ETTR in Practice

[chanh_nguyen1]

<p align="left">Instead of hijacking the entertaining "Debunking ETTR" thread, I'd like to start a discussion not on the merits of ETTR but how to do it. First, the capture of the image: Spot meter the brightest part of the sence and push exposure up until the highlight is just about to be clipped. Am I correct in thinking that this simple method will capture the optimal data to produce (after the appropriate post processing) the image with the least noise for the given camera at the given ISO? (Of course, if my scence contrast is more than my camera DR, and if for subjective reasons I want to bring out low light details, I can elect to clip the brightest part; but that's a different issue and the exposure in that case will be even more TTR)</p>

<p align="left"> </p>

<p align="left">Next is the PP of the captured data to produce the image. Let's assume that the contrast is less than camera DR and ETTR is +2EV above the aesthetically desirable exposure. I would think that in order to get an image with the same tones and colors as the one exposed 2 stops below, all I have to do is move the Exposure Slider in ACR or LR down by 2 stops. But this is not the case for my D200. I see color shifts and the histograms are different. From what I read about Exposure Slider, it seems that its adjustment should have the same effects as changing the actual camera exposure. What am I missing here?</p>

<p align="left"> </p>

<p align="left">Making other adjustments I can get the images to be close enough (for me). But that's is a trial and error process I don't like. I am looking for an understanding of the PP steps to create an image captured by ETTR without color and/or tone shifts. Thanks in advance for your input.</p>

[ben_goren]

<p>Chanh Nguyen wrote:</p>

 

<blockquote><p>First, the capture of the image: Spot meter the brightest part of the sence and push

exposure up until the highlight is just about to be clipped. Am I correct in thinking that this simple

method will capture the optimal data to produce (after the appropriate post processing) the image with

the least noise for the given camera at the given ISO?</p></blockquote>

 

<p>That depends entirely on the scene. No spotmeter in the world, and no in-camera histogram either,

will possibly tell you when the miniature rainbows in the droplets in the sunlit spider’s web shot

against the shadowed forest are at the edge of clipping.</p>

 

<p>On the other hand, an ambient / gray card exposure is all but guaranteed to give you a workable

exposure, especially if (in this particular example) you go so far as to err on the side of underexposure.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[roger_smith4]

<p>" Spot meter the brightest part of the sence and push exposure up until the highlight is just about to be clipped. Am I correct in thinking that this simple method will capture the optimal data to produce (after the appropriate post processing) the image with the least noise for the given camera at the given ISO"</p>

<p>Basically. This is a good reason to bracket. I usually shoot hand reflectance +1 and bracket in 1/2 stops from there.</p>

[digitaldog]

<p>You have to test both the meter, metering technique (understanding what a reflective meter pointed at an object is telling you), the sensor and the raw converter. <br>

An incident meter would make this easier, unlike a reflective meter, it tells you the light falling on the subject instead of the reflective properties (assuming something in the neighborhood of 13-18% gray) just to begin to decide where to start the bracketing for analysis of the ETTR results in the raw converter. In my article, I setup a lighting situation with a very good, very white and neutral subject among others (see: http://www.babelcolor.com/main_level/White_Target.htm). Using that as the white I wish to retain, not clip, and using the incident meter (this was shot with strobe where each corner of the subject was even within 1/10 of a stop), I bracketed from the meters recommendation (again, based on film with a non linear behavior), in the finest increments the camera allows as high as +3 stops. After bringing them into the raw converter of choice (Lightroom), and looking at each bracket while examining the Exposure slider, I found I could shoot at +1.5 stops over its recommendation and still keep from clipping the babelcolor tile. This is the ideal and max ETTR compensation for this chip as +2 stops, the tile was clipped and no amount of rendering options would change this (sensor overload). </p>

