Wikipedia:Featured picture candidates/delist/Image:Yarra Panorama.jpg

Yarra Panorama edit

Yarra

Edit by Diliff. A re-processing of the RAW files from scratch and re-stitched with slightly better perspective correction. This image is also slightly darker as I noticed the contrast was lacking slightly in the original. The shadows are still there as I'm not prepared to remove them.

Edit of Diliff's restitch to lighten it a bit, by Fir0002

Edit of Diliff's restitch with an anti-noise median by User:Fcb981

Reason: Can you see those shadows under the bridge? I love this image but those are major technical flaws.
Nominator: Arad

Delist or Replace with Diliff edit — Arad 00:32, 24 January 2007 (UTC)[reply]
~~Keep~~ I do see the shadows and also the blurred people on the bridge but I think the merits of the shot outweigh the flaws. The control of exposure is great, as is the subject and composition. I'd like to see this one stay -Fcb981 01:03, 24 January 2007 (UTC)[reply]
Comment. Image:Melbourne yarra twilight.jpg, also a featured picture, is very similar to this one. --KFP (talk | contribs) 01:06, 24 January 2007 (UTC)[reply]
Delist Technical flaws this severe are unacceptable. I wouldn't have approved this anyway even without the flaws --⁪froth^T 05:31, 24 January 2007 (UTC)[reply]
~~Reluctant delist. Composition is awesome but blurry people, focus and compression problems are too much to overlook. —Dgies^t c 05:35, 24 January 2007 (UTC)~~[reply]
- It's improved somewhat, but the blurred people still bug me. Changing to no vote. —Dgies^t c 02:34, 26 January 2007 (UTC)[reply]
Comment. I love it when people make blatantly incorrect assumptions. Why do you think those shadows are faults? They are not. That is how it looked. There are numerous downlights on the arch of the bridge. The arch stretches diagonally across the bridge, meaning that their position varies and creates a number of independent shadows as the light passes by the bridge rails. That is what has caused the 'major technical flaws'. I don't mean to be blunt here but why don't you ask the original photographer about them before making assumptions? Its one of my earlier panoramas and I will happily admit is in't among my best, but I don't see any major compression or focus problems. Diliff | (Talk) (Contribs) 10:10, 25 January 2007 (UTC)[reply]
- Compression problems are noticeable in the haze near the lights on the distant bridge, and on the narrow spire on the right bank. What I was calling focus problems (in the trees) are on second thought probably related to wind movement and exposure length. Really both of these are fairly minor and my biggest problem is the smeared people on the bridge. —Dgies^t c 16:04, 25 January 2007 (UTC)[reply]
  - I see some artifacts in the sky but I think they're quite minor but possibly made worse by some posterization which I could likely fix with a reprocess from RAW and re-stitch. As for the people, I don't really see them as a significant problem since they're incidental to the scene, but thats just how I see it. Diliff | (Talk) (Contribs) 16:44, 25 January 2007 (UTC)[reply]
    - You know that I admire your works. I was sure enough that you'll check this page and comment on the delisting as you did. Even if those are not technical, they are weird, unusual shadows on water which are almost 3D for some reason. Even the water changes the color from brown to black. Anyway, i don't think this is FP with those unusual shadows. But If you can make the compression problem better, then why not? Maybe you can also remove those shadows, even if they are natural? --Arad 21:52, 25 January 2007 (UTC)[reply]
      - Well, I disagree with removing things from photos in principle though. The shadows are real and are part of the reality of the scene. As always, FP is a subjective thing but I can't see how 'strange' (they're perfectly normal if you consider the physics) shadows on the water should affect it. Obviously nobody even noticed them during the original nomination and it seemed that most people were voting to delist based on the incorrect notion that they were technical flaws in the stitching or something. Diliff | (Talk) (Contribs) 22:42, 25 January 2007 (UTC)[reply]
      - Also, you could tell the same to those people in commons, when everyone were opposing because of those shadows, which means if it was nominated today, it wouldn't make it. --Arad 21:55, 25 January 2007 (UTC)[reply]
        I did tell them at the time of the nomination on commons! I quote: "It is a reflection of numerous lights from behind the bridge through the geometric shapes of the bridge supports. Thats why there is overlap of shapes in the reflections. Definitely not an optical artifact of the lens". Diliff | (Talk) (Contribs) 22:47, 25 January 2007 (UTC)[reply]

Yeah, the shadows looked natural enough when I first glanced at it but I was willing to take the noms word that they were flaws. as I said before the composition, control of exposure etc. out weighs the blurred people on the bridge. also the second edit exaserbates the artifacts. nuke it.-Fcb981 06:12, 26 January 2007 (UTC)[reply]

