Expose to right, it is as simple as

[dreed]

stamper, yes, there are contrast curves applied by default (*), but those curves are applied after the lightness-dependent portion of the color profile.  The Exposure control, on the other hand, is applied before the lightness-dependent portion of the color profile.  This means that if you've used ETTR in the camera or bracketed exposures in the camera, and want to normalize the results afterwards, you should use the Exposure slider because that will get the image values in the appropriate tonal range before the 3D color table is applied.

Riddle me this: why do ACR/LR appear to ignore the "exposure compensation" setting (-1,+2/3,etc) in the raw file? (Well, I might add that at least Canon cameras store this in the CR2 file, I don't know about other vendors.)

Why do I ask this? Because an application of "Auto Tone" in LR almost never moves the exposure compensation like you are suggesting above. Although maybe I'm thinking that "Auto Tone" should be "Auto Expose" of the digital negative?

[Pelao]

Sort of.

You may want to increase exposure up to about 3 stops, or not at all. It all depends on the scene. The important point is to place as much of the data to the right of the histogram, but without clipping any important highlights. The problem (until manufacturers automate this) is that the histogram on your LCD is not based on the real raw data in its very large colour space. This means it's going to yell "clipping" when you still have extra headroom available.

For this reason it's worth doing some testing to see what your particular camera's characteristics are. Keep adding a third of a stop well into clipping (on the LCD histogram) and then look at the files in your raw converter and see when they really do clip. This will tell you how much extra headroom you have.

You can also set your camera to autobracket.

Michael

I remember reading a post somewhere, I think it might have been LL, where there was specific advice on settings for the LCD to help compensate for this difference. Basically, it made the LCD image look really like something you would reject, but helped offer a more accurate histogram.

[digitaldog]

I remember reading a post somewhere, I think it might have been LL, where there was specific advice on settings for the LCD to help compensate for this difference.

You can compensate a tad to get a bit closer but its still a mile off.

[bjanes]

It seems to me that Michael is giving us the recipe for a metaphorical pie, while the engineers are telling us how to make flour and grow cherries. If I want to make a pie, I need a recipe, not a treatise on milling. I think Michael made the recipe pretty understandable, and it allows me to function better as a photographer. Throwing a lot of engineering stuff at me does not, because not only do I not understand it, I don't really care to. That's what engineers are for. No offense.

John,

A certain amount of technical understanding is needed to make the best use of your camera. The term ETTR is somewhat misleading, since it suggests that a histogram to the right will give optimal results. It is the number of photons collected that determines the shot noise, which is the primary determinant of SNR in digital cameras. Read noise becomes important in determining noise floor from which dynamic range is determined. I think that we nearly all agree that for a given ISO, the histogram should be to the right. At base ISO an accurate histogram to the right will indicate that the sensor is near saturation and one will obtain optimal SNR and DR. However, what about histograms above base ISO?

When shutter speed and f/stop considerations indicate that the histogram is not to the right, one can increase the ISO setting on the camera to increase the amplifier gain and histogram will move to the right, but exposure will not change because of the above mentioned restraints. An ideal digital capture will lose 1 stop of DR for each doubling of ISO, since only half the number of photons are collected for each doubling of ISO. However, many older digital cameras have less than ideal electronics and read noise decreases with increasing ISO. and the slope of the DR vs ISO curve may be less than 1/2 at low ISOs. This is primarily because of increased read noise at low ISO. Increasing the ISO will decrease the read noise and give a lower noise floor, but there reaches a point where increasing the ISO has a minimal effect on read noise. Beyond this point, increasing the ISO will give a better histogram but will do nothing for dynamic range, but will merely limit highlight headroom and increase the chance of blowing highlights.

One can use the dynamic range vs ISO plot on DXO to determine this point of diminishing returns. When the curve becomes linear, increasing ISO will have minimal effect on read noise. As shown on the plot below, this occurs at about a measured ISO of 700 for the Nikon D3 (the ISO indicated by the camera will be higher with this camera). With this camera there is really no point in increasing the ISO above 800, since you will only decrease highlight head room. Expose as much as possible at a camera ISO of 800 and use the raw converter to increase exposure.

With the D7000, the curve is linear from base ISO and there is no real need to increase the ISO to obtain a "better" histogram. One can expose as much as conditions allow and make up the difference in the raw converter. The appearance of the histogram is irrelevant. The scientific explanation for these considerations is given by Emil Martinec. The article was written before the D7000 was available.

In summary, a bit of engineering theory may make you a better photographer, but you may keep your head in the sand if you so desire.

