to even lighting of the building.
 * Articles this image appears in:Trevi Fountain
 * Creator:User:Diliff

--jjron (talk) 08:46, 6 May 2008 (UTC)
 * Support as nominator Diliff  | (Talk)   (Contribs) 11:56, 26 April 2008 (UTC)
 * support original, support edit colours are pale, particularly the sky so the image looks flat and very cold, I actually mistook it for a painting until i read the description. Framing is nice though and looking at it in the article, it certainly adds to it. I am also extremely impressed that this is the resolt of image stitching, it is entirely seemless. While I feel it needs more depth, it's very very good already. Capital photographer (talk) 12:17, 26 April 2008 (UTC)
 * Yeah, you can't make a scene something it is not though. You can see from the people on the edges that there is plenty of saturation in the image - its just that the fountain itself IS somewhat pale. If I'd taken the photo in strong sunlight, it would no doubt look warmer but instead, there would be heavy shadows and probably uneven lighting. As you no doubt know, lighting in phototgraphy is often a compromise and indeed it is in this case. I think that if I were to edit the hell out of it to try to get more oomph and colour, I'd probably ruin it in the process, as I've already massaged it enough to get it looking like this. Diliff  | (Talk)   (Contribs) 12:28, 26 April 2008 (UTC)
 * The big issue for me is that the building and sky are both quite pale and bright and the sky lacks contrast and saturation. The result is in image where the building blends too much into the background and the image is inbalanced, with a darkish colour foreground and a very pale and bright top part of the frame. Lighting is a tough thing and the lighting you took this shot in is actually ideal, minimizing possible shadow and glare. We all have our own eye for photography, so I opened this image in Photoshop and made some changes using nik Viveza to show what is possible and what I personally feel would work better, to better express what I'm saying. — Preceding unsigned comment added by Capital photographer (talk • contribs) 13:43, 26 April 2008 (UTC)
 * Your welcome to your opinion and I'm sure feedback is appreciated, but I'm not sure it's entirely good form to replace the original with your edit in the article while it's under discussion here and Diliff hasn't even had a chance to reply. --jjron (talk) 14:15, 26 April 2008 (UTC)
 * Have you updated your cache? I reverted the image a while ago after testing it in the article. I would certainly encourage the nominator to test the edit in the article to see how it looks though. Capital photographer (talk) 16:14, 26 April 2008 (UTC)
 * Not that I'm personally offended that you replaced my image in the article, albeit briefly, but why did you need to see how it looked in the article, when your edit is only moderately different, and you could see it just as clearly in this nomination. And even if you did 'need' to see it in the article, you could always preview it rather than save the change... And for that matter, and I know I'm being picky (sorry!), but you should probably try to make all your changes in a single edit, rather than going back over and over. ;-) As well as creating potential edit conflicts when others are trying to contribute, it also just looks a bit messy in the edit history.
