Commons:Featured picture candidates/File:Reichstagsgebäude und Paul-Löbe-Haus, Berlin-Mitte, 170327, ako.jpg

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

File:Reichstagsgebäude und Paul-Löbe-Haus, Berlin-Mitte, 170327, ako.jpg, featured[edit]

Voting period is over. Please don't add any new votes.Voting period ends on 12 Apr 2017 at 05:43:02 (UTC)
Visit the nomination page to add or modify image notes.

Reichstagsgebäude und Paul-Löbe-Haus, Berlin-Mitte
  • Category: Commons:Featured pictures/Places/Architecture
  •  Info The east side of the Paul Löbe building in Berlin-Mitte, facing the river Spree, during the blue hour. On the left the Reichstag building can be seen. Both buildings belong to the German parliament ("Bundestag"). If you look carefully through the Paul Löbe building (direction west) you can see both the Federal Chancellery and the sunset. Taking this photograph was quite challenging. I've planned to take this photo since a while. Finally I had a free evening with suitable light and weather conditions. When I arrived at the place a really hughe amount of photographers (let's say around 20-30 persons) were already lined up at both sides of the river (behind me there was the Marie-Elisabeth-Lüders building which is also a well known photo subject). The light conditions changed very quickly so it wasn't easy to find the right settings to have a short exposure on the one hand (to avoid that the first frames of the stitched mosaic are differently exposed compared with the last frames) and to get the water of the river Spree smooth on the other hand (for this purpose I'd prefered a longer exposure even more but then the light situation would have change inbetween the single exposures too much). For the same reason I used my 35mm lens in this case instead of the 50mm: I wanted to have less frames to take. The dynamic range of the scene was very high so I had to do it using HDR technique. To have a short exposure I decided not to take five exposures as usual but only three exposures (-2 EV, 0 EV, +2 EV) for each frame. For the stitching itself I had to play around a little bit with different kinds of projections. In the end I've chosen a "Vedutismo"-projection because the rectilinear projection caused extreme stretching for such a wide field of view. Regarding the crop I decided to use a 16:9 ratio because 2:3 would have ended in too much empty space at the top and the bottom. Placing the buildings right in the middle of the picture follows the rule of thirds. Finally, after some hours of work, I'm personally quite convinced by the result. I hope you agree. I'll be thankful for any comment. All by me. -- Code (talk) 05:43, 3 April 2017 (UTC)[reply]
  •  Support -- Code (talk) 05:43, 3 April 2017 (UTC)[reply]
  •  Support -- Ikan Kekek (talk) 05:53, 3 April 2017 (UTC)[reply]
  •  Support --Uoaei1 (talk) 05:58, 3 April 2017 (UTC)[reply]
  •  Support Outstanding image! Thanks for the explanation, much appreciated! --Frank Schulenburg (talk) 07:38, 3 April 2017 (UTC)[reply]
  •  Support I also appreciate the explanation. --cart-Talk 08:29, 3 April 2017 (UTC)[reply]
  •  Support Thanks for both your pic and the explanation! --Martin Falbisoner (talk) 10:38, 3 April 2017 (UTC)[reply]
  •  Support Great pic, reminds me a bit of one of my all-time favourites --A.Savin 12:42, 3 April 2017 (UTC)[reply]
  •  Support --Cayambe (talk) 13:13, 3 April 2017 (UTC)[reply]
  •  Support ~ Moheen (keep talking) 15:02, 3 April 2017 (UTC)[reply]
  •  Support --Yann (talk) 19:18, 3 April 2017 (UTC)[reply]
  •  Comment very impressive - is there anything you can do about artefacts in the sky? Charles (talk) 22:11, 3 April 2017 (UTC)[reply]
  •  Support, the explanation is nice but the finished image still speaks for itself. Daniel Case (talk) 06:10, 4 April 2017 (UTC)[reply]
  •  Support -- Colin (talk) 11:40, 4 April 2017 (UTC)[reply]
  •  Support --Agnes Monkelbaan (talk) 18:08, 4 April 2017 (UTC)[reply]
  •  Support It is a very impressive picture, and I appreciate the thought and planning that went into it but I am just curious: What is the point of stitching a panorama from a whopping 54 frames instead of using a single wide angle lens (or maybe using just two or three frames) if the final version is going to be downsampled to only 20 megapixels anyway? Why not upload the full 50+ megapixels? Also, I slightly prefer the rectilinear projection for architecture (as it does not play tricks on the eyes regarding which lines are straight and which are curved), but I see that the rectilinear image doesn't include the interesting church Reichstag building on the left, and I would support either version. dllu (t,c) 18:53, 4 April 2017 (UTC)[reply]
On Megapixels