Regards,

Bill

[hjulenissen]

Isn't the simplest solution to this conundrum two sets of firmware, one for the users who need to have the current jpg emulation since they lack the sophistication to delve deep into how ETTR works and maximizing the image and a second for the professional user who wants these options in the camera menu?  The only problem with camera companies doing this is that the professional base is always going to be significantly smaller than the casual user.  I cannot believe the amount of programming to achieve this would be that significant (but then I'm not a software engineer).  At least this approach would let both parties have their cake to eat.

I think that mainstream DSLRs will continue to have "PictBridge" shortcut buttons, but not functions like this.

The solution seems to be something like remote-controlling your camera from your smart-phone, or some (under-dog?) manufacturer choosing to make available cameras based on something like Googles Android and letting people like LL nerds write simple "apps" that takes control of exposure and simple gui functionality. Only then will the functions that needs 2 weeks of implementation but which caters to <5% of the buyers be available. I look forward to that day. Not because I want to play games on my camera, but because I think that flexibility and openness is a good thing.

-h

[Craig Arnold]

I was lucky enough to get one of the first X100s outside of Japan in early March.

I was very excited when a few weeks later ACR support was added for the camera. Except, as a long time LR user and having shot nothing but RAW with my Canons for the last 7 years - I don't shoot RAW with the Fuji.

The JPG engine with Auto-DR + Auto-ISO switched on is doing something remarkable. I think it is doing exactly what the engineers here are discussing behind the scenes.

I cannot generally even get close to the JPG performance of the Fuji engine when I shoot RAW and process in LR. Or rather lets say, with a fair amount of work on each image I can get equivalent results 50% of the time. 49% I cannot get it as good and maybe 1% I can improve on the JPG by shooting in RAW, usually it's a WB issue. That's why there is a RAW button on the X100, because they had the A-team on the JPG engine and the B-team on the menus.

In auto-DR you never see clipping in the JPG if you are using A-priority or Program mode. Extensive use of the fill light slider is mandatory in the high-DR modes, but there is so little noise in the shadows it's barely believable, this with the NR switched right down in camera.

Hence my suspicion that it's slightly ironic that Michael has used Fuji X100 images in this article because I reckon Fuji has done exactly what he is berating camera manufacturers for not having done all these years. Never a clipped highlight and tons of shadow detail.

[As an aside, there is a significant benefit to shooting in JPG with the X100, which is that the camera never slows down or becomes unresponsive when writing the JPG files to the card, but writing RAW files is slow, even with a fast 45Mb/s card. It's like a different camera really.]

I have a fair number of sample X100 photos on my blog pages. Probably over-processed for the general tastes of this website, and I'm much more an admirer of Bill Brandt than Ansel Adams, but original files available if anyone wants them. Just PM me.

[deejjjaaaa]

Never a clipped highlight

what is that ?

[Adam L]

Craig,

I believe that the increased DR in the Fuji is accomplished by pushing up the ISO.   I say this because if you set DR in the auto setting and ISO in the auto setting the minimum ISO is overwritten based on the DR setting. 

I'll have to compare the jpgs to the raw files.  Honestly I don't look at them at all. 

[douglasf13]

John,

A certain amount of technical understanding is needed to make the best use of your camera. The term ETTR is somewhat misleading, since it suggests that a histogram to the right will give optimal results. It is the number of photons collected that determines the shot noise, which is the primary determinant of SNR in digital cameras. Read noise becomes important in determining noise floor from which dynamic range is determined. I think that we nearly all agree that for a given ISO, the histogram should be to the right. At base ISO an accurate histogram to the right will indicate that the sensor is near saturation and one will obtain optimal SNR and DR. However, what about histograms above base ISO?

When shutter speed and f/stop considerations indicate that the histogram is not to the right, one can increase the ISO setting on the camera to increase the amplifier gain and histogram will move to the right, but exposure will not change because of the above mentioned restraints. An ideal digital capture will lose 1 stop of DR for each doubling of ISO, since only half the number of photons are collected for each doubling of ISO. However, many older digital cameras have less than ideal electronics and read noise decreases with increasing ISO. and the slope of the DR vs ISO curve may be less than 1/2 at low ISOs. This is primarily because of increased read noise at low ISO. Increasing the ISO will decrease the read noise and give a lower noise floor, but there reaches a point where increasing the ISO has a minimal effect on read noise. Beyond this point, increasing the ISO will give a better histogram but will do nothing for dynamic range, but will merely limit highlight headroom and increase the chance of blowing highlights.

