Debunking ETTR

<p>As some of you are aware, I firmly believe that the common wisdom of ETTR is bad advice that, when practiced

as commonly described, is far too prone to overexposure and heartache.</p>

 

<p>Instead, I advocate starting with a traditional “proper” exposure, such as determined by metering

off a gray card. If the histogram shows no blocked shadows or clipped highlights (or blinkies in the preview), leave

well enough alone. Only if the dynamic range of the scene is such that shadows are blocking or the highlights are

clipping is it profitable to deviate from such an exposure. When making adjustments, be aware that shadows get

noisy and lose detail and that highlights suddenly transition to no recoverable detail whatsoever. Either make

whatever sacrifices you feel best fit the scene; alter the contrast using traditional techniques such as altering the

lighting; or use HDR.</p>

 

<p>The common wisdom behind ETTR is that, due to the linear nature of encoded data, the lower stops carry less

information. By using a brighter exposure and compensating in post-production, one can capture additional

information, leading to more shadow detail and less shadow noise. This effect is carried throughout the dynamic

range, of course, but is most pronounced in the shadows where the least amount of information is recorded.</p>

 

<p>I do not dispute the theory behind ETTR nor claim that the effect does not exist. What I <em>do</em> dispute

is the real-world usefulness of the technique as well as the degree of enhancement to be gained. Considering that

application of the technique requires one balance right on the cliff of overexposure and that additional post-processing is required, I consider the benefit to be far overwhelmed by the risk and therefore the technique not

worthy of recommendation.</p>

 

<p>Worse, even the theory can only possibly apply at base ISO, as it is exactly equivalent to lowering the ISO.

Above base ISO, it simply doesn’t even apply. The signal amplification will have a far more detrimental

impact on image quality than any linear digital quantization effects could possibly have. (Of course, pulling detail

from underexposed high-ISO shadows is properly known to be insane, but, if you have the dynamic range for ETTR

at high ISOs, you also have enough light to shoot at a lower ISO.)

 

<p>One should certainly be aware of the theory so one might apply it in suitable circumstances — such as,

for example, when photographing the proverbial black cat crossing the asphalt road at midnight. But, at other times,

the closest that the theory should get to practical exposure is the observation that shadows tend to get noisy and

lose detail — and equal weight should be given to the observation that clipped highlights are

unrecoverable.</p>

 

<p>Enough preamble. A theory is useless without tests. To that effect, I performed the following experiment.</p>

 

<p>I shot a ColorChecker with a 5DII and the 180 L in open shade. I defocused the picture so as to smooth out the

texture of the card itself and thereby make noise easier to distinguish; otherwise, the macro clearly resolves each

little bump and pit in the surface.</p>

 

<p>I made two exposures. (Many, really, but selected only two for this demonstration.) One was metered off the

middle gray patch of the ColorChecker (which, according to my i1 Pro specrophotometer is XYZ: 19.2 20.1 16.5, D50

Lab: 51.9 -0.95 0.28 on this particular ColorChecker); the other was overexposed by exactly one stop. I used f/8 and

ISO 100 for both shots; the metered exposure was 1/30 second and the overexposed shot 1/15 second.</p>

 

<p>With the following exceptions, I processed the two pictures identically in Adobe Camera Raw. White balance I

set (to 6850K, -4 tint) from the neutral square under the yellow square. All development sliders were set at the

factory defaults, except shadows were set to 0 (and the other exception noted below). I selected the flat tone curve.

All sharpening and noise reduction sliders were set all the way to the left. I used the “Camera Faithful”

picture style. Everything else was set to factory defaults. Images were opened as objects at their native resolution in

8-bit sRGB space. The full-frame pictures I Bicubic resampled to 700 pixels wide; the 100% crops are as-is. For both

I stretched one canvas before placing the two side-by-side. Finally, I used the “Save For Web” export

at JPEG quality 69.</p>

 

<p>Lab 52 0 0 = sRGB 124 124 124. The middle gray patch with the metered exposure read roughly 143, 143, 143.

