Wikipedia:Featured picture candidates/St Peter's Square
From Wikipedia, the free encyclopedia
[edit] St Peter's Square in Vatican City, Rome
- Reason
- A detailed, clear, high resolution mosaic image of St Peter's Square showing the Vatican and its position within the city of Rome. Although it appears that the image is tilted, I have spent a lot of time studying reference points such as the horizon and corresponding points in the square and it is near enough to exactly level. However, it seems that St Peter's Basilica is not facing exactly straight toward the square and the avenue behind it, but I haven't been able to find a high enough resolution to confirm it. This is an example I did find though showing the same angled roofline.
- Articles this image appears in
- Rome, Vatican City, St. Peter's Basilica and Saint Peter's Square.
- Creator
- User:Diliff
- Support as nominator — Diliff | (Talk) (Contribs) 12:58, 17 May 2007 (UTC)
- Support. Soo many details, I wish there was a 100MP version though ;-) --Dschwen 13:48, 17 May 2007 (UTC)
- Support As above. How do you focus each segment? do you auto focus, use a tiny apature and the same focal setting. Basicly, what is the best stratagy for focusing a stitched panorama and how did you keep both the statues in the extream forground and the square in focus. thanks -Fcb981 14:38, 17 May 2007 (UTC)
- Good question.. From memory I focused on something in the middle (ie around the distance of the obelisk) of the frame and set the lens to manual focus, stopped the lens down significantly (around f/8 or so) and just scanned from bottom left to bottom right with around 25-50% overlap between frames, then up a row and repeated the process. As it was a 5x6 segment panorama (30 frames at 13mp each) with up to 50% overlap, I suppose it should be around the 200mp range, but I downsampled it substantially which would obviously have some impact on perceived sharpness. There might be more DOF issues at original 100% res. Diliff | (Talk) (Contribs) 16:04, 17 May 2007 (UTC)
- Which lens did you use? Did you perform the stitching at a high res version (that would need insane amounts of RAM, wouldn't it?). Do you have a 100% crop lying around, I'd be interested to see it. --Dschwen 16:52, 17 May 2007 (UTC)
- You didn't read the image description, did you? ;-) It was a 70-200mm f/2.8L. I think each segment was at around 70mm. And yes, it does use an insane amount of ram. By default I usually do my big stitches with a factor of 4 downsample (ie reducing width to 50%) in order to speed it up. The quality is basically the same as stitching at 100% and downsampling afterwards. When I really want to maximise image quality I'll export every frame to 16 bit uncompressed TIFF files but stitching that sort of image grinds my computer to an absolute halt (usually an overnight job even with 2gb ram) with some serious pagefile shuffling. This is mainly useful when the blender doesn't do a good job of sky gradation as processing the images in 16 bit tends to avoid posterization. I'm moving to the US in about a month and will be looking to splurge on a new PC and 30" 2560x1600 monitor... I'm looking at 4gb ram - maybe 8gb ram. With any luck that'll help to resolve the memory issues with stitching. Anyway, I'm digressing... Diliff | (Talk) (Contribs) 17:01, 17 May 2007 (UTC)
- Whoops, sorry. Don't you carry your 85mm prime around anymore? Well, I guess at that amount of downsampling it doesn't matter... --Dschwen 18:25, 17 May 2007 (UTC)
- I do, but stopped down that much the 85mm prime probably doesn't look any different to the 70-200mm f/2.8L anyway - they're both excellent lenses. I checked the EXIF and it was 135mm focal length anyway, not 70mm. Diliff | (Talk) (Contribs) 19:07, 17 May 2007 (UTC)
- If you used so many segments, then with all these people and cars moving around, how come we don't see any shadows anywhere on this photo? And I liked to know what program you used to stitch the photos. Thanks a lot! Amazing photo again from David. --Arad 21:04, 18 May 2007 (UTC)
- Well, there are shadows but because of the angle of the sun they're almost directly behind the people. I use PTGui to stitch the images and smartblend to blend them. It is pretty good at removing duplicates. I tried to take all the segments as fast as possible so that people/cars didn't have the chance to move far. When they're nearby the blending software will usually only keep one version of them! Diliff | (Talk) (Contribs) 21:32, 18 May 2007 (UTC)
- Thanks again! Which one removes the people? Smartblend? I'll get that program! But didn't you said you won't give Wikipedia high res images anymore? ;-) --Arad 23:19, 18 May 2007 (UTC)
- Well, technically compared to the 200MP original this is just a lowres thumbnail ;-). j/k. --Dschwen 07:24, 19 May 2007 (UTC)
- Thanks again! Which one removes the people? Smartblend? I'll get that program! But didn't you said you won't give Wikipedia high res images anymore? ;-) --Arad 23:19, 18 May 2007 (UTC)
