Debunking ETTR

lukerossknecht · March 14, 2010

All of these posts yelling about how only idiots can't control ETTR-related color shifts, yet I haven't seen anyone explain what causes it or how to control them.

I agree: it's apparent that ETTR results in shadows with more detail, but what to do about color? All the ETTR vs. non-ETTR comparison images that have been posted show them. Is there a guide out there or is it verboten to even ask?

digitaldog · March 14, 2010

I agree: it's apparent that ETTR results in shadows with more detail, but what to do about color?

</blockquote>

Alter the myriad of available rendering settings, assuming you are totally convinced that the color is “wrong” (since raw data is essentially Grayscale data, and you have control over tone and color rendering, I don’t know what’s “right” other than the color appearance you desire). IOW, just because the sole adjustment for ETTR appears to be exposure, doesn’t mean that’s the only rendering control you should alter.

Ray House · March 14, 2010

Michael, I agree with you as to bracketing, however there are times when I can't bracket and often I will meter in manual and check my histogram. I'd buy Ben a beer just for posting a thread that gives us much to think about. My conclusion...Shoot RAW, check my histogram for highlights, adjust exposure and do the photography while trying not to be too anal about it cause I'm not that good anyway! When it comes to "worry" or "enjoy", I choose "enjoy".

michael_young3 · March 15, 2010

Consider it debunked.

Attached are LR2.6 screen shots showing the full frames and a detail of high and low bracketed shots. AEB was set to 2/3 stop; the low stop was set to center meter. As circumstances would have it, the metered, "low" shot was dead on, and what I consider ETTR. The right side of the histogram nudged right up on the edge without a spike, and the LCD did not flash. The high shot is then 1-1/3 stop overexposed.

The only processing was to equalize a selected highlight tone to match in both shots by adjusting Exposure. The adjustment for a tone match was exactly -1.33. Highlight detail in the overexposed shot did not improve by using Highlight Recovery. Its primary effect was to further lower overall contrast. Highlight recovery was restored to 0 for these screen shots.

The big difference is in noise. The two shots effectively compare ISO 400 to ISO 160 noise performance. Overexposure wins big here. See the second screen shot.

The downside is the lower contrast in the overexposed shot. It is lower than I like. If this had been a real shot, I would have taken time to boost the midrange and highlight contrast.

The color shift is minimal, and I would not have bothered to correct one or the other. Judge for yourself.

michael_young3 · March 15, 2010

The big difference was in noise. Effective ISO for the overexposed shot is 160, compared to 400 for the straight shot.

The obvious conclusion: select ISO 160 rather than overexpose. This will obviate any color or contrast concerns. Since there was no other observable difference, the lower ISO will have produced the better, more easily manageable image.

peter_stacey1 · March 15, 2010

Michael Young wrote:

Consider it debunked.

</blockquote>

Michael, ETTR does not promote over-exposure - ie. pushing pixels to saturation unnecessarily.

Your first shot is the ETTR shot and at the same time, the EV shot. In this case, they are one and the same thing.

If your original exposure maintains detail throughout (no blown highlights) and you are optimizing the RAW data by maximizing the amount of signal you are recording, then that is what ETTR is. The first shot, not the second.

Your post gives good evidence towards not blowing out highlights, but that is standard good practice for a scene where the dynamic range is within the range of the camera.

If the scene dynamic range is outside the range of the camera, then you have a choice to make in relation to blocking the shadows, blowing the highlights or combining bracketed exposures.

Michael Young wrote:

The obvious conclusion: select ISO 160 rather than overexpose.

</blockquote>

For that shot, I think the obvious choice would be to shoot with a tripod, use the native ISO of the camera and still shoot it the way you shot it in the first place.

So if the native ISO of your camera is 100, then there'd be no need to shoot higher then that for this situation. If it's 200, then that would also be a better choice than 400.

What made you choose 400 for this scene? That seems rather sloppy unless you could only hand-hold your camera and need it in order to get the DOF/shutter speed you wanted. However, a good tripod and lower ISO would be the obvious way to improve the shot.

william l. palminteri · March 15, 2010

<a href="/photodb/user?user_id=555571">Ray House</a> <a href="/member-status-icons"><img title="Subscriber" src="http://static.photo.net/v3graphics/member-status-icons/sub8.gif" alt="" /><img title="Frequent poster" src="http://static.photo.net/v3graphics/member-status-icons/1roll.gif" alt="" /></a>, Mar 14, 2010; 11:29 p.m.