<p>Once you know the absolute max you can compensate away (increase in exposure over base recommended exposure), you can, based on the shooting situation (scene, time you have to meter, etc) how much you can safely apply ETTR. For example, in my recent trip to the Galapagos, shooting nature and animals quickly and with differing lighting situations, I set my 5DMII to +½ stop knowing I wasn’t going to (and didn’t) clip anything I wanted to retain based on my understanding of how the in-camera meter works when you point it at an object. In a studio situation, I’d increase that far more, probably bracket since there is zero harm in doing so. </p>

<p>Key is understanding that if you follow the camera meter situation to produce a nice looking JPEG, you are under exposing the raw data. Understanding the limits of the sensor, you may not produce the max benefits of ETTR, but you’re not practicing the noise inducing ETTL (true under exposure). </p>

<p>Again, despite the nearly 200 posts elsewhere, the photographers who <strong>understand basic photography 101</strong> is that you must understand the media you are shooting onto (transparency vs. Neg, JPEG vs. Raw), and how a reflective meter is easily fooled based on its simplistic assumptions about gray** (and that its assumption that correct exposure is equal for a media that has a tone curve and one that doesn’t). One must also understand the <strong>relationship between exposure and development!</strong> Expose for the shadows, develop for the highlights works based on the film media and its classic H&D curve. Ignoring developing and exposure, something those who treat JEPG and raw the same don’t understand, is folly. Those that find ETTR too difficult are really saying, they find exposing too difficult. Give them a few rolls of E6 film and lets see what they come back with (no clip testing allowed!). Pro photographers purchased their film of choice in bulk (bricks), tested the ISO suggested and the lab they would use to process the film to properly expose and develop that media with no surprises. That’s what you need to do with ETTR as well.<br>

In the end, there is no difference in terms of the craft of photography be it shot on film or digital, just the media and development can’t be treated the same way because the two are vastly different. ETTR is simply correct exposure and development for a different media than what we used to implement with film. </p>

<p>** just point a reflective meter at a black cat on coal, then a pure white dog on snow and look at what you end up with. Gray dogs and cats, not white or black. An old photo trick to “convert” your in-camera reflective meter to behave more like an incident meter was to meter off your hand and open up a stop. Again, a major key is understanding the meter before you even consider the differences between the raw, JPEG, film media.</p>

[ben_goren]

<p>Andrew Rodney wrote:</p>

 

<blockquote><p>Using [http://www.babelcolor.com/main_level/White_Target.htm] as the white I wish to

retain, not clip [….]</p></blockquote>

 

<p>The BabelColor target is excellent. For all practical purposes, it is the brightest, whitest, non-emissive, non-specular, non-fluorescent object one is likely to come across.</p>

 

<p>Notice those caveats?</p>

 

<p>Hang that BabelColor target from a Christmas tree, use ETTR to ensure it won’t clip, and all

the lights on the tree will clip and come out white. That shouldn’t be surprising.</p>

 

<p>But the tinsel will clip, too,

much more than it needs to. Worse, the clipping of the tinsel will be invisible on the in-camera

histogram because it’s such a miniscule fraction of the total pixels. You might not be able to spot the clipping of the lights, either, depending on how many and how big you’ve filled the frame with them.</p>

 

<p>But it doesn’t end there. The angel your kid made from plain white office paper will have the

blue channel blown out from the OBA in the paper. Good luck spotting that on the histogram. The glitter will be colorless, and the day-glo ink

won’t glo nearly as much as it should. And, again, none of that will show up until you get back to your computer and load the picture into your RAW converter.</p>

 

<p>Then there’s the foil on the bows, the polished brass on the mantlepiece, the glossy black paint on the electric train engine…</p>

 

<p>And your chances of spotting any of that clipping on the camera itself are close to nil. But, hey! the shadows under the tree will have a stop less noise! Woo-hoo!</p>

 

<p>None of that would be (much of) a problem with the Olde Skool method of metering off a gray card.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[digitaldog]