 * One of the FP criteria is does it contribute to the article. Personally, I can't tell here if it would or not. Images appear at a different size and with different text rapping in articles. Also, leaving it in the article for a brief period allowed time for a reversion. Given that it looked good in the live article and was not reverted, I consider it a success and that it does contribute to the article. Capital photographer (talk) 07:57, 27 April 2008 (UTC)
 * Well, you wouldn't expect it to be reverted within 3 hours unless it was objectively inferior or vandalism. As I said, it is so similar to the original that most people simply would not notice or care. Eventually someone with a keen eye would probably notice but that is a bit beyond the scope of a 3 hour test. The active contributors here at FPC are just as competent in determining an image's contribution or potential contribution to the article. Anyway, I'm not really having a go at you, just letting you know the way it usually works here. Diliff  | (Talk)   (Contribs) 10:05, 27 April 2008 (UTC)
 * Now, 'finally' getting to the critique of your edit. Aesthetically, I don't mind the change you've made, as it does make the sky somewhat more dramatic and contrasty against the building, but you've introduced (or possibly just exacerbated) posterisation in the sky. Also, you say you haven't altered the fountain/building in any way, but you have. The windows are now tinted a bit overly blue, and the plaque at the top of the building is clearly blue-tinged now. Seems like that little tool of yours has been strong on the blue saturation. As I said, I don't mind what you've tried to do in theory, but in practice it seems a bit heavy-handed. Diliff  | (Talk)   (Contribs) 19:04, 26 April 2008 (UTC)
 * The tool I used has a saturation adjustment function, but I didn't use it. Just corrected levels and turned the brightness of the sky and reflections of the sky down 25% and contrast up 8%. You are correct, the reflection of the sky in the windows and water is part of the layer mask used to correct the sky. The sky is reflected in the windows and water... so if the sky changes, so to does any reflections of the sky. I didn't consider this a change to the building and the building itself was not deliberately altered, only the sky and reflections of the sky.Capital photographer (talk) 06:15, 27 April 2008 (UTC)
 * Well ok, perhaps it was just a brightness change rather than saturation change (darkening highlights to midrange often does increase the perceived saturation), but you ignored the point I made about the plaque of the building also being blue tinged now. That is clearly not a reflection of the sky. It evidently had a small amount of blue in it to begin with, since I'm assuming you essentially darkened anything with blue in it, assuming that it must be the sky?), but it is looking a bit over the top and unrealistic now. Regardless of whether it was deliberate or not, it is an unfortunate mistake. ;-) I would have made a mask of just the sky, rather than a mask of the colour of the sky. But then, the more I've looked at both images, the more I believe the changes just aren't that necessary. Diliff  | (Talk)   (Contribs) 10:03, 27 April 2008 (UTC)
 * I checked mask coverage in PS and it has only included the sky, windows and water, not the plaque or building. However, the mask also adjusts surrounding luminance and has darkened the plaque and top of the building a bit. On my display, it looks neutral, white-grey, but on a non-calibrated screen, there may appear a blue or even mild magenta tinge.Capital photographer (talk) 10:45, 27 April 2008 (UTC)
 * Are you sure your screen is really calibrated properly? If you're not convinced, eye drop various areas on the plaque and you will see that significant portions of it are blue-tinged. Particularly above the top row of text - hold the eye dropper on it and drag along the top and you will see the blue is almost constantly more luminent. They are on my image version too, but to a lesser extent, and less visible as they are brighter. I suspect that your screen isn't as well calibrated as you think, if you're unable to see it. I can see the blue clearly in the plaque on two different monitors - my contrasty, punchy (and admittedly not professionally calibrated) 30" Dell screen and a cheap flat HP business laptop screen. Can anyone else comment on this? I need a third opinion here. ;-) Diliff  | (Talk)   (Contribs) 11:06, 27 April 2008 (UTC)