When it comes to "worry" or "enjoy", I choose "enjoy".

</blockquote>

Ray, I'm with you!

I enjoy the entire process, from location scouting to final print.

Bill P. vs. the Forces of Doom

Tim_Lookingbill · March 15, 2010

Michael, ETTR does not promote over-exposure - ie. pushing pixels to saturation unnecessarily.

</blockquote>

Then why have the "R" in ETTR? Over-exposing compared to what? What the camera considers a proper exposure? No one in this thread has advocated pushing pixels to full saturation or clipping. We expose as far to the right without clipping any data outside of spectrals. I may be misunderstanding the way you've worded your comment.

Shooting ETTR which I often practice I get Michael's same loss of local contrast in 200RGB highlight region detail which ACR's Highlight Recovery won't fix requiring tightly grouped curve point adjustments. This may have something to do with ACR/LR's default tone curve not lining up with the non-linear behavior of the sensor's electronics when pixels get close to full saturation.

I notice this erratic sensor behavior examining my incamera histogram as it inches closer to the right with each 1/3 stop increase shooting high contrast outdoor scenes similar to Andrew's sunlit pavement and dark door sample.

Starting with my camera's somewhat under exposed settings the highlight peak in the histogram will nudge gradually to the right with each 1/3 stop increase where the closer to the right it gets this gradualness decreases and the peak will abruptly jump close to clipping. Shooting warm colored pavement I know to back off even when the incamera histogram shows the peak just barely touching the right side of the histogram or else I'll get off color blotches where one or more channels have clipped detail in the pavement.

This erratic/jumpy/non-linear behavior is what makes ETTR risky.

ben_goren · March 15, 2010

Peter, it would seem you missed the obvious. In Michael’s example, it was the ETTR version that

was noisier. ETTR is supposed to reduce noise, but it had the exact opposite effect.

Bill, as far as I’m concerned, the primary purpose of photography is to have an excuse to

get outside and look at beautiful things. The tech stuff is fun, and I’m still amazed every time I

see the print drop into the catch cloth.

ETTR has caused me lots of grief over the years, as I’ve gotten what the ETTR proponents

say should be the perfect exposure, only to have it turn out to be useless or the worst of the bracket.

I’ve often felt a bit of shame at somehow failing at it.

Now that I’ve abandoned this quixotic chase for an ideal exposure balanced right at the cusp

of overexposure, my hit ratio has dramatically improved and my enjoyment has gone up. Get the

exposure in the camera closest to what you want in the final print and stop worrying about theoretical

imperceptible enhancements in IQ that, as often as not, aren’t even there in the first place. If you have problems with dynamic range, use ±1 EV HDR. If even that won’t cut the mustard, you’re doing something like shooting the entrance to a mine with the sun in the frame, and you want to image sunspots and the back wall of the mine at the same time…and you should know full well what kind of insanity you’re getting yourself into.

I mean, come on! The IQ and DR of the 5DII and the 7D and the 50D (and their Nikon counterpoints) isn’t good enough already, that

we’ve got to agonize about squeezing yet another stop out of them? All those people who brag

about how they get great prints out of ISO 1600 shots must really be on crack to put up with all the

noise those shots have — and however did we manage in the days of the 300D?

Puhleeze.

It’s time to stop chasing that last fraction of a percent of technical perfection, accept that the tools are more than plenty good enough, and worry less about the rendering of the bark and much more about the picture of the forest. Open your eyes.

Cheers,

b&<div></div>

martin_howard1 · March 15, 2010

william l. palminteri · March 15, 2010

<a href="/photodb/user?user_id=820080">Ben Goren</a> <a href="/member-status-icons"><img title="Subscriber" src="http://static.photo.net/v3graphics/member-status-icons/sub7.gif" alt="" /><img title="Frequent poster" src="http://static.photo.net/v3graphics/member-status-icons/3rolls.gif" alt="" /></a>, Mar 15, 2010; 12:09 p.m.

