How far to the right are you?

[purplehaze]

A deliberately provocative title to get your attention. I just read a very interesting Luminous Landscape article yesterday that I see Allan referenced in this thread without sparking any discussion. The article gathers a lot of principles I have read elsewhere to argue for bracketing exposures much higher than what the camera meters, either by spot or matrix metering. I have started experimenting with the author's methods, around the house (because winter) and finally see what some have you have been trying to drum into me about proper exposure mitigating the effect of high ISO. I am curious to know how many of you already use this technique and what reservations you might have, if any.

[Polar01]

Janis, I am still working my head around the concept, up to Toronto today with some friends, so I was able to work more on the theory from the article. Below is a shot I did from today, f/5.6 ISO 800 1/30th which is 2 full stops overexposed.

Cheers: Allan

[Stagecoach]

I have read this article before Janis and have one significant problem with it;

The author states that the blinkies on the back of MY (referred to as YOUR in the article) camera occur at 1 stop before the highlight warning in MY software.

The author does not know what camera I use, how it is adjusted or performs or what software I use and when it 'highlights' with the RAW images I feed into it.

Case closed

[FrankMi]

For the right image, ETTR makes good sense. For a golden hour scene, I find I not only prefer to bracket the exposure, but I also ETTL to keep the sky from being blown out. Because of the +1EV, -1EV and -3EV bracketing, I still get the shadow detail and gain a beautiful sky.

Although photographers have their favorite way of doing things, there isn't a one size fits all.

[purplehaze]

Allan, that's a challenging scene. I have been working with much simpler ones.

Grahame, if I'm reading him right, I think the author is speaking there in generalities, and his method really does encourage you to figure out the optimum white point for your camera and software. Trouble is, someone like me would have to buy his book to find out how. I gather this fellow has a system, but I'll have to read him several more times to understand it.

[NorthernFocus]

People need material for websites/blogs. They have to write something. There is virtually no rule of thumb or technique that works in all situations, with all equipment, for all subject matter, etc. In direct response to the question, I do generally apply ETTR techniques. How that is accomplished varies based on equipment, subject/BG, lighting conditions, etc. To attempt to go beyond explaining the concept and to suggest detailed procedures to a broad audience is nonsensical. Advising someone who shoots the same equipment, sure. But advising the public at large, well...

[Stagecoach]

Grahame, if I'm reading him right, I think the author is speaking there in generalities, and his method really does encourage you to figure out the optimum white point for your camera and software. Trouble is, someone like me would have to buy his book to find out how. I gather this fellow has a system, but I'll have to read him several more times to understand it.

Janis,

I think one of the problems with all these different methods is that whilst they may be 'technically' accurate with the specific procedure they suggest how often in real world shooting do you have time to actually set things for the optimum results wrt exposure.

Today I have shot two subjects, one I had all the time in the world to get the optimum settings in all areas, the other subject having frantically changed lenses, set what I hoped would achieve my priority which was sharpness I managed to burst off 6 shots before it flew off. The modern camera using it's auto exposure provided excellent exposure of the image.

I agree it's useful to know the workings and limitations of your camera and knowing how it's 'indications' affect the end result you can achieve in post but all these methods seem to come up with the same result as far as I can see.

Grahame

[purplehaze]

how often in real world shooting do you have time to actually set things for the optimum results wrt exposure

Exactly the question I am asking myself, particularly as I am slow to begin with. But the technical stuff, especially in the second article, has really hit home with me, as it has made me understand how much I stand to gain by being a little less afraid of blinkies in the camera, and bracketing, bracketing, bracketing.

[rpcrowe]

Grahame wrote, "how often in real world shooting do you have time to actually set things for the optimum results wrt exposure"

All canon DSLR cameras and some other models of various brands can set up a three stop bracketing when AEB is selected when the camera is placed in burst mode, three bracketed shots are fired each time the shutter button is depressed. The camera then stops firing until the next time the button is depressed.

