Talk:Exposure value/Archive 1

Need for another edit

This article has been raised in a discussion on DPReview [1]. The criticism is valid, Table 1 is non-sensical, and cuts across the already given definition of EV. I suspect that it should be relocated to later in the aricle, after the usage of EV as a measure of illuminance has been introduced, and then correctly labelled as 'EV₁₀₀. Further, the beginning of the article is repetitive. The section 'Representing camera settings:EV' repeats the section 'Formal definition', only makes its case less well. I propose to move that section up to just after 'Formal Definition' and reword as 'Motivation' and to move table 1 and surrounding text down, and correct the label on it. Don't want to put the work in and get reverted, so I'll wait to see if anyone has other ideas. Bobn2 (talk) 20:56, 5 September 2017 (UTC)

Untitled

Exposure value sounds right (I don't know for sure), but Canon calls it evaluation value in the explanation for Exposure Compensation.

Light gathered

In trying to understand all the values, I must admit Im a bit confused now. The F-Number article tells that the higher the number, the smaller the hole the light comes through. That makes sense (given the numbers f/8 for bright and f/2 for dark of the human eye). Now Table1 in this article says for e.g. f/8 the EV at 60s is 0 and at 1/8000 second is 19. So taking this table only, the lower the EV, the more light has been gathered (same aperture + longer exposure makes more light in total). But then I totally do not understand Table2. shouldn't it show the light amount gathered vs EV? So following that Table, the higher the EV, the more light will be gathered, contrary to what I got and explained from the other parts. Could anyone clear this up? especially in the article?

Table 2 explains what a meter does. Luminance is how much light is coming from the scene; when it's low, you need a slower shutter or lower f-number, hence a lower EV, to properly expose a photo. Dicklyon 15:56, 18 August 2006 (UTC)

Link to C.R.I.S.

I restored the link to C.R.I.S. Under “kyoritsu test equipment,” the links for the Multi Camera Testers (e.g., EF-1) make reference to EV and K, and the links for Calibrated Light Sources (e.g., LBF-2000) make reference to LV at ISO 100 and K. The link is included to support comments in the text.

I also made slight changes to section headings for consistency with the text ... which I forgot to note in the edit summaries.

JeffConrad 22:00, 19 August 2006 (UTC)

Jeff, thanks for that explanation. I won't delete it again, but would suggest we put a little explanation by the link, and maybe link deeper if possible to the relevant pages. I found nothing on a quick first look there. Dicklyon 22:07, 19 August 2006 (UTC)

The Flash makes it hard, but you can google up some URLs to use instead: [2] and [3] (both are in criscam's "old" section; someone needs to explain to them about URLs so they won't rely on the just the flash carp. Dicklyon 22:12, 19 August 2006 (UTC)

Dick, I'm hesitant to link to an "old" page for fear that it might vanish. Does the comment that I added at least partially address your concern? JeffConrad 05:37, 23 August 2006 (UTC)

Partially. But don't be surprised if someone comes along and deletes it as link spam. Dicklyon 05:57, 23 August 2006 (UTC)

Dick, see if the 24 August 2006 edit helps. If so, I'll make the same change in the Light value article (I need to fix "C.R.I.S.S." anyway). JeffConrad 02:03, 24 August 2006 (UTC)

I like that much better. Dicklyon 02:12, 24 August 2006 (UTC)

Edits of 7–18 October 2006

A long discussion with Marc Lacoste has raised several issues about the style of this article (and by implication, similar articles on photographic terms). Some of the issues probably apply to Wiki articles in general, but because a few are a bit esoteric, this may be the place to start.

The current version includes a combination of Marc's and my recent edits; if the format seems reasonable, perhaps it should be incorporated into similar articles. It does involve some effort, however, so if people don't like it, I'd rather not bother.

Citation of National and International Standards as Sources

The authoritative sources for several of the photographic term articles are national and international standards, such as from ANSI/ASA and ISO. Such references are comparatively uncommon, and are given little coverage in most style guidelines. As nearly as I can tell, they are not covered at all in the Wikipedia Manual of Style ("WMoS"). With author-date ("Harvard") referencing, the Chicago Manual of Style, 15th ed., ("CMS") would use citations such as (ISO 1974) or ISO (1974), depending on context. Nearly universal practice within organizations that publish such standards is a style such as (ISO 2720:1974) or ISO 2720:1974, and this style also appears to be common in the photographic literature (see, for example, Ray (2000 305–309), listed in the References fo the Exposure value article; the chapter I mention here actually was written by Ralph Jacobson). This form also seems to find use in Wiki articles; see, for example ISO 31. I prefer this form, which I have used in several of the photographic-term articles on which I recently have worked; it would seem to provide everything that strict CMS style offers, with the additional benefit of the specific reference for those who may be familiar with it.

Citation of Self-Published Material as Sources

I have avoided citation of self-published articles by Doug Kerr, Gordon McKinney, and me as References, because Wiki guidelines expressly forbid it. Although I agree with Marc's contention that Extenal links tend to collect all manner of garbage, I'd rather adhere to the Wiki guidelines because I am afraid that, once the door is opened to self-published material as sources, there will be no such thing as a reliable source. I'd rather we monitor external links and remove obvious link spam.

Jeff, I think there's a different standard for sources versus external links. I agree that self-published materials as sources is seldom OK (the guidelines don't say it's forbidden, though); there are some gray cases, though, such as David Jacobson's Lens FAQ; it's self-published (I think), but has been through some years of use, review, and revision, so it's pretty reliable. I think some of your stuff, and Doug's, are as good, but have not been subject to that kind of community exposure and feedback, so are not as good as sources. Nonetheless, they don't have to pass as high as standard to be used as external links. If they are non-commercial credible opinions or analyses related to the article topic, that's usually enough. Of course, you should hesitate to link your own papers. But if you choose to link some others, I'm going to add some of yours. Dicklyon 02:14, 13 October 2006 (UTC)

I think we agree; I like to think that I hold external links to a fairly high standard, even if it less than the guidelines for a reliable source. There may indeed be a few exceptions, such as Jacobson's FAQ, which for many is still the gold standard of introductory treatments of the subject, and possibly Merklinger's The INs and OUTs of FOCUS, which has been around for a while and has been extensively quoted. Nonetheless, I don't think either has the same status as works by Ray, Stroebel, or Zakia, and I would cite either as firmly establishing a position only reluctantly—how does one make the call on which self-published materials are reliable? If Jacobson, why not Atkins, Koren, or Rockwell (or even Conrad or Lyon)? There is a lot of nonsense on photo-related web sites, and once the bar is lowered, it may be impossible to prevent free fall.

Perhaps I'm shameless, but I'm not above linking to one of my papers (I'm sure not the only one to do such a thing), especially if it explains a point in more detail than may be appropriate for a Wiki article. JeffConrad 09:36, 13 October 2006 (UTC)

I hear you. Sometimes those Lyon papers are just what we need! Dicklyon 15:01, 13 October 2006 (UTC)

Let's put it this way: there are many far less appropriate links. It's the reader's choice to follow or not follow an external link, especially when it is not represented as authority; if the link to a potentially useful site or article is omitted, the reader does not have the choice. The key, I think, is the exercise of reasonable judgment when including links. When no endorsement is implied, the call isn't as critical. JeffConrad 23:17, 13 October 2006 (UTC)

Mathematics as Intimidating for Some Readers

The definition of exposure value is imprecise without the math, but precision may not be necessary for all readers, and may be offputting. I'm inclined to put the formal definition in a separate section; would this suffice to keep the article accessible?