Bill, as far as I’m concerned, the primary purpose of photography is to have an excuse to get outside and look at beautiful things. The tech stuff is fun, and I’m still amazed every time I see the print drop into the catch cloth.

</blockquote>

Ben, I'm with you.

Let's get outside and take some more photos!

Bill P. vs. the Forces of Doom

william l. palminteri · March 15, 2010

<a href="../photodb/user?user_id=820080">Ben Goren</a> <a href="../member-status-icons"><img title="Subscriber" src="http://static.photo.net/v3graphics/member-status-icons/sub7.gif" alt="" /><img title="Frequent poster" src="http://static.photo.net/v3graphics/member-status-icons/3rolls.gif" alt="" /></a>, Mar 15, 2010; 12:09 p.m.

ETTR has caused me lots of grief over the years, as I’ve gotten what the ETTR proponents say should be the perfect exposure, only to have it turn out to be useless or the worst of the bracket. I’ve often felt a bit of shame at somehow failing at it.

</blockquote>

Ben, I'm going to take a guess here, since you haven't posted a "failed" example.

That "failed" example might not live up to what the "experts" are looking for, but it may be a very pleasing photograph.

It's a misconception that all available values must be used at all times.

Just because you can capture them and print them does NOT guarantee a pleasing result.

It guarantees that you got all the information available at the time, but it does not guarantee an artistic photograph.

A symphonic piece does not require that all the pieces be playing all the time.

Bill P. vs. the Forces of Doom

ben_goren · March 15, 2010

Bill,

That’s exactly the lesson that’s been so hard for me to learn — in large part

from all the hype over ETTR.

Waaaaay up above, Philip Partridge linked to <a href="http://www.color-

mag.com/spotlight1/sl1.html">an article</a> featuring the work of Pablo Aguinaco. The purists will

decry all the blocked shadows in his work, completely missing the point that it’s those rich,

deep, featureless tones that so effectively set off the amazing colors he captures. If he had tried to

protect shadow detail in the manner commanded by ETTR proponents, the pictures would have been flatter and had far less

impact.

I want my photography to resemble Philip’s much more than I want it to resemble the ETTR

examples in this thread.

This is from some years ago, my first outing with my then-brand-new original 5D and original TS-E

24. Of the hundreds of exposures, this was the only one that really turned out, and it’s still one

of my favorite pictures.

Is there noise in the shadows? You betcha. Lots of it. I even had to fight sensor banding. Would I

prefer there not be noise? Sure, of course. Do I think it ruins the picture? Not by a long shot.

Cheers,

b&<div></div>

peter_stacey1 · March 15, 2010

Tim Lookinghill wrote:

No one in this thread has advocated pushing pixels to full saturation or clipping.

</blockquote>

No, I think we are in complete agreement Tim. My view of what ETTR means, seems to be the same as your understanding.

However, unfortunately Michael did clip the image he used as an example of ETTR and used it as evidence to claim that ETTR is debunked. The numbers are fairly clear, especially after the exposure slider has already been bought back down 1 1/3 stops:

You can't recreate clipped data and this isn't the application of ETTR. It's just unfortunately a poor exposure and Michael's original exposure is much more pleasing on the eye, because it is much better.

Ben Gosen wrote:

Peter, it would seem you missed the obvious.

</blockquote>

Ben, if you think the 1/25th second shot (not a good shot) is noisier than the 1/60th shot (the best exposure), then there is something seriously wrong.

The 1/60th shot is the better image because of the two, it's the only one that is exposed well. There isn't any need to look at noise, etc. because the other shot is clipped and should be deleted and forgotten.

There is a lot of false information in this thread and I don't think people mean for there to be. It's just a lack of understanding and poor technique being used as an example of poor theory.

The theory is fine and it works in some situations, but definitely not all. When it can help, it's a useful tool for capturing optimum RAW data.

Regards,

Peter

digitaldog · March 15, 2010

Ben, I'm going to take a guess here, since you haven't posted a "failed" example.

</blockquote>

Yup, but a lot of failed thinking on the subject and (forgive the pun) noise. I think he’s the only one who believes he’s debunked anything.