With the Canon 7D and many other Canon DSLR cameras since the 40D; you have user selected modes. I have one of my three user selected modes set up for bracketed exposure - so I can select bracketed burst exposure with a single twist of the mode dial which takes almost no time or effort.

There is a caveat when using bracketed exposure on a Canon DSLR camera, the factory default setting discontinues bracketed exposure every time the camera is shut down. You need to go into the menu to keep bracketed exposure selected until you tell the camera to stop bracketing.

BTW: For most night exposures a combination of Auto Exposure Bracketing at one stop intervals and Exposure Compensation of minus one stop will give you a series of three exposures, one of which is usually in the ballpark for most any night scene...

[george013]

Why do I always think somebody is trying to sell me a second hand car reading those articles?

What's the difference between spot-metering the lightest part of a view and take care that that's placed at the right in the diagram or using a matrix-metering and do the same for the maximum of the view?

What is clipping? Yes, the pixel-value is getting a maximal or minimal value. But that value s not always the same with different software or camera's.

Although I see the benefits of ETTR, I totally disagree with some arguments used to come to that concept. And off course you need some space available in your histogram.

I don't know what happened with the photo of Allen(polar01). He used matrix metering and an EV +2. To me the idea of ETTR and other concepts like that is to overexpose if possible and correct that in PP. So the PP is en elementary part of the idea. In Allen's photo the blacks became grey. Ok, with details.

George

[jcuknz]

I am a right-wing/socialist so I am always right
But seriously I read "exposing to the right' when I found it some time after it came out and thereafter set my camera to under expose by one stop, first thing I do with each new camera I get. If it still blinks at me I decide if I am bothered about those areas being washed out ....I did shoot RAW plus JPGFINE for awhile but I notice somehow the camera has reverted to JPGFINE so I am leaving it at that
My monitor is uncalibrated so I never comment on washed out photos as I cannot reduce its density ... my photos look good on it and elsewhere so I really do not know what to think

To Post #2 I added a curves adjustment layer and moved the bottom to the first quarter and then moved the centre point slightly to the right .. IMO As the people are the important aspect of the shot they are better.

[george013]

I am a right-wing/socialist so I am always right
But seriously I read "exposing to the right' when I found it some time after it came out and thereafter set my camera to under expose by one stop, first thing I do with each new camera I get. If it still blinks at me I decide if I am bothered about those areas being washed out ....I did shoot RAW plus JPGFINE for awhile but I notice somehow the camera has reverted to JPGFINE so I am leaving it at that
My monitor is uncalibrated so I never comment on washed out photos as I cannot reduce its density ... my photos look good on it and elsewhere so I really do not know what to think

To Post #2 I added a curves adjustment layer and moved the bottom to the first quarter and then moved the centre point slightly to the right .. IMO As the people are the important aspect of the shot they are better.

As a left-wing socialist I say red is beautiful.

All the camera's I had where standard -0.67. It has to do with the light at day-time and the callibration of the lightmeter.

George

[Inkanyezi]

Why do I always think somebody is trying to sell me a second hand car reading those articles?

My guess is that part of your conclusion is that it's a very similar thing going on. Someone is trying to sell an idea. The problem is that the idea is not explained very clearly, or maybe sometimes that the problem of understanding is in the other end.

What's the difference between spot-metering the lightest part of a view and take care that that's placed at the right in the diagram or using a matrix-metering and do the same for the maximum of the view?

This difference is huge. Using matrix metering in order to expose the RAW file to the right is impossible, if the matrix metering does not have a way to assess the brightest part of the scene. So in essence, ETTR can generally not be done by using matrix metering, while with spot metering, there is a possibility to do that -- provided you have a fairly neutral highlight to measure. It works perfectly with clouds on a sunny day, while a field of poppies needs special attention, because your meter reading will be off.

Colour is a special problem for the light meter. There are essentially two ways of assessing light level for TTL metering, one integrated for all colours, and the other measuring the three primary colours separately. Most digital cameras of today measure the three colours as a matrix, but integrate them to average the value, which makes it impossible to assess saturation of a single colour. Hence, bracketing might be a sensible idea when you take a reading from a saturated colour, or you could use your experience to make a qualified guess.