 * It is sometimes difficult to teel between a mild blue and mild grey, so I will conced that depending on people's eyes and displays, it may look bluish to some. I loaded it into Photoshop CS3 and used the eyedropper. Running it over the area, the colours ranged from light to mid grey and a couple were borderline. Here are the colours I am getting: #ced3d6, #dfeaee (very borderline, looks mostly grey on my display with a very faint hint of blue), #cfd4d7 , #ccd9e1 (again, borderline) , #e3e8ee , #aabbc5 (grey but leaning to the blue region of the spectrum), #b7c5ce. On inspection with the eye dropper, some could be classed as grey with a slight blue tint. Mostly grey though. Capital photographer (talk) 12:01, 27 April 2008 (UTC)
 * (reset indent) Sometimes I think it is better to explain visually than numerically, as hex colour codes are unwieldly. I know that HOW blue something looks will be determined by the monitor as well as the colour values, but I've cropped and increased the saturation of the plaque so I can show you what I mean here. Now I'm not saying you've created this blue - I admit it is there in my version too - but you have darkened it in your edit, which has made it much more visible and unrealistic looking. Thats all I'm saying. I'm not sure its really borderline blue-grey, but as I said, different screens will show different things. I'm not sure that I'd trust a laptop screen for accuracy though. I've never seen one I was very impressed with. Diliff  | (Talk)   (Contribs) 13:41, 27 April 2008 (UTC)
 * Yes there is definitely a blue cast to the plaque in the Edit. Mfield (talk) 14:21, 27 April 2008 (UTC)
 * I had a chance to check it using a CRT display today, yes there is a blue tinge. I would guess the displays on the Macbook Pros, despite calibrating, have issues with blue and grey because it certainly doesn't look blue on the MCP, but unmistabaly light grey-blue on the CRT. Given I didn't adjust saturatation and further testing of the original indicates blue hues were present in that area of the image, my brightness adjustment has merely increased the appearence of the blue rather than created it so I guess the image is more or less true to the original in terms of the colours present but not so in how pronounced they are. Capital photographer (talk) 03:14, 28 April 2008 (UTC)
 * What calibrator did you use? I ask because I have a MBP and it took some messing to get it to calibrate correctly. Using a Huey Pro, I found I had to apply the calibrator to the screen horizontally rather than vertically. Others have reported the same thing. I think its something to do with the way the LCD panel is constructed. With the calibrator vertical, the display calibrated with an unmistakeable tint. Mfield (talk) 03:28, 28 April 2008 (UTC)
 * The MCP 15" uses LED backlighting so it's suppose to be more accurate however I have heard rumours that the panels inside (like many notebooks) have lower colour range than desktop panels and, like most notebooks, it uses dithering. Anyway, I use a ColourVision Spyder2Pro, which isn't actually made for new LED backlit display which have a different backlight temp to regular LCDs. I think I'll need to do more tweaking, the prints I've done have being coming out ok but this image has shown there is definatly room for improvment. Capital photographer (talk) 03:38, 28 April 2008 (UTC)
 * The edit is even more unrealistic than the original. The sky is a dark grey blue and looks like there must be a storm on the way, yet the building is glowing bright as day. Kaldari (talk) 18:45, 29 April 2008 (UTC)
 * I agree that the sky is more unrealistic in the edit, but having a bright foreground and dark looming clouds in the background is perfectly possible without manipulation. The sun often breaks through clouds as a storm passes through. A moot point as it isn't in this case, but I just wanted to counter the argument that if it looks fake it must be fake. Diliff  | (Talk)   (Contribs) 19:12, 29 April 2008 (UTC)
 * Wow, the unprocessed original is different from what the processed original made out. Would have done it differently seeing that. Capital photographer (talk) 07:14, 30 April 2008 (UTC)
 * I wouldn't say it is unprocessed. I never added that caption, Kaldari did. I would just say it is less processed, or perhaps badly processed. But you can see how the lighting conditions were difficult. The fountain is surrounded on all sides by buildings, so it is darker at the bottom than at the top, and the sky is far brighter than the building. But to compensate for this isn't really stretching reality too far I don't think. What would you have done differently? ;-) Diliff  | (Talk)   (Contribs) 09:04, 30 April 2008 (UTC)