One can use the dynamic range vs ISO plot on DXO to determine this point of diminishing returns. When the curve becomes linear, increasing ISO will have minimal effect on read noise. As shown on the plot below, this occurs at about a measured ISO of 700 for the Nikon D3 (the ISO indicated by the camera will be higher with this camera). With this camera there is really no point in increasing the ISO above 800, since you will only decrease highlight head room. Expose as much as possible at a camera ISO of 800 and use the raw converter to increase exposure.

With the D7000, the curve is linear from base ISO and there is no real need to increase the ISO to obtain a "better" histogram. One can expose as much as conditions allow and make up the difference in the raw converter. The appearance of the histogram is irrelevant. The scientific explanation for these considerations is given by Emil Martinec. The article was written before the D7000 was available.

In summary, a bit of engineering theory may make you a better photographer, but you may keep your head in the sand if you so desire.

Regards,

Bill

Exactly, Bill.  The idea of ETTR is becoming more and more nuanced, and IMO outdated, as sensor technology improves, and I think it is oversimplified.  The color issues that happen when you expose mid tones a couple of steps past midpoint, plus the near-ISO-less cameras that we are starting to see on the market really put a bunch of asterisks next to the idea of ETTR.  It is certainly time to start understanding that exposure is independent of ISO, and more recent cameras are better off boosting gain in the raw converter.

[Craig Arnold]

what is that ?

Most of my processed pics have crushed blacks and blown highlights. a.k.a. black and white. All the blame goes to me though, Fuji almost always (glances around nervously) get it perfect.

I'll dig up the original and post it on flickr. But probably not tonight, wasn't planning on firing up the Mac today. It's getting late.

[NikoJorj]

The idea of ETTR is becoming more and more nuanced, and IMO outdated, as sensor technology improves, and I think it is oversimplified.

True, even if as far as I'm concerned I still have difficulties understanding what these hue problems in the midtones really are...

But there will still be one basic truth in ETTR, it's the "expose for the highlights" thing. In digital, clipping siome channel can seldom be recovered, and more often means a lost image.
ETTR means expose border to clipping, but without clipping anything important, and I'd think that's something useful to repeat and important even (or mostly maybe) for the beginners.

[dreed]

Michael,

Will or do any of the upcoming tutorial videos demonstrate the practice of ETTR with post processing in ACR/LR?

[Schewe]

Will or do any of the upcoming tutorial videos demonstrate the practice of ETTR with post processing in ACR/LR?

Yep, in Camera to Print and Screen...

[Schewe]

But there will still be one basic truth in ETTR, it's the "expose for the highlights" thing.

No...you still don't fully grok it. It's expose "properly" based on the scene contrast range and the dynamic range of your sensor. It's also useful to fully understand what your meter is telling you and to know the difference between a scene with low and high contrast.

If the contrast of the scene is beyond the dynamic range of the camera, then you have to choose the relative importance of highlight vs shadow detail and expose to maintain the detail that's important to you...either that, or combine multiple exposures to get the final contrast range you want. Don't forget that ETTR also applies when you do HDR combos because you'll generally want to use the lighter exposures in the deep shadows and darker exposure for the highlights.

[Graeme Nattress]

Schewe's point is important. Often people will take ETTR too literally and protect even specular highlights from clipping and under-expose enough to plunge the shadows into the noise floor - sacrificing the meat of the scene for highlights that don't really have any detail in them anyhow....

Graeme

[Anders_HK]

Even more ridiculous is that an MF digital back like my Hassy one can only shoot in RAW, but I believe the histogram you see on the LCD is derived not from the RAW but from the JPEG preview. How dumb is that?

Nope, likely Hassy works same way as does Leica S2, Leaf and Phase One and derives the histogram from the raw data. See here http://www.luminous-landscape.com/forum/index.php?topic=49859.0

[Anders_HK]

To further my post #36 above there are in essence two means to meter precisely for ETTR without trial and error. Michael’s is for automation. The other is using a spot meter when knowing (having predetermined) the sensor’s precise limits of DR and simply knowing how many stops above and below mid tone it is capable to capture. The mid tone is because that is what we set aperture and shutter to on lens/camera. Thus when spot metering one would not necessarily need to refer to a histogram since one can already visualize the scenes by measuring tonal values using the spot meter (same as Ansel Adams did). The two methods are for different purposes, same as different types of metering (scene vs. spot) have been used for film capture for years. The “new” is that they need to be applied to the media used, thus to digital as compared to B&W, slides, Polaroid etc, all of which have different DR.

