Talk:High-dynamic-range imaging

From Wikipedia, the free encyclopedia
  (Redirected from Talk:High dynamic range imaging)
Jump to: navigation, search
WikiProject History of photography (Rated B-class)
WikiProject icon This article is within the scope of WikiProject History of photography, a collaborative effort to improve the coverage of History of photography on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.

Precision vs. Dynamic Range[edit]

The editing section confuses precision with dynamic range as if they were the same concept and is incorrect. One can have a high precision image (16 or 32 bit) shot from a low dynamic range camera. Also one can have an 8-bit high dynamic range image. The number of stops recorded in an image is separate from how precisely the brightness levels are represented. Precision and dynamic range are related in practice though, as is even stated elsewhere in the article, in that higher precision is more important in HDR images than SDR in that it allows sufficient gradations to be stored for each stop to avoid banding artifacts. — Preceding unsigned comment added by (talk) 08:20, 27 February 2012 (UTC)

Digital only?[edit]

Under the header "Photography", the article reads "[..] images must necessarily be digitized for processing." But! this very article mentions an early creator of analogue HDR, and I am sure there are many more. I think this only contributes to the common misconception of high dynamic range photography being all about the over-saturated colors. Polymeris (talk)

And the awful over-saturated example photograph of the German river scene does nothing to dispel that perception. (talk) 18:34, 11 October 2012 (UTC)
Yes, back in the day when photography required mastering recalcitrant materials and chemical processes, we would sometimes take multiple photos of the same scene and mask and combine them to synthesize an expanded dynamic range image. The idea of this now being blessed with it's own catchy acronym and pseudo-scientific sounding name is so post-modern, don't-cha think? — Preceding unsigned comment added by (talk) 02:11, 7 November 2013 (UTC)

Added more example exposures[edit]

I've added some example exposures of the HDRI. I can't get the positioning looking nice though. If someone else wants to have a go then be my guest. —Preceding unsigned comment added by Deanpemberton (talkcontribs)

I think it wasn't you who added the picture at the beginning of the Article, but IMO, that should be removed. The colors are vomitive, and it isn't a representative example of what HDR is. While I don't like it very much either, the NY shot further down the page is much better in this regard. At the very least, I think, they should be swapped. ~ Polymeris (talk)

Scanning film[edit]

I tracked down where the unreferenced scanning film section was added: [1]. Ren Kusack made the addition. I left a request on his talk page for references. —Darxus (talk) 08:46, 2 January 2011 (UTC)

H-D-R imaging?[edit]

I cannot see why the article title was changed to hyphenate HDR, i.e. from “High dynamic range imaging” to “High-dynamic-range imaging”. This is not how HDR imaging is commonly written, see for example Reinhard et al.’s HDR book. Tony1 just said this was a “MoS requirement”, but I could not find anything about this in WP:TITLE. I would like to see the old (and correct) title restored. Any other opinions? — Richie 13:23, 19 February 2011 (UTC)

WP's Manual of Style has sections on the use of professional-standard typography, including hyphens and dashes. Obviously, an article title should not be different from the article text in this respect. Omitting hyphens might be the practice of some specialists (as in many scientific/technical fields), but here we write for a readership with a wide range of expertise—critically, for the general reader too. Without the hyphens, I myself was asking how to parse this four-word nominal group: is there such a thing as high imaging? Range imaging? In fact, it's imaging over a high dynamic range, yes? Please account for both the needs of non-experts and the directions of WP's MoS. Tony (talk) 14:09, 19 February 2011 (UTC)
Although the change was at first jarring to me, too, I think Tony's rationale is correct. In my experience it is common for engineers—and hence engineering publications—to ignore style conventions, especially when it comes to hyphenation of multi-word adjectives and capitalization. For example, engineers often presume anything that has been formed into an initialism should naturally be capitalized when the initialism is expanded, e.g., "High Dynamic Range" everywhere. To conclude, I think Tony's view is correct that "high dynamic range" should be unhyphenated when it serves as a noun and hyphenated when it serves as an adjective. Robert K S (talk) 15:11, 19 February 2011 (UTC)
Thanks for clarifying the rationale, and I do agree on typographic grounds. I was just concerned by the deviation from the way that practitioners use the term: without hyphens, or just as an initialism, as pointed out by Robert. — Richie 16:56, 4 March 2011 (UTC)
There's only one problem with this explanation: WP:HYPHEN doesn't require hyphenation, it permits it. This is sound; hyphenation (as MOS also says) is slowly falling out of use in such contexts, where genuine ambiguity does not exist. (Tony is, as I would expect, correct on what the syntactic grouping is; but it doesn't need hyphens to support it.) 05:28, 12 April 2011 (UTC) — Preceding unsigned comment added by Pmanderson (talkcontribs)