Adjusting the slider exposure to -0.45 resulted in a proper exposure. This is consistent with my experience of how

this particular camera tends to overexpose by roughly a third of a stop, give or take, and is consistent with reports

I’ve heard about other 5DIIs. Therefore, the one frame I processed with -0.45 exposure and the other -1.45

exposure.</p>

 

<p>I will not claim that there is no difference. I expect most people who examine the following pictures with a critical

eye will probably be able to tell which is which. I <em>will</em> state that, as far as I’m concerned, the

difference is entirely insignificant. Remember, these were processed with no noise reduction whatsoever and tone

curves (<i>etc.</i>) that emphasize shadow noise much more than those normally used in artistic work. Not to mention that the subject itself is as textureless as a photographic scene ever gets. Even in this

worst-case scenario, were I to make 24″ × 36″ prints from these exact files, you’d still

have to stick your nose in the print to even have a hope of maybe being to tell which is which.</p>

 

<p>First (assuming I don’t screw up the image attachment, as I’m wont to do) is the full-frame shots.

I’ll post the 100% crops of the darkest neutral square immediately after.</p>

 

<p>Cheers,</p>

 

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

<p>Here’s the promised 100% crops of the darkest neutral patch.</p>

 

<p>Cheers,</p>

 

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

<p>It would be useful to fully understand the principles behind ETTR if you wish to debunk them...</p>

<p>ETTR has NOTHING TO DO with the amount of levels in the shadows...it has nothing to do with the base ISO (vs altered ISO) is has EVERYTHING to do with the fact that more photons equals less perceptible noise (which means a better signal to noise ratio).</p>

<p>Your test is fine...it proves what you want to prove., but it does NOT prove that ETTR is wrong (or dangerous).</p>

<p>ETTR has EVERYTHING to do with the screen contrast range and the dynamic range of your sensor...if your scene contrast range is equal to or greater than your sensor, ETTR doesn't apply...forget about it and deal with trying to figure out what is the most important tones in your image and expose for them.</p>

<p>However, if the contrast scene of within the dynamic range of your sensor (or lower–as is often/usually the case with soft overcast light) then you are a fool to ignore ETTR...</p>

<p>More photons=better noise</p>

<p>That's the basis of ETTR, has nothing to do with "shadow bits.</p>

<p>You should also test out the fact that HTTR can also actually benefit when "modest" increases of ISO are used...with today's cameras you can prolly go upwards of ISO 800 and STILL take advantage of ETTR if the scene contrast range is below the sensor...</p>

<p>I thought this article was convincing (as a non-expert):<br>

http://chromasoft.blogspot.com/2009/09/why-expose-to-right-is-just-plain-wrong.html</p>

<p>I'm not versed on the theory, but there seems to be a few inconsistencies in your logic:<br>

<em>"... leading to more shadow detail and less shadow noise. This effect is carried throughout the dynamic range, of course, but is most pronounced in the shadows where the least amount of information is recorded."</em><br>

You just shot a flat evenly lit object to test a theory that applies most when there is shadows.<br>

<em>"... and that additional post-processing is required" </em>and <em>"With the following exceptions, I processed the two pictures identically in Adobe Camera Raw. ... one frame I processed with -0.45 exposure and the other -1.45 exposure."</em><br>

So your additional processing was one slider set to -1. That doesn't seem like much aditional processing.<br>

Your other point rephrased is "don't miss a shot due to overly complex techniques". I agree on this point.</p>

<p>Jeff Schewe wrote:</p>

 

<blockquote><p>ETTR has NOTHING TO DO with the amount of levels in the shadows […] [it]

has EVERYTHING to do with the [scene] contrast range and the dynamic range of your

sensor[.]</p></blockquote>

 

<p>Except, of course, that those statements are perfectly contradictory. Shadow dynamic range is

defined as the limit where details are swamped by noise.</p>

 

<blockquote><p>[W]ith today's cameras you can prolly go upwards of ISO 800 and STILL take

advantage of ETTR if the scene contrast range is below the sensor.</p></blockquote>

 

<p>That statement is exactly equivalent to claiming that an exposure of, say, 1/100 s @ f/8 @ ISO

200 will give superior results to an exposure of 1/100 s @ f/8 @ ISO 100 when the ISO 200 shot is

underdeveloped in the RAW conversion by one stop. And, if that’s the case, then, clearly, 1/100

s @ f/8 @ ISO 400 underexposed by two stops will give better results still, and one would expect the

best results from a 5DII to be had at 1/100 s @ f/8 @ ISO 6400 underexposed by six stops. I

don’t really need to take and post actual pictures debunking that, do I?</p>

 