Has anyone used this or something like it.  It says you can shoot a series of exposures to determine the DR of your camera.  If you knew that, couldn't you exposure compensate more accurately?  Wouldn't it even calculate the exposure settings for you?

http://www.sekonic.com/Products/L-758DR/Overview.aspx

Have not used it, but per http://www.sekonic.com/Portals/0/Products/Sekonic_Brochure.pdf

"The Sekonic L-758dr incorporates a precise 1-degree, reflected light spot meter."
"The key to working with the L-758dr’s spot meter is knowing the dynamic range of your camera. That is, the tonal range your camera is capable of recording. Program this into the L-758dr and you can easily know which details will be properly imaged and which will be over or under the range of your camera."

It reads as it functions to give a programmed guide using the limit points of the DR capability of the sensor. What Sekonic enables to determine are the upper and lower clip points from Ansel Adam’s zone system using their ($$) calibration target:

  Zone 0 = Pure black (no detail)
   Zone X = Pure white: light sources and specular reflections (blown out)

As is e.g. described in Ansel’s books also a concrete wall or similar can be used as a neutral target and using the meter to determine +1, +2 etc test exposures. Ansel determined intermediate zones at each stop in order to help him precisely visualize the image at capture. Thereby the zone system enabled him to at capture know how he could best expose for achieving the optimum print and for the adjustments he would later make during the printing process (or as with digital we do during processing). I pointed out zones I, II, VIII and IX as valuable to digital in my post above because those are the ones that have information (as compared to 0 or 255 which have no texture or tonal info).

  Zone I = Near black, with slight tonality but no texture
   Zone II = Textured black; the darkest part of the image in which slight detail is recorded
   Zone VIII = Lightest tone with texture: textured snow
   Zone IX = Slight tone without texture; glaring snow

In contrast to the Sekonic L-758DR the now old Pentax digital spot meter works fine and is far more simpler.  

Regards
Anders

[ErikKaffehr]

Hi,

Nick Rains just posted on this discussion that Leica S2 derives histogram from JPEG, he should have good info, as he is at Leica factory right now and talking to their engineers.

Best regards
Erik

Nope, likely Hassy works same way as does Leica S2, Leaf and Phase One and derives the histogram from the raw data. See here http://www.luminous-landscape.com/forum/index.php?topic=49859.0

[ErikKaffehr]

Hi,

This is a good point. It also shows a limitation of the histogram, namely that you don't know what part of the image parts of the histogram correspond to. Sony/Minolta has also blinking highlights.

Anders "HK" has a good suggestion to use spotmeter and overexpose a predetermined number of stops on highlight detail. I used to that when I shot slide film. Negative film was exposed for shadows while slide film was exposed for highlights.

With digital we want to have maximum information and that really means that we can capture as many photons as possible.

On the other hand, there is no reason to overdo! I don't think that exposure differences below one stop really affect shot noise that much and mothern cameras seem to be pretty good at keeping read noise unvisible.

Best regards
Erik

Schewe's point is important. Often people will take ETTR too literally and protect even specular highlights from clipping and under-expose enough to plunge the shadows into the noise floor - sacrificing the meat of the scene for highlights that don't really have any detail in them anyhow....

Graeme

[John Camp]

John,
A certain amount of technical understanding is needed to make the best use of your camera. <big snip> In summary, a bit of engineering theory may make you a better photographer, but you may keep your head in the sand if you so desire.
Regards,

Bill

No, it won't make *my* photography any better -- I concede that it may make yours better. When I'm shooting, I don't have time to think about that stuff, because I'm thinking about too many other things. If all the engineers and Michael get together and decide Michael's article wasn't accurate, I'd be extremely interested in knowing that. What I then want from them is a recipe, or a prescription, or rules-of-thumb, for what is roughly, probably, maybe, the best practice under a set of given conditions, in and out of trees and hot sunlight, into heavily shaded doorways, through windows into the street, and from the street through windows to the inside. I don't expect the rules to give me perfect exposures, or even the best possible under the conditions, I just want them to be very good. I use automatic settings and autofocus a lot, because sometimes that's the best I can do with a reasonable chance of success. When possible, I'd like to go to the best fast manual set-up I can get, to see if I can kick up the quality a notch. I use rough ETTR rules some of the time, using ISOs of 160-800 (RAW) on several Panasonic m4/3 cameras with zoom lenses and a Pentax K5 with pancake primes and exposing to the right, as had been recommended here, and trying to get back as much as I can in Lightroom. That has actually seemed to me to work, though my shots fail a lot because of camera shake, poor framing, inaccurate focus and so on. If the engineers here would like to give me a set of rules-of-thumb (which is what I've done with Michael's articles -- I've boiled them down to rules) I would be very interested in seeing them, and grateful to anyone who'd provide them. That's about the most I can work with. I really don't have time for analysis or calculation, or even a lot of rules. Five rules would be about as many as I can handle; any more than that, and they'd slow me down too much.
 
JC