Exposing to the right, or ETTR, is an approach to photography that is as helpful as it is controversial. On one hand, exposing to the right is yet another technique to remember while shooting, and it can potentially ruin your exposure if utilized incorrectly. On the other hand, at least in theory, ETTR is the epitome of digital exposure. With proper ETTR, your images have as much detail in the shadows as they possibly can, without any of the highlights losing information along the way.
It is important to know that this article is not for beginners. If you do not fully understand the fundamentals of exposure, then this this article will be more confusing than helpful. But if you already know the basics and you are looking to get the highest-quality images possible, ETTR is something that could benefit your photography.
1) Overview
The basis of ETTR is simple: optimize your exposure, and get the highest-quality image possible. Most people expose a scene so that the image looks how they want — and, at face value, this makes sense. Optimum exposure is different, though. Instead of exposing the scene “correctly”, it is better to expose a scene to be as bright as possible, without blowing out the scene’s highlights and losing all of that data. Then, in post-processing, you darken the image so that it looks how you want.
By darkening an exposure in post-processing, you are effectively using a lower-than-base ISO. It brings similar benefits, too — a decrease in image noise, richer colors, and a greater dynamic range. Images exposed using ETTR are more malleable in post-processing, making it easier to produce the photo you have in your mind’s eye.
Look at the histograms below (from the photo of the columns at the top of the page):
The first histogram is the “correct” exposure for the columns photo — everything is at the brightness level that I want. The second histogram, though, is from the ideal exposure. It can be darkened to match the histogram on the top, and none of the highlights in the image are completely white. As such, the second histogram contains more details than the first.
2) Image Quality Comparisons
Of course, ETTR is only useful if it has noticeable effects on an image. The comparison below shows an example of the difference between a matrix-metered exposure and an image with +1.3 exposure compensation, shooting a relatively low-contrast scene at ISO 100 on a D800e.
Now, when I reduce the second image’s exposure in Lightroom, the photos look essentially the same:
However, the second image has some more detail than the one on the left, at least in theory. With contrast, sharpening, and vibrance changes in Lightroom, the differences are visible in the crop below:
The above photos show why there is so much controversy behind ETTR. Yes, there are differences in the images above — the crop on the left is certainly noisier than the one on the right — but this is a 100% crop. If the images above are four inches (10 cm) wide on your screen, they would be crops from a six-foot wide print (about two meters).
In fact, the differences between the images are all but invisible on a print that is less than two feet wide. For many photographers, the image quality gains simply are not worth the hassle of exposing to the right.
Still, there are clear differences between the two exposures above. Photographers worry so much about capturing optimum data in the first place, and exposing to the right undeniably does improve the resulting image. In this scene, there was only a difference of 1.3 EV between the metered exposure and the ETTR exposure — in scenes where this difference is greater, the improvement from the ETTR shot will also be more obvious. And, if you ever print large or use an older-generation camera, these differences start to become more relevant.
3) Putting it into Practice
The hardest part about ETTR is that you must be careful to avoid exposing too far to the right. The danger is that you accidentally render an image’s highlights completely white (which means there is zero / no data there), even if you intend to shoot a darker exposure. And, although it is fairly easy to recover dark shadows in post-processing, it is nearly impossible to recover completely blown out highlights. This is what turns many people away from using ETTR, which is unfortunate. There are a few ways to determine the proper ETTR exposure and still avoid blowing out the highlights.
One way to do so is to look at the photo’s histogram when reviewing the image in-camera. Your goal is to take the brightest photo possible that does not push too far to the right of the histogram, as shown below.
Unfortunately, the histogram on your camera is not as accurate as it looks. Current cameras are incapable of showing the RAW histogram of an image, even if you shoot in RAW (which you should, if you use ETTR). Instead, the histogram is based on the processed JPEG image that is embedded into RAW files. This means that although the camera might indicate that you have pushed your exposure too far, there is potentially more headroom for recovery in post-processing.
There is truly no reason for camera companies to avoid implementing a RAW histogram option, which has been one of the most-requested features amongst landscape and studio professionals for more than a decade. At this point, it is truly absurd that photographers cannot judge their RAW histograms until loading their photos onto a computer (see Iliah Borg’s excellent article on culling RAW images vs JPEG). The first mainstream camera manufacturer to get this right will earn a lot of respect in my book.
Still, although the camera-generated histogram is not completely accurate, it does serve as a useful guide for what the proper ETTR exposure could be. And, as I describe later, there is one (albeit cumbersome) way to fix the inaccurate histogram issue.
4) ETTR Method One
One method of aiming for a proper ETTR image is to take a photo, then analyze its histogram in-camera. If none of the histogram is bunching up against the right, you can increase the exposure. Stop increasing the exposure once the histogram says that your highlights are blown-out to white, and you’re good.