- Well, there are shadows but because of the angle of the sun they're almost directly behind the people. I use PTGui to stitch the images and smartblend to blend them. It is pretty good at removing duplicates. I tried to take all the segments as fast as possible so that people/cars didn't have the chance to move far. When they're nearby the blending software will usually only keep one version of them! Diliff | (Talk) (Contribs) 21:32, 18 May 2007 (UTC)
- If you used so many segments, then with all these people and cars moving around, how come we don't see any shadows anywhere on this photo? And I liked to know what program you used to stitch the photos. Thanks a lot! Amazing photo again from David. --Arad 21:04, 18 May 2007 (UTC)
- I do, but stopped down that much the 85mm prime probably doesn't look any different to the 70-200mm f/2.8L anyway - they're both excellent lenses. I checked the EXIF and it was 135mm focal length anyway, not 70mm. Diliff | (Talk) (Contribs) 19:07, 17 May 2007 (UTC)
- Whoops, sorry. Don't you carry your 85mm prime around anymore? Well, I guess at that amount of downsampling it doesn't matter... --Dschwen 18:25, 17 May 2007 (UTC)
- You didn't read the image description, did you? ;-) It was a 70-200mm f/2.8L. I think each segment was at around 70mm. And yes, it does use an insane amount of ram. By default I usually do my big stitches with a factor of 4 downsample (ie reducing width to 50%) in order to speed it up. The quality is basically the same as stitching at 100% and downsampling afterwards. When I really want to maximise image quality I'll export every frame to 16 bit uncompressed TIFF files but stitching that sort of image grinds my computer to an absolute halt (usually an overnight job even with 2gb ram) with some serious pagefile shuffling. This is mainly useful when the blender doesn't do a good job of sky gradation as processing the images in 16 bit tends to avoid posterization. I'm moving to the US in about a month and will be looking to splurge on a new PC and 30" 2560x1600 monitor... I'm looking at 4gb ram - maybe 8gb ram. With any luck that'll help to resolve the memory issues with stitching. Anyway, I'm digressing... Diliff | (Talk) (Contribs) 17:01, 17 May 2007 (UTC)
- Which lens did you use? Did you perform the stitching at a high res version (that would need insane amounts of RAM, wouldn't it?). Do you have a 100% crop lying around, I'd be interested to see it. --Dschwen 16:52, 17 May 2007 (UTC)
- Good question.. From memory I focused on something in the middle (ie around the distance of the obelisk) of the frame and set the lens to manual focus, stopped the lens down significantly (around f/8 or so) and just scanned from bottom left to bottom right with around 25-50% overlap between frames, then up a row and repeated the process. As it was a 5x6 segment panorama (30 frames at 13mp each) with up to 50% overlap, I suppose it should be around the 200mp range, but I downsampled it substantially which would obviously have some impact on perceived sharpness. There might be more DOF issues at original 100% res. Diliff | (Talk) (Contribs) 16:04, 17 May 2007 (UTC)
- Support, and at least according to Google Maps' satellite image, there is a slight angle. --KFP (talk | contribs) 15:12, 17 May 2007 (UTC)
- Support, perfect detail and you can see many famous sights.--Svetovid 18:02, 17 May 2007 (UTC)
- Support - "Excellent" Booksworm Talk to me! 19:14, 17 May 2007 (UTC)
- Support - An awesome shot! Great detail! Nothing to criticize! In short, this should be FP. --Gabycs 19:38, 17 May 2007 (UTC)
- Support, you make Wikipedia prettier. And you remind me of how little I know about photography. gren グレン 20:12, 17 May 2007 (UTC)
- Support Stunning and flawless. As for the tilt, bear in mind that your instrument(s) are probably more precise than the ones they had when the built the plaza.--HereToHelp 20:41, 17 May 2007 (UTC)
- Support, Crystal clear even if full zoom. Most amazing picture I've seen on Wikipedia so far. How are there no stitching errors due to all those people? Centy 21:59, 17 May 2007 (UTC)
- Support Very sharp and detailed. -- RM 12:20, 18 May 2007 (UTC)
- SupportExcellent picture --St.daniel Talk 23:39, 18 May 2007 (UTC)
- Comment Can't fault this for sheer clarity and definition and for this and other reasons I'm not opposing it, but the apparent off-centre and almost-front-on lighting really ruin it for me. mikaultalk 23:45, 18 May 2007 (UTC)
- OpposeThis is a perfect picture from a perfect photographer and because it is otherwise so perfect that shadow in the lower middle bothers me.
Mbz1 00:37, 19 May 2007 (UTC)Mbz1
- Support, sensational. -- Phoenix2 (talk, review) 01:56, 19 May 2007 (UTC)
- Support Talk about a super high-res pic! I can actually see the faces on the ant sized people. Jumping cheese Cont@ct 04:25, 23 May 2007 (UTC)
- Support , perfect example of an encyclopedic photograph.--McKDandy 19:20, 23 May 2007 (UTC)
- Support Another amazing pic Diliff. Cacophony 23:00, 27 May 2007 (UTC)
- Support It does have some problems but I still think that it is a great picture --TrentpnGB 03:05, 25 September 2007 (UTC)
Promoted Image:St Peter's Square, Vatican City - April 2007.jpg MER-C 12:00, 30 May 2007 (UTC)