Jeff, to tell the truth, I am not keen on the math the way it is now. I preferred what I had back in June, which started off with just EV = AV + TV and then the (logarithmic) definitions of TV and AV based on exposure time and aperture. I do appreciate the logic of your approach, but it's a lot more complicated. It brings issues of light meter calibration, photometry, and sensitometry right to the front, and defers the definition of EV to pretty deep in the intro, and involves squares and divisions with too many symbols for a naive user to get their head around. So, my suggestion would be to make a new section for all that stuff, and to put EV = AV + TV and not much else in the intro. This is the simplest valid definition, right? The rest is more about how to choose a good EV. Dicklyon 02:22, 13 October 2006 (UTC)

Isn't the simple (and hopefully intuitive) definition given in the article's opening sentence? The problem I see with with introducing AV and TV is that we dabble in APEX, a concept that essentially was stillborn. For practical purposes, AV and TV are superfluous except as historical footnotes. Having said that, perhaps the intial definition could deal only with the left-hand side of the exposure equation, with the more complete definition moved to an advanced section. Perhaps the APEX reference also should be moved to the advanced section. I wonder if squares and division are any more intimidating than logarithms (which still give me trouble ...). It sounds as if some simplification is indicated, however. Would putting the math in a section separate from that with the two images (with which it doesn't really belong) be a start? Perhaps the EV vs. camera setting table could be introduced before the math and without any reference to it. To my mind, the mathematical definitions and the relationship to luminance and illuminance are somewhat advanced topics that really aren't needed by most practical photographers.

On choosing a good EV: one must of course know the proper EV, either from a meter reading or from another source, such as a table. I added a link to Fred Parker's EV table, but Marc suggested that Wikipedia should contain such a table directly. Perhaps this article is the proper place for it? JeffConrad 05:25, 13 October 2006 (UTC)

I agree that choosing a good EV is important, and your expert knowledge of that metering topic is appreciated. But to me, the definition of EV is just TV + AV, and you can't really talk about it without knowing a bit of that APEX stuff. The logs can be intimidating, it's true, so maybe small 1D tables to define AV and TV would be a way around that.

I would argue that the definition of EV is

$\mathrm {EV} =\log _{2}{\frac {N^{2}}{t}}$,

which doesn't require mention of AV or TV. I'm working on a version that goes no further than this in the introduction. As nearly as I can tell, the concept of EV predated APEX by several years (some Hasselblad lenses had EV markings around 1957, while APEX was not proposed until 1960), so we should be able to handle the definition without mentioning AV or TV. I wish we could find a reader from Germany who was an actual witness (and can provide reliable sources, of course ;-) ). JeffConrad 06:30, 13 October 2006 (UTC)

Really? It had not occurred to me that EV could pre-date APEX. Anyway, that definition is OK, but it's going to scare some people. I thought using APEX was a way to make it much more accessible, but maybe I'm out of order. Dicklyon 06:48, 13 October 2006 (UTC)

I'm relying on comments in the forward to ASA PH2.5-1960 (long unavailable, unfortunately). As nearly as I can tell, the idea of EV was to simplify choosing among combinations of camera settings that would give the same exposure; the APEX proposal extended the concept, stating "There appears to be merit in the proposition that the exposure value concept be extended to include all the parameters of the camera exposure equation."

Some people are scared by any math, but I don't think my definition is any more intimidating than yours of 29 June 2006. In any event, I don't know how to simplify mine any further. JeffConrad 07:15, 13 October 2006 (UTC)

Table of Exposure value vs. Camera Settings

I extended the EV vs. camera settings table to EV −4 to cover the light from a full Moon; it could be argued that I should have gone to EV −6. The logarithmic EV scale was not chosen without reason; when linear sequences are used, some of the values can be come quite large, such as the exposure time for EV −4 at f/64. Accordingly, I expressed exposure times longer than 60 seconds in minutes, and appended the symbol 'm' to those values to so indicate. Apparently the meaning of the 'm' in this context isn't clear to everyone, even with a note in the table captions. The official SI abbreviation is 'min'; it's easy enough to use it, though the presentation of the data becomes less attractive. Alternatively, the really long exposure times simply could be eliminated, as I have done with really short exposure times less than 1/8000 seconds.

JeffConrad 01:57, 13 October 2006 (UTC)

I think the fix here is just to put the "m" note by the table, not in a footnote. Jeff, thanks for all your hard work and serious attention on this article. It has benefitted enormously from your work (even if I don't like your mathiness up front so much). Dicklyon 02:26, 13 October 2006 (UTC)

This sounds simple enough (Marc had the same suggestion); it also follows style convention for tables. With the note that close, it probably doesn't need to be clickable, though that's not difficult if it's needed. The question now is the footnote style ... conventionally, notes for table data cells are alphabetical, though I'm not sure that applies to table captions. I don't think a superscripted number is a good idea because we already have a numerical series of notes. Perhaps an asterisk is the solution ... I'd first like to see if there are any other comments, though. As Marc pointed out, we 'Merkins sometimes forget that what is intutive to a native-English speaker is not so for others.

As noted in my comment in the earlier section, it sounds as if the math needs some rework. JeffConrad 05:21, 13 October 2006 (UTC)

Edits of 14–15 October 2006

Dick, I left AV and TV out of the definition, in trying to cut the math to the absolute minimum. Put them in if you insist, but I still think they are irrelevant to this article (and they are well covered in the APEX article). The EV concept indeed was introduced to simplify setting camera controls, but without logarithmic markings on the camera controls, AV and TV are of no value. Presumably, when EV was introduced, the photographer would obtain the EV from a meter or tabulated exposure guide, and transfer that value to the guide. To the extent of my knowledge, the first use of EV was on Hasselbald lenses: one would depress the EV-lock button to change the EV, and then release it. The shutter time then would change with the f-number to maintain constant EV. There was no mention of AV or TV—the shutter and aperture markings were arithmetic. I don't think I've ever seen tables of AV and TV outside of obsolete ANSI standards and perhaps a photo book or two (I guess the Minolta Flashmeter III did specify f-number indication range in AV).

Well, I still don't see how you can omit EV = TV + AV in the section on EV and APEX. The whole section uses the complicated multiplication, squaring, division, log type math instead of the additive math from the system that it's talking about. Personally, I think APEX is useful and easy to understand, and even though it didn't catch on it's worth demonstrating a bit. Dicklyon 00:41, 15 October 2006 (UTC)

It depends on how complete this section needs to be. I tried to keep it to an absolute minimum, showing only how EV could be used as a measure of luminance or illuminance (necessary because nearly every equipment manufacturer uses EV in that sense). In particular, I had hoped to avoid essentially a complete duplication of the APEX system article, which covers the subject in depth. If people think it important, I can include some of the material from that article, but I'm always concerned about maintaining consistency in essentially parallel articles. Perhaps I also need to make it clear that EV as an indicator of luminance or illuminance is the only element of APEX that has survived. JeffConrad 02:03, 15 October 2006 (UTC)

Dick, I've added a one-sentence summary of the purpose of APEX, as well as a direction of the reader to that article. Does this help? JeffConrad 02:53, 15 October 2006 (UTC)

Dick, one other minor point that I missed last time: in light of the last sentence of the first paragraph in the article, isn't the sentence, "Each such increment is known as a change of one "stop" or one "step" in exposure" in the Formal definition section redundant? JeffConrad 09:49, 15 October 2006 (UTC)

Redundant, yes, a bit. However, I think having it there will be very helpful to people who don't fully appreciate that that's what the log2 implies. These are the people who won't remember from the lead paragraph a line that would help them interpret the definition. Dicklyon 16:39, 15 October 2006 (UTC)

I still think the inclusion of the APEX values is superfluous. We now define EV in terms of two other terms that we haven't defined, I don't know how that can be done. I also question use of "definition"; ASA PH2.5-1960 said, "The equation then takes the form ..." I also question the statement, "proper exposure value can be computed from a film speed and a light meter reading by EV = SV + BV"; I've never seen a meter that indicated in terms of APEX BV (I changed the acronym to LV for consistency with the rest of the article). Even the ANSI exposure guides don't list the values in the form indicated by APEX BV or IV (but that's another issue for another time and place ...) Also, why would we include BV/LV but not IV? I won't beat this to death, but I think we open a can of worms by mentioning the APEX stuff here. I don't see how the APEX values for the RHS quantities can possibly help the reader determine or understand exposure. If the material does remain, it think some rearrangement is indicated so that the details develop in sequence (unfortunately, that also would require some additional math). I also wish we could be consistent in the use of "stop" (the popular expression) and "step" (the formally correct expression, preferred by most of the ISO standards). JeffConrad 21:49, 15 October 2006 (UTC)

I've made a slight rearragement (we need the exposure equation before we can refer to it), trying to keep mention of the APEX stuff to an absolute minumum. I still don't think this material adds anything, but if you think it's needed, I won't bring it up again. One little issue that we get into by going back in time the conflict between current and previous standard letter symbols (e.g., $L$ vs. $B$). We also are using acronyms vs. the quantity symbols that APEX employed. JeffConrad 22:23, 15 October 2006 (UTC)

I reformatted the tables using inline styles; the coding is cluttered and ugly, but I don't know of a good alternative except for adding several new table styles to the Common CSS, which I doubt would amuse anyone.