nearly useless[edit]

I turned to this article for information abut HDR techniques and algorithms. It has next to nothing on such things. I also feel some of the material is misleading or even incorrect. WilliamSommerwerck (talk) 18:26, 22 December 2011 (UTC)

Agree. The big picture / 30,000' view is that it exaggerates differences. That is totally missing from the article. North8000 (talk) 11:22, 24 May 2012 (UTC)
Actually, to refine my comment, the article actually accurately covers the imaging process itself. Such represents the capture and storage of more information than is available from "standard" methods. Where the article goes wrong is that it incorporates creation of printed images in a way that exaggerates such differences as being a part of the process. North8000 (talk) 00:21, 25 May 2012 (UTC)

Seeking consensus on the images used to illustrate the subject[edit]

The quantity of photos in this article is far too great. Photos are supposed to provide concise illustration of the subject, not every possible permutation. I am also concerned about the technical quality of many of the photos, many of which are very amateurish examples of HDR. Please forgive me if I offend the creators of any of these photos. I propose the following:

  • The nav box should be removed as it does not contain the subject article. This should not be controversial and is backed by policy, so I will make that edit.
  • Eliminate the next four photos HDR image + 3 source pictures (Cerro Tronador, Argentina).jpg BrnoSunsetHDRExampleByIgor.jpg Leuk01.jpg New York City at night HDR edit1.jpg in the right column, but consider the top-most photo montage as a candidate for the examples section. The second photo montage has a low gamma and poor contrast; the third and fourth photos have clipped highlights and shadows, and is over-saturated. There are not good examples of HDR imaging. We should find a good representative candidate photo for the article main photo (top right).
  • Completely eliminate the gallery per WP:NOTGALLERY.
  • Keep the Gustave Le Gray photo. Gustave Le Gray - Brig upon the Water - Google Art Project.jpg
  • Keep the first example set of photos and eliminate the rest.
  • The time lapse video may merit discussion as well.

I would like feedback on this proposal and, depending on that, we can either make the changes or move to a formal RfC. - MrX 14:05, 24 November 2013 (UTC)

Under the assumption of silent consensus, I have made some of the changes that I mentioned yesterday. I removed the gallery based on commonly accepted practices and since there is a link to a multitude of HDR images on Commons. I have also removed the second example and formatted the first. - MrX 14:16, 25 November 2013 (UTC)
Belatedly, I support your proposals. I'd probably go one step further and remove the photograph of Leukbach (3rd thumbnail) as, IMHO, it looks like one of those HDR-style images that are popular these days. People already confuse that sort of garish tonemapping (often of LDR scenes) for HDR. nagualdesign (talk) 15:24, 25 November 2013 (UTC)
Thank you for your support Nagualdesign. - MrX 16:23, 25 November 2013 (UTC)
  • I have removed the other two superfluous images. I am going to see if I can find a more representative image on Commons for the lede, but the current one, while not exceptional, is at least technically closer to a representative tone mapped image resulting from an HDR process. I think it would be ideal to find a more recognizable subject, without the heavy orange tonal artifacts, and with a broader range of colors in general. - MrX 21:41, 26 November 2013 (UTC)
Agree. Images are immensely important and useful tool here, but those were superfluous. North8000 (talk) 22:59, 26 November 2013 (UTC)
Thank you. I'm putting together a straw poll of possible lede images, in my sandbox, which I will post here when I'm done. Of course, everyone is free to nominate their own choices as well. - MrX 23:34, 26 November 2013 (UTC)