<blockquote>

<p>

<p>Hang that BabelColor target from a Christmas tree, use ETTR to ensure it won’t clip, and all the lights on the tree will clip and come out white. That shouldn’t be surprising.<br>

But the tinsel will clip, too, much more than it needs to. </p>

</p>

</blockquote>

<p><strong>Nothing clips</strong> that a photographer who can control the process doesn’t <strong>allow to clip</strong>. Its as simple as that. </p>

<blockquote>

<p>None of that would be (much of) a problem with the Olde Skool method of metering off a gray card.</p>

</blockquote>

<p>I don’t know what makes you think this is old school or correct. It isn’t any more than allowing the meter to produce gray dogs and cats on snow and coal. Do you intent to make this thread go on forever based in misunderstanding or simple photo technique? </p>

[ben_goren]

<p>Andrew, are you claiming that, if one is confronted with a scene of a white cat on snow, and one sets

the meter on the neutral position when the meter is pointed at an 18% gray card in the same light as the

cat and snow, that the cat and snow will be rendered as gray?</p>

 

<p>If not, why are you demonstrating an intent to make this thread go on forever based on an

<strong>intentional</strong> misunderstanding?</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[digitaldog]

<blockquote>

<p>Andrew, are you claiming that, if one is confronted with a scene of a white cat on snow, and one sets the meter on the neutral position when the meter is pointed at an 18% gray card in the same light as the cat and snow, that the cat and snow will be rendered as gray?</p>

</blockquote>

<p>I know I’ll slap myself for trying to discuss this...<br /> No, I never said that. I’m talking about reflective metering on the subject. BUT, your technique is still not sound for raw capture because (after all the posts elsewhere, its ironic I have to say this again), metering off the gray card, which in essence forces the in camera reflective meter to now “act” like an incident meter, as I pointed out above, is incorrect for raw capture OTHERWISE, in the piece I wrote, the +1.5 stop image OVER this recommendation would be incorrect and its not. <strong>ETTR is correct exposure</strong>. It appears to the untrained, those that think the meter is infallible and not to be questioned, those that think its always based on film, raw, JPEG, to be correct. That’s a bad assumption! In some cases it is correct, in some cases its not. The incident meter in my article, and your idea of using a gray card for raw capture results in under exposure for raw data. It would be fine for JPEG and, depending on the stock and the development, might be fine for film.</p>

<p>Again, until you clearly express your goals in exposing, that means what I wrote above, defining the metering, metering technique, mode of capture (film, raw, JPEG) AND development, all the questions are so ineffectively presented, a correct answer isn’t possible.</p>

<p>Ask the question such as “<em>are you claiming that, if one is confronted with a scene of a white cat on snow, and one sets the meter on the neutral position when the meter is pointed at an 18% gray card in the same light as the cat and snow, shooting (JPEG), with a pretty much standard develop preset in Lightroom....</em>” then the question can be answered. Get it?</p>

[blumesan]

<p>Andrew & Ben,<br>

I'm disappointed to see you two rehashing the same argument. For the record, I am firmly in Andrew's camp. I have "calibrated" my camera and am pleased with the results. It seems to me that the OP has a pretty good understanding of ETTR and a reasonable method of practicing it. I was hopeing for an answer to his basic question:</p>

 

<blockquote>

<p>I am looking for an understanding of the PP steps to create an image captured by ETTR without color and/or tone shifts. Thanks in advance for your input.</p>

</blockquote>

 

[digitaldog]

<blockquote>

<p>

 

 

<blockquote>

<p>I am looking for an understanding of the PP steps to create an image captured by ETTR without color and/or tone shifts. Thanks in advance for your input.</p>

</blockquote>

 

 

</p>

</blockquote>

I thought I got you going into the right direction above by suggesting you setup a series of tests to gauge what your meter is telling you, what your sensor provides and what develop settings might be necessary to produce a desired rendering. Exposure is only one such setting, but the one that does the biggest job of "normalizing" the rendering due to the ETTR. You will need to alter other settings to session to taste. Then just make a preset to get you 85-90% of what you'll want for all such subsequent ETTR exposures. The "color shift" is a red herring, there's no more or less a shift in terms of what you may or may not find an appropriate rendering than any other preset rendering. YMMV.