I'm working on a table of exposure values based on the withdrawn ANSI exposure guides PH2.7-1973 and PH2.7-1986, and I hope to have it done in a day or two. Other sources, such as Kodak film data, have minimal data and several inconsistencies. The ANSI subcommittees that developed these guides were chaired by H.R. Condit and Leslie Stroebel, respectively, so they should be authoritative. Whatever I come up with will be an extreme simplification (the ANSI guides are 64 pages and 80 pages). I have some concerns about this, but most other tables, including the two in the external links, make similar simplifications.

I'm not sure that such a table belongs here, but I don't know where else to put it. It certainly doesn't belong in Light value. The more I look into this the more I am convinced that Light value, luminance value, incident-light value, or whatever are simply elaborately disguised forms of EV (when incident-light value is given for night scenes that consist primarily of light sources, one suspects that the values given do not derive from fundamental laws of physics). JeffConrad 22:40, 14 October 2006 (UTC)

I've added the table, for what it's worth. The location in the article probably is debatable, but I've assumed that a reader is primarily interested in using EV to set the proper exposure, and have tried to provide the necessary information for that first, with the stuff that is more of academic interest later. Considerably more subjects could have been included in the table, but the article already is getting quite long; hopefully, I have included situations of the most interest. JeffConrad 09:49, 15 October 2006 (UTC)

On second thought, I wonder if including the Further reading section listing Kodak Existing-Light Photography book is appropriate for this article. It contains a wealth of information about determining exposures for low-light situations, but this really isn't an article about exposure determination. Perhaps an article on that topic would be the appropriate repository for the tabulated exposure values as well existing-light data. JeffConrad 21:49, 15 October 2006 (UTC)

Edits of 18–19 October 2006

I made a preliminary attempt to eliminate redundancy in some of the articles on photographic terms. Much of the material on exposure meter calibration was removed, referring the reader to the Light meter article. At the time the material originally was added to this article, it was not included in the Light meter article; now it is redundant here. At Dick Lyon's request, some of the APEX material was restored, and a specific referral to that article (which now is more comprehensive than it was when that article was written.

This article still is a bit long; in particular, I'm not sure that the exposure guide ("Tabulated exposure values") really belongs here. Moreover, I think the link to Doug Kerr's APEX article is redundant (it's also in the APEX article). For some reason, we tend to mention APEX the minute exposure value comes up; the reality is that, while exposure value continues in common use, APEX. whatever its merits, was a concept proposed 45 years ago that never caught on. Had exposure mmeters been added to cameras just a few years earlier than they were, APEX never would have been proposed. JeffConrad 01:59, 19 October 2006 (UTC)

It's a big improvement. Still too long, maybe; I wouldn't object to more trimming. By the way, spurred by your comment I reviewed the APEX article, and made a few edits to the section about EXIF. Take a look. And I completely removed this bit that didn't fit with EXIF, so see if you think it needs to go back into some other section (sorry, looks like I truncated it some place):

The relationship between $I_{v}$ and illuminance similarly depends on the speed scaling constant $N$ and the incident-light meter calibration constant $C$:

$I=NC2^{I_{v}}$

ASA PH2.12-1961 recommended a value for $C$ of 20.8 ± 5. Using the nominal value, an $I_{v}$ of 0 was equivalent to 6.18 footcandles. Comparison with the recommendations in ISO 2720:1974 suggest that this value of $C$ applied to a cosine-responding (flat) sensor rather a cardioid-responding (hemispherical) sensor, but unfortunately, ASA PH2.12-1961 made no mention

My point being that APEX is alive and well in EXIF, if not exactly in full original form. Dicklyon 05:36, 19 October 2006 (UTC)

It's alive, after a fashion (I originally added the EXIF material); whether it's well remains to be seen. See my latest edit of the APEX article—there is a bit of a conflict. I cannot see a home for the deleted material, so I've left it out.

My edit also finds yet another awkward example of reference to standards. In this case, everyone knows the standard as EXIF 2.2, yet the official designation is JEITA CP-3451, which I doubt anyone but the developers would recognize. The first citation that I used (JEITA 2002) is the way The Chicago Manual of Style would do it; I personally find it a bit awkward though. JeffConrad 09:52, 19 October 2006 (UTC)

Tables and Formulas

I like having the formulas in this article, but I'm somewhat confused about the relationship between the tables and formulas.

It is stated that the EV in Table 2 is with reference to ISO 100, and that formulas are provided for conversion to different ISO values. Shouldn't the EV in Table 2 remain consistent regardless of what ISO one is using in their camera? For example: If the full moon has an EV = 15 (as per Table 2) shouldn't it always have this same value since it's supposed to represent the actual amount of light present? Maybe I'm misunderstanding EV, as I'm thinking that it's an absolute value, but perhaps it's actually a relative value?

I was expecting Table 1 to be referenced to ISO 100 instead, and that the formulas can be applied to correct for various ISO as required. Meaning if one wants to photograph a scene that has an EV = 12 using ISO 100, one would use the f/ and shutter speed as per Table 1, but if one wants to photograph this same scene using ISO 200, one would have to correct using the ISO formula(s)?

My confusion arose when I used these same formulas to create similar tables specific for my camera, but the numeric results didn't make sense to me, hence my confusion. Can someone clarify this for me?

--Dennis Lee 18:52, 12 November 2006 (UTC)

You need to re-read the article. EV is an indication of a combination of f-stop and shutter speed; not a measure of light level. To use it to refer to a light level, you need to add an assumed ISO speed and an assumed meter calibration. At other ISO speeds, different f-number and shutter speed will be required to give a correct exposure. Dicklyon 20:07, 12 November 2006 (UTC)

I think I understand now, I was confusing EV with LV (Light Value). Thanks. --Dennis Lee 23:18, 12 November 2006 (UTC)

Light value is an even more confusing topic, since it has no standard definition. Dicklyon 00:13, 13 November 2006 (UTC)

I'm not sure if this is the right place, but I think currently the correction for higher ISO values in the formula is incorrect. It says: log(N^2/t)+log(S/100) = EV(S), but I think the correct one is: log(N^2/t)+log(100/S), because when rewriting this to have t=100*N^2/S*EV(S) makes t inverse to S, which is more like reality (EV=1 at ISO=1600 will be effectively the same as EV=5). --SimonOosthoek

You have misinterpreted. It says the exposure you should use should satisfy EV(100)+log(S/100) = log(N^2/t). Here, S is inverse with t as you require. I think you'll find that to be consistent with the examples. Dicklyon 21:37, 21 June 2007 (UTC)

Thanks for the clarification, I see where I misinterpreted --SimonOosthoek

It almost sounds as if we should mention, once again, that a greater EV indicates less exposure. This behavior seems to confuse a great number of people. JeffConrad 06:35, 23 June 2007 (UTC)

Hello,

Check my numbers. If N=16 and t=1/60, you get about 14 for EV(100). My analog computer on my spot meter
incorporates the ISO too so let me set this at s=400. That is EV(400)= 12 NOT 16 like the formula on the page shows. There
should be a minus where there is a plus. It even shows up when I hold the 2 dials on my analog computer dial and turn
them. As ISO goes up, EV goes down. I recommend that the writer not debate with me and try out some examples first. I
admit that I am not so confident or arrogant to change someone else's page. Redding7 (talk) 18:08, 8 February 2009 (UTC)ChrisRedding

The formulas for adjusting EV for different ISO speeds apply only to the right-hand side of the meter calibration equation near the beginning of the section Relationship of EV to lighting conditions. If you set N = 16 and t = 16, you have EV 14, period. If you change the ISO speed on your camera, you will alter the exposure but not the EV. The values for EV₁₀₀ in Table 2 derive from fixed lighting conditions; let's oversimplify a bit and assume that each condition is a fixed, uniform luminance. Then

$2^{\mathrm {EV} }={\frac {LS}{K}}\,.$

If L is constant and S increases, EV also increases.