That "failed" example might not live up to what the "experts" are looking for, but it may be a very pleasing photograph.

</blockquote>

Pleasing is in the eye of the image creator. That’s why way, way up there I asked Joe about his example images and what the OP swears is damage (his words, not Joes), something the OP does a lot here. I think its necessary at this late stage of all the treads to separate pleasing because its subjective, its really up to the image creator and its useful if the image creator uses the tools available in the raw converter to produce his or her goal of pleasing. Our debunking OP has some serious inadequacies using the raw converter tools for one as has been pointed out. Most of the examples that are properly exposed and rendered to match are an attempt at an apples to apples rendering to illustrate the effect on noise and detail using PROPER ETTR vs, what the camera meter suggests. Interesting that in terms of what is considered “normal” exposure, most appear to be using some undefined in-camera metering instead of an incident reading. You think that has some bearing on the exposure? You bet.

Just because you can capture them and print them does NOT guarantee a pleasing result.

</blockquote>

Agreed! That’s why in the context of this “debunking” of proper exposure for raw data, we should stick to that technique and leave pleasing for another post.

A symphonic piece does not require that all the pieces be playing all the time.

</blockquote>

True indeed. But one hopes that the instruments are all in tune and that the techniques of the various musicians are proper, much like exposure.

I wonder why this tread continues when it seems clear to the majority of readers and posters of the thread that understand ETTR’s goals and have shown more illustrations than necessary, that nothing was debunked other than the OP’s understanding of the process of photography, exposure, rendering and so forth.

digitaldog · March 15, 2010

Waaaaay up above, Philip Partridge linked to <a rel="nofollow" href="http://www.color- mag.com/spotlight1/sl1.html" target="_blank">an article</a> featuring the work of Pablo Aguinaco. The purists will decry all the blocked shadows in his work...

</blockquote>

Nonsense! Here you go again putting words in people’s mouths. There is absolutely NOTHING wrong clipping shadows. One of Jeff and my closest friends is a very well know portrait photographer who had no regard for shadow detail, that’s his style (http://greggormanphotography.com/). Suggesting that anyone who promotes proper exposure for raw data would also decry blocked shadows is yet another of your silly and easily dismissed statements (like raw isn’t linear data or digital cameras don’t produce noise, ETTR is about clipping highlights). Go back to work, get something accomplished and leave proper photography techniques for those that understand when and why they should be used. I can’t fathom with your utter lack of understanding on so many points why you’d even use a DSLR and capture in raw.

michael_young3 · March 15, 2010

Your post gives good evidence towards not blowing out highlights, but that is standard good practice for a scene where the dynamic range is within the range of the camera.

</blockquote>

Peter, listen to yourself squirming in the wind, denying what there isn't to deny. I pointed out there were no differences in highlight details, even though I expected them to have blown out. They might have, but it wasn't important. The details I showed were in the midtones. The highlight detail were indistinguishable to the eye.

The shots were right on or blown out only relative to some idea I held in my head. I could have re-shot it to match that idea. As it were, I had no need for a shot of junk on my desk. They would do for an ETTR versus non-ETTR shot based on some slightly lower tone value. Nothing has changed.

What made you choose 400 for this scene? That seems rather sloppy unless you could only hand-hold your camera and need it in order to get the DOF/shutter speed you wanted. However, a good tripod and lower ISO would be the obvious way to improve the shot.

</blockquote>

It was a valid test shot, Peter. Unless you want to extend ETTR to preclude shooting at anything other than base ISO. Sloppy , indeed!

(Ben wrote: )

In Michael’s example, it was the ETTR version that was noisier. ETTR is supposed to reduce noise, but it had the exact opposite effect.

</blockquote>

No, that's not correct. The higher exposure has lower noise. There is no confusion, unless one wishes to confuse things with labels, such as ETTR, non-ETTR.

peter_stacey1 · March 15, 2010

Michael Young wrote:

Peter, listen to yourself squirming in the wind

</blockquote>

Michael, I think you've taken offence, where none was intended and in any case I'm not really the kind of person who would squirm over anything, least the wind.

The first shot in your comparison is a nice shot (and I understand the subject was for test purposes, not artistic) and even in the images posted, I can see a difference in the highlight detail.

