Commons:Featured picture candidates/File:Ethiopia - Lake Assale.jpg

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

File:Ethiopia - Lake Assale.jpg, featured[edit]

Voting period is over. Please don't add any new votes.Voting period ends on 2 May 2023 at 02:54:33 (UTC)
Visit the nomination page to add or modify image notes.

Lake Assale - an Ethiopian salt lake, 130 m below sea level
  • That would be redundant, wouldn't it? And beyond whether all photos are "photo art," let's say for the sake of discussion that this is indeed heavily processed (although I don't know whether it was). Photographers have been heavily processing images for well over 150 years, haven't they? -- Ikan Kekek (talk) 13:40, 23 April 2023 (UTC)[reply]
  • You're probably right. Anyway - really full explanation on this one, so plan to support. But is the licence OK - the copyright info on the EXIF looks restrictive even though it has a CC 4.0 banner. Charlesjsharp (talk) 18:55, 23 April 2023 (UTC)[reply]
  • I'm now going to  oppose per A.Savin's comment below. --SHB2000 (talk) 01:11, 24 April 2023 (UTC)[reply]
  •  Support per Lifar. A red dead pixel at the bottom right and a white one at the left should be fixed.--Ermell (talk) 07:54, 23 April 2023 (UTC)[reply]
  •  Support Is the purple color realistic ? I don't know, but the wow is there. PS : I don't see dead pixels mentionned by Ermell. --Sebring12Hrs (talk) 11:22, 23 April 2023 (UTC)[reply]
  •  Comment first of all, Thanks to Ikan Kekek for the nomiation! I really appreciate it. Regarding the notes and other question: I fixed the tilt, I found the dead pixels (coming froming the 25sec expos., not using the reduction for it) ... the colors: I got curious, because usually I don´t do a lot of processing, except using hdr´s from time to time if light makes it easier or I can´t use filters. Usually other photographers tell me to add more saturation and so on. I uploaded the original RAW image here:
    Ethiopia - Lake Assale (raw image, no editing)
    Details of the shot: tripod, ND-filter, GND-filter, ISO 100, 25sec, f/8,0. The low ISO and low shutter speed explains the nice soft colors. And yes, it was a really blue to purple hour. Thhe whole set of images are crazy. Detail of editing: Exposure +0,63; contrast +24; highlights -100; shadows +22; whites -40; blacks -3; clarity +8; vibrance +12; saturation +32; no tone curve edits; no HSL or Color edits. More questions? Ask. :) --Snowmanstudios (talk) 16:40, 23 April 2023 (UTC)[reply]
Discussion
  • I don't know about this photographer's editing processes, but Camera Raw and Photoshop/Lightroom always output "Uncalibrated" for me when I save a JPG even though I have the colour space set in program (Google search suggests this is pretty common). Presumably the tint comes from the auto WB in camera? BigDom (talk) 05:55, 24 April 2023 (UTC)[reply]
  • Watching this image through 3 different devices, the aspects are very far to each other. And that's because the colors space has not been set, IMO. I have exactly the same camera, and my original files are all sRGB, either recorded in CR2 or in JPG. As I always shoot in CR2 +JPG simultaneously, it is easy to verify. Moreover the metadata of the original JPG coming out from the camera have no "Tint" section. Read the color space section on Commons:Image guidelines#Technical details: "Untagged non-sRGB images will not render correctly except by chance." -- Basile Morin (talk) 08:02, 24 April 2023 (UTC)[reply]
  • Due to obscure historical reasons the ‘Color space’ tag has only two possible values: ‘sRGB’ or ‘uncalibrated’. The value ‘uncalibrated’ (a very misleading name) just means that the colour space has not been defined as sRGB; every picture with another colour space will be labelled as ‘uncalibrated’. This photo contains an embedded colour profile, namely Adobe RGB. The Adobe RGB (1998) colour space has its problems, of course, but it is widespread and a valid choice – it represents e.g. rich cyans, turquoise and green tones, shades of orange and generally dark and saturated colours better than sRGB. So on the colour space/profile side everything is technically OK, the image is tagged correctly. (I have verified this with Photoshop 24.4.1, Apple’s ‘ColorSync Utility’ and exiftool.) The varying display on different devices must have some other reason – often devices are not able to display all AdobeRGB colours, and even today software is often still ignoring or mis-interpreting embedded colour profiles, assuming sRGB regardless of the actual colour profile embedded in a photo. --Aristeas (talk) 08:52, 24 April 2023 (UTC)[reply]