The EV scale on a meter calculator dial (as on my Pentax Spotmeter V) indicates EV₁₀₀ rather than EV; the value on the calculator dial thus corresponds directly to the values in Table 2, so you don't need to worry about conversion other than setting the proper ISO speed. If I have S = 100, and set N = 16 and t = 1/60, I also get EV 14. If I change the speed from 100 to 400, the indicated settings change to N = 16 and t = 1/250, but the EV₁₀₀ doesn't change. When using the meter, the value on the EV is presumably transferred from a meter reading of an object; the luminance of the object doesn't change just because you change the ISO speed for the camera, and the setting on the EV scale shouldn't, either. JeffConrad (talk) 04:37, 9 February 2009 (UTC)

Hi Jeff, I think you got my number. The spotmeter V EV is EV(100)!!! Boy am I stupid!! Redding7 (talk) 16:50, 9 February 2009 (UTC)Chris Redding

I thank you Jeff. Redding7 (talk) 17:20, 9 February 2009 (UTC)Chris Redding

It would be nice if camera and meter people would adopt the “EV₁₀₀” notation and use it when that's what they mean, but I doubt that this will happen. It's worth noting that the value also depends on the calibration constant K, so the values in Table 2 really are EV_100/12.5 or something like it. But this would be too much for most readers to digest, so I avoided it. You might note that your Spotmeter V has a K of 14, so if it's in calibration, it should indicate an EV 1/6-step less than the values in the table (at least for a single object of uniform luminance). In practice, a 1/6-step difference doesn't merit a great deal of worry. JeffConrad (talk) 08:21, 10 February 2009 (UTC)

I would like to resurrect the question of the sign for the ISO term in the EV equation. I have been trying to use the Ev calculation as a way to "correct" exposures taken for panoramas, when I have used a camera that does not allow me to fix the exposure for all shots. I suppose the first question is whether or not it is valid to calculate the Ev for two photos, then use software to adjust one of those photos up or down to match the other. I have done this using photo editing software, and visually, these look correct. My confusion comes from using the Hugin panorama application, which calculates Ev for automatic exposure optimization. Their algorithm subtracts the ISO term from the EV100 term, as opposed to adding it. Can anyone comment on this? Thanks! Sydthackerr (talk) 11:16, 28 March 2011 (UTC) Jack Gould

We have several “EV equations”; the only one I can find with an “ISO term” is

$\mathrm {EV} _{S}=\mathrm {EV} _{100}+\log _{2}{\frac {S}{100}}\,,$

which is correct as it stands. Were the sign on the second term negative, an imaging medium “faster” than ISO 100 would require more exposure, which makes no sense (recall that greater EV indicates less exposure). Note that the adjustment we describe is for different ISO speeds; I don’t see how this would normally be an issue with panoramas—although EV might differ among the various images, the ISO would not normally change. I’m not familiar with Hugin, and a quick glance at the documentation doesn’t indicate how or where it is doing what you describe. JeffConrad (talk) 22:44, 28 March 2011 (UTC)

Though it seemed obvious to me, perhaps something I could have added is that the equation assumes the same scene luminance, simply adjusting the EV for ISO speeds other than 100. Again, this usually would not apply to a pano in which some images had different exposures because of variations in lighting or metering. JeffConrad (talk) 03:54, 29 March 2011 (UTC)

Exposure Value Chart

===EV Graph===
Exposure Value Chart

Lazaro Gimena, it is generally considered a bit of a WP:COI or WP:OR to publish your own original creations in wikipedia this way. I move it here, so other editors can consider it. If someone finds it to be useful, gets you fix the errors, and wants to include it in the article, we can consider that then. Please do not put it back yourself. Dicklyon 03:42, 9 February 2007 (UTC)

I tend to agree with Dick. To me, this seems to present the concept of exposure value as more novel than it is; the concept has been around for almost 50 years, though I'll concede that in recent years, the average photographer's familiarity with the basics is not what it once was. I cannot see how over- and under-exposure as shown in the graph relates to the rest of the graph. Perhaps I miss something, but I don't think this graph contributes anything that isn't already covered in the article. To my mind, the extra ornaments tend to distract rather than elucidate.

This graph implies the more traditional table of EV presented as f-number vs. shutter time. The article had such a table until the edit of 7 October 2006; if people think it would be helpful to include the traditional arrangement as well as the current, I'd be glad to restore that table, formatted to match the current article style. JeffConrad 08:14, 9 February 2007 (UTC)

Lázaro Gimena 12:43, 9 February 2007 (UTC)
First of all, I want to thank expressly, for not deleting the EV Chart. I have to beg your pardon, because I am new in "wiki" and it's been a hard work for me, just to understand how it works. But I insist of my gratitude.
Respect to the EV Chart, that is what it is, a chart; not a table, not a graph, not a picture, I will explain.
If you have the time to have a deep insight on it, you will realized that most basic concepts related in camera shots are simply explained and related.
Normally, "new" photographers, confuse all these values (see light gathered coment above as an example).
The fact is that the camera, grosso modo, have to receive an "amount of light" (light flow) to fix the image. It depends on the environment (outdoor or indoor, bright or dimmed scene, for example) to set the shutter speed and aperture area, obviously.
For the same light state (EV number given by the photometer) there are different options to capture this mentioned light measured: faster and bigger or slower and smaller (in terms of shutter speed and aperture).
There are also other factor that is involved, the "film (or LCD) sensitivity" which it is also expressed in the chart.
Respect to the JeffConrad's comment, I have to say that I agree that design, colors and "ornaments" could confuse at the first sight and can be improved.
But they are not casual:

• decolorize from dark (grey) to bright (white) represents EV value (more or less light). Diagonal straight (thanks to algorithms) lines represents same light states or EV numbers. That's why if you take the shot above the line, the photo will be overexposed (more light than needed) and if it is below it will be underexposed (less light than needed).
• I agree, clouds and sun, seem to be not very serious, but representative of some normal lighting conditions (Table 2).
• It can also be seen that for a same state, the aperture chosen (with its corresponding speed-time) influence the depth of field represented by such kind of bow tie (wide or narrow).
• Below the shutter speed, there is a representation of time (represented with yellow in different widths) shots and how it influence the nitid or blurred photo and also the need of tripod use.
• There is also a rectangle that simulates the capacity that some cameras have (bracketing) to shoot varying a little bit the light conditions to avoid over and under exposure (it really takes three photos, normal, over and under, and you choose the best).

In the EV Chart there is no way to be confused once understood. Besides, there are not special maths procedure (at least superficially).
Thank you again, and hope this comment will clarify the EV Chart and could be useful for readers.
P.S. I've correct the errata.
Lázaro Gimena 12:43, 9 February 2007 (UTC)

Referencing styles

The WP:MOS allows both author-date ("Harvard") and endnote-style referencing; one style is no more proper than the other. The MOS specifically forbids changing from one style to another without prior discussion (see WP:CITE#HOW). We went through this same thing a while back; please don't do it again without some discussion and reasonable consensus.

The author-date system here essentially conforms to that described in the Chicago Manual of Style, 15th ed. I'm not especially fond of the coding style, either, but at present, the "Harvard" referencing templates have some issues. I'm certainly open to discussion on how to improve the current format, but far less receptive to capricious changes. JeffConrad 23:27, 16 March 2007 (UTC)

If the citation style is a real problem, then fine. You didn't have to revert both of my changes. My first edit didn't touch the references. This article still needs a clean up.