<p>Clearly, ETTR can only possibly be effective at reducing base ISO at the cost of an equivalent

number of stops of highlight dynamic range. And, hey, wha’d’ya know, the 5DII already

does that for you! Turn on ISO expansion and you can shoot at ISO 50. The manual even tells you that

you lose some dynamic range. And if ISO 50 <em>still</em> isn’t clean enough for you, well,

then, use ETTR and pull that all the way down to ISO 25! Never mind, of course, that you’ve

now lost two whole stops of highlight dynamic range, almost guaranteeing that you’ll be blowing

those highlights. And what’s the number one problem people have when using ETTR in the

field? Oh, yeah…blown highlights…but it can’t possibly be because ETTR reduces

dynamic range by a stop for every stop you overexpose, could it?</p>

 

<p>Joshua McKinney wrote:</p>

 

<blockquote><p>You just shot a flat evenly lit object to test a theory that applies most when there is

shadows.</p></blockquote>

 

<p>In this context, “shadows” refers to the darker tones of an image, regardless of the

nature of the illumination. It would equate best with zone II in the classic Adams ten-zone system.</p>

 

<p>Roger, thanks for the link. He did an even better job of explaining the points I’m trying to

make than I did.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<p>I had been practicing it for years, and then I found out today it has a name. Good heavens. And a fan club. :)</p>

<p>Ben and I took part in a conversation earlier today. I know exactly where he's coming from. Part of the context was how good are our tools, how much can we trust them, and how close can we edge to that precipice without falling in. I'm pretty sure I contributed to the madness, but I can blame it on the weather. It was wet, gray, and drizzly this morning, right on schedule for March in Chicago. After reading for a bit, I pointed the camera out the filthy window, and promptly got a flatter than pancakes, +2 EV shot that only a linear RAW conversion could revive. And so here we are. ;)</p>

<p>The thing of it is, I'm very sure all of us practice ETTR even if we don't call it that by name. You take a shot, peek at the histogram or preview, and shoot it again if you don't like what you see. It's no different than waving the old SpotMeter around, only faster and more certain. Back when our only choices were the 10D and the then new 300D, we took a pounding on usenet from the film-forever diehards. The term "chimping" came into being in those days. (Where are they now, BTW, now that the monkeys run the nuthouse?) Chances are very good that if you shoot RAW, and jab irritatedly at the camera back buttons after each shot, you're practicing ETTR. Ben's probably one of us, whether he owns up to it or not.</p>

 

<blockquote>

<p>Instead, I advocate starting with a traditional “proper” exposure, such as determined by metering off a gray card.</p>

</blockquote>

<p>Ben, how did you establish that this is the right exposure? It was the correct exposure with film, whose characteristic curve was S-shaped and right exposure of 18% gray was in the middle to avoid the nonlinear extremes. With raw, the curve is no longer a curve, it is a line. The right exposure is an exposure that gives highest quality data that can survive post-processing, no?<br>

I don't agree with your definition of right exposure, but I have to admit that I usually don't use ETTR either. The reason is that my converter cripples the look if I do ETTR and fix the exposure in PP using the Exposure and other sliders. The problem is however not ETTR, but support of correct exposure in converters and perhaps cameras or my technique.</p>

<p>Ben,</p>

<p>It would be useful if you stuck to the basic premise of ETTR...if you have a scene contrast range below the dynamic range of the sensor, would you agree that increasing the exposure (exposing to the right) produces an image that has a higher signal to noise ratio?</p>

<p>Ignore for the moment F stops and shutter speeds and ISO...</p>

<p>Do you understand that MORE PHOTONS equal BETTER SIGNAL TO NOISE? (lower perceptible noise-REGARDLESS of the ISO)</p>

<p>If you don't understand this, debating with you is a waste of time.</p>

<p>And in point of fact, it has been shown that (some cameras at least-recent ones in particular) the analog to digital conversion of the ISO amplification is actually better than post processing exposure compensation-particularly at somewhat lower ISO's (the Nikon D3/D#x and Canon 5D mII are examples where the A to D conversion of lower 3200/6400 "real ISO's" is superior to increasing the exposure compensation in software.</p>