  • @Aristeas: it is perfectly possible with Photoshop to attribute a sRGB color profile, so that the exif are correct to use on Commons: Edition -> Attribute a profile, then select sRGB. I made a try and it works. Also Commons:Image guidelines#Color space recommends: "It is safest to use sRGB: which is the default on most computers, including Windows and Mac OS X 10.6 and later. Images in other color spaces will not render correctly at all in many web browsers". You are certainly aware of this because on your pictures explicitly Adobe RGB you include a warning: "Please view and edit it with an application that respects colour spaces and ICC profiles, else the colours of the picture may look strange." -- Basile Morin (talk) 10:22, 24 April 2023 (UTC)[reply]
  • Certainly ;–). I just wanted to point out that this photo does not miss colour space information (as that stupid ‘uncalibrated’ name suggests) but has an embedded ICC colour profile. Whether one wants to use colour spaces/profiles besides sRGB is something each editor must decide. Personally I use AdobeRGB when I notice that converting an image to sRGB would cause a loss of colours which is visible on good screens; these colours may not be visible on all devices (this is why I include that warning). Personally I think when person x uses a browser or setup which does not respect colour profiles, this is a problem of person x. Wikimedia Commons is a media archive with many uses, not only for on-screen display; many of the colours contained in AdobeRGB, but missing in sRGB can even be printed in ordinary offset printing, therefore I regard these extra colours as really useful, unlike the many other possible colours in more exotic colour spaces which often can neither be viewed on usual devices nor printed. – That sentence from Commons:Image guidelines#Color space is a bit dated; it refers to old software like Mac OS X 10.6 and Firefox 3.5 (oh yes, I remember them well, but I am old, too ;–). Since a couple of year at least on macOS most applications, including browsers, automatically respect colour profiles and try to display photos containing colour profiles as accurately as possible. I have been told that it is possible to achieve the same on Windows. So IMHO that sentence should be updated to the hint that whenever one wants to upload a picture in any other colour space besides sRGB, one should always make sure that it contains the correct colour profile; and that for simple images which do not need colour accuracy, e.g. for maps, one should best just use sRGB. --Aristeas (talk) 11:11, 24 April 2023 (UTC)[reply]
  • Well, that's your own interpretation, that I respect but do not share. For 2 reasons, 1) Macintosh environment represents about 10% only of the whole market, 2) if Adobe RGB was really better, with no drawbacks, then this profile would be recommend here. On the contrary, the aspect of non-sRGB images is often too far from what we expect (here). Secondly, the CANON 5D mark IV instruction manual confirms page 217: "Adobe RGB: This color space is mainly used for commercial printing and other industrial uses." [...] "The image will look very subdued in a sRGB computer environment""... -- Basile Morin (talk) 12:12, 24 April 2023 (UTC)[reply]
  • And that’s your interpretation; I think we can differ here and continue with two different views. :–) – I have referred to macOS only because I use this system for photo editing and therefore know it best in respect to colour reproduction. But I have good news: the same seems to be true today at least when viewing photos with Firefox on Windows 10 and Linux, too, without any additional configuration. I have taken the time to view some of my photos in AdobeRGB and sRGB on my wife’s office computer (a good, but fairly standard Windows 10 office laptop with a simple display) and on the Linux machine I use for programming (Linux Mint with default settings, a simple office monitor). On both machines Firefox handles AdobeRGB photos automatically correctly, wow! Of course it cannot display all colours, because the screens of both machines cannot reproduce them, but it displays them as good as possible, namely exactly as a correctly converted sRGB file. This means that today Firefox applies some colour management by default. (Firefox is just an example. If Firefox can do this automatically, every photo editing should be able to do it, too.) So we have obviously made some progress. – But even when this does not work on every computer, my view is that by publishing a photo on Wikimedia Commons I offer it not