ETTR post processing

I've been trying to follow the threads on "exposing to the right" to limit noise, but stumble when it comes to post processing.

I use Canon DPP and Photoshop CS2 and just don't know how to recover just the highlights without a lot of work in layers. It appears that in other software (ACR?) there is something called a "Luminance" slider which does this but I don't have this in CS2.

Does either DPP or CS2 have anything to save me spending days on a single image (as I do now), or must I resign myself to it?

I'm happy with the quality of the results using my present labour intensive methods, but would love to make life easier.

First thing you want to do is lower the exposure in raw conversion to expose the highlights properly. Then the question is, how do you alter gamma and black point to balance the mid and shadow tones at the new exposure? See, you spent all of the ETTR effort to get better shadows, so they can stand up to this kind of manipulation. Why try to "recover" highlights that are actually present in the data, when you can simply lower the exposure to see the highlights properly? Otherwise, you start getting unnatural looking tonemapping in the highlights.

kirk

Kirk,

The problem is that pushing the exposure to the right might recover some shadow detail, but its at the cost of the now over exposed highlights. They need to be recovered, which I currently do by splitting the highlights from the darker areas in an image and working with them on separate layers before recombining them. And its the recombining which is taking time, hours of painstaking, detailed work.

In articles and threads here, people glibly talk about "recovering the highlights" as if its a 10 second job.

Well, it is supposed to be a 10 second job.

We mean ETTR to use the exposure settings of the camera, and the exposure settings in RAW conversion literally.

Add 1 stop in camera, remove it in post, using minus 1 exposure on the exposure slider in any RAW converter.
If the highlights were blown during the shoot, then you went to far to the right.

But the general idea, as Kirk points out, is not to try and adjust only part of the tone, but the entire RAW exposure.

The first article on this;
http://www.luminous-landscape.com/tutorial​s/expose-right.shtml

this one shows a solid example of it working;
http://schewephoto.com​/ETTR/

Another good one about exposure in digital in general;
http://www.cambridgein​colour.com …l-exposure-techniques.htm

A a strong counter argument that you might find of interest here;
http://chromasoft.blog​spot.com …-is-just-plain-wrong.html

The Highlight, Shadow, White, and Black sliders introduced in ACR7.x (CS6) work so much better than the old Recovery, Fill Light, Blacks, and Brightness did.

The sliders may work better, but the image data is no longer intact in its original tonal relationship once you start using "recovery" to recover things you do not need to actually recover. When you use exposure and recovery in Lightroom/ACR, the sliders to which you are presumably referring, it is doing things to your image data without really telling you - in PV2012 even the exposure slider is building in some sort of roll off toward the highlights, so that clipping seems to disappear magically in LR when, in other converters, clipping will be present until you lower the exposure.

Try a couple of your ETTR images in Lr/ACR with PV2010 and PV2012 - set up the image in PV2012 so that there is no clipping and then switch to PV2010 and set the same exposure - clipping will likely be indicated. It is the same raw file and you did nothing to it, so something funky is going on under the hood without your explicitly being able to control it. PV2012 is also making adjustments differently to each image, as the tonal controls adapt to a certain extent to the image content. How can this help the user be able to use the controls in a predictable way?

See:

http://www.adobepress.​com …les/article.asp​?p=1930485

In the above linked blurb by Martin Evening on the Exposure control in PV2012, his ultimate advice is not to worry too much about clipping using the exposure control in PV2012, but use the exposure control to get the image to be the correct "brightness" - you now use the Exposure Slider to SET THE MIDTONES. This seems pretty loosey goosey advice, but what other advice can he give with a tool whose behavior is not predictable in any specific way other than by eye, on a case-by-case basis. If it is working for you, that's great - just be aware that LR/ACR is making moves on your highlight data (and your image data in general) that are not explicitly revealed to, or controlled by, you.


With respect to the extreme ETTR example posted on Schewe's site in the above link (the image of Niagara Falls) - that the data was able to be rescued (recovered) is made dramatically clear - however, that data is of pretty low quality unless you like your water to have neon green and blue in it. So, ETTR is useful, but it cannot perform miracles.

kirk

Here's a quick example of the visual difference between boosting exposure in PV2012 versus 2010. The image is a Sekonic gray target with each step in the gray patches being 0.3 EV apart. The scene was incident light metered prior to shooting. The target is not perfectly uniformly lit, so take each gray step with a grain of salt (look around the black border of the target, you will see an observable difference in tonal level of the border).

Each PV image was generated in ACR 7.4.0 (CS6) with default settings and Adobe Standard camera profile. Which way does your camera work if you add 2 EV to the exposure?

kirk

I just Googled "PV2012" The first hit was a Victoria Bampton page, that gave me more of an indication of why this is so much more complex in LR.
http://www.lightroomqu​een.com …8/pv2012-develop-sliders/

I'm saying it is 10 seconds, but that's in a program like DPP, where you have the one slider for exposure.

kirkt wrote in post #16017091
Which way does your camera work if you add 2 EV to the exposure?

Depends on what film and development I'm using

CyberDyneSystems wrote in post #16016678
Well, it is supposed to be a 10 second job.

We mean ETTR to use the exposure settings of the camera, and the exposure settings in RAW conversion literally.

Add 1 stop in camera, remove it in post, using minus 1 exposure on the exposure slider in any RAW converter.
If the highlights were blown during the shoot, then you went to far to the right.

But the general idea, as Kirk points out, is not to try and adjust only part of the tone, but the entire RAW exposure.

Then I've completely misunderstood the process. By overexposing I am automatically taking the highlights, in the above example, 1 stop too high. And that needs pulling back. What am I not grasping?

You overexpose the image, but without blowing out (important) highlight detail.

Lowner wrote in post #16017368
Then I've completely misunderstood the process. By overexposing I am automatically taking the highlights, in the above example, 1 stop too high. And that needs pulling back. What am I not grasping?

Select a neutral camera profile, turn contrast, sharpening and saturation way down, and then turn on your highlight warning. When you get the blinkies, back off till they go away (unless it's a hightlight that can or should be blown).

Your camera has "headroom" in the highlights - that is, there is useable data in the raw highlights that may appear to be clipped on the JPEG preview on the back of your camera, or the JPEG-based histogram and blinkies on the clipping warning on the camera display. You can take a series of images of a scene with detail in the highlight areas, with a dynamic range that fits in your camera's sensor. You progressively overexpose one or two stops (in 1/3 or 1/2 EV increments, for example) above the metering method you choose. Then, take those raw files into your raw converter of choice and turn on the clipping indicator. For the particular color space that you are using to convert into, the highlight warning will likely indicate some degree of clipping. Back off the exposure slider until clipping disappears. If the image was overexposed "correctly" then the highlight data should be intact - if the image was overexposed too much, you will see a mess in the areas for which data exceeded the headroom of the camera. Once you process a few sets of these images, you will learn about how much you can overexpose and still pull that overexposure back down in raw conversion to get the highlight data.

The "camera style" and metering style that you use may often dictate how "overexposed" your image is relative to how much you are able to pull the image in raw conversion. Pick a style and metering method that works for you and then find that relationship between the camera's warnings and how much exposure you can pull back in raw conversion. It is typically 1 to 2 stops.

kirk

René Damkot wrote in post #16017325
Depends on what film and development I'm using

The PV2012 appears to be building in a film-like curve but without really making it accessible.