The second is just not a good shot. It's blown out and clipped in the highlights; and whether or not you are concentrating on the mid-tones, using that image in trying to debunk ETTR doesn't help you, because it isn't an ETTR image - which is about capturing optimum RAW data (best S/N ratio for the file), not about clipping information when you don't have to.

No offence was intended in my post about the comparison, it's just that the technique in the second version is not a good example of ETTR.

For this particular shot, the first image was the money shot and that's where it can stop. No need for anything other than the way you metered and exposed the image.

I agree with your conclusions for your test about lowering ISO instead of clipping information as a way to produce a better more manageable image. The first shot is clearly superior to the second and could have been even improved further by lowering the ISO more (100 on your 7D), using a sturdy tripod, remote shutter release, mirror up and then metered the way you approached it.

If both shots were captured that way and had useable data throughout, your test would have been a better example of what you were trying to show.

Regards,

Peter

digitaldog · March 15, 2010

Here we are, a few posts short of 200 and where do we stand?

The OP has proposed a theory about ETTR. In terms of good science and recommended techniques in any field, its useful for someone to propose a theory to a peer group. That’s where we are although based on his understanding of basic photography, raw processing, proper testing protocols and not as important, the ability to listen his peers, retest and be objective, I don’t see much need to go farther, the debunker has pretty much been debuted, again based on peer review. I skimmed through many of the posts here to paste what I think are very salient comments on the debunking exercise. Again, based on peer review, having them all in one place is interesting reading (hopefully mostly to the OP). I’ve excluded myself (there’s more than enough meat here). But lets look at what the OP wrote in post #1 way, way, way up there.... Before I post the replies from the peers in the next post, lets look at some of the comments the OP has made:

I firmly believe that the common wisdom of ETTR is bad advice

I advocate starting with a traditional “proper” exposure, such as determined by metering off a gray card

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. (this is a personal opinion and those who’ve seen the correctly produced ETTR examples can agree or disagree).

</blockquote>

Major misunderstandings and misstatements:

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.

Clearly, ETTR can only possibly be effective at reducing base ISO at the cost of an equivalent number of stops of highlight dynamic range

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

The problem is 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.

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. (a straw man point until you define proper expsoure)

I started this thread with, the difference is imperceptible. (subject to each users opinion of proper ETTR examples)

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…. (this is fully dismissed below)

I don’t dispute that there’s a scientifically-measurable difference. But I think I’ve already pretty emphatically demonstrated that there’s no practical real-world human-discernible difference.

Tell me, Andrew, have you stopped beating your favorite underaged male prostitute? A simple yes or no answer is fine.

Even though the Camera Faithful profile is the most colorimetrically-correct one that one can select (by a huge margin), it is still pretty far off in absolute terms. (that’s nonsense since its output referred).

Digital exposure works the same way as film exposure because the digital cameras are designed to mimic film. (well JPEG is)

The ETTR dogma is based on the assumption that RAW data is linear. It gets repeated over and over and over again.So, now, when you add ETTR exposure compensation on top of all the rest, you get quite radical color shifts, loss of contrast, loss of detail, and the like. ETTR fanatics claim that they’re increasing information content by avoiding the darkest bits with the least amount of information and encoding the darkest parts of the picture in lighter bits which have more resolution. But that comes at the price of dramatically reducing the information content of the picture as a whole. Essentially, you’ve just turned your 14-bit camera into a 10-bit camera by telling it to not use those nasty lower four bits.

Joe made it clear that the damage done to the highlights by ETTR is unacceptable, and that shadow noise just isn’t a problem that so desperately needs to be solved. (he didn’t say that, I can’t find it…)

Overexpose with ETTR and you probably won’t ever even realize what you’re missing.(That of course wouldn’t be ETTR, it would be over exposure, he doesn’t get it).Can somebody explain to me what the obsession is with shadow detail and noise that none of the ETTR proponents ever think to look at what happens to the highlights? According to ETTR, I should have let the bits of sky blow out in order to get rid of the barely-visible-at-100% noise in the shadows. (no it doesn’t Promote that).