only for on-screen Wikipedia use and display on smartphones etc., but also for many other uses which may benefit from using a larger colour space. – The quotation from the 5D III manual certainly refers to out-of-camera JPEG files. When we set a colour space in a camera (many models allow to select either sRGB or AdobeRGB), this concerns only the JPEG files created by the camera. The raw image files from good cameras contain a wider palette of colours than sRGB or AdobeRGB encompass; in the raw converter we can select which colour space should be used for editing (if we select here sRGB, we already throw away many colours), and finally, when saving the result, we can again convert the file to another colour space (this is IMHO a good place to convert to sRGB, if we want so). An AdobeRGB file will only look very subdued when an application does not match/convert the colours, but just assumes that the file is in sRGB (this is what the old browsers did). As described above, this is not longer the case: today even on Windows or Linux machines with default setup, Firefox displays the photos with contrast, white and blacks, by converting the colour space, and many other applications should be able to do that, too, at least every photo editing application. (And if not – that is not my problem, sorry.) Best regards, --Aristeas (talk) 14:27, 24 April 2023 (UTC)[reply]
  • Then there is also ProPhoto RGB color space which is larger than Adobe RGB :-) Not a reference, probably for good reasons. In my experience, the "uncalibrated" (= non-sRGB) color space is most often detectable because of a visual discomfort. That means the colors display weird on the screen. Sometimes oversaturated. I also don't think that we can draw generalities from a few local observations. As the guidelines warns: "except by chance" -- Basile Morin (talk) 15:11, 24 April 2023 (UTC)[reply]
  • The ProPhoto RGB color space and the Wide-gamut RGB color space are excellent working colour spaces, but it makes no sense to use them with 8 bit per colour channel, for a proper representation of the large amount of colours provided by these colour spaces at least 16 bit per colour channel are necessary. And this is also the reason why they are useful for editing (and maybe archival purposes), but not (yet) for the final photos uploaded to Commons: the traditional JPEG format does not support 16 bit per colour channel, we would have to use JPEG2000 (not supported here?!) or the good old TIFF (still not supported well by Mediawiki software). – Oh yeah, I know that warning on Commons:Image guidelines#Color space and the hint in your camera manual well. I have read them a hundred times. The wording goes back to the 2000s, maybe even to the late 1990s; these warnings have been repeated, almost word for word, again and again; but people forget to check them, they just repeat them, “to be on the save side“. These warnings are still valid today regarding files without an embedded colour profile – these will display more or less correctly only when the colours correspond to sRGB. But are these warnings really valid for a reasonable part of today’s devices with photos that contain a correctly embedded colour profile? You are so right that we should not draw generalities from a few local observations. But why then do you draw generalities from your local observations? I have provided evidence that today even some cheap, mainstream devices without any configuration are not necssarily subject to the faulty display of non-sRGB photos, as it was the case in the past and as the warnings from the 2000s suggest. You say that I have just catched some lucky exceptions, and that most devices still ignore or misunderstand colour profiles. That seems a tiny little bit unlikely, but if you really want a definitive answer, we must stop here and ask somebody who knows exactly about the state of art of colour profile handling in today’s mainstream software on today’s mainstream OS on today’s mainstream computers, tablets and phones. Do you know somebody who is really competent? Maybe some imaging engineer at Adobe or Microsoft? --Aristeas (talk) 18:35, 24 April 2023 (UTC)[reply]
  • No, you extrapolate my position here. All the screens don't display well the uncalibrated photos, and even though you're convinced of the contrary, the consensus does not currently support your impression. You consider "applications should be able to do that, and if not that is not my problem, sorry". But a majority of experts in 2023 seem to think the other way: images should be sRGB, and if not, that is not their problem.
