bkellyusa Loc: Nashville, TN

This is the version of ETTR that I understand and use. Anything more difficult than this seems unnecessary. I also wouldn't use higher ISO to raise the exposure level since that seems counter-productive.

rmalarz Loc: Tempe, Arizona

This and your previous post erroneously connect file bit depth with ETTR. ETTR, again in unrelated to bit depth or dynamic range, it is related to exposure. The technique, through exposure, places the brightest part of the scene at the maximum, or near maximum, capability of the sensor's range. That is, just before the data overflows, if you will, the sensel capturing it. EBTR refers to past the jpg indicating overexposure by blinkies, the right hand side of the jpg histogram. This still leaves the ability of the RAW data to successfully capture details in areas the jpg thought were overexposed.

In practice, one would spot meter a white, or highlighted part of the scene. Knowing that the meter is indicating an exposure that would place that part of the scene in middle gray, to use a Zone System term, one would increase the exposure the necessary number of stops to Place, another Zone System term, to the brighter end of the tonal spectrum. In doing so, the other values of the scene are "dragged" upwards along the scale, as well.

In post, the normalization of the histogram will produce an image with no blown highlights, details in the shadows, and a spectrum of tones to the limits of the camera's dynamic range. The system is independent of the camera's Dynamic Range or bit depth of the image. It will work regardless.
--Bob

selmslie Loc: Fernandina Beach, FL, USA

Sorry if I overlooked this statement earlier.

My initial post proved that additional exposure reduced the noise. It was not ETTR that did it.

You missed the part where I explained that, "There is some reduction in noise from image #1 to #2 because of the additional exposure. The same exposure and result would have happened if the ISO had been set to 2000 without ETTR."

In other words, an exposure at ISO 2000, 1/40 sec @ f/9 would have looked identical to ISO 6400, 1/40 sec @ f/9 after moving the Exposure slider in post processing to the left by 1.67 Ev - same noise level, same resulting tonal distribution.

It would have looked identical because the ISO 6400 exposure with 1-2/3 ETTR did not blow any highlights. The ISO 2000 version would have had more latitude on the right. I also tried an ETTR of 2 stops (equivalent to ISO 1600, 1/30 sec @ f/9) and the highlights were indeed blown.

The exposure at ISO 800, 1/15 sec f/9 did not have any blown highlights. Neither did the exposure of 1/2 sec at ISO 100.

Acufine3200 Loc: Texarkana USA

I'm new here--but have been shooting since mid-'70's. After this I may be asked to leave. Honestly I don't get OP's need to make this such a big deal. Even with film I generally bracketed when possible. As a photojournalist my ultimate goal was the best image possible--regardless of whether I used Ansel Adams' Matthew Brady, or heck, Jadon Schneider's approach.

Now we have this 2-3" monitor on the back of even the cheapest cameras which affords us an opportunity to check for exposure after shooting. Digital media is relatively cheap...shoot away. I'd hate to limit myself to only one understanding of exposure theory, and get home to find I missed an incredible shot because I'm bound to ETTL or ETTR. That said, if one is in the field shooting 4x5, or 8x10 film, yeah there's a need to allow for more science.

But, with digital (similar to film) most of us have learned not all camera sensors record alike. And, since all of us have a differing idea of how we like our images, this is a good.

selmslie Loc: Fernandina Beach, FL, USA

No, it puts it near the maximum numerical value that your raw file can record, 16383. But at ISO 400 you can only use 25% of the sensor's physical capacity.

A gain of 4x (ISO 400) magnifies the 25% level recorded by your sensor to 100% to reach 16383 or a range from 8192 to 16383 in the brightest zone. At 25.0001% the highlight is blown, the maximum recorded value stops at 16383.

At ISO 100 the gain is 1x and 100% of the sensor's physical capacity is used. At 100.0001% both the sensor's and the raw file.s capacity are both exceeded. The sensor's "buckets" overflow and the raw file value is maxed out at 16383.

selmslie Loc: Fernandina Beach, FL, USA

My intent is actually the exact opposite of making this a big deal. I am trying to make the point that exposure does not have to be difficult or controversial.

There is no need to get deep into the technical aspects of ETTR/EBTR. It has no demonstrable benefits that you can't get by simply lowering your ISO. Exposure simply does not have to be rocket science, especially with today's crop of cameras.

Common sense and bracketing should be all we need to worry about.

My entire reason for starting these threads is to encourage skepticism. Don't do something because someone tells you it's going to make your images better. Try it for yourself. If you can't see the difference then it's not worth pursuing.

a6k Loc: Detroit & Sanibel

I am both a newcomer to UHH and an amateur who wishes he knew more. I've been taking photos since 1966 and did some fairly good B&W work back then. I customized the Adams Zone system to work for me with 35 mm B&W film and polycontrast paper. I built my own darkroom densitometer for spot readings of projected negatives. That's just to explain that I'm neither an expert nor a neophyte but something between. I'm fairly new to digital photography so some of this is confusing me.