[ben_goren]

<p>[sigh]</p>

 

<p>Andrew, I’m sorry, but you’re so far worng I simply don’t know where to

start.</p>

 

<p>Chanh asked for practical advice in implementing ETTR. You outlined a cockamamie procedure which,

simplified, says to determine the amount one can overexpose an image by determining at what point a

target specifically designed to be non-specular, non-emmissive, and non-fluorescent blows out.</p>

 

<p>If you had bothered to even glance at the documentation that ships with the Watch Your White

target, you would have seen that they suggest using it to guage the amount of OBAs in paper by

seeing which is brighter, the target or the paper. That right there should have clued you in to the fact that there

are plenty of things in typical photographic scenes which are, in fact, brighter than the Watch Your

White target. I gave a whole list of them, a list which you’ve studiously ignored. And I’m

sure you’ll continue to ignore the list and blather some more about how I should shoot JPEGs

with a P&S because I’m so stupid.</p>

 

<p>So, here’s some more of your ignorance that you continue to parade about.</p>

 

<p>If you had a clue, you would have noticed that your gray card, when exposed with any variety of

incident metering, opens with default RGB values that map to within less than half a stop of an L* value

of 50. Middle gray, in other words. But your Watch Your White target, which has a D50 reflective L*

value of 99.5 when measured with a spectrophotometer, is rendered in the scene with an L* value of

about 90.</p>

 

<p>“Aha!,” you think, “there’s some wasted dynamic range there, since

clearly nothing can be whiter than this target!” And yet you somehow manage to be so

completely oblivious to the fact that the world is filled with all sorts of things that are, in one way or

another, indeed brighter than the target.</p>

 

<p>It just so happens that those sharp guys who design cameras know what they’re doing.

They’ve given you a very reasonable amount of headroom to still be able to capture all those

things brighter than the Watch Your White target. In fact, if you use incident metering, RAW clipping will

happen at roughly the same point where we as humans start to have trouble discerning color ourselves

(assuming a scene lit with reasonably uniform lighting and ignoring the effects of adaptation). And, it just so

happens, the shadow areas will only start to get (insignificantly) noisy at levels where we have a hard time seeing

details anyway.</p>

 

<p>Or, as I’ve been advocating for the whole time, if your dynamic range fits within the

sensor’s capabilities, shoot a standard exposure and be done with it.</p>

 

<p>But wait, there’s more! Your method is completely oblivious to the effects of color

temperature. That “white” paper with lots of OBAs will certainly blow the blue channel in

open shade with your naïve idea of how to implement ETTR, and probably at least come close to clipping under studio strobes. But it might not clip in

incandescent light, and probably won’t clip in candlelight. Even the Watch Your White target will

clip at different amounts of overexposure depending on color temperature, though it should never clip with a correct incident exposure.</p>

 

<p>So, unless you’ve got a spotmeter that acts as a spectrophotometer with a field of view

measured in seconds, not degrees, combined with an insane set of charts that would make those needed for the

Zone System look trivial, your advice for how to implement ETTR is so far off, even in theory, that

about all I can do is shake my head in wonder.</p>

 

<p>Chanh, if you’re still worried about reducing shadow noise by another stop, my best advice would be to photograph a gray card at varying exposures until the clipping point, and then doing so under lots of different lighting conditions. You’ll then have an idea of how many stops above middle exposure you have until clipping, as well as how clipping shifts with different light. If you want to use the Watch Your White target, ignore Andrew’s advice and base your range on the difference between standard / middle gray and overexposure of the target itself (filling the frame with the target — use a macro lens), not his insane unmeasured offset of middle gray + target + overexposure, or whatever the hell he thought he was aiming for.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[digitaldog]