<p>Boy, you sure have spent a lot of time thinking (and testing and writing) about this subject...sure would be useful if that time was productive (meaning you learned something from the effort).</p>

<blockquote>

<p>That statement is exactly equivalent to claiming that an exposure of, say, 1/100 s @ f/8 @ ISO 200 will give superior results to an exposure of 1/100 s @ f/8 @ ISO 100 when the ISO 200 shot is underdeveloped in the RAW conversion by one stop. And, if that’s the case, then, clearly, 1/100 s @ f/8 @ ISO 400 underexposed by two stops will give better results still, and one would expect the best results from a 5DII to be had at 1/100 s @ f/8 @ ISO 6400 underexposed by six stops. I don’t really need to take and post actual pictures debunking that, do I?<br /> Clearly, ETTR can only possibly be effective at reducing base ISO at the cost of an equivalent number of stops of highlight dynamic range.</p>

</blockquote>

<p>ETTR goes the other way: overexposure, rather than underexposure, to encode the luminance values in a more favorable, higher numeric range. This would be the other part of my undoing this morning. I'm pretty sure ETTR as practiced only intended for us to open the aperture a little or give a little extra exposure time, to bias the exposure toward the top of the dynamic range. I took it a step further. While keeping the same shutter speed and aperture, I increased ISO<em> </em> to force overexposure. How else would I get +2 EV overexposure handheld in the gloom?</p>

<p>This demonstrates two things. First, noise is related to the amount of light hitting the sensor, not strictly speaking the ISO speed. By keeping the shutter speed and aperture the same, the sensor saw the same amount of light, and so the noise stayed the same. (Normally, increasing ISO allows us to use less light, either a faster shutter or smaller aperture. Less light, more noise. That's how noise came to be associated with ISO. It's the other side of the same coin.)</p>

<p>By keeping the exposure -- aperture and shutter speed -- the same but increasing the ISO, the camera encoded the same A/D samples with higher numbers. Higher numbers are good, because more bits are available to represent the number. This is actually the easy part to understand, despite the long explanations. From Photography 101, each doubling of light is a stop of exposure. The sensor is assumed to be linear; doubling the light doubles the quantized value. From Computer Science 101, doubling a binary number requires one more bit (binary digit) to represent the number. Together, each stop of exposure above pitch black successively gives us one more bit to represent values in that range. Up to 12 bits are available to describe the values in the highest bin. More bits means more precision and finer gradation. So, pushing the image as farther up the exposure curve serves us by retaining more detail and gradation. And it does so with no increase in noise. You CAN have your lunch and eat it too, but you pay for it with slower shutter speed or larger aperture.</p>

<p>Of course, everything can be taken to extremes. Andrew related that 1.5 stops was about as much useful overexposure as he could use. +2 EV was too much for me to process well. In reality, if you have room to comfortably nudge the exposure higher by a stop or so, doing so can improve the quality of the image without any cost in noise. Just don't go changing the ISO as I did.</p>

<blockquote>

<p>This demonstrates two things. First, noise is related to the amount of light hitting the sensor, not strictly speaking the ISO speed.</p>

</blockquote>

<p>First and foremost you need to understand that yes, more photons hitting the photo sites (sensels) is always going to be better than less. The more photons of light, the better the signal to noise ration. That fact is not open for debate...</p>

<p>The next question regarding camera ISO amplification VS post processing exposure boots (setting the camera for the same F-stop and Shutter Speed and only changing the ISO) really comes down to a camera by camera variation...newer cameras are better at ISO amplification and generally, increasing the ISO to get a better depth of field or a shorter shutter speed (to make the overall image sharper) is a good thing to do...within limits.</p>

<p>Know that all of this goes right out the window when the question becomes, do you fustz around with exposure or do you GET THE SHOT?</p>

<p>You can ETTR, change the ISO and F-stop and Shutter speed till you are blue in the face, but f you are screwing around with camera settings while the light or the subject goes away, more the fool you...</p>

<p>Bottom line, ETTER is useful when the scene contrast range is less than your sensor and IF you have the time to think about things when you are shooting...if you need to increase the ISO to get the F-stop or Shutter Speed to "get the shot" do so...</p>

<p> </p>

<p>Mirek Elsner wrote:</p>

 