One workaround for the problem of saturated colours is to measure incident light. But it will not work for neon signs or traffic lights, and incident light measurement will not provide "optimum" exposure according to the idea of ETTR always. When the whole scene is very dark or bright, an incident reading will suggest a sub-optimal exposure according to the ETTR idea.

Essentially though, there is no light meter that will provide a correct value for a saturated primary colour. A colour that is saturated, in absence of the other two, reaching the maximum value, will only be represented as a third of that value by the meter. If you use that reading without correction, it will be hopelessly blown out.

What is clipping? Yes, the pixel-value is getting a maximal or minimal value. But that value s not always the same with different software or camera's.

No, there is never any clipping to the left, beyond the minimal value, there is no clipping. You should discard the idea of clipping to the left, because what takes place there is different from clipping.

It is also independent of the software as long as the RAW data is concerned, although the sensor of the camera determines the value where clipping will take place. Clipping takes place when the A/D data reaches the maximum possible value.

Clipping only takes part on the right side of the histogram. There are two different views on clipping. Where the jpeg clips, is not necessarily (or usually) where the RAW data hits the roof, but the jpeg clipping largely depends on the software algorithms. Mostly algorithms make the jpeg clip much before than RAW data clips. The idea of exposing to the right is built upon this premise.

The left side is a bit different. While there is an absolute maximum value for any digital file, depending on the number of bits, and also an absolute maximum analogue voltage that can be stored in a sensel, although there is a minimum voltage and a minimum digital value, it is not related to the light levels of the darkest parts of the image, and there is not an absolute limit at the left side for values that can be extracted.

There is however a practical limit, which depends on how much noise you might accept. Noise is always present. In the brighter values, it is drowned in image data, while in the darkest parts, noise will drown the image data. This is also a foundation for the philosophy of exposing to the right. More image data in the darker parts of the picture will discern more detail above the "noise threshold".

Although I see the benefits of ETTR, I totally disagree with some arguments used to come to that concept. And off course you need some space available in your histogram.

What histogram? The histogram shown on the back of the camera, based on its jpeg conversion? Or do you mean the RAW data histogram, which is an entirely different thing, only available to you when you open the RAW file in your software?

The ETTR idea discards any extra room at the right side of the histogram, implying that it shall end just there, and nowhere else. The left side is left to its own premises in a way, although you may use ISO setting to gain the possibility of a faster shutter speed, accepting a higher noise threshold. The ETTR idea is to utilise the available room optimally, placing highlights at the absolute maximum value that the system can register, thus allowing as much space for the darker values as available at the dynamic range chosen by ISO setting.

I don't know what happened with the photo of Allen(polar01). He used matrix metering and an EV +2. To me the idea of ETTR and other concepts like that is to overexpose if possible and correct that in PP. So the PP is en elementary part of the idea. In Allen's photo the blacks became grey. Ok, with details.

The idea of ETTR and "other concepts like that" is NOT to over-expose. The idea is to expose correctly for the workflow that is chosen. ETTR in no way promotes over-exposure. That is a total misconception of the idea. ETTR means that you place the white point when you expose the image.

Much of the concept is confusing because of the fuzzy concept of "exposure" in the software. It is easier to understand if you crap the idea of correcting exposure in PP. Exposure takes place in the camera. Once you have captured your image in the camera, there is no way to change exposure. Your exposure is the exact amount of light you registered with your sensor, defined by your aperture, shutter time and the light from the scene. It will not be altered in post production, it cannot.

[george013]

My guess is that part of your conclusion is that it's a very similar thing going on. Someone is trying to sell an idea. The problem is that the idea is not explained very clearly, or maybe sometimes that the problem of understanding is in the other end.

I stick to the first, bad explanation.

This difference is huge. Using matrix metering in order to expose the RAW file to the right is impossible, if the matrix metering does not have a way to assess the brightest part of the scene. So in essence, ETTR can generally not be done by using matrix metering, while with spot metering, there is a possibility to do that -- provided you have a fairly neutral highlight to measure. It works perfectly with clouds on a sunny day, while a field of poppies needs special attention, because your meter reading will be off.