Lede image candidates - straw poll[edit]

Here are some possible candidate images for the lede, selected based on these criteria:

  • Represents the technique with a full tonal range from light to dark, without large areas of black or white clipping (loss of shadow detail; blown highlights)
  • Colors are not exaggerated, over-saturated or grossly unreal
  • Local contrast is not extreme, especially in skies resulting in a gritty gray look (some of the nominated images push the limit of this)
  • The image is largely free of halos
  • The subject is recognizable
  • Confirmed as having the source images available to prepare a collage as the current lead image has.

For those interested, please vote for the one that you think best represents HDR Tone Mapping, or nominate your own choice by adding it to the gallery. You can also add criteria to the above list if you think there is anything that I missed. - MrX 15:56, 3 December 2013 (UTC)

Votes only (chose one image; comments go in the next section)
  • No. 9 Alfie↑↓© 22:36, 3 December 2013 (UTC)
  • 9 - MrX 22:48, 3 December 2013 (UTC)

I would oppose any image that does not meet the "sources images confirmed as being available" criteria. North8000 (talk) 16:18, 3 December 2013 (UTC)

While I agree that may useful, I don't think all of the prior lede images had source exposure or were in a montage format. I'm not entirely opposed to it, I just don't think it's entirely necessary. Of course, you can nominate a montage if you like. - MrX 22:48, 3 December 2013 (UTC)
  • I'm voting for 9 as well. I was leaning in that direction, and Alfie's vote has tipped the scales for me. - MrX 22:48, 3 December 2013 (UTC)
Two problems with image 9: There doesn't seem to be any originals available (and I think a montage is the only way to truly show what HDRI is, otherwise you get misconceptions like the one below), and the image shows considerable distortion. That said, I do think that an interior shot with windows may be the way to go. A 'room with a view' is the quintessential application of HDRI. nagualdesign (talk) 04:44, 27 December 2013 (UTC)

There doesn't seem to be clear consensus in this discussion, but what I glean from it is that the lede image should be a montage (or at least have available source images) and that an interior shot would be desirable. I will browse through Commons again to see if I can find any new candidates. Failing that, I may upload one or two of my own. I shot a Greek Orthodox church interior a few months ago and haven't yet post processed it, so that may work. I also have some interior shots of a Ritz Carlton with window views.- MrX 00:44, 2 May 2014 (UTC)

HDR image processed in single file[edit]

HDR image processed in single file by using Dynamic Photo HDR software

User:Alfie66, how could you see this image as bad example? You should discuss with me before you claim as "bad" and revert my edit. --Anton·٠•●♥Talk♥●•٠· 04:36, 25 December 2013 (UTC)