  • @Dllu: The stitched image has a field of view equivalent to a 3mm lens. Obviously, such a lens doesn't exist so it simply wouldn't be possible at all to take such a photo as a single frame. It would look rather ugly, too. You can have a look at the PTGui preview for a rectilinear version of the same view and I don't believe you'd find it better than this "Vedutismo"-version. I usually don't use stitching to increase the megapixel count but for other reasons: First I'm using the panorama head to simulate a tilt-shift-lens. When taking a multirow panorama and stitching it with PTGui I can easily get the verticals rectilinear without any loss of quality. Then I'm using the panorama head to get a field of view that I wouldn't get using a single picture. You can see here an example of more or less the same view taken with a 24mm lens. Third, I can easily improve the image quality by downsampling the stitched image to more or less the same size a single picture would have if it came directly out of my camera (~20MPix). Everything is sharp then and you don't get any complaints about noise and so on. All in all I prefer image quality over image size. I could of course upload a version with some more megapixels but I honestly don't see the use of it. You can print this version in any size you want and even on the screen you won't see anything in the 50MPix version that you can't see in the 20MPix version. By the way the buliding on the left isn't a church. --Code (talk) 04:32, 5 April 2017 (UTC)[reply]
  •  Comment Thanks for your explanation. I see the motivation for the Vedutismo projection now. Regarding downsampling, Commons:Image guidelines explicitly states "do not downsample", with the reason being that future devices may support more pixels. Indeed, there is already a 33 megapixel monitor. Regardless of whether the human eye can see so much detail, I still think that for a project like Wikimedia Commons, which seeks to document human knowledge, one should store as much information as possible. dllu (t,c) 04:47, 5 April 2017 (UTC)[reply]
  • I don't think it should be called "downsampling" when the result of my work has exactly the megapixel count a single picture out of my camera would have. As stated above I don't use stitching to increase the megapixel count but to get a wider field of view and straight verticals without losing quality or image size compared to a single exposure out of the camera. And as I already said I don't care much about Megapixels but I do care about visual quality. And we all know very well that many of our fellow Commoners would complain about noise and sharpness regardless of whether they look at a 50 or 100MPix picture. --Code (talk) 05:12, 5 April 2017 (UTC)[reply]
  • Dllu, I don't think the "do not downsample" guideline is always appropriate for stitched photos. With a rectilinear projection especially, the stretching that occurs is effectively upsampling, and so that's why I think some downsampling can be justified to get back to reasonable pixel sizes in the corners. I don't see the point in uploading >100MP images if they are soft at 100%, demonstrate lots of noise or have their pixels hugely stretched in the corners. All of Diliff's stitched photos are downsized significantly, though not nearly as much as this one here. I think, Code, you could have uploaded a significantly larger version than 20MP without any issues. Some of my stitched images are not downsized at all, and I was happy with the sharpness. Others are downsized, from memory, to 75%. Possibly if my image was handheld (and thus not fully sharp) or contained extreme stretching then a 66% downsize might sharpen things up sufficiently, but I can't recall whether I've gone that low. My guess here is that you generated an image well over 100MP and downsized to only 20MP. Your sharp lens, HDR technique and tripod will reduce noise and ensure a sharp image. So if we consider stretching... A cylindrical projection does not stretch the horizontal, only the vertical. So how do the pixels at the top of this image compare with the source pixels in each of your photos? IMO you should aim to be about the same, then you can claim the downsizing is appropriate, and the centre will be extra sharp. Another test I've used is to compare a full size JPG with a downsized JPG that is then upsized and sharpened -- is there any loss of fine detail? If there isn't then the level of downsizing is not too bad. It seems here your significant downsizing is only to avoid pixel peeping comments, and that makes me sad, because it is completely avoidable in other ways, and I think to be honest you are a respected enough photographer that you might be surprised to not get them. -- Colin (talk) 07:54, 6 April 2017 (UTC)[reply]
  • @Colin: I'm a little bit busy these days so I probably won't be able to do it before the weekend but I'll upload a larger version soon (maybe via OneDrive first so that you can compare both versions). --Code (talk) 08:26, 6 April 2017 (UTC)[reply]
  • The bigger one is definitely better. I'm surprised "full size" is only 66MP. I guess you must have a lot of overlap in your frames. When I upsize the smaller one and compare the two, there's a lot of extra detail in the larger one. Possibly a little over-sharpened. The extra size will also be useful if people want to crop it differently. I don't think, if you overwrote it, that anyone would complain about sharpness at all. -- Colin (talk) 19:02, 7 April 2017 (UTC)[reply]
Confirmed results:
Result: 18 support, 0 oppose, 0 neutral → featured. /George Chernilevsky talk 16:10, 8 April 2017 (UTC)[reply]
This image will be added to the FP gallery: Places/Architecture