Commons:Featured picture candidates/File:Vista de Moros, Zaragoza, España, 2015-01-05, DD 19-36 HDR PAN.JPG

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

File:Vista de Moros, Zaragoza, España, 2015-01-05, DD 19-36 HDR PAN.JPG[edit]

Voting period is over. Please don't add any new votes.Voting period ends on 26 May 2015 at 17:48:30 (UTC)
Visit the nomination page to add or modify image notes.

View of the small village of Moros, province of Zaragoza, Aragón, Spain. The whole village of Moros lies on a hill, with the most relevant buildings in the top (church and former town hall), the residences in the middle and the sheep pens at the bottom. The current population of Moros is 441 people (35% of the population one century ago, that's why many houses are abandoned). The picture is the result of the blend of 24 pictures (panorama stitch of 8 + 3 frames of each for HDR purposes). It was necessary because the sun (shone on the left) made the clouds very bright, causing an overexposed sky within one frame. There is already a FP of this object but I am so amazed about it that I went back to shot at dusk.
  • Category: Commons:Featured pictures/Places/Architecture/Cityscapes
  •  Info View of the small village of Moros, province of Zaragoza, Aragón, Spain. The whole village of Moros lies on a hill, with the most relevant buildings in the top (church and former town hall), the residences in the middle and the sheep pens at the bottom. The current population of Moros is 441 people (35% of the population one century ago, that's why many houses are abandoned). The picture is the result of the blend of 24 pictures (panorama stitch of 8 + 3 frames of each for HDR purposes). It was necessary because the sun (shone on the left) made the clouds very bright, causing an overexposed sky within one frame. There is already a FP of this object but I am so amazed about it that I went back to shot at dusk. All by me, Poco2 17:48, 17 May 2015 (UTC)[reply]
  •  Support -- Poco2 17:48, 17 May 2015 (UTC)[reply]
  •  Oppose. I agree, it's a fantastic view but the image quality is just too poor. The church almost looks like a ghost (lack of contrast) and the mountains in the background have banding, shifting colour (patches of green, blue and red) and there is a significant blending error along the horizon on the left side. I think this scene at night definitely has the potential for a great FP but I don't think you've managed to nail it. I think it's a combination of processing problems and the choice of exposure (needed a much longer exposure to capture the details of the shadows). Diliff (talk) 21:36, 17 May 2015 (UTC)[reply]
  •  Oppose (for now) per Diliff. Great composition but all in all the picture looks flat. Seems like contrast is missing or something went wrong with HDR. I hope you can redevelop it. --Code (talk) 04:38, 18 May 2015 (UTC)[reply]
    • I'd be happy if the image could be improved with redevelopment too, but my guess is that it still won't be FP quality because there is significant banding and colour blotches in the shadow detail which is already very dark. The only way to avoid this banding is to avoid pushing the shadows too much, but if they are any darker, they will be practically invisible. That's why I said in my oppose that part of the problem is the choice of exposure. The darkest bracket needed to be much brighter to avoid this. Maybe seeing the detail of the mountains is not absolutely essential though. Diliff (talk) 13:13, 18 May 2015 (UTC)[reply]
  •  Comment This "stitching error" at the left horizon, my educated guess is that the lighting situation changed during the image series. On top of it a fairly sharp transition line makes this effect stand out.--KlausFoehl (talk) 11:22, 18 May 2015 (UTC)[reply]
  •  Oppose Heavy ghosting on the horizon when viewed at full scale.--KlausFoehl (talk) 11:22, 18 May 2015 (UTC)[reply]
  •  I withdraw my nomination Feedback follows Poco2 18:12, 18 May 2015 (UTC)[reply]
    @KlausFoehl, Code, and Diliff: : I have withdrawned the nomination because I agree that the quality of the result is really improbable and I will definitely do it better next time. Indeed, all frames were too dark and I had to artificially lighten it up to much. The reason why I didn't go over 30 seconds to get brigther frames: I wanted to have lots of detail and therefore opted for a stitched panorama of portrait-format frames. As the shot was also a HDR taking the 3 frames and rearranging the camera took me approx. 1 minute. The lighting one minute later was pretty different but I hoped that no so much and the transition from frame to frame would be "smooth". If I had spent 2 minutes for each shot (to get brigther captures), under these lighting conditions the result of each HDR would have been dramatically different. My question now to you: how would you have handled it? is there an easy way to let whatever software align the exposure over different frames so that there is no visible transition from one frame to the other? Poco2 18:38, 18 May 2015 (UTC)[reply]