Without commenting on the quality of the image, it has no encyclopedic value beyond what is already included in the article. It also has clipped shadows and highlights, making it no different than a standard exposure image. Technically, it is not an HDR tone mapped image, it is simply a tone mapped image.- MrX 12:31, 25 December 2013 (UTC)
Did not you see other images, which are used in this article, has clipped shadows and highlights? The image is placed as a sample of single HDR processed image rather than different EV value images. Is there any standard for HDRI? --Anton·٠•●♥Talk♥●•٠· 13:02, 25 December 2013 (UTC)
AntanO, IMHO your argument is in direct conflict with the meaning of High-dynamic-range imaging. Starting with the word "imaging" (not merely image processing). North8000 (talk) 13:05, 25 December 2013 (UTC)
Anton, can you be specific about what other images you think are a problem, and why? The idea behind HDRI is to reproduce a broader reproduction of luminosity than afforded by the medium by combining images. The sample that you used seems to simply add some local contrast to emulate the effect. I don't see how that example advances the understanding of the subject.- MrX 14:18, 25 December 2013 (UTC)
Anton, a HDR image (not the tonemapped output) must, by definition, have High Dynamic Range. Which is to say, it must have a greater dynamic range than can ordinarily be captured with current technology. The resultant HDR image cannot even be viewed or printed using current technology (save, perhaps, for some OLED screens with ultra-high contrast ratios). For these reasons, HDR imaging is currently the process of taking several diffent exposures of the same scene, combining them with special software, then tonemapping them to produce a normal dynamic range image that retains detail in the highlights and shadows. Sometimes this produces halos and other artifacts, which have become synonymous with HDR. Processing a normal image to intentionally produce HDR-like artifacts, high saturation and/or high contrast is not the same thing at all. Although the crepuscular rays in this image are pretty, it doesn't belong in an HDR article. nagualdesign (talk) 22:36, 25 December 2013 (UTC)
Nagualdesign, there are serious flaws in your arguments.
The idea of HDR imaging is that the resultant image is viewable on the output medium, within its output DR. Standard display DR is about 8 stops, even less with prints, yet the resultant images make the appearance of high DR by using HDR tonemapping techniques that essentially compress DR but preserve the visual impression. If one happens to have a high-DR display, that's even better, because then heavy DR compression is not needed, but one can't expect all internet users to have them, and conventional prints will never allow wide DR.
I don't see combining multiple images in "HDR imaging" definition! It only says high dynamic range imaging! Multiple exposures were necessarily required some time ago, when the DR of the digital imaging sensors was barely wide enough for decent capture of the DR needed for standard displays. Still holds for cameraphones ;) But single-exposure HDR technique got wings with sensors like the one in Nikon D90, with about 12 stops by DxOMark, that is almost 10 useful stops, and is a bit more than the 8 stops viewable standard. It required multiple raw conversions at different software exposure compensation levels initially, because of the DR limitation of the raw conversion software (finite precision computation, restricted rendering options), but recently many raw converters can handle wide DR in a single conversion. A bit of this is even already available in in-camera DR "extensions", such as Nikon ADL, Sony DRO etc (not to be confused with multiple-exposure HDR modes, which tends to produce dull-looking results, due to restricted in-camera HDR processing), see for example