<p>Chanh & Mike, Ben’s going to once again, hijack this tread with more gobbledygook nonsense. The rant above is as easily dismissed by the peers as they dismissed the last 270+ post discussing the so called ETTR debunking. Contact me off list if you need more direction in setting up an effective ETTR workflow. I simply can’t continue to waste my time with him, he’s beyond hope.</p>

[chanh_nguyen1]

<p>Thank you all for your responses. However, the exposure part, which is actively discussed here, is not a major concern for me. I am comfortable with biasing the exposure to the right without clipping when I want to. I may not get the optimal exposure from the noise reduction standpoint, but I am OK with that. The part I am still groping in the dark by trial and error is Post Processing of the ETTR raw capture. Let me frame my problem more specifically.<br>

I have a low contrast scence that a "normal" exposure gives me just the right colors and tones after applying various adjustments in LR. I just want to lower the noise of the image. A +2EV ETTR in this particular case results in no clipping. But I couldn't reproduce the colors and tones simply by moving the Exposure Slider 2 stops before applying the same adjustments. It appears that LR exposure compensation is not equivalent to scaling the number of photons across the entire color and tonal range as is the case with changing exposure time. Is this true? If so, what does the Exposure Slider do? If not, why don't I get the same colors and tones? In other words, shouldn't 0 EV = +2 EV (camera) -2 EV (LR) when no data are clipped?</p>

[blumesan]

<p>Chanh,<br>

It seems to me that you have answered your own question:</p>

<blockquote>

<p>I have a low contrast scence (sic) that a "normal" exposure gives me just the right colors and tones <strong>after applying various adjustments in LR</strong>.</p>

</blockquote>

<p>The operative phrase is "after applying various adjustments in LR." It is an inconvenient but not surprising fact that a +2EV ETTR and a -2EV correction in LR does not produce exactly the same image that one would obtain with no EV change at either point. This will simply require you to make some changes to your customary adjustments in order to get "just the right colors and tones". </p>

[blumesan]

<blockquote>

<p>Hang that BabelColor target from a Christmas tree, use ETTR to ensure it won’t clip, and all the lights on the tree will clip and come out white. That shouldn’t be surprising.</p>

</blockquote>

<p>Ben, I doubt that surprises anyone with a rudimentary knowledge of exposure. No matter how you meter the scene (gray card, Babel, etc.) point sources of light are almost guaranteed to clip. Certainly that can be avoided at the cost of grossly underexposing the entire scene. Personally I prefer to allow the specular highlights to clip and use an optimum exposure for the important elements in the scene.</p>

[ben_goren]

<p>Chanh,</p>

 

<p>As you have discovered, current popular RAW tools are designed to simulate film in the darkroom,

and not to do simplistic bit-fiddling of the type ETTR evangelists naïvely think

“should” happen. Simply put, LR and ACR <em>will</em> cause various sorts of non-linear changes in both tone and color when applying ETTR principles, in much the same way that overexposing many types of color film will also cause non-linear tone and color changes. With the most neutral RAW

development settings combined with modest amounts of ETTR, the effects are minimized.</p>

 

<p>As I expressed numerous times in that other ETTR thread, I don’t think there exists a

workflow today that can correctly implement ETTR.</p>

 

<p>Your absolute best bet would be an ICC profile-based workflow starting with a linear RAW

development. But even this is fraught with problems. First, a standards-compliant ICC profile will clip

anything with a reflective L* value above 100, much as Andrew encourages you to do anyway. Argyll

CMS allows you to create a non-compliant profile that will scale values past 100 rather than clip them,

which helps. ICC input profiling is great for art reproduction, good for product photography, worth

considering for studio photography, and an exercise in frustration in all other kinds of

photography…for a variety of reasons it’s best not to get into now. (Of course, once the

image is in the computer, you’d have to be insane to not use an ICC-based workflow for the rest

of the process.) And, even for scenarios where it makes sense, it’s notoriously difficult to

implement in practice.</p>

 