<blockquote><p>Ben, how did you establish that this is the right exposure?</p></blockquote>

 

<p>I didn’t; the camera manufacturers did when they designed the systems to mimic film as

much as possible. Never mind the actual-bits-on-the-disk, the fundamental nature of the medium;

that’s been modified so heavily elsewhere in the chain that it’s basically irrelevant. And

your concluding paragraph illustrates this perfectly:</p>

 

<blockquote><p>I have to admit that I usually don't use ETTR either. The reason is that my converter

cripples the look if I do ETTR and fix the exposure in PP using the Exposure and other sliders. The

problem is however not ETTR, but support of correct exposure in converters and perhaps cameras or

my technique.</p></blockquote>

 

<p>The problem <em>is</em> ETTR, because it’s trying to solve a problem —

excessive shadow noise and insufficient dynamic range at base ISO — that doesn’t exist

in today’s cameras, and hasn’t for years and years.</p>

 

<p>Jeff Schewe wrote:</p>

 

<blockquote><p>It would be useful if you stuck to the basic premise of ETTR...if you have a scene

contrast range below the dynamic range of the sensor, would you agree that increasing the exposure

(exposing to the right) produces an image that has a higher signal to noise ratio?</p></blockquote>

 

<p>I thought I had made that as clear as possible when I wrote:</p>

 

<blockquote><p>I do not dispute the theory behind ETTR nor claim that the effect does not exist. What

I do dispute is the real-world usefulness of the technique as well as the degree of enhancement to be

gained.</p></blockquote>

 

<p>Jeff continued:</p>

 

<blockquote><p>And in point of fact, it has been shown that (some cameras at least-recent ones in

particular) the analog to digital conversion of the ISO amplification is actually better than post

processing exposure compensation-particularly at somewhat lower ISO's (the Nikon D3/D#x and Canon

5D mII are examples where the A to D conversion of lower 3200/6400 "real ISO's" is superior to

increasing the exposure compensation in software.</p></blockquote>

 

<p>I can interpret that in one of two ways. First, that you’re stating it’s better to use

a higher ISO to get a proper exposure than to shoot at a lower ISO with the intention of adding

exposure compensation in post-processing. And, not only could I not agree more, I have to wonder

where you could get the impression that I or anybody else has ever advocated otherwise.</p>

 

<p>But the other interpretation would be that you’re claiming that superior results can be had by

keeping the shutter and aperture the same, boosting the ISO, and compensating in post-processing

with negative exposure compensation. If that’s your claim, I’ve got some prime Arizona

beachfront property to sell you….</p>

 

<p>Michael Young wrote:</p>

 

<blockquote><p>I'm pretty sure ETTR as practiced only intended for us to open the aperture a little or

give a little extra exposure time, to bias the exposure toward the top of the dynamic

range.</p></blockquote>

 

<p>That may well have been the original intentions lost in the mists of time. If we defang ETTR to the

point of, “When in doubt between two almost-equivalent exposures, go with the brighter

one,” it’s not such a bad thing.</p>

 

<p>It’s also, once again, in the real world, completely unnecessary. As I mentioned, my 5DII

tends to meter such that it typically overexposes by about a third of a stop. Most other modern DSLRs

do the same. There’s your slight bit of bias towards the top of the dynamic range, right there,

built into the onboard meter. What more could you want?</p>

 

<p>Michael continued:</p>

 

<blockquote><p>By keeping the exposure -- aperture and shutter speed -- the same but increasing the

ISO, the camera encoded the same A/D samples with higher numbers. Higher numbers are good,

because more bits are available to represent the number.</p></blockquote>

 

<p>In theory at the dawn of the digital age, yes, sure. Today? Not a chance.</p>

 

<p>The 5DII records that data using 14 bits, and all those cameras that don’t also encode at 14 bits encode

at 12 bits. Even 12 bits is sufficient that quantization effects are swamped by analogue noise even in

theory. With a 14 bit camera…well, if you can figure out a way to take a picture and do some sort of

post-processing that reveals posterization due to insufficient bit depth, I’d be quite curious to

see how you did it.</p>

 

<p>Indeed, I just tried it in ACR: I opened a RAW file of a night scene shot at ISO 200, and set all the

sliders to their brightest positions. Even at 400% pixels I saw no evidence of posterization.</p>