Either which way you go, exposure affects the whole image. Doing a spotmetering and guess how much I've to correct the exposure using my histogram on the back of the camera is for me the same as using a matrixmetering and look at the histogram how much I have to the right. And the same if you have clipping in the histogram and you must expose less. Your judgement is based on the histogram of your camera at that moment.

What histogram? The histogram shown on the back of the camera, based on its jpeg conversion? Or do you mean the RAW data histogram, which is an entirely different thing, only available to you when you open the RAW file in your software?

At the moment I shoot, that's the only histogram available.

Clipping only takes part on the right side of the histogram. There are two different views on clipping. Where the jpeg clips, is not necessarily (or usually) where the RAW data hits the roof, but the jpeg clipping largely depends on the software algorithms. Mostly algorithms make the jpeg clip much before than RAW data clips. The idea of exposing to the right is built upon this premise.

I don't think that ETTR is based on the maximum possible exposure before clipping. And knowing that pixelvalues are changed with JPG-compression doesn't give me a tool to handle that. I've to stick with the available histogram.

No, there is never any clipping to the left, beyond the minimal value, there is no clipping. You should discard the idea of clipping to the left, because what takes place there is different from clipping

Depends on how you approaches clipping. Call it a value where the pixel can only distinguish itself on one side, highlights can't go higher anymore, lowlights not lower. For the image it is less important that the lowlight is based on the noiselevel and the highlight is based on a saturated pixel.

The idea of ETTR and "other concepts like that" is NOT to over-expose. The idea is to expose correctly for the workflow that is chosen. ETTR in no way promotes over-exposure. That is a total misconception of the idea. ETTR means that you place the white point when you expose the image.

I think we read different articles.

George

[Inkanyezi]

I don't think that ETTR is based on the maximum possible exposure before clipping. And knowing that pixelvalues are changed with JPG-compression doesn't give me a tool to handle that. I've to stick with the available histogram.

The idea of ETTR is indeed based on maximum RAW exposure before clipping in RAW. There is a tool invented to handle that in the camera, UNI-WB. With UNI-WB set in your camera, you get a set of useful histograms of the three colours. The jpeg it presents however is just horrible. People will look like they're related to the Hulk.

Sticking to the available histogram, with white balance in the camera correctly set for the lighting, means you will not use the ETTR strategy, and it is your choice. Then you do not expose to the right, and that's it. I won't object in any way, it's fine with me. Most people won't cling to the idea of ETTR anyway.

[george013]

The idea of ETTR is indeed based on maximum RAW exposure before clipping in RAW. There is a tool invented to handle that in the camera, UNI-WB. With UNI-WB set in your camera, you get a set of useful histograms of the three colours. The jpeg it presents however is just horrible. People will look like they're related to the Hulk.

Sticking to the available histogram, with white balance in the camera correctly set for the lighting, means you will not use the ETTR strategy, and it is your choice. Then you do not expose to the right, and that's it. I won't object in any way, it's fine with me. Most people won't cling to the idea of ETTR anyway.

ETTR is a technique of using space left in your histogram to overexpose the picture, without clipping of course, and than correcting that in PP. Your gain is more captured detail in the blacks, and less noise in the black as I read.

I had to read about Uni-Wb. What I understood is that it is showing the colors without white balance correcting.
http://www.dslrbodies.com/cameras/ca...-is-uniwb.html

George

[bambleweeney]

Sticking to the available histogram, with white balance in the camera correctly set for the lighting, means you will not use the ETTR strategy, and it is your choice.