In Michael’s example, it was the ETTR version that was noisier. (no it wasn’t according to Michael)

All these photographers telling the world how ETTR is superior and anybody who dares suggest otherwise is a poppy-head — but none of them showing why that is the case. (lots of examples above bud).The mantra is to let clip those highlights you don’t care about in order to preserve the shadows. (you admit you don’t care about the highlights).

</blockquote>

digitaldog · March 15, 2010

Now onto the peer review. I’m not suggesting we are all 100% in agreement here, but most seem to be in disagreement with the debunk. And yes, they are taken out of context by author and in a rough chronological order by poster:

JEFF said It would be useful to fully understand the principles behind ETTR if you wish to debunk them... Jeeesh, no wonder you have a hard time understanding ETTR, you don't know much about post processing...if you EXPOSE TO THE RIGHT, you'll need to bring the tone curve down using something other than simply the Exposure slider...think the point curve or parametric curve in Camera Raw (or hey, try the Brightness slider and see what THAT does).The utter irony is the verve and vigor in which the OP keeps coming back and proving that ETTR is actually useful (in the proper situations) and still claiming that that his "proof" totally debunks the practice... Should you keep ETTR in your knowledge toolbox to use when it's useful and beneficial? Well, if you care about image quality, you would be a serious fool not to...obviously, the OP doesn't really care about image quality...your milage may vary (YMMV)... ETTR has NOTHING TO DO with the amount of levels in the shadows... 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... Should you keep ETTR in your knowledge toolbox to use when it's useful and beneficial? Well, if you care about image quality, you would be a serious fool not to...obviously, the OP doesn't really care about image quality...your milage may vary (YMMV)...

SASVATA said In other words, you should always shoot at the lowest ISO you can without clipping the highlightsLike 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.

MICHAEL said: Every stop brighter theoretically improves S/N by 3 dB. When talking about communication channels, 3 dB is HUGE. Ben, there's not even a question. Bottom has way more color noise. Texture details in the vertical sashes are clearer in this order: abdc

I'm OK with there being an optimal exposure, and OK with careful conversion and processing to bring out the best details. Gosh, how new is that? And having done this for more than a weekend, it's abundantly clear that ETTR boils down to this: bracket your shots, and favor the higher exposure because it holds a little more detail.

If the higher exposure hasn't blown your highlight details, use that for all the reasons ETTR is good. If, OTOH, the high exposure blew out your details, toss a nickel in the kitty and buy Ben a beer.

MIKE (summed it up well) said: To follow Ben's original post: (a) the theory of ETTR is correct and differences can be discerned (b) one downside to an over-rigid adherence is a potential for ruining a shot by over-exposure* © a second downside is that while working about the best degree of ETTR you risk missing the shot while you firtle with the exposure settings (d) the camera electronics/software are so good now that differences made by ETTR are reduced to the point that the practical risk of over-exposure outweighs the practical benefits.

*so learn to expose properly

GK said: No mystery, I think there is a very subtle improvement in the shadows in the right hand shots.

FRANCISCO said: For this, you should know the ratio between the ETTR exposure and the "normal" exposure. Beause of the linear nature of raw data, what needs to be done is just to scale the raw data by that ratio (ETTR/normal) before any raw processing (something like pre-raw conversion) and then you'll have the mid tones where the (existing) raw converters expect them, so you will not have the hue and tone issues. This procedure will keep the improvement in S/N ratio that ETTR offers. (the OP actually agreed).

SCOTT said: First off to those who see colors different when changing their exposure, learn to use your raw converter better, you should not see color shift unless you are clipping one or more color channels.

CRAIG said: Absent clipping highlights or changing ISO, a "ETTR" image will contain more data than a "normally" exposed image. As your signal-to-noise ratio is roughly proportional to the square root of the counted photons, counting the most photons possible *must* make a *better* image in that you have recovered the maximum amount of information in the scene. This is an indisputable fact, so let us not even argue this. ETTR *must* lead to better contrast and better color as there is more information in the file. Failure to achieve these results is a failure of the post processing, not of the original image capture. Again, this is a simple fact.

DAVE said (to the OP): It has become apparent that your entire difficulty with the principle has nothing to do with ETTR being wrong....and everything to do with the fact that you don't know how to control exposure and color in raw software. All you examples posted clearly show this.