where you can notice that the input DR captured gets to about 12 stops. Still, manual processing from raw still produces better results than these automatic options, because about two stops appear to get lost.
Current best sensors offer two more stops, about 14 stops by DxOMark, and that's pretty much high dynamic range already. Multiple exposures can get you a bit cleaner shadows, but can't help you record unlimited DR, because internal reflections inside the lens and camera bright up the darkest parts of the image. Another issue is that the wider the scene DR, the more difficult is tonemapping to produce an image that still looks kinda natural (it can never be the same, but it can appear much closer to what we see than a default high contrast 8-stop camera rendering with blown highlights and sunken shadows). So it is very much valuable to have a HDR solution that can process moderately higher dynamic range with little fuss.
A raw file coming from a wide-DR sensor and containing an image of a wide-DR scene can be most certainly classified as a form of a HDR image, just because it contains a high dynamic range image!
Also, the number of bits used for storing HDR images is not directly relevant. In-camera multi-exposure HDR is generally acknowledged as a HDR technique, although it uses simple exposure fusion techniques without involving a high-bitdepth HDR master image (see Enfuse). Also (but not related directly), JPG files may hold more than 8 stops of DR (which may be difficult to display properly on standard monitors) despite having a bitdepth of 8 bits, due to the nonlinear storage technique.
For the record, I am not discussing the gothic form of false HDR tonemapping :) Not that I would want to deny anyone artistic value of such images, and the label I've used is a joke because I don't know a better non-joke term, but there are two characteristic differences. Firstly, they are (mostly) made from low-DR input. Secondly, the tonemapping is not conducted with a realistic intent (although a certain degree of artefacts such as halos is often present also in images rendered with a realistic intent - the higher the scene DR, the more difficult to avoid them, and keep some life in the image).
Hi Anton, see WP:BRD. BTW, if you produced the image with Dynamic Photo HDR, why does its Exif state Adobe Photoshop CS3? Alfie↑↓© 01:16, 26 December 2013 (UTC)
I know what does HDR or HDRI mean? I do not have conflict knowledge of HDR. But, the theory/definition of HDR changes as per many theories. Therefore, I was intended add a picture for “single image processed HDR”, and I was planned to elaborate in the article by mentioning single image processed HDR could be done by special software like Dynamic Photo HDR and in-built features of photo editing software like Photoshop. If anyone feels it should not be here and irrelevant to “encyclopedia”, it’s up to you. I don’t want to insist or beg anyone. --Anton·٠•●♥Talk♥●•٠· 16:49, 26 December 2013 (UTC)
With all due respect, "single image processed HDR" is meaningless. Raw files certainly have a greater dynamic range than, say, jpegs, and processing a digital negative 2 or 3 times to produce different exposures before combining them with software (HDR or otherwise) can help to retain highlight and shadow details, but a raw file isn't classed as HDR by any standards. To put it another way, a 32-bit HDR image has more than 260,000 times the number of possible tonal values than a 14-bit RAW image per subpixel. nagualdesign (talk) 21:34, 26 December 2013 (UTC)
//"single image processed HDR" is meaningless.// If you like to continue your theory, I don't wanna argue further. --Anton·٠•●♥Talk♥●•٠· 03:17, 27 December 2013 (UTC)
I think the reality is that it is no one specific meaning, leaving it to use however anyone wants for whatever their purposes. North8000 (talk) 11:40, 27 December 2013 (UTC)