<p>If an ICC input profile won’t work for you, your next best bet would be to again start with a

linear RAW conversion. Open it in Photoshop and apply (don’t convert) a synthetic ICC profile

with linear gamma. You’d probably also need the profile’s primary coordinates to match the XYZ

coordinates to eliminate color shifts…maybe. At this point, in theory, you should be able to use a curves adjustment with a straight

curve that simply clips the black and white points wherever you want them clipped. But now another

glaringly obvious problem arises: you’ve now got a file with the RGB values scaled…but

what colors do those RGB values represent? What color space? The answer is that there is no answer.

Just pick one and…well, you’re in a similar boat as you were with LR, but hopefully not

<em>quite</em> as obviously leaky. You will, though, no matter what, need to manually apply an appropriate gamma curve. Your best bet might be to manually create an ICC profile just for that image…and, if you think that sounds like an insane thing to do, you’re exactly right.</p>

 

<p>You could probably also get to that same point — an untagged RGB file with the RGB values

scaled to not clip yet no way of knowing what colors those RGB values represent — in a

somewhat more automated manner by using any standard graphics programming library, if

you’re so inclined. At that point, you really should be starting with the RAW data yourself and, essentially, program your own RAW converter. LIBRAW (and others) have already done all the heavy lifting, so it shouldn’t be as bad as you might think, assuming you’re a competent programmer.</p>

 

<p>And, by now, it should be somewhat obvious why ETTR is bullshit even in theory if you care about

accurate color rendition, as well as why there aren’t any tools to implement it.</p>

 

<p>Except, of course, there <em>is</em> one: in-camera ISO.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[ben_goren]

<p>Mike,</p>

 

<p>The point is that, as Andrew has rather crudely demonstrated, you’ve got at least a couple

stops of headroom over 100% reflective objects when you start with a proper exposure.</p>

 

<p>Will that prevent the sun or a laser pointer from clipping? No, of course not. But it <em>will</em>

prevent all but the hottest part of the Christmas tree lights from clipping, and give you what you almost

certainly were looking for anyway: intense color surrounding small white dots. Andrew’s recipe

for overexposure will give you large white blobls. There’s no need to underexpose, severely or

otherwise.</p>

 

<p>Cheers,</p>

 

<p>b&</p><div></div>

[digitaldog]

<blockquote>

<p>The operative phrase is "after applying various adjustments in LR." It is an inconvenient but not surprising fact that a +2EV ETTR and a -2EV correction in LR does not produce exactly the same image that one would obtain with no EV change at either point. This will simply require you to make some changes to your customary adjustments in order to get "just the right colors and tones".</p>

</blockquote>

<p>I’d agree with all points. In my article, the numbers were “close” to a point (up to +1.5) but didn’t correlate exactly and did, as pointed out there, require other rendering settings to produce a match to the “normal” exposure (assuming that’s the rendering goal, it might not be).</p>

 

<blockquote>

<p>Personally I prefer to allow the specular highlights to clip and use an optimum exposure for the important elements in the scene.</p>

</blockquote>

<p>Absolutely! And many images without a good specular highlight can often appear “muddy”. The same could be true for not clipping some black pixels. Hence, the controls provided for two decades in Photoshop and found in Adobe raw processors to see and control this. </p>

[ben_goren]

<p>…and there you have it. Andrew <em>knows</em> that ETTR produces an incorrect color

rendition <em>but doesn’t care.</em></p>

 

<p>He’d rather have an image with inaccurate color, so long as it has a stop less shadow noise.