• Join up the multi-line paragraphs into single lines like my first edit. It's the style generally followed on Wikipedia. Breaking a paragraph into multiple lines makes it more troublesome to edit and diffs aren't as easy to read. Just let everyone's web browsers do the line wrapping.
• Replace all the HTML entities with their simple ASCII/Unicode equivalents. e.g &ldsqo;/&rdsqo; => "
• I don't like the way the references are wikilinks that look just like all the other links in the article. That's confusing and deceptive, especially when the majority of other Wikipedia articles uses a different system.
• I also don't like the mis-use of the reference tags to add "notes". Just put that stuff right in the text. The Sekonic "note" is just pointless.
• I think the article also makes rather heavy use of HTML and table markup to tweak the appearance.

Still on the HTML issue, was this copied from a web page? Simplifying the markup would make the article easier to edit. As it is now, it's a mess. All I want to do is make this article more consistent with the rest of Wikipedia. --Imroy 23:55, 16 March 2007 (UTC)

Whether my choices were the best is open to question, but I assure you that none was capricious.

• Ease of editing is in the eye of the beholder. I find the single-line paragraphs (such as here) much more confusing to edit. I also often use an external text editor, and for me, it's much easier to break the line. I rely on the UA to do the filling. Some of this may simply be habit; for years, I used troff, and failure to break input lines generally was considered bad style because it made editing with vi more difficult. I carried the practice over into HTML coding.
• I strongly disagree with replacing the HTML quote entities with ASCII quotes. Opening and closing quotes aren't the same; although they may look similar in the default Wiki sans-serif font, the neutral quotes look awful if a user has specified a serif font in his user CSS file, or if the user prints a page using a serif font.
• The Harvard templates give a similar appearance to the reference links used here. I originally had the references without links, but objections were raised, so I added the links. I'm open to a better approach. I've played with revised "Harvard" templates, but thus far, have found it quite a difficult task--programming in assembly language is easier.
• The use of reference tags for footnotes is not a misuse--see WP:FOOT. As with many other articles, this one seems (to me) to benefit from having both references and footnotes. One of the primary reasons the Chicago Manual of Style suggests the author-date reference system is that endnote-style references and footnotes really don't mix, even in a paginated medium. In a non-paginated medium such as WP, the combination of footnotes and endnotes is simply a mess. Incidentally, the purpose of the Sekonic footnote is to support the claim that Sekonic use a K of 12.5. There is so much nonsense bandied about when discussing meter calibration that an authoritative reference seemed appropriate.
• I agree that I've used a lot of table markup. I dislike including all this junk as much as anyone, but HTML table facilities are quite primitive, and wikitables even more so. For all of the terrible things said about troff and tbl (many of them justified), the tables were quite parametric, and could be accomplished with comparatively little markup. I originally used an HTML table, but someone wanted the wikitable style, so I used it. I made the decision to favor readability of the article over code minimization. Ideally, we would have both, but changes to the wikitable template would be required, and I didn't see that as my role to implement.

I'm as much a believer in clean code as anyone else, but not at the expense of proper presentation. Again, I'm open to improving the appearance and structure of the article, but I'd like to have some discussion prior to sweeping arbitrary stylistic changes. JeffConrad 01:17, 17 March 2007 (UTC)

I should have mentioned that my issue with Imroy's first edit was more with the elimination of the HTML entities than with reformatting the paragraphs. A few common characters (opening and closing quotes, and horizontal punctuation such as hyphens, minus signs, and em and en dashes) look the same in typewritten copy (i.e., ASCII) but are quite different in typeset copy. I think using the proper characters precludes ambiguity and improves readability, and I think readability takes precedence over simpler coding. Some others may disagree, however. JeffConrad 23:08, 17 March 2007 (UTC)

Exposure for circuses

Sources vary, but the one (ANSI PH2.7-1973, Photographic Exposure Guide) with the most specific recommendation for circuses, gives a value equivalent to EV 8 for ISO 100. A later version of that guide, ANSI PH2.7-1986, gives a value of EV 8–EV–9 for "Brightly lit sports events, stage shows, and the like" (included in the article's table), but makes no mention of circuses. The Kodak Professional Photoguide (Publication R-28, 1986), and Kodak's Existing-Light Photography (Publication KW-17, 1991) suggest values closer to EV 7. The recommendation could be broadened to include values from R-28, but doing so would properly require every other value to be re-examined as well. There probably are additional sources that merit equal consideration, so going beyond the current sources could be a bit of a project.

For now, I've revised the table to mention that the values for circuses and ice shows assume floodlighting. JeffConrad 02:22, 20 April 2007 (UTC)

EV as an indicator of camera settings

I don't get it. According to Table 1. faster shutter speed results in higher exposure value. While IMHO it should be the opposite. If we get 0 EV with 1/100 then setting shutter to 1/200 should give us -1 EV. The table is completely wrong i think. Can anybody correct me if I'm wrong? Jan Winnicki * 11:01, 30 September 2007 (UTC)

No, the table doesn't say that faster shutter speed results in higher exposure value; it says that faster shutter speed is a higher exposure value; that's what it means. Think of it this way: in higher light levels, your meter tells you to use a higher EV setting, to let less of that light into the camera, so you set a faster shutter speed. Dicklyon 15:59, 30 September 2007 (UTC)

The table is correct, but a current problem is that table is introduced with text: "EV corresponds simply to a combination of a shutter speed and an aperture setting, independent of any ISO setting." That is obviously incorrect, the opposite is true. Independent means not dependent, but EV obviously is dependent on ISO. The statement may only mean ISO is not a direct term in the EV formula, but it confuses novices, they imagine it means EV is independent of ISO, which is obviously wrong. If we know how to use EV, we all already know this.

Obvious evidence: Have a Sekonic meter? Bright sun will meter EV 15 at ISO 100, and the same light meters EV 18 at ISO 800. That is a dependence. Then the whole purpose is to look up either EV in table, which then shows proper settings for exposure at that ISO. EV is dependent on ISO, because ISO is already factored into the camera setting combination used by the EV formula. The settings in each next table row matches an exposure at 2x value of ISO, a first basic of photography.

The issue: I edited it to fix it, and someone reverted it. How is this solved? Wikipedia is currently incorrect. WF222 (talk) 16:19, 9 February 2016 (UTC)

It was correct as it stood. I think you are conflating actual settings and “recommended” settings—Table 1 deals only with the former. As indicated in ASA PH2.5-1960 and ANSI PH3.49-1971, EV can be applied to either side of the exposure equation. When applied to the left-hand side—as is done in Table 1—it refers only to the actual shutter speed and aperture, independent of ISO or light level. For example, settings of 1/100 s and f/16 correspond to EV 14.64—period. This EV will not, in general, result in the “correct” exposure unless it matches the EV recommended by the right-hand side of the exposure equation, illustrated by the values in Table 3. The “recommended” EV can be also be dictated by experience, as given in Table 2.

JeffConrad (talk)

Upon further thought, I wonder if the confusion arose because, as we have the article laid out, it’s confusing. Perhaps it would be more clear if the dual meaning of EV—as is described under APEX system#Exposure value—were stated near the beginning of the article. Much of this could be accomplished by simply moving Relationship of EV to lighting conditions to right after EV as an indicator of camera settings. Perhaps the dual meaning should be briefly explained before any of the tabulated values are given.

JeffConrad (talk) 08:38, 2 December 2016 (UTC)

Hasselblad gets all the credit?

Weren't Kodak Retinas the first cameras to use the LVS as it was originally known? Al Olson, who could probably provide authoritative sources, has written that the LVS was used on the IIc (1954) and the IIIc (1954). LVS was renamed EVS in '57 with the IIIC (1957). Doug Kerr has added the Ib (1954) and IB (1957). The Vitessa L (1954) has settings that are numerically equivalent to EV and an uncoupled light meter that gave a readout in EV (source: my Vitessa L). Seikosha introduced the MXL shutter in 1956. This had permanently-coupled LV/EV settings. The Aires 35-II-L, Konica III, Olympus LS/2C and Ricoh 500 (with "Triggermatic" rapid wind) all had the MXL. The MXL could be set to half LV settings. Because of the shutter speed series (eg 1/2, 1/5) the integer LVs did not all match to a standard aperture. (source: a 1957 article in US Camera about the Aires, Olympus, Ricoh and Konica cameras that used the Seikosha shutter with EV markings. The article refers to 'light value system also known as the exposure value system'.) Helen Bach 22:06, 11 October 2007 (UTC)

Dimensional consistency in definition of EV

Here is something more mathematical to think about: The definitions EV = log₂(N²/t) or 2^EV = N²/t are mathematically inappropriate: A scientist or a mathematician would not take the logarithm of a quantity with physical units (in this case, seconds).