Now let's focus on this picture in particular please if you accept, otherwise the debate will go on forever -- Basile Morin (talk) 23:33, 24 April 2023 (UTC)[reply]
  • “ All the screens don't display well the uncalibrated photos” – maybe all the ones you tried. All the ones I have tried display them fairly well, namely as good as possible. “the consensus does not currently support your impression” – there is no consensus, only us two differing here. “But a majority of experts in 2023 seem to think the other way” – There are no such experts, or at least you have not mentioned them. You have quoted sources which just repeat the state of affairs of the 2000s, no current sources giving evidence about the current state of devices used today. And I see no experts – neither you (sorry) nor me is an expert. As I said above, get in contact with some colour management experts at Adobe or Microsoft, maybe Google, and report back what insights they have to offer. Now that would be interesting and convincing. “Now let's focus on this picture in particular please if you accept.” An excellent suggestion. You do not want to discuss your point of view. I cannot ignore the arguments I have mentioned. So this is not a targeted discussion, we just keep repeating our points of view. So let’s stop here and Let us part in peace. I have already suggested the same thing above. --Aristeas (talk) 13:56, 25 April 2023 (UTC)[reply]
  1. ✓ Done sRGB converted now. Thank you very much, Snowmanstudios.
  2. "The Adobe RGB (1998) colour space has its problems, of course" => Yes! Totally agree :-) We can discuss this elsewhere!
  3. Consensus: see how the softwares (designed by engineers) translate the non-sRGB pictures.
  4. 2000 => No, 2023. Canon 5D mark iv instruction manual edited in 2016.
  5. Thank you :-) Basile Morin (talk) 15:06, 25 April 2023 (UTC)[reply]
  • You seem not willing to close the debate, but repeat the points you have given above – thank you, that is not necessary, we can read them above. The sentence from the Canon 5D mark IV manual is just repeated from older manuals, see e.g. the original 5D I manual, page 67, which is “© Canon Inc. 2005” (p. 180); so that wording really dates back to the 2000s, as I said. And you totally ignore several of my points, e.g. about Commons being a free media archive for many purposes, not a simple Wikipedia picture repository. We can continue thus forever; that’s pointless. I’m out. Now you can declare yourself as the winner of the debate as often as you want, I will not care to contradict, because everybody can judge themselves from the arguments whether that claim is justified. --Aristeas (talk) 16:06, 25 April 2023 (UTC)[reply]
  • I can't believe I missed this chance to nerd out, I hope you guys are not done yet! :) Basile, my understanding is that indeed most, but not all, current software is able to accurately display AdobeRGB images i.e. to convert them on the fly to sRGB with minimal loss of colour fidelity. I've tried Aristeas' example below on a trio of platforms (Windows, Mac, and iOS) and a couple of browsers each, and image (A) displayed correctly on all six combinations (with just a hint of extra saturation). Defaulting to sRGB for taking images and uploading them somewhere is certainly the best option for most applications, but there are exceptions. For instance, photographers and graphic designers that work with both print and online media use a monitor with nearly 100% coverage of both Adobe RGB and sRGB (e.g., these), so that they get the best out of both worlds (wider gamut of colours for Adobe RGB, less posterisation in 8-bit exports for sRGB). I welcome Aristeas' initiative to offer images with different colour spaces where necessary, allowing the end user to choose the most appropriate one for their intended purpose. I would follow his lead but alas, my monitor and calibration device are sRGB only --Julesvernex2 (talk) 17:01, 25 April 2023 (UTC)[reply]
  1. Archive
  2. Archive -- Basile Morin (talk) 08:28, 27 April 2023 (UTC)[reply]
  • More good news, before somebody calls out “Firefox is niche, too”: AdobeRGB photographs also display well with Microsoft Edge on the same Windows 10 machine. Even better, they even display well in the nameless default browser installed on my smartphone, wow. I often think the world goes worse everyday, but at least colour profile handling has really improved since the 1990s. ;–) --Aristeas (talk) 14:50, 24 April 2023 (UTC)[reply]