He also seems to have no understanding of ICC profiles, or why RGB values are perfectly meaningless without

an associated color profile.</p>

 

<p>If you like the ETTR look, go for it, just the same way you might like the look of cross-processing

or film overexposure or whatever else. But don’t kid yourself into thinking that you’re

enhancing image quality, and certainly don’t kid yourself into thinking that ETTR is, in any way,

shape, or form, “correct.”</p>

 

<p>Treat it as a creative tool, but it is most emphatically <em>not</em> a best practice.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[Dave Luttmann]

<p>Ben,</p>

<p>Give it a rest already. Andrew is a recognized expert in this area. Others have pointed out the errors in your logic. And for some reason you continue to feel the need to justify what you're saying....and in so doing....making a further fool of yourself.</p>

<p>This whole ETTR and not blowing highlights concept is accepted for a reason....it works.</p>

<p>You're wrong. Move on please.</p>

[ben_goren]

<p>Dave,</p>

 

<p>Andrew may be a recognized expert, but he’s demonstrating shocking ignorance in basic

matters, along with quite the cavalier attitude to color in general. In the article he wrote himself, he

acknowledged color shifts but didn’t think they were worth worrying about.</p>

 

<p>Frankly, I don’t give a damn about who is and isn’t an authority. Experts are human

and prone to error. Just look at how many published articles in the peer-reviewed literature later get

ripped to shreds.</p>

 

<p>What I care about are results from repeatable processes. And Andrew’s processes are not

only based on bad theory (that, for example, the Watch Your White target is the brightest object in a

scene) but he himself has demonstrated their shortcomings (the inaccurate color he reported

finding).</p>

 

<p>Besides which, he never avoids an opportunity to sidestep my criticisms of his theory and

technique and turn it into a personal attack on me. That’s hardly the behavior of a so-called

expert worthy of respect.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[digitaldog]

<blockquote>

<p>Just look at how many published articles in the peer-reviewed literature later get ripped to shreds</p>

</blockquote>

<p>The Debunking ETTR is a prefect example! </p>

[ben_goren]

<p>Andrew, in case you’re not following the other thread any more, I asked a specific question

relating to the practice of ETTR. I’ll repeat it here for your benefit:</p>

 

<p>One of the stated goals of the DPWG charter is to “[identify] limitations of ICC color management

with respect to digital photography use cases, and [develop] recommendations to the digital

photography community to address these limitations.”</p>

 

<p>In your article, you identified a limitation of ICC color management with respect to a digital

photography use case when you observed that your implementation of ETTR produced color shifts.</p>

 

<p>What is your recommendation, as a member of the DPWG, to the digital photography community to

address this limitation?</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[roger_smith4]

<blockquote>

<p>He’d rather have an image with inaccurate color, so long as it has a stop less shadow noise. He also seems to have no understanding of ICC profiles, or why RGB values are perfectly meaningless without an associated color profile.</p>

 

</blockquote>

<p>Give it up, Ben, you've stopped making sense a while ago. It looks like you have an axe to grind against a particular technique because you don't like it. That's your choice but there's no need to force it on others.</p>

<p>Let's take a foggy day on a lake. You can easily give this greater exposure using ETTR to reduce shadow noise and clip no highlight data. Why would you do that? Say there is a figure in the foreground and you want to increase contrast to make him stand out more. Doing that can exacerbate noise. </p>

<p>In other higher dynamic range situations you may choose to shoot multiple exposures and blend together to expand dynamic range and cut noise. In both cases there may be other unintended consequences to image color and tone, but the great thing about shooting raw is that you can correct them. Raw files also don't have ICC color spaces and nothing is baked in until you render the raw and save it as another type of file, so I don't see your concern there.</p>

[joshroot]

<p>Howdy folks, ETTR called me and asked for a nap. It said that these threads were spiraling down into uselessness and personal attacks.You want to yell at each other? Do it via private email.</p>

<p>And so, that is what is going to happen. Give it a rest for a while or find yourselves booted from the forum with a "time out".</p>

<p>Go shoot photos, thread closed.</p>