~~Keep~~. As explained, the "shadows" are not a technical flaw. howcheng {chat} 23:37, 25 January 2007 (UTC)[reply]
- Replace with edit 1. howcheng {chat} 18:46, 26 January 2007 (UTC)[reply]
Keep, but I propose that Edit 1 replace the FP as I feel it is an improvement in artifacts (or lack thereof), sharpness, perspective correction and contrast. I put a bit more effort into this attempt than last time. Diliff | (Talk) (Contribs) 01:01, 26 January 2007 (UTC)[reply]
~~Keep~~ Keep but replace …with edit Fir0002 (third version from top). Very impressive picture. Made me click on it for a larger view and spend some time studying it. P.S. ~~I like the original better only because it's easier to see the cloud detail on my LCD monitor.~~ I have a gamma-adjusted monitor and have observed that it shows shadow detail better than many barbarian-adjusted monitors that I happen upon. Many LCD monitors make shadow detail look very dark so the clouds in edit #2 will probably still be visible. On my monitor, edit #2 looks wonderful. If there is poor support for the Fir0002 version, then any of these is fine with me; just go with the consensus here. ~~Greg L 02:45, 26 January 2007 (UTC)~~ Greg L 21:49, 28 January 2007 (UTC)[reply]
change to edit Fcb981. yes it is my own edit but I think it does a good jod of killing the distracting artifacts in the sky. there is bound to be a little loss of detail on the very tops of the buildings that extend into the sky because I did a color select then feathered it 15px before doing a 7px median. -Fcb981 06:34, 26 January 2007 (UTC)[reply]
- I don't know exactly what technique you used on your edit but it has messed up the bright highlights in the sky and the water. Switch between both images and you will see obvious manipulation of them. For example, the spire on the right hand size is very dull at the top. There are similar changes all over the image. Do you really think there ARE artifacts in the sky? Could you be more specific? I don't really see them. Diliff | (Talk) (Contribs) 08:12, 26 January 2007 (UTC)[reply]
- Square reigons of pixles in the low light gradieants in the sky... maybe my lcd display creats them but I havent experienced them mostly, they become very apperant in Fir0002's edit. I'm sure you can see them there. there is also noiese in the clouds on that Edit. Although, clearly if the noise isn't visible to two other people it's more likely me than the picture. and yes there is some change to the skyline but if I had run a median or despeckle on the entire image you would have lost loads of sharpness. It's a nice picture in any event -Fcb981 15:39, 26 January 2007 (UTC)[reply]
  - It sounds like your LCD screen's gamma might be slightly off. However, I do believe that you're right about it being there even though its essentially not visible to me. That sort of thing (compression artifacts in dark areas) is actually often very difficult to avoid.. I saved it at a high quality/low compression setting in Photoshop so artifacts were minimised but JPEG compression loves to turn what it considers dark detail-less areas into squares. Short of bumping quality levels as high as they can go (and tripling file size), such things are almost impossible to avoid with lossy file formats (although I'm sure if a new lossy standard were to be developed and popularised today, quality would be far better). You could introduce luminance noise into the sky to force the JPEG algorithm to save all detail and not try to cut corners, but then you'd be left with... *drum roll* luminance noise! Working with lossless formats is the only way to avoid it really and obviously that is impractical for the web. My take on it is this: yes it exists; no it isn't that important. Diliff | (Talk) (Contribs) 16:33, 26 January 2007 (UTC)[reply]

I've just looked at Fir's edit and the artifacts are clearly visible in that version. That's because they're present (but unavoidable at reasonable file sizes and nigh-invisible) in the original JPEG and exacerbated by the processing that Fir's applied. I often find that with LCD screens (I regularly use three different ones), even if they're properly calibrated you can often see flaws in images which aren't visible on a CRT, due to the change in brightness/contrast with viewing angle. That can sometimes be handy when you're trying to make sure that an edit hasn't, for example, introduced clipping in dark or bright areas. More often, though, I find that using an LCD makes very minor quality issues seem distracting. If we want high-resolution images that remain accessible to people using slower internet connections, we sadly have to accept a (small) degree of lossy compression. --YFB ¿ 17:06, 26 January 2007 (UTC)[reply]

Replace with Diliff's Edit 1 - I find it mildly ridiculous that an image of this standard is even being considered for delisting, given some of the other images people have voted Keep for recently. Diliff's explanation makes sense and the shadows are only distracting if you're looking for them; removing them to suit the tastes of the ill-informed would be A Bad Thing. I don't see artifacts on my (properly calibrated) monitor and Fcb981's edit has introduced unsightly effects along the skyline. --YFB ¿ 12:23, 26 January 2007 (UTC)[reply]
Replace with Diliff's Edit 1. strong ditto YFB. --Dschwen(A) 19:05, 28 January 2007 (UTC)[reply]
Replace with Diliff's Edit 1 or my edit per above --Fir0002 08:32, 29 January 2007 (UTC)[reply]
Comment At least I got what I wanted, an improved version. --Arad 22:31, 31 January 2007 (UTC)[reply]

Replaced with Diliff's edit. --KFP (talk | contribs) 12:56, 1 February 2007 (UTC)[reply]