  • Here is a quick test everybody who is interested can try: (A) AdobeRGB file with correct profile, (B) sRGB file with correct profile, (C) AdobeRGB file with sRGB profile (intentional error). Please open (A), (B) and (C), each in full size, in three tabs or windows. Then compare the colours. If (A) looks like (B), your machine cannot display the extra colours of the AdobeRGB colour space, but correctly matches them to sRGB – fine, colour management is working. If (A) looks like (B), but with a deeper blue in the sky, then congrats, your machine can even display some of the extra colours in AdobeRGB. If (A) looks like (C), your machine does not understand colour profiles at all, but handles all RGB files as sRGB. This is what the old warning meant when speaking about “subdued” or faded colours. And if (A) looks totally different from (B) and (C), the colour management is working, but buggy. --Aristeas (talk) 19:00, 24 April 2023 (UTC)[reply]
  • And, Basile, I absolutely did not want to criticize or attack you with my comment. I just dared to comment because that stupid “uncalibrated” thing pops up again and again, it is really very confusion, so I wanted to give a hint what it means and what not. If you feel attacked, I apologize. --Aristeas (talk) 11:37, 24 April 2023 (UTC)[reply]
  • Well, this is interesting. I see it the same way as @Aristeas. The light and color surly depends on cloud coverage, sun, dust and probably many other things in the atmosphere/sky. So shooting two images at the same time doesn´t mean the colors will be same, right? Have you ever seen a sunset/sunrise in central Europe impacted by the Saharan Air Layer. I guess there are plenty impacts like this. IMHO the colors of the sky and therefore the landscape look different every single day. We all see this every single day beeing outdoors taking this beautiful images and uploading them here. The color space is AdobeRGB, yes. Anyway I can uploaded a new version with sRGB later this day if you prefer. Snowmanstudios (talk) 12:44, 24 April 2023 (UTC)[reply]
  • For the unedited version ... yes, Sorry Basile Morin. You are right. The tint +32 came from the last edit of the processed image and wasn´t set back to zero in the virtual copy in Lightroom. I fixed this together with the color space for this image as well. Snowmanstudios (talk) 12:42, 24 April 2023 (UTC)[reply]
  • Thank you, Snowmanstudios! My emotion is honestly higher with the natural colors coming from the original picture. Then I would support a version based on this shot. The purple tint adds nothing in my view, and I regret it even spoils a bit the realistic aspect of this beautiful scenery. -- Basile Morin (talk) 23:33, 24 April 2023 (UTC)[reply]
  •  Comment This sentence: "Copying, duplicating, saving as a digital file, printing, publishing in form of media including web, manipulating, transmitting or reproducing without the prior written permission of Thomas Fuhrmann is strictly forbidden and would constitute a breach of copyright." is refuting the terms of the current license CC4.0 and should be removed from the metadata exif -- Basile Morin (talk) 00:49, 2 May 2023 (UTC)[reply]
  •  Support It is a long exposure photo (25 sec.). Probably made with a gray filter. Then you consciously get different colors and it becomes more like a painting and not an exact print of reality.--Famberhorst (talk) 05:22, 24 April 2023 (UTC)[reply]
  •  Support--Agnes Monkelbaan (talk) 09:01, 24 April 2023 (UTC)[reply]
  •  Support Beautiful. And, to my eye, the colours look perfectly attainable. Lighting can change drastically, depending on atmospheric conditions, the angle of the camera towards the sun, and the sun's elevation, so comparing this image with others of the same location doesn't tell much (e.g., this and this were shot at the same location within less than 5 minutes). The camera's white balance setting and the Raw processor demosaicing algorithm will further influence colours, making it very hard for anyone but the photographer to assess if the resulting image is faithful to what his eyes are seeing. In a nutshell: are these colours possible? Yes. Were they like that when the image was taken? If Snowmanstudios asserts they were, then they were. Does it really matter? Not for me, but I understand why others think differently. --Julesvernex2 (talk) 10:29, 24 April 2023 (UTC)[reply]
  •  Support Great capture that deserves a support. The suggestion to update the misleading hint about color profiles is very constructive, in my opinion. -- Radomianin (talk) 12:20, 24 April 2023 (UTC)[reply]
  • As requested, can you upload RAW file, though it seems that proves nothing if filters were used. In photo competitions (which this sort of is) I guess that the use of filters would normally be disclosed? Charlesjsharp (talk) 16:16, 24 April 2023 (UTC)[reply]
Confirmed results:
Result: 20 support, 3 oppose, 1 neutral → featured. /Basile Morin (talk) 05:21, 2 May 2023 (UTC)[reply]
This image will be added to the FP gallery: Places/Natural#Ethiopia
Retrieved from "https://commons.wikimedia.org/w/index.php?title=Commons:Featured_picture_candidates/File:Ethiopia_-_Lake_Assale.jpg&oldid=757739635"