    Hard for me to say exactly how I would have done it differently, because I don't know all the details of how you did it. What aperture and ISO were you using? What lens and camera were you using? Your EXIF information is completely stripped out so I have no idea. What I do know is that if speed of capture is important, it's usually better to shoot higher ISO than to underexpose and push. High ISO will of course have more noise, but you probably wouldn't have the same amount of banding in the shadow detail and you'd have a lot more flexibility in reducing the brightness later if you needed to. With HDR bracketing, you can be a bit lazy with your exposures too, because if you overexpose and need to reduce the exposure later, it won't hurt because you will also have darker images in the bracket to work with. I often shoot 5 bracketed images (maybe -3EV, 0EV and +3EV would have been enough for this scene, but maybe not). 5 bracketed images is often more than is needed, but your longest exposure in a 5 image bracket may actually be the same length as the longest exposure of a 3 image bracket anyway. The main difference would therefore be that you'd have 2 additional darker images to work with, which would preserve more highlight detail if you need it. Because the two shorter images are probably a fraction of a second, there's no problem with it taking longer to shoot 5 images instead of 3. The only downside is of course more processing time on your PC, and more space on your precious memory card being used. If you go for 5 bracketed images, you can also reduce the Exposure gap between exposures, which would again give you a bit more flexibility in processing later. Anyway, you could probably have gone with a wider aperture to shorten the exposure times too, but I'm only guessing as I don't know what aperture you used and how that would impact DoF and sharpness. I have the same problems with lighting changing quickly even with my interior stitches though. Sometimes I'm half way through a 10 minute panoramic shoot when the sun goes behind the clouds and ruins everything. So I start again, and then half way through the second attempt, the sun comes back out and ruins the second attempt. Etc. I much prefer overcast days for interior photography because at least I know the light won't change. ;-) Finally, what I would do is to process the HDR differently. I don't know exactly what HDR workflow you're using at the moment, but it doesn't seem to output realistic looking images to me. Diliff (talk) 18:56, 18 May 2015 (UTC)[reply]
    @Poco a poco, Code, and Diliff: : I have not much experience with hugin producing HDR, but I use it a lot for standard panoramic images. I you select the right options, alignment usually works pretty well, this being a necessary prerequisit for further steps. One can chose the order, first make HDR from stack the blend, or first stitch and then create HDR. If you prefer to do the HDR with a different programme, you may use hugin to stitch dark images separately, then bright images etc. and then use the stack of panoramic images in your favourite software. The seam between images is automatic and usually pretty good but can change, however one can save and load if a dark pair and a bright pair should require the very same seam. Still keep in mind that blending with enblend, incorporated into hugin, uses data from both images in the overlap region (for experts: blending width is proportional to spatial frequency wavelength). So hugin with built-in enblend and enfuse might be worth a try. -- KlausFoehl (talk) 09:46, 19 May 2015 (UTC)[reply]
  •  Info You can check here --Mile (talk) 19:21, 18 May 2015 (UTC)[reply]
    Diliff: the panorama was stitched from 8 HDR pictures (3 frames each) with this settings: 70 mm f/11 ISO100 30 sec, 20 sec, 5 sec. Seeing the result I agree with you that increasing the ISO would have been a way to improve the result. I am always reluctant to do that, and more with a lower ISO optimistic enough. I enfuse the HDR first and stitch the TIFF files together with PTGUI to generate a huge TIFF (this one 700 MB). To be honest I am quite happy with the result using Enfuse, apart from some interior images (as you pointed out in the spheric panoramas talk page), and would like to stick to it. I will play a bit around with some settings, though.
    Thanks for the link Mile. I had already read that page some time ago. My main problem is the exposure of the whole frame. I tried, of course, to adjust in LR each frame to the existing light to get them balance but the result was disappointing, as you would expect from a manual approach. I will work on this topic and let you know if I find a good way to arrange this. Poco2 21:01, 18 May 2015 (UTC)[reply]
    With single exposures, I'd be a bit reluctant to go to higher ISOs also, but honestly with HDR, it's less important. HDR merging and tone mapping always uses the 'best data' from each exposure and that is not very noisy even up to about ISO 800. That's why with my interior panoramas, I frequently shoot them at ISO 640 or 800 and there's really not much noise at all, except maybe in the darkest shadow areas. I would have bumped the ISO to about 200 and the aperture down to about f/8. The foreground may not have been perfectly sharp (not sure how far away it is but it looks fairly close), but you would have been able to reduce the shutter speeds by about 4x, or alternatively taken images that were +4EV which would have given you much better exposed shadows. Oh well, next time. :-) Well, if you're happy with Enfuse, keep going with it, but it's really not the best in my opinion. The algorithms it uses were good 5-10 years ago but have become a bit dated IMO. There is also the additional problem that you're blending first and then stitching. This is problematic and you can end up with different images that are not blended in the same way. Even if you use the same settings for each panorama segment, the actual data in each image determines how it blends them together and if you have one segment with mostly sky and another segment with mostly mountainside, they won't be processed with the other in consideration, and they won't blend well together when stitched). That may be part of the cause of why you have a blending error in the panorama. I can only recommend that you reconsider your workflow... That's all I can do. :-) Diliff (talk) 21:57, 18 May 2015 (UTC)[reply]