 

<p>One final observation. ETTR has been “common wisdom” since the dawn of the digital age. Do y’all <em>really</em> think that the camera manufacturers are unaware of it? No, of course not. So, being aware of it…what do you think they’d do? I would sure hope that they’d analyze it, determine what merit, if any, it has, and see if they could come up with a way to incorporate the benefits into the system such that the end user doesn’t have to bother with it.</p>

 

<p><strong>And that’s exactly what they did.</strong></p>

 

<p>ETTR can only even theoretically reduce noise at the expense of highlight headroom. Modern cameras have so little noise at base ISO that the fundamental problem itself has been practically eliminated, but the designers didn’t stop there. At the simplest level, they biased their meters for a slight bit of overexposure rather than underexposure. They increased the bit depth to the point that quantization errors don’t even theoretically exist. They gave us expanded ISO, where the lower-than-base ISO does, in camera with superior analog signal processing, exactly what ETTR is intended to achieve. They also gave us highlight tone priority that does the exact opposite of ETTR — you can capture increased highlight detail (at the expense of the shadows), something that no ETTR-equivalent could even theoretically do.</p>

 

<p>Really, folks. What more could you possibly want?</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<p>Ben, I think ETTR is meant to be practiced without changing ISO. In other words, you should always shoot at the lowest ISO you can without clipping the highlights, since the higher ISO you choose, the more amplification is needed and you amplify the noise just as much as the signal. Given a fixed, low ISO for your needs, the more you shoot to the right (overexpose), again without clipping the highlights, the better S/N ratio you have, and when you bring the exposure down in processing you end up with less noise.</p>

<p>Shash,</p>

 

<p>That’s exactly correct. And yet, as you can see from the example pictures I started this thread

with, the difference is imperceptible.</p>

 

<p>Curiously enough, nobody has yet tried to identify which of those two is the ETTR version and which

is the version with the standard exposure….</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<p>I meant to add..</p>

<p>Now, whether you have bedunked this or not, depends on what the histograms looked like for the before and after shots. If the histograms were wide and you you moved the right edge a bit more to the right, yes, I would expect the effect to the small. On the other hand, if you had a narrow histogram and you could relatively move the right edge of the histogram a larger distance, the noise reduction should be better. It would all depend on how far to the left the left-most edge of the histogram was to begin with. Can you post the before & after histograms? In any case, I wouldn't expect ETTR to make night and day difference, and less than the effects of going to lower ISO (up to base ISO).</p>

<p>Shash,</p>

 

<p>I’m afraid I must be missing some point you’re trying to make. What difference does

the overall histogram matter? It’s been universally agreed upon that ETTR is only for those parts

of the image where one wishes to preserve — indeed, maximize — details. One could

have completely blown overcast skies and a completely blocked mine entrance and still practice ETTR

for the wildflowers between. Besides, the ColorChecker should have the best-known histogram of any

object in the history of photography.</p>

 

<p>If you really want, I’ll post the histograms later today. Heck, I could even put the original

RAW files somewhere for you to play with. But, if that’s what you want…wouldn’t

you be even better served by doing your own experiments?</p>

 

<p>And, in any event, the two sets of images above are identical except that the one was overexposed

by a stop and compensated for in post-production with the exposure slider in ACR. That’s the

exact procedure ETTR calls for. And, if you can’t tell a difference in that bend-over-backwards

scenario…well….</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<p>Left is noticeably noisier than the right.</p>

 

<p>Like Michael says, left is noticeably noisier than the right. In other words, the whole point of moving to the right is to move the left-most parts of the histogram as far right as possible (again, without clipping the right-most part). In theory, every "bin" that moves to the right gets less noise, but the further right you start with, the effects become less and less noticeable. The two histograms would show us how far left the left-most histogram was when you started and how far right it moved when you finished.</p>

<p>No, I am not interested in my own experiments, I am happy shooting. I was just responding to your post, and trying to understand what it all meant, and realized that we didn't have an essential piece of the information (histograms) for the understanding.</p>

<p>As far as the colorchecker goes, I am not sure what the brightness values of the colors are. If you take away the color information, I have a feeling they will all bunch up to a very narrow range of brightness values.</p>

<p>Michael,</p>

 