Better examples?[edit]

Could someone more experienced create better example images? The current examples don't seem to do HDR jsutice, even simple correction of the -2 stop image gives much better results than these two examples.--2A00:1028:83D4:436:CD3B:25AA:61D9:B047 (talk) 12:32, 3 September 2014 (UTC)

I have replaced the lead image per the above discussion. The examples are supposed to illustrate HDR, not necessarily be the most aesthetically pleasing. Perhaps someone will be able to find or produce better examples.- MrX 14:40, 4 September 2014 (UTC)

Requested move 5 July 2016[edit]

The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review. No further edits should be made to this section.

The result of the move request was: No consensus to move, although it looks like the general notion of high dynamic range might deserve a broad-concept article. — JFG talk 00:00, 25 July 2016 (UTC)

High-dynamic-range imagingHigh dynamic range – Per WP:COMMONNAME since it is used by the vast majority of articles and by organizations such as ATSC, CEA, DVB, ITU, UHD Alliance, and Ultra HD Forum. I rarely see HDRI used outside of Wikipedia. GrandDrake (talk) 23:55, 5 July 2016 (UTC) --Relisting.JFG talk 09:29, 14 July 2016 (UTC)

  • Support per WP:COMMONNAME. Pageview stats plainly show this article to be the PRIMARYTOPIC of the two relevant pages on the DAB page. The other two entries are something else entirely.- MrX 00:18, 6 July 2016 (UTC)
  • Oppose since high dynamic range is also a useful concept in audio, radio, and other fields, and this added amgibuity does nobody any good. And thousands of books do use "HDR imaging". Dicklyon (talk) 21:19, 10 July 2016 (UTC)
The vast majority of people looking for high dynamic range will be looking for information on images and video so this would fall under WP:PRIMARYTOPIC. Also almost all of the articles from reliable sources I have seen refer to it as HDR which can be seen in a comparison of HDR to HDRI on Google News. --GrandDrake (talk) 22:10, 10 July 2016 (UTC)
I don't see why you would want to drop the noun part of the title, even if you would decide the primarytopic applies; better to let it redirect from the shorter title to the one that includes the noun that the topic is about: imaging. Dicklyon (talk) 22:18, 10 July 2016 (UTC)
I've taken the liberty of redirecting high dynamic range to the primary topic here, leaving a dab hatnote. fgnievinski (talk) 03:45, 11 July 2016 (UTC)
Why would you do that? That's a primarytopic grab without consensus. Dicklyon (talk) 03:49, 11 July 2016 (UTC)
Suggest rewriting it as a WP:DAB page. I convert bad redirs to those all the time.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:38, 11 July 2016 (UTC)
From what I can see the only reason there isn't a main article on HDR is because this article was originally about compositing multiple images together to create HDR. --GrandDrake (talk) 23:59, 11 July 2016 (UTC)
There's no such page as High dynamic range audio, so why would that be a consideration? Also, "thousands of books do use "HDR imaging"" is no more relevant or credible than the similar unsupported claim that "thousands of books do use "HDR"- MrX 21:16, 11 July 2016 (UTC)
  • Oppose per Dicklyon (who is a subject-matter expert), and per GrandDrake's objection to conversion of this from a noun to an adjective title, which we would normally not do, as covered at the WP:AT policy. Nom's rational that GHits for "HDR" prefer this topic over others is irrelevant, since the acronym form is not under discussion as a potential article title.  — SMcCandlish ¢ ≽ʌⱷ҅ʌ≼  20:38, 11 July 2016 (UTC)
WP:TITLEFORMAT is for deciding on questions not covered by the five principles. Having competing HDR articles on Wikpedia leads to awkward statements such as "The Ultra HD Forum will help navigate amongst the standards related to high-dynamic-range imaging (HDRI) and rendering (HDRR)" which isn't supported by any of the references in that article. It would be far better to have a main article on HDR which could than have subarticles on specific topics (images, video, computer rendering, etc...). Having a main article for HDR would be more concise and more natural. --GrandDrake (talk) 23:59, 11 July 2016 (UTC)
It makes perfect sense to have separate articles on HDR capture and photography versus HDR computer graphics, no matter what you call them. If you'd like a main summary-style article, it could be HDR imaging, with HDR photography as one of the subs. Dicklyon (talk) 00:47, 13 July 2016 (UTC)
It doesn't make sense for Wikipedia to have two completely separate articles on HDR which is causing people to link to both of them (and in the case of Ultra HD Forum to use terms that aren't used by any of the references). It would be more concise, natural, and recognizable to have a main article on high dynamic range that could than have subarticles on images, video, rendering, and anything else that is needed. High dynamic range is by far the most commonly used term in reliable sources and is the only term I have seen used in recent HDR standards such as Dolby Perceptual Quantizer and Hybrid Log-Gamma. --GrandDrake (talk) 04:19, 13 July 2016 (UTC)
A top-level summary-style article that covers imaging, audio, and other applications might make sense, but to move this one doesn't. Dicklyon (talk) 04:39, 13 July 2016 (UTC)
Dolby Perceptual Quantizer and Hybrid Log-Gamma are HDR standards that are based on human vision. Also in the context of PRIMARYTOPIC do you believe that most people who search for high dynamic range are looking for the Frostbite game engine mixing technique that allows louder sounds to drown out softer sounds or that they are looking for XDR (eXtended Dynamic Range) which is a duplication process that can be used in the production of pre-recorded audio cassettes? Based on PRIMARYTOPIC I think it would be better to move this article and than make changes to it. --GrandDrake (talk) 23:31, 13 July 2016 (UTC)
Here is a comparison of high dynamic range to high-dynamic-range imaging on Google News. I see no reason for Wikipedia to have two completely separate articles on HDR (High-dynamic-range imaging and High-dynamic-range rendering). It would be more concise, more natural, and more recognizable to have a main article on high dynamic range with subarticles on specific topics. --GrandDrake (talk) 23:11, 12 July 2016 (UTC)

The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.