The noise in a given sensor's output in a given shot is the result of many possible factors but the most obvious and perhaps the most influential is the "gain" applied to the pixels which have a natural sensitivity that does not change (seems to be ISO 100 on most of the cameras I've looked at). It makes sense, then, to reduce the need for gain as much as possible. So the way to do that is to hold the ISO down and use some combination of slower shutter or bigger lens opening, where available. The limit on this is that the brightest "zone" that should contain "information" should be given no more light than the sensor can handle. In this kind of example, the constraints are the available light and the available lens opening. The practicality of shutter speed enters into this in a similar way. I don't see how any of this is at all in doubt. So that is question #1.

Question #2 is overlapping: what is the practical difference between ETTR and giving more light with lens/shutter combinations? Does any of this change the value of keeping ISO as low as you can?

cactuspic Loc: Dallas, TX

Amphoto1, I agree with your correction. What I meant to say was that in recent years, ETTR had become more important to Canon shooters. The only reason I raised it is that the OP was shooting recent low noise Sony sensor with which the benefits of ETTR would be less apparent. Also, the Sony sensors tend to be more ISO invariant than the Canon sensors. (I apologize for not quoting directly but I could not get the part of the quote I wanted using the quote reply button)

selmslie Loc: Fernandina Beach, FL, USA

It sounds like you have grasped my point.

Changing the ISO only changes the gain. It does not change the sensitivity of the sensor. That's built in to the camera.

Using a higher ISO only encourages you to cheat the sensor out of the full exposure it was designed to receive - at base ISO.

At ISO 100 and f/11 you can capture a landscape in broad daylight at 1/200 sec. You would probably use a tripod, maybe a remote release and you will not be worried about the rocks, buildings, clouds and mountains moving - 1/200 is fast enough.

When the light drops it will be pretty easy to lower the shutter speed to increase the exposure.

When it comes to moving subjects you might need to increase the shutter speed and possibly raise the ISO or open the aperture. But people are not then going to be looking for noise in those images. They will be more interested in whether you captured the action.

rmalarz Loc: Tempe, Arizona

"My entire reason for starting these threads is to encourage skepticism. Don't do something because someone tells you it's going to make your images better. Try it for yourself. If you can't see the difference then it's not worth pursuing".

That is exactly what I did. Numerous tests, using photographic based test equipment. As a result, I found that ETTR/EBTR was a factual, proven process. It's quite easy to learn to do and quick to perform in the field. It's that simple.
--Bob

selmslie Loc: Fernandina Beach, FL, USA

Then it should be simple for you to prove that ETTR provides benefits that cannot be obtained more easily just by lowering the ISO.

Why can't you do that? As I suggested before:

1. It can't be proven because it's not true.
2. You don't know how to prove it.
3. You don't understand the issues.

Scotty

rmalarz Loc: Tempe, Arizona

If you wish to express it numerically, fine. That is the exact reason for exposing to the right. It places that scene value of 16383 right at, or slightly below the sensor's capacity of 16383. Thus, no blown highlight.

Again, you're showing a hang up on base ISO. Photography is a system management exercise everytime someone takes a photograph. It's a constant trade off, Does one want depth of field, then shutter speed is affected. If things are too poorly lit, then possibly more grain in the photograph. It's a constant system management exercise.

Now, if you wish to continue pushing a rope up a hill, go right ahead. I'm going to continue to use a system to determine the optimum exposure that I know works. How do I know? I simply tested it. Recorded the results and accepted that it is a superior method to exposing a digital photograph.
--Bob

robertjerl Loc: Corona, California

All well and good in very good light. But you lower the ISO you also lower the shutter speed, often not a good idea with BIF etc. Esp hummingbirds. Like everything else, learn when and where each thing works best and follow that. But don't do a blanket statement that your favorite technique is ALWAYS right.
As I taught my students in units on test taking skills: When there is no room for exceptions the answer is probably false or that choice is probably not the best one.

Oh, can you prove it is not true, under all conditions? And I don't mean opinions/beliefs that you think are facts (true facts, there are also false facts)

Exp "It is a nice day." even if 99.9999999999% of people agree it is still an opinion, there will be someone who doesn't like the weather. But "It is sunny." is a statement of fact, it may be either true or false but after checking only a liar can disagree with everyone who checked to see if the sun is shinning.

selmslie Loc: Fernandina Beach, FL, USA

Absolutely not! We have seen enough numbers.

I have been asking you to demonstrate it visually, with images, like I have done.

If you can't do that then refer to my previous post. Maybe you can come up with a fourth excuse.

TriX Loc: Raleigh, NC

Staying out of the ETTR debate because after examining this carefully and at length, I've drawn conclusions based on my own camera and its characteristics, and no need to convince anyone else - each person should determine this for themselves. I am however, interested in your comment concerning ISO invariance, a subject that's directly related to the discussion at hand. There have been two definitions propagated in this and other threads: one being a straight line DR vs ISO curve with a constant downward slope in DR with increasing ISO, and the other being the idea that the there's no difference in noise between underexposing and bringing up in post vs increasing ISO in-camera. Which were you referring to?