<p>You are, of course, correct that the left is noisier.</p>

 

<p>Now, to practicalities.</p>

 

<p>This was shot with all noise reduction turned off. Do you think you would still be able to tell the

difference with even a minimal amount of noise reduction?</p>

 

<p>This was shot with a perfectly flat tone curve. Do you think you would still be able to tell the

difference with the default ACR settings (black to 2, medium contrast curve)?</p>

 

<p>If I were to make 12″ × 18″ prints of these images exactly as they are, do you

think you would be able to see any noise in either?</p>

 

<p>And if I were to make 24″ × 36″ prints and put them in two different rooms, do

you think you’d be able to tell which is which?</p>

 

<p>According to UPS, my TS-E 24 II should arrive in a few hours. Assuming it does, I have a couple

test shots in mind that should serve double duty: first, to play with the new lens and make sure

there’s nothing worng with it; second, I have a couple more ETTR-debunking shots in mind.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<blockquote>

<p>more photons hitting the photo sites (sensels) is always going to be better than less. The more photons of light, the better the signal to noise ration.</p>

</blockquote>

<blockquote>

<p>Bottom line, ETTER is useful when <strong>the scene contrast range is less than your sensor</strong> and IF you have the time to think about things when you are shooting...if you need to increase the ISO to get the F-stop or Shutter Speed to "get the shot" do so...</p>

</blockquote>

<p>I know that contrast range has nothing to do with the overall photons available in such a scene. They may be related in some way or another. Not sure. An overcast sky and/or heavily shaded scenes such as a darkened hill with the sun blocked on the other side by the hill providing more of a diffused low light still has enough photons to produce a reasonably noiseless signal at base ISO. A tripod and/or Shake Reduction would have to be used.</p>

<p>What I want to know is how far to ETTR such a scene. Relying on the camera's metering of such a scene usually creates a histogram on my camera with a uniform ant hill shape somewhere just left of center. I could manually adjust the exposure to allow that ant hill histogram to slider further and<br>

further to the right but how far?</p>

<p>Clearly this amount of lighting doesn't have as many photons to record than if the sun was blazing down on it so am I gaining anything by ETTR'ing to make that histogram move farther to the right? Or am I just amp-ing up the noise due to the lack of photons even without increasing ISO?</p>

<p>I don't know, Ben. Noise and detail are funny that way. I scrutinized carefully the patches when you posted. It wasn't that easy to see. A little later, they were still up in the browser when I Alt+Tabbed through going somewhere else. The left looked grainy as heck out the corner of my eye when it flashed by. (Besides, it was 50-50. I could have just guessed and done almost as well. ;)</p>

<p>I'd love to see those shots. The Gen1 24 is my most favorite of all lenses. There's something a little odd about the pictures they make. The 24 TS-E shots can look absurdly sharp sometimes, but nothing I can put my finger on when I look close. It's something like the noise and detail mentioned above. It's something you feel with your mind rather than see with the eyes. It might be just the contrast, or more likely the unusual way all TS-E focus. It all adds up to sharper than even tacks.</p>

<blockquote>

<p>Like Michael says, left is noticeably noisier than the right. In other words, the whole point of moving to the right is to move the left-most parts of the histogram as far right as possible (again, without clipping the right-most part).</p>

</blockquote>

<p>Exactly. When you expose properly, there are benefits albeit, some may find them small. </p>

<blockquote>

<p>You are, of course, correct that the left is noisier.</p>

 

</blockquote>

So much for debunking.... Even when the methodology above (how it was exposed in the first place, the rendering settings, the degree of ETTR versus non ETTR exposure among others) is so questionable.

 

<blockquote>

Now, to practicalities.

</blockquote>

Practicalities have nothing to do with the <strong>science</strong> of ETTR. You are more than welcome to not use ETTR, but to show that <strong>proper</strong> exposure produces less noise, then go off and say things in another post like “<em>But aiming for a theoretical half-stop reduction of noise in uncommon scenes <strong>with cameras that are already noiseless</strong>…</em>“ when its absolutely false to suggest they are (and now you’ve illustrated above how well you can contradicted yourself), or “<em>You’re obsessed with shadow detail, as are all other ETTR evangelists</em>“ when ETTR has nothing to do with shadow <strong>detail</strong>, and further misunderstanding the use of <strong>proper exposure</strong> by saying “<em>Aiming for ETTR with a bride in a white dress on a sunny day is just begging for disaster</em>“ indicate to me and hopefully others, you have either an agenda or a very unscientific mindset.