TOM said: But, if you do ETTR correctly there is no high bit information that is sacrificed. This does not make sense. If it does not clip, you must acquit.

KYLE said: I think we've all established that the basis for this thread is a lack of understanding/ability.. why are we still arguing it? It has gone from rather amusing, to somewhat dull and repetitive.

Ok seriously? All of these 150+ posts and you -still- don't understand the actual idea of ETTR?

You're arguing your idea of it.. the problem is, your idea of it is wrong.

Nobody is saying blow out highlights at all costs to preserve shadow detail. We're saying push them up if you can to help get more out of the image. At no point has anybody said otherwise except you.

KELLY said: Unless one knows the *film type or sensor type* and what the *scenes range in fstops* that one is using/shooting; there is no way one can logically have a dialog about proper exposure. Mapping a scene to a plate was FULLY understood back in the civil war era when folks had darkrooms in covered wagons; it seems abit obfuscated purposely with the newbie lingo and dialog on this thread.

Whether noise or grain in the shadows is an issue depends on the iso setting; type of sensor; technology of the sensor; even temperature of the sensor. This was well understood many decades ago. In film and plates it goes back to the beginning too.

AS Andrew stated; alot of this is "basic Photography 101".

ANDREW ROBERTSON said: I don't think anybody is arguing that you should expose to the right and obliterate highlights that are important to the aesthetic you wish to create. If this is your premise, that expose to the right *forces* you to obliterate the sky in favor of pumping up the shadows, you are mistaken.

TIM said: The one stop ETTR version was quite clean as well as a bit over exposed at ACR defaults, but the two stop where the image was quite washed out with just tiny spots clipping had severe color hue shift toward red in the shadows once I normalized it but no noise. The Solux lamp often gives me a slight reddish cast so this might have contributed to this red shift.

I would definitely see an advantage with ETTR trying to capture for accurate reproduction of fabrics, fine art paintings and other subjects that require the maximum capture of all available detail. The quality of lighting would need to be dealt with though probably a custom WB even with flash or professional lighting.

Then why have the "R" in ETTR? Over-exposing compared to what? What the camera considers a proper exposure? No one in this thread has advocated pushing pixels to full saturation or clipping. We expose as far to the right without clipping any data outside of spectrals. (Peter agreed saying: No, I think we are in complete agreement Tim. My view of what ETTR means, seems to be the same as your understanding.)

PETER said: Ben,

You need to go back through the thread and read what has been pointed out several times about the theory of exposing to the right, most eloquently in parts by Jeff Schewe.

ETTR has nothing to do with clipping the highlights and is just one of many ways you can approach the capture of optimal RAW data. It applies when the scene dynamic range is within the sensor dynamic range and requires no clipping of any data.

Michael, ETTR does not promote over-exposure - ie. pushing pixels to saturation unnecessarily.

Your first shot is the ETTR shot and at the same time, the EV shot. In this case, they are one and the same thing.

If your original exposure maintains detail throughout (no blown highlights) and you are optimizing the RAW data by maximizing the amount of signal you are recording, then that is what ETTR is. The first shot, not the second.

Your post gives good evidence towards not blowing out highlights, but that is standard good practice for a scene where the dynamic range is within the range of the camera.

RAY said: The discriptions for using ETTR I've read and that are common to me say simply slide the histogram to the right using exposure that doesn't overexpose. I'm wondering where the OP gets his "common wisdom" and "commonly described" information.

LUKE said: I agree: it's apparent that ETTR results in shadows with more detail, but what to do about color?

michael_young3 · March 15, 2010

There is only one other point to be made: the initially unnoticed rainbow in Joe's tall, cool glass shot. Just slightly higher exposure obliterated it to an uninteresting, simple white caustic. Lower exposure would have accentuated the serendipitous colors. You can nail the shot and exposure for what you see, but you sometimes don't see everything.

william l. palminteri · March 15, 2010

<a href="../photodb/user?user_id=820080">Ben Goren</a> <a href="../member-status-icons"><img title="Subscriber" src="http://static.photo.net/v3graphics/member-status-icons/sub7.gif" alt="" /><img title="Frequent poster" src="http://static.photo.net/v3graphics/member-status-icons/3rolls.gif" alt="" /></a>, Mar 15, 2010; 12:59 p.m.