I'm sorry but I don't understand how nailing my white balance is going to give me an ideal exposure (ETTR not required). I've got a bin full of images where the white balance was correct but the highlights were blown out or the histogram was crammed into it's left hand side.
I spot the brightest part of the scene (often the brightest patch in a cloudy sky) and overexpose the spot reading by two or two and a bit stops. This takes care of the highlights, I'm not sure whether I'd call it ETTR or not. If the shadows are still totally blocked then I'll bracket for those and blend in Photoshop. This is generally the case for shots against the light, otherwise I'll use an incident light meter which gets it right more often than my in camera metering. Colours remain a problem, somebody mentioned a field of poppies which are little buggers for clipping the red channel and like most bright pure bright colours should sound alarm bells and compel you to check the individual histogram channels. I've got one or two poppy shots where the general exposure was spot on but the red channel was clipped. Exposing properly for the reds meant underexposing the rest of the scene a little. Choice then is yours as to where you put your exposure or whether you bracket or not. Each scene is different, the correct exposure depends on your interpretation of the scene, no one rule suits all occasions. ETTR is useful under some circumstances and not in others and personally I'm really (too) fussy about getting it spot on in camera. I get really (too) disappointed when I have to use Photoshop to correct my mistakes before processing the image.

[Inkanyezi]

ETTR is a technique of using space left in your histogram to overexpose the picture, without clipping of course, and than correcting that in PP. Your gain is more captured detail in the blacks, and less noise in the black as I read.

The fundamental flaw in this reasoning is using the term "over-exposing".

How you should best expose depends on your workflow. If you always use the jpeg straight out of the camera, the best strategy cannot be ETTR. Neither will it be the best strategy if you always use the jpeg from the camera, intending to correct it for exposure by editing. Strictly, the ETTR strategy is for RAW shooting and adjusting tonality and colour in the RAW conversion software. When using the jpeg from the camera, it is a matter of judgement whether you should allow highlights to clip.

A RAW file that is exposed to the right will need another procedure for conversion, compared to a file that is exposed for best jpeg rendition in the camera. They mostly will not be exposed the same, as the ETTR file in many cases would be exposed somewhat more, ranging from about one and up to about three stops. In rare cases, the ETTR file will be less exposed than the SOOC file. Those rare cases always include at least one saturated colour, mostly red.

Now in practical photography, most people don't see a need for ETTR, as their camera usually has sufficient dynamic range for most scenes, and they accept noise in the shadows. Fine! All of us don't share the same beliefs and ideas, and there are different ways of using the available dynamic range of our image chips. Noise in the shadows can be allowed, clipped highlights as well.

Some photographers, maybe most of them shooting landscapes, want to make the most out of the image quality their camera can give them. If there's a possibility to get more shadow detail without noise, they can use ETTR to achieve that goal. If at the same time they don't want to clip the highlights, ETTR is definitely a way to achieve both of those requirements when possible.

But exposing to the right is not to over-expose. It is to expose correctly, for a particular workflow, where the image is adjusted in conversion to the tonality you want it to have. If your workflow is different, ETTR is not for you.

[chauncey]

Any method of ETTR is inherently good providing that one has an understanding of the process...
the goal being to capture as much dynamic range as that camera is capable of and then to expand
that dynamic range in PP.

It's a no-brainer if you shoot Canon, it has a displayed histogram in live view mode, making that
process a whole heap easier.

[Manfred M]

Janice - I hate when budding authors promise you the world will be a better place when you read their book. I find it especially troubling when the budding author treats his potential readers as morons and misrepresents the facts (or at least uses examples that are bound to show his or her point of view). In tricky situations, I will shoot what my histograms, rather than what my light meter tells me. When doing portraits and object shots, even outdoors, I will often use an incident light meter, which gives me far more accurate readings than my camera's built in light meter.

Do I use ETTR, yes. Do I do so often, NO!. In fact I use it quite seldom. That is because most of my images have a decent distribution of dark to light tones and moving my exposure to the right will simply cause me to blow out my highlights.

ETTR does make sense when the tonal range is biased towards the dark tones, especially when shooting at higher ISO values. When I did some testing at low ISO (my mantra regarding ISO is "shoot as low as you can go"), I found no difference when I used this technique. At higher ISO values, especially when pushing at the limits, yes, it does reduce the amount of digital noise, even with a modern camera.