 * Well, the processed, "unprocessed original" appears much warmer. Where as with my edit, I was keeping it cool and crisp as per your edit which has a cool overcast lit pallette. The "unprocessed original" actually looks unnaturally warm for the setting, so I am puzzled by which is more accurate. I guess if I had seen the "unprocessed original" my edit would have tended warmer instead of cool. Capital photographer (talk) 09:10, 30 April 2008 (UTC)
 * I agree, it does look warmer, but I'd say that it isn't correct, if you look at other images of the Trevi Fountain. At the time I processed that image I was on a poor quality laptop screen and quite a lot of my images were not very accurate, looking back. I don't suggest that you base any changes on that 'unprocessed' version. About the only conclusion you can draw is the brightness differences, not the colour balance. I shot it in raw. If you let the camera decide white balance, each frame would be different, so you have to pick the settings and apply it to the whole series, and I evidently picked wrong. So as I said, I wouldn't call it unprocessed or authentic in all respects. Diliff  | (Talk)   (Contribs) 09:46, 30 April 2008 (UTC)
 * BIG Support Regardless of the colour of the sky, this picture truly stands out amongst all I have seen. The stitching is well done, as I can't seem to find any areas of problem. Kudos to the creator.  crassic ![ talk ] 14:57, 26 April 2008 (UTC)
 * On a technical level, I agree with 110% Capital photographer (talk) 16:18, 26 April 2008 (UTC)
 * Support Both  Beware  ofdog  16:51, 26 April 2008 (UTC)
 * Support edit Easily one of the best images I've ever come across on Wikipedia. Seegoon (talk) 17:05, 26 April 2008 (UTC)
 * question Is it just me or is the image slightly squished vertically? I wouldn't know it from the fountain, but the people look slightly... wide to me. Am I seeing things? Matt Deres (talk) 19:18, 26 April 2008 (UTC)
 * No, you're right about that. In rectilinear projection, because all lines remain straight, elements on the periphery of the image are distorted. With a view as wide as this, it is impossible to avoid, except by bending straight lines (see the Roman Baths nom for an example of that - you can see that the straight lines of the rectangular baths are bent). It comes down to which projection is more aesthetically pleasing to the eye, and whether straight lines are important to the composition. In this case,IMO they are. Diliff  | (Talk)   (Contribs) 20:51, 26 April 2008 (UTC)
 * Support. Glorious. Both versions are fine. - Darwinek (talk) 19:53, 26 April 2008 (UTC)
 * Support original Excellent. My only observation/niggle would be a slight tilt of the side roof sections of the building down in the center, but I am fully aware of how difficult it is to get this many images to stitch whilst correcting distortion. What software did you stitch with as a matter of interest? Mfield (talk) 21:19, 26 April 2008 (UTC)
 * Yep, I'm aware of that stitching niggle too. I tried to correct it but for some reason if I made any change to straighten the horizontal lines, the vertical lines shifted, and I figured the verticals were more noticable. You're right though, it is one of the more difficult scenes to stitch and correct perspective distortion on. As I mentioned in the image notes, the horizontal angle of view is about 100 degrees, which is borderline extrem for rectilinear. Diliff  | (Talk)   (Contribs) 22:19, 26 April 2008 (UTC)
 * Oh, and I use PTGui to stitch it as per my user page info. Diliff  | (Talk)   (Contribs) 00:15, 27 April 2008 (UTC)
 *  Conditional Support. Could someone fix the tilt? Aside from that, it is a stunning, informative, and high quality picture. Nautica Shad e  s  23:19, 26 April 2008 (UTC)
 * I just read the comment above, and it would seem that the tilt is unfixable. If so, that's fine. If I'm misreading it, then please tell me. Nautica Shad e  s  23:26, 26 April 2008 (UTC)
 * Well, I tried to fix the horizontal tilt (its not so much tilt as a concaving of the horizontals), and correcting it caused tilt on the verticals, so while I wouldn't say its definitively unfixable, I wasn't able to. Its very slight though. Diliff  | (Talk)   (Contribs) 00:15, 27 April 2008 (UTC)
 * Support Very nice work, can only image how difficult stitching this scene must have been. --Krm500 (talk) 23:50, 26 April 2008 (UTC)
 * Support either with preference to photographer's choice. --H92110 (talk) 07:37, 27 April 2008 (UTC)
 * Support technical and artistic merits, high encyclopaedic value. Guest9999 (talk) 13:19, 27 April 2008 (UTC)
 * Oppose. Overprocessing/HDR effect makes the subject look flat and unrealistic. Would prefer a single exposure version. Kaldari (talk) 16:47, 27 April 2008 (UTC)