If anything has been debunked, its your logic, process and understanding of exposure and noise. As I said in another post, and shouldn’t have to be repeated again, <strong>ETTR is not over exposure, its idealized exposure for raw</strong> data capture. You’ve illustrated it results in less noise. You can decide that you’ll be sloppy with exposure when, as Jeff points out, you have the time to nail it, then “fix” this by moving sliders in a raw converter, but the facts remain. ETTR, an idealized exposure for raw results in less noise. Use it, don’t use it but stop trying to make up the facts (like cameras are noiseless).

You may not be able to tell on a print, a chrome shot at proper exposure versus one shot -1 stop and pushed 1 stop but few well trained photographers would consider the later approach to exposing a chrome ideal or recommended. I see little reason to promote sloppy techniques here, especially with such sloppy science.

<p>Tim, the only way to shift the ant hill without changing ISO is to let in more light. The noise is unchanged because ISO didn't change. Every stop brighter theoretically improves S/N by 3 dB. When talking about communication channels, 3 dB is HUGE. Not sure why the effects here are subtle enough for any meaningful discussion, but maybe that's the message of itself.</p>

<blockquote>

<p>What I want to know is how far to ETTR such a scene. Relying on the camera's metering of such a scene usually creates a histogram on my camera with a uniform ant hill shape somewhere just left of center. I could manually adjust the exposure to allow that ant hill histogram to slider further and<br />further to the right but how far?</p>

</blockquote>

<p>You have to test the limits of your camera as I tried to do with the 5D. Then you have to, as Jeff rightly suggests, decide if you have time to capture the image ideally (say in a studio) where you can pull out an incident meter and actually gauge the light falling, not reflecting off the subject. And when in doubt, bracket in + exposure increases. </p>

<p>Do I use ETTR all the time? No way. On my trip to the Galapagos last month, the light was changing a lot, the animals moved all over, it simply couldn’t find the time to use an Incident meter. I know what my in-camera meter assumes based on what I point it at (something black vs. white versus gray). I upped my exposure compensation only +½ based on what I know about my camera, shot raw and processed everything in Lightroom without anything clipping I didn’t want to clip. No surprises. Did I get the max ETTR? Probably not, certainly in all 1200 captures. Often the ISO was high as well. I’d rather have an image with noise than no image at all. But I ensure I didn’t ETT<strong>L</strong>! That’s almost always the case with a “normal” exposure setting, even controlled by a user who understands the feedback the meter suggests based on what its pointed towards. And that’s the key. You have to understand not only exposure and how meters report what it reports, you have to understand the role that exposure plays on a JPEG versus a raw file. Not the same. You have to use some intelligence and when pointing the camera at a whiteish seal lion, know the meter will under expose because unlike you the image creator, its kind of stupid. </p>

<p>As I tried to explain to Ben numerous times in other posts, one needs to understand the role of exposure on the data you wish to capture and not treat raw and JPEG identically. One has to understand what a reflective meter reports based on the object you point it at. This has been the heart of photography long before anyone knew what a pixel was! </p>

<p>One thing is clear. If the camera manufacturers didn’t ask us to treat exposure based on the camera LCD histogram like a JPEG when we shoot raw, if they actually provided us a histogram or clipping indicators based on the raw capture, all this would be much easier. </p>

<blockquote>

<p>But I think I’ve already pretty emphatically demonstrated that there’s no practical real-world human-discernible difference.</p>

</blockquote>

<p>No, its only impractical to you. You feel its “better” to incorrectly expose the data, be it a lack of understanding basic photography practices or laziness and now you want us to agree to that. Everything is close in hand grenades. What you haven’t told us is why you propose sloppy exposure techniques? The noise is less, you agree. Why go out of your way to produce a noisier capture? </p>

<p>You didn’t debunk anything so far, but you did illustrate that using the tools in a less than optimal way produces less than optimal results. </p>

<p>Which capture has better data from the same equipment, the one where you under exposed or the one where you properly exposed for the raw data? A simple yes or no answer is fine. </p>