To put it another way, consider the definition 2^EV = N²/t = LS/K in terms of the exposure equation. An exponent (in this case, EV) of some power (in this case, two) cannot have and should not have physical units. Exponents should be pure numbers. EV should be measured in stops, and stops should not depend on units. Stops are simply the number of multiples of two in a (pure) number. Nevertheless, the right-hand side of the definition N²/t has units of reciprocal seconds. Logically, the definition is incorrect.

(Examples of empirical laws with an exponential or logarithmic form abound: the Beer-Lambert Law, the Weber-Fechner Law, first-order chemical kinetics, biological growth. In each case, we have the logarithm of a ratio in which the physical units cancel.)

This would not be a problem if there were a t_o in the exposure equation and particularly in the definition of EV with physical units that would cancel the units of time. For instance, 2^EV = N² / (t/t_o) = L(t_o)S/K is a correct mathematical statement.

Note well that N = f/D_aperture and S = H_o/H_sp are pure numbers. (H_o is mathematically the speed point exposure for S = 1.) L(t_o)/K is also a pure number, though you might not recognize it as such because the exposure meter constant K is regularly quoted incorrectly with units of lux or lumen per square meter instead of lux seconds or lumen seconds per square meter. The definition of A_v is not a problems because N² or A² is a pure number. (Although using N as the aperture number is a problem because the APEX system uses N for scaling constant and A for the aperture number.) The definition of T_v is a problem that could be rectified using T_v = log₂(t_o/t) (where t is used instead of the T used in the APEX notation.) (Note that T_v is then the relative shutter speed in stops.) S_v is not a problem because film speed is a pure number. B_v is a problem and should be defined so that we are calculating the logarithm of the pure number log₂(L t_o/K). The APEX scaling constant N = 2^1.7 should be eliminated as unnecessary (in the face of a clever choice of K).

The authors of ANSI PH2.5-1960 should be taken as the authors of the APEX system. They were no doubt influenced by the German D.I.N. sensitometric naming system that used logarithms and that could very well have been around prior to the 1950s. Unfortunately, the authors were not strict theoreticians. The traditional definition of the exposure value should be referenced, but since this is an encyclopedia, the mathematically correct expressions should be used. As to what comes first (the topic of this discussion), it seems appropriate to start with the (correct) definition in terms of aperture number and exposure times (or relative shutter speed), to recognize that this ratio appears in the common (midtone) exposure equation whose purpose is to give a pair of camera settings, and to note that the binary logarithm converts these values into stops--particularly stops that can be added or subtracted: EV = stops(A²) + stops(t_o/t) where t_o/t is simply the denominator of the fractional exposure time. I don't see any reason to introduce the APEX monikers. You could allow that the various terms that go into the calculation are those quantities in stops. It should mention that the purpose of the EV is to make exposure calculations simpler. It does when you have stop tables for A2, t_o/t, S/K and measured values of Lt_o. Table 2 in the article would be more useful for calculations if it gave values of Lt_o instead of EV.
--Calculist (talk) 20:42, 15 October 2008 (UTC)

Indeed, the Germans who inspired the concept of EV were not strict theoreticians; the exposure formula in its various forms has never been other than a rule of thumb. But to devise a “correct” system of definitions would constitute original research, which we cannot do. I've added a mention that the definition of EV is dimensionally inconsistent; see if that addresses the problem. Because EV isn't a physical law, I've also added that exposure time is in seconds.

I agree that we don't need the APEX monikers, as I had previously suggested. They've never seen any practical application and serve more to distract than inform. But that suggestion did not meet with much favor, so I dropped the matter.

On the issue of strict correctness, EV differences are measured in steps (divisions of the exposure scale) rather than stops (devices that control the amount of light transmitted). But the proper terminology seemed to confuse most people, so it was relegated to a footnote. JeffConrad (talk) 22:05, 15 October 2008 (UTC)

As an alternative to your wording "so properly, the denominator is the numerical value of the exposure time in seconds" (which took me a moment to recognize what you meant), what about "so properly, the denominator should be divided by a standard exposure time, such as one second, to cancel the physical units in t and to leave a dimensionless number (in the same way that the aperture number N, the ratio of the focal length and the aperture diameter, is a dimensionless number)."

I had similar misgivings about the wording; I've tried to simplify it a bit—see what you think. I think discussing a standard exposure time t₀ is more likely to confuse than inform; perhaps more to the point, though, I'm not sure that dimensional consistency really applies to other than physical laws, and if we get too carried away with correctness here, we may imply that EV is more than it really is—just a number that conveniently encodes combinations of camera settings. JeffConrad (talk) 23:30, 16 October 2008 (UTC)

Steps and stops: that is something that even I have avoided.

My other suggestion is a different form of the EV table. I find a table of EV values with aperture numbers and "shutter speeds" along top and side useful. I have, for instance, along the top, the representative aperture numbers, the actual aperture numbers, and the number of stops (steps). In that way the reader can see the two components that add together and the stripes of equal EV values for equivalent camera settings. I can give you a copy if you want it.

We originally had the table in that form (as in this 12 September 2006 version), but some people objected, and we put the table in its current form, which, in my experience, is less common than the one you propose. There is some argument for the current form, however; if one first determines the EV required by certain lighting conditions, putting EV as the stub column does make for a fairly logical index into acceptable combinations of camera settings. JeffConrad (talk) 23:30, 16 October 2008 (UTC)

Again and finally, a table of Kodak Lvs is easier to use in calculations than a table of Evs. You have to be careful, though. If you or Kodak has used the values from the ANSI Exposure Guide (1986), there is that inappropriate scaling factor N = 2^1.7 that needs to be removed. --Calculist (talk) 21:50, 16 October 2008 (UTC)

This article is about exposure value, so I don't see how we could give the LVs more than a mention. If we do so, they probably also should get mentioned in the Light value article. I don't have an authoritative source for Kodak LVs, but if you have one, I'd be glad to help get it added, especially to the Light value article—we've covered almost all other uses of the term, so addition of the Kodak values would be good for the sake of completeness, as well as to demonstrate that the concept has been around for quite some time.

Tables in ANSI PH2.7-1986 gave incident-light value I_v, defined as in ANSI PH2.5-1960. All values that I used in Table 2 of this article were converted to EV for ISO 100 using

$\mathrm {EV} =I_{v}+S_{v}$;

the scaling factor (strictly, 2^−1.75) was incorporated when calculating the speed value S_v. JeffConrad (talk) 23:30, 16 October 2008 (UTC)

I've slightly tweaked the explanation for the dimensional inconsistency, and moved it to a note to avoid needless repetition in the article. I'm still not sure whether it adds to or detracts from the article; if the former, we should add similar notes to APEX system and similar articles; if the latter, we should remove it from this article. The photographic literature is filled with sloppy use of terminology; as Dicklyon mentioned in Talk:Shutter speed, it is not our prerogative to change that terminology, however wrong we think it may be. We probably could address most of the technical inaccuracies by adding notes such as the one I added to this article, but in doing so we could be opening a can of worms. Opinions, anyone? JeffConrad (talk) 21:57, 17 October 2008 (UTC)

I agree with Calculist; it takes only one line to clarify that $t_{0}$ = 1 sec. User:Nfette —Preceding unsigned comment added by 68.226.31.48 (talk) 17:41, 23 December 2010 (UTC)

It would be easy to add, but I fear we’d be venturing into WP:OR. The way we have it is how it’s done by every published source I’ve ever seen. Lack of dimensionless function arguments is actually quite common; for example, glass manufacturing usually deals with the log of viscosity, and we speak of log exposure in photography. Although this practice goes against what many of us have learned, our role here is to document rather than criticize or “improve” a formula or “educate” those who we think misuse them. We may already be pushing it with the note I added in response to Calculist’s comment. JeffConrad (talk) 00:23, 24 December 2010 (UTC)

EV of the Sun Surface

I've added an EV value for the Sun surface: EV 34.