 * Thats quite ridiculous though. Where have you got the impression it is HDR? It is a single exposure per segment. The only changes made were a slight contrast enhancement of the building, darkening the sky slightly and brightening the foreground. Have a look at the previous version. Diliff  | (Talk)   (Contribs) 19:19, 27 April 2008 (UTC)
 * After looking at the original version, I see why the current version looks unrealistic. The building in the current version is quite bright with perfect contrast despite the fact that the sky is quite dark. This is a common feature of HDR images, thus why it looks similar to them. Kaldari (talk) 18:51, 28 April 2008 (UTC)
 * Support both well done. —αἰτίας •'discussion'• 23:03, 27 April 2008 (UTC)
 * Oppose a very nice image, but the sphere-like crown and angel's heads at the very top are just too distorted, as are the people on the sides. —Pengo 05:12, 28 April 2008 (UTC)
 * The crown is clearly not spherical, it is quite elongated to begin with .. The people on the sides are rather incidental to the subject in question,surely? With such a wide view, could you really expect anything else? Ah well, you can't please everyone. Diliff  | (Talk)   (Contribs) 10:26, 28 April 2008 (UTC)
 * Oppose pincushion distortion Thisglad (talk) 10:07, 28 April 2008 (UTC)
 * Do you really expect distortion correction to be pixel-perfect? Would you oppose every other image nominated here for the same reason? Are you aware that a 'normal', uncorrected photo of this scene would be far more distorted, due to the angle of view? *sigh*. Diliff  | (Talk)   (Contribs) 10:17, 28 April 2008 (UTC)
 * The distortion is barely noticeable and is certainly less than what a wide angle lens would produce in a single frame. Capital photographer (talk) 13:05, 28 April 2008 (UTC)
 * The crown thing looked spherical in one of the alternate (web) shots, but I admit I didn't look very closely. I first noticed the angel's heads seeming very long though, and the people on the sides being very stretched, and I find it disorientating. Of course ideally the shot would be from far away enough as to not to require any correction whatsoever, (rather than a wide angle), although I'm guessing shooting from further away wouldn't be possible. —Pengo 14:51, 28 April 2008 (UTC)
 * No, its not really possible to take it from further, unless you'd prefer to look at a photo of the back of somebody's head, instead of the fountain. ;-) Its hard enough just getting this photo without people in the way. I had to stand up on top of a small pillar just to avoid them. This gives you an idea of what I mean. Diliff  | (Talk)   (Contribs) 15:56, 28 April 2008 (UTC)
 * Oh, and I didn't put two and two together at first, but looking at the crown at the top again, I realised it is from the Coat of Arms of Vatican City, and you can clearly see it is not spherical there. Diliff  | (Talk)   (Contribs) 08:40, 29 April 2008 (UTC)
 * Support per nom. Is the yellow building on the left tilted in real life?  Spikebrennan (talk) 14:27, 28 April 2008 (UTC)
 * What yellow building? You mean the orange-peach coloured one? I don't think it is tilted in the photo. It is deceptive because the drain pipes are tilted, but the building itself isn't. Not significantly anyway. Diliff  | (Talk)   (Contribs) 16:10, 28 April 2008 (UTC)
 * Support Original improved classic shot of classic building lifts the article - don't see that it needs fiddling about with any more. I suspect any "tilt" that people see along the top of the building is an optical illusion caused by the diffential in the adjacent building heights. Motmit (talk) 20:46, 28 April 2008 (UTC)
 * Support original. Full size doesn't suffer from low contrast/saturation like a small thumb. --Janke | Talk 08:12, 29 April 2008 (UTC)
 * Support original per nom and discussion above. Matt Deres (talk) 00:02, 30 April 2008 (UTC)
 * Comment. I just uploaded this section of the photograph to show the tilt I am speaking of. Once again, it is very minor and I support the image's promotion regardless, but some have said that they do not see it. Nautica Shad e  s  22:02, 30 April 2008 (UTC)
 * Support. A lot.   Spinach Dip  02:06, 2 May 2008 (UTC)
 * Support Original per Janke. Spencer  T♦C 19:10, 3 May 2008 (UTC)
 * Oppose top three versions. It would be good to fix the tilt before promoting. I think the tool you're looking for is called perspective correction. Papa Lima Whiskey (talk) 00:14, 4 May 2008 (UTC)
 * I have a sneaking suspicion that Diliff has heard of and is quite proficient with perspective correction ;-) -Fcb981 (talk:contribs) 16:34, 4 May 2008 (UTC)