Bill,

That’s exactly the lesson that’s been so hard for me to learn — in large part from all the hype over ETTR.

</blockquote>

I know, right?

That's why I look at the photos of the people who write this stuff,. to see the results of all the "expertise".

You posted a GREAT photo, I love it!

The histogram says it's a disaster, which is why I look at photos, not histograms.

Those hard blacks REALLY push the fireworks forward!

I'm amazed that nobody has addressed DTTL, the second part orf the ETTR equation.

I want my photography to resemble Philip’s much more than I want it to resemble the ETTR examples in this thread.

</blockquote>

Ben, I want your photography to be what YOU want it to be.

Bill P. vs. the Forces of Doom

david_ralph1 · March 15, 2010

This post was started with a title that is just in bad taste, by someone advocating not bothering to get the best image when better is possible.

In this post the OP amounts to a troll, to bait people.

Jeff Schewe's responses have amounted to casting pearls before swine. For those that do not know Jeff Schewe's accomplishments, perhaps read one of his books, such as his update of Bruce Fraser's bible on sharpening, or perhaps check out the plug-ins for Photoshop by his software company Pixelgenius (Photokit Sharpener, without which I do not want to use Photoshop, has been incorporated by Adobe in Lightroom 2), or try out one of the video training series that Jeff Schewe has done with Michael Reichmann at the Luminouslandscape.com.

That attitude and tone of the OP in his troll title, and the lower standard he would have others adopt, is just plain rude.

Just my $.02

dan_south · March 15, 2010

Here we are, a few posts short of 200 and where do we stand?

</blockquote>

That's a GREAT question. I still haven't seen a SINGLE image that convinces me of the superiority of ETTR. Not ONE! Talk is cheap. Demonstrate.

or try out one of the video training series that XX has done with XX and XX.com

</blockquote>

Okay, well, since you brought it up, I paid real money for one of those "courses." In my opinion, the topics were well-organized, but the delivery was meandering and choppy enough to be distracting. The presenters assume that the audience already knows a good deal about post-processing. Comments such as these abound:

"Everybody knows that XXX is the best way to do YYY."

"Uh-huh, that's correct."

OR

"Lightroom has a wonderful feature called Clarity."

"Yes, I really like Clarity."

Instead of taking the time to explain fundamental concepts that would have made the material accessible to a wide audience, the presenters spent conspicuous amounts of time on name-dropping - why was it necessary to use an unflattering photo of Jay Maisel to demonstrate Lightroom's editing capabilities? - and alternately admiring, interrupting, and arguing with each other. If this had been some free video on YouTube, I wouldn't complain. It's just not up to par with most of tutorials/podcasts from Adobe, Kelby, Alain Briot, el al. (not even the FREE ones). Some rehearsal, some judicious editing, and just plain sticking to the script (if there was one) might have made the presentation a lot easier to follow. Again, this is one person's opinion of one course. Your mileage may vary, but I didn't get a whole lot of useful information out of this material.

peter_stacey1 · March 16, 2010

Dan South wrote:

<blockquote>superiority of ETTR. Not ONE! Talk is cheap. Demonstrate.

</blockquote>

This thread isn't about the superiority of ETTR, it's about the apparent inferiority of the technique.

A proper discussion of the advantages and disadvantages of different RAW capture techniques might be best had in a

separate thread, but not this one, which is trying to show an inferiority.

There's nothing to prove in relation to superiority anyway, since superior would never be a good description. It's just one of

a number of ways to approach the capture of optimum RAW data and only applies in some situations. When it does apply,

it's useful, but at other times it's not.

And that is possibly the heart of the matter. The OP acknowledges the theory but doesn't seem to know when to use it so

has decided that the theory is bad. If you don't know something, blame the technique. That's much easier than trying to continue to improve.

Sign In

Debunking ETTR

Recommended Posts

Link to comment

Share on other sites

Top Posters In This Topic

Popular Days

Top Posters In This Topic

Popular Days

Posted Images

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Create an account or sign in to comment

Create an account

Sign in