I realize this seems somewhat bizarre. Hardly anyone takes pictures of the sun surface with a DSLR, and it actually can't be done without special equipment, since even ISO 50, f/32, and 1/3200 sec, which is about the bright limit for most cameras, is only good for an EV 24 scene. You need a Baader film (optical density D=5, which corresponds to 17 stops) to cut the light in order to take a sensible picture or view the sun safely.

Bizarre or not, I've found this value useful in the design of cameras. An interline CCD for instance, might have a specification of 1000x antiblooming. That means that when the wells overflow, 99.9% of the electrons end up in the overflow drain, and 0.1% end up in the adjacent wells. Because the sun is often 1,000,000x brighter than the rest of the scene, there can be a significant overflow current into those adjacent wells. You might then ask what exposure setting would be necessary to eliminate the big streak through the picture when the sun is in the frame. The calculation has many details, but one of the inputs to the calculation is the surface brightness of the sun. And if you try to reproduce the problem in the lab, it turns out that many things that are bright, such as halogen spotlights pointed directly into the camera lens, or the Space Shuttle rocket plume, are not nearly as bright as the surface of the Sun. Focussed laser beams can be.

I also realize that the value I've supplied does not come with a reference. It can be independently derived from the other data in the article: the full moon surface is listed as EV 15, and a scene under full moonlight is listed as EV -3. A scene under bright sunlight is listed as EV 15, 18 stops brighter than full moonlight. Because the moon and sun are nearly the same angular size in the sky, the 18 stop brighter lighting must be coming from an 18 stop brighter surface. Thus the surface is EV 15+18 = EV 33. My own measurements give a value of EV 34, and I can't account for the 1 stop difference. If someone wants to change the listed value in the article to EV 33 I'm fine with that.

Iain McClatchie (talk) 09:02, 17 December 2008 (UTC)

And what if we're not fine having your WP:OR instead of WP:Verifiability? Notice that an official source is cited for the table data, but your new item is not in that source, so it would need at least a footnote to explain how it's verifiable; but if you're not even sure which is the right answer, it's hard to see how you will be able to cite something reliable. Dicklyon (talk) 13:00, 17 December 2008 (UTC)

The argument is reasonable, if somewhat oversimplified. A note explaining it would definitely be needed, though it's tough to say whether such a note would suffice absent support from a verifiable source. Using various methods, I get values between about EV 33.1 and EV 33.5 for an altitude of 40°. Describing any of these methods in sufficient detail to support the value obtained would require a note comparable in complexity to a small article, so I don't think that's the right approach.

I'm not too concerned that there is a range of values; this consistent with the caveat stated below Table 2. But I agree with Dick that adding it without support from a verifiable source is marginal. In any event, the Sun doesn't have phases, so the “Full” line is inappropriate. JeffConrad (talk) 23:35, 17 December 2008 (UTC)

I forgot to mention that your “measured” value is definitely WP:OR and therefore inappropriate; using the derived value of EV 33 remains marginal. I'm not suggesting that your measured value is wrong, but simply that WP guidelines disallow putting it in the article. JeffConrad (talk) 01:36, 18 December 2008 (UTC)

Moon, New York, 6 March 1865 - Moon photo of Lewis Morris Rutherfurd, mentioned in the 1888 Britannica

Here's a source that talks about the sun relative to moon brightness as well as absolute numbers. Not connected to EV, but it would be a simple step... Dicklyon (talk) 06:01, 18 December 2008 (UTC)

The magnitude difference given there, 14.29, corresponds to almost exactly 19 stops (1 magnitude being a ratio 10^0.4 of 1.33 stops), which would give your EV 34 number if the subtended angles are equal (and since the sun size within the range of variation of moon size, that's close enough to ignore). Jeff, should we consider adding this odd one after the chart, with citation to this book and a brief explanation of the change of unit? Dicklyon (talk) 06:38, 18 December 2008 (UTC)

By the way, this 19 stops, or factor of just over 500,000, is also in pretty near agreement (within 1/4 stop) with ratios around 600,000 reported in the 1888 Britannica. Dicklyon (talk) 06:45, 18 December 2008 (UTC)

And one more thing: if EV 15 is the right exposure setting for ISO 100, typical scene in daylight, the 19 stops less light might suggest using EV -4 in moonlight; yet the ANSI exposure guide suggests -3 to -2, which would seem to be a stop or two underexposed. I believe the explanation of the discrepancy is that in daylight you're generally trying to capture some shadow detail, and white highlights, while for a moonlit scene, unless you want it to look like a sunlit scene, you want black shadows and excellent highlight detail closer to the midtones. They seem to have worked that into the table, which is why Iain's 15-(-3) didn't lead to the number that his measurments and the sources agree on. So these numbers shouldn't be interpreted as implying a definite illuminance level, but rather just what they are, suggested exposures. When you take pictures of the moon, like the Rutherfurd one that the Britannica talks about (bottom of col. 1 p.802), an exposure guide is what's relevant (Iain, I have the original print of this one if you'd like to see it, the frontispiece of the book by Vogel). For the sun, it might be best to leave the quantification in some other form; unless you want to take pictures of it. Dicklyon (talk) 07:08, 18 December 2008 (UTC)

I think the currently accepted ratio (for a full moon) is about 402,000, or 18.6 steps; it's fairly simple to derive this from consideration consideration of several parameters for which the values are generally agreed upon. Even so, some massaging is needed. The Moon's reflectance is about 15% at opposition, but we never see it at opposition except during a lunar eclipse. The reflectance falls off rapidly as the Moon moves away from opposition, and a more typical value is about 12%. I think it's easy enough to find a source for the latter value, but in most cases the sources are telling only part of the story.

The luminance ratio is the same as the illuminance ratio, but using the scene EVs from the table involves a bit more than may be apparent. The EVs for scenes (based on the illuminances from the Sun and Moon) include the contributions of direct illuminance from the bodies and indirect illuminance from the sky. It seems reasonable enough to me that both the direct and indirect contributions scale linearly with the intensity of the light sources, but I don't think I've seen a source that explicitly says this, and I'm not sure what seems obvious to me would meet WP:Verifiability.

We could avoid some of this problem by using the absolute values. Even so, there are a few issues. The value given in the source in the link is for the Sun at zenith, which is never observed outside the tropics. Because the airmass at zenith is 1, the reduction from 1.85×10⁹ cd/m² to 1.4×10⁹ cd/m² evidently assumes an extinction coefficient of about 0.28, and we'd need this value to arrive at an EV of 33.2 (which I assume we'd round to EV 33) for a solar altitude of 40°. It's easy enough to show how this is obtained, but it's a matter of what liberties we may take and how much we need to explain them. It would help if there were a more expansive article on atmospheric extinction, but at present, such an article doesn't exist.

One other problem we might face: it's hard to find two sources with the same values for some parameters. I got a value of 1.6×10⁹ cd/m² for the Sun's zenith luminance from the IES Handbook (though I seem to have lost the copy of the relevant page), and values for the extinction coefficient range from the “standard” value of 0.21 to almost 0.40 for some urban areas. Perhaps, though, we simply need to accept that, as with so many things photographic, the EV for the Sun isn't something that's meaningfully taken to six significant figures.

Most values in the ANSI exposure guides are rounded to the nearest whole EV (actually, they're given as L_v and I_v, respectively, in the 1973 and 1986 guides). Conversion to EV requires the S_v for ISO 100 speed. Assuming that, at the time, ISO 100 was actually $128/^{3}{\sqrt {2}}$, this works out to 4.92, so there's some additional rounding in Table 2 of the article.

If we do add the value, I think it should be in the table, else no one is likely to find it. The real issue is what we would need to support whatever value we might add. JeffConrad (talk) 08:37, 18 December 2008 (UTC)

Expanding briefly on Dick's final comment: in a few simple circumstances (e.g., a broad light source or a single scene element with uniform luminance), it's possible to directly relate EV to scene luminance, but for the most part, Table 2 in the article is an exposure guide. Most of the values are taken from the 1986 ANSI exposure guide, which give incident-light values. These obviously make little sense when applied to a light source such as the Moon, or to a dark scene (such as a night street scene) with a few small light sources; the tabulated values are really meant to suggest camera settings that will result in a satisfactory picture. JeffConrad (talk) 09:12, 18 December 2008 (UTC)

Absent some additional ideas or a published value for a reasonable solar altitude, I think we should remove the EV for the Sun from the table. As I indicated previously, there probably are many ways in which a reasonable value could be provided and supported, though doing so would require getting into astronomical photometry in some detail, and that isn't the purpose of this article.

I have a problem basing relative luminance on Rougier's 1933 value of −12.46 for the Moon's visual magnitude, because that value differs significantly from more recent values (e.g., Allen (1973) gives –12.73 and Lumme and Irvine (1982) give −13.04). The latter value is based on zero phase, which, as mentioned, can't be observed from Earth; under more typical conditions, the value would decrease to about −12.8. The difference then would be about 18.5 EV; if combined with the unrounded value of EV 14.9 for the full Moon's luminance, the rounded value for the Sun's luminance would be EV 33, so I don't think giving a value of EV 34 would be appropriate.

It is fairly simple to show that a celestial body's luminance is related to its direct illuminance by

$L={\frac {E}{\pi \sin ^{2}s}}\,,$

where L is the luminance, E is the direct illuminance from the body, and s is the body's angular semidiameter. It then could be shown that the luminance ratio for two bodies is the same as the illuminance ratio if the semidiameters are same, provided that it could be shown that the effects of indirect illuminance from the sky could be ignored.

It would be better to start from absolute values, either published or derived from basic parameters. The extraterrestrial solar illuminance is

$E_{\mathrm {S} }=K_{\mathrm {S} }E_{\mathrm {S} }^{\prime }\,,$

where K_S is the luminous efficacy of sunlight and E′_S is the extraterrestrial solar irradiance (i.e., the solar constant). The extraterrestrial luminance is then

$L_{\mathrm {S} }={\frac {K_{\mathrm {S} }E_{\mathrm {S} }^{\prime }}{\pi \sin ^{2}s_{\mathrm {S} }}}\,.$

Except for an astronaut, there is little value in an extraterrestrial luminance. There's also not much value in a terrestrial value at zenith; the article is about photographic exposure, not theoretical astronomical photometry. It would seem reasonable to give a value for same altitude, 40°, as used for the Moon. Consequently, we would need to account for atmospheric extinction, even with a published value (unless it were given for 40°). The effect of atmospheric extinction is

$L=L_{\mathrm {xt} }\,\mathrm {e} ^{-cX}\,,$

where L_xt is the extraterrestrial luminance, c is the extinction coefficient, and X is the airmass. The complete expression would be

$L_{\mathrm {S} }={\frac {K_{\mathrm {S} }E_{\mathrm {S} }^{\prime }}{\pi \sin ^{2}s_{\mathrm {S} }}}\,\mathrm {e} ^{-cX}\,;$

we'd need values for K_S, E′_S, s_S, and c, and would need to calculate X (though at an altitude of 40°, almost any method would suffice). Using values of 94.2 lm/W, 1366 W/m², 0.26656°, and 0.21, respectively, and using the secant formula for X, the result is

$L_{\mathrm {S} }={\frac {94.2\times 1366}{\pi \sin ^{2}0.26656^{\circ }}}\,\mathrm {e} ^{-0.21\times 1.556}=1365\;\mathrm {Mcd/m^{2}} \,,$

equivalent to EV 33.4. The Budding and Demircan extraterrestrial luminance of 1850 Mcd/m² would give EV 33.3.

But we'd need to source or derive everything used above, and a derivation would seem to be pushing it with respect to WP:OR. Even if a derivation were acceptable, it would make for the mother of all notes. And it's way off the topic of the article.

Again, short of finding a good published value for the Sun's luminance, I think it would be difficult to justify adding it to Table 2. If no one objects, I'll remove this value. JeffConrad (talk) 05:52, 21 December 2008 (UTC)

I agree. It's an interesting and useful estimate, but not suitable for wikipedia. Dicklyon (talk) 07:31, 21 December 2008 (UTC)

A reread of Kodak Technical Publication P-150, Astrophotography Basics, Solar Eclipse (the complete publication is available in PDF) gives the formula (p. 11 of the PDF)

$f^{2}=St\,10^{7-D}\,;$

where f is the lens f-number, S is the ISO speed, t is the exposure time in seconds, and D is the filter density. With some rearrangement, ISO 100 speed, and a D of zero (i.e., no filter), this gives

${\frac {f^{2}}{t}}=2^{\mathrm {EV} }=10^{9}\,,$

which works out to EV 32.9, and supports a value of EV 33 for the table. Or does it? As we've discussed, Table 2 is primarily an exposure guide, and actually attempting an exposure without a filter would probably fry the sensor as well as the photographer's eye. What P-150 really suggests is using a 5.0 density filter and EV 16.3. Again, this article is about photographic exposure rather than the luminances of solar system objects, so unless we decide to expand coverage to astrophotography, including filter recommendations, I think it may be a good idea to simply leave the Sun's EV (and the attendant caveats) for another topic. JeffConrad (talk) 04:02, 23 December 2008 (UTC)

If anyone is reading this, PLEASE DO NOT LOOK AT THE SUN THROUGH YOUR VIEWFINDER.

That said, taking a shot with the sun in frame does not fry the sensor. CMOS sensors can point at the sun all day long, for months, without obvious ill effect (I haven't tested this down to f/1.4, though, and I haven't tested it on DSLRs, which have fancy autofocus sensors and specific warnings not to point them into the sun). CCD sensors produce the previously mentioned large blooming artifacts.

I'll note that both of you have managed to come up with better support for a sun EV value than I did. Is there no way to put this useful information into Wikipedia somehow? Can I make a "EV of the Sun" article, and include your discussion here, and then reference that from this article?

Iain McClatchie (talk) 21:12, 3 January 2009 (UTC)

Iain (what a funny name!), you need to get it published in a realiable source first; then we can write an article on it, sourced to that. Dicklyon (talk) 21:35, 3 January 2009 (UTC)

Iain, it's fairly easy to support a value for the luminance of the Sun; I could probably cite at least half a dozen credible sources (which nonetheless vary slightly, though differences in EV would be minor). Absent a source that specifically gives an EV for stated conditions (e.g., altitude of 40°), the conditions used to derive a value would need to be stated, as I've said. As an example, Allen's Astrophysical Quantities (3rd ed, 1973) gives the Sun's luminous intensity as 2.84×10²⁷ cd; with a solar radius of 6.955×10⁸ m, it's easy to get an extraterrestrial luminance of 1869 Mcd/m². A terrestrial value would require assumptions about air mass and extinction coefficient that would need to be stated and perhaps justified. The ANSI exposure guide doesn't state how the values for the Moon were obtained, but a read of the guide suggests they were determined from measurement or practical experience or combination thereof rather than from astrophysical considerations. Consequently, a direct comparison might be difficult, though I may be splittting hairs here.

I come back to the basic question: is the real objective to give a recommendation for photographic exposure or to give a value for the Sun's luminance? If the former, it should be related to conditions under which someone might actually take a picture, i.e., using a filter, as I mentioned. If the latter, such as for designing a sensor, wouldn't the photometric (or even radiometric) value be more useful? I agree that the information is useful, but think it might better fit in an article on physical and photometric properties of the Sun. I have generally been less concerned about using EV as a measure of luminance than have some authors, but I'm not sure the concept makes much sense outside of a photographic context (or perhaps using an exposure meter as a poor man's luminance photometer). JeffConrad (talk) 02:22, 4 January 2009 (UTC)

Or one could make an article on solar photography from these book sources. Dicklyon (talk) 02:37, 4 January 2009 (UTC)

Or one could add it to the Astrophotography article, which currently says nothing about the Sun. JeffConrad (talk) 07:14, 4 January 2009 (UTC)