Instead of a histogram, it is also possible to use so-called “blinkies” to judge your image’s exposure. With this option enabled, the pure-white areas of your photo will blink between white and another color — typically black or red, but it depends upon your camera.
Blinkies are easier to notice than a histogram, but they are a more limited method of judging your exposure because they only tell you one value: white. Blinkies are valuable, just not to the same extent as a histogram is. Plus, just like histograms, blinkies are based off the JPEG that is embedded in the RAW file, not the RAW data itself.
Using the JPEG embedded in a RAW file certainly isn’t the perfect method of judging exposure, whether you use a histogram or blinkies. Still, it is perhaps the easiest ETTR method to implement, since it requires almost no extra work. Plus, it still leads to a better exposure than blindly using the matrix meter without any compensation.
5) ETTR Method Two
The other method is a bit more complicated, and it requires that you do some testing. First, change your metering mode to spot metering, then find the brightest part of the scene that has to retain detail. If you test your camera beforehand, you will know the amount that you can increase exposure compensation by and still be able to recover information from this area in post-processing. For example, with my Nikon D800E, I can increase exposure at least by +2.3 EV above spot meter’s recommendation for the brightest part of the image.
If you use this method, you don’t need to rely on the less-than-accurate histogram in your camera, although it puts more weight on your personal ability to judge the brightest part of a scene.
The first step is to find how much brighter you can expose your image compared to the spot meter’s recommendation. Set the camera to manual mode and spot metering. Point the metering point on the brightest part of the scene, and record the exposure that the camera tells you to set. Then, take several photos, ranging from +1/3 EV to +4 EV. Open the images in your image editor, then darken them all to match the spot-metered exposure. Check the exposure compensation of the brightest photo that still retains highlight detail — with my D800E, it was +3 exposure compensation.
However, to give yourself some breathing room, you don’t want to choose the value that is right on the limit. It is best to decide on an exposure compensation that is -1/3 or -2/3 from the value you just found. In my case, I decided that a +2.3 exposure compensation is ideal.
To use ETTR from now on, shoot in manual mode with spot metering, permanently using the compensation that you just found. Scan your photo with the spot meter, and manually set the exposure to render the brightest area at what the camera says is 0 EV (since it is automatically factoring in your positive exposure compensation). Congratulations — your images are now exposed to the right.
6) Bracketing
Regardless of the method you choose, it is always a good idea to bracket your scene if you are exposing to the right. I recommend that you shoot one image that is 2/3 EV more than your standard ETTR exposure, and one that is 2/3 EV less. If you are implementing the second method of ETTR (assuming that your “standard” compensation is, like mine, +2.3 EV), your bracketed images would be +1.7, +2.3, and +3 EV.
It is important to have a lower exposure than your normal value, in case you accidentally exposed too far to the right. On the other hand, you may find that your +2/3 EV exposure did not blow out any highlights, and you can use it instead of your standard ETTR exposure.
7) ETTR with High-Contrast Scenes
A common misconception is that ETTR says a photographer should always use positive exposure compensation, meaning that the resulting photo is brighter than what the camera’s matrix meter says. Often, this is not the case. For example, see the image below:
The extreme contrast in this scene fooled my camera’s matrix meter into suggesting a much brighter exposure than ideal. I took this image with a -2/3 exposure compensation, and some of the highlights are still blown out (especially along the left-hand side of the tree in the middle). Using ETTR with this image, I actually should have shot at one stop less than what the matrix meter recommended.
The lower a scene’s contrast, the more likely it becomes that your camera will recommend an exposure that is too dark — one that does not contain all the possible detail of an ETTR exposure. However, with a contrasty scene like the one above, your camera could do the opposite, recommending an exposure that is brighter than ideal.
8) What is UniWB?
As discussed above, the histogram that you see on the back of the camera is not technically accurate — often times, a highlight that your histogram shows as pure white can actually be recovered in post-processing. If you can’t trust the histogram, it becomes much more difficult to know if your ETTR exposure is ideal until you open the photo on your computer.
Compounding the issue, modern cameras have several “picture control” options for JPEG shooters (or “Picture Styles” for Canon cameras). Since the camera’s histogram is based off of the JPEG preview, each of the different picture controls will show a different histogram, even when you are shooting RAW. Plus, your white balance setting will also affect the histogram you see, even though the white balance setting does not affect a RAW file in a destructive way. The complicated part is that none of these histograms is actually the same as the RAW histogram — the one that you actually want to see.
In an effort to fix this dilemma, Iliah Borg created UniWB, or unity white balance. This white balance makes it easier to see an accurate histogram in-camera, but it creates its own set of problems along the way. To get an accurate RAW histogram, your only option is to load a UniWB on your camera and to implement a completely flat tone curve — in other words, settings with no contrast whatsoever.
The only issue with UniWB and a flat tone curve is that they make the image on the back of your screen look horribly green and completely free from contrast.
This problem is important for me, and many other photographers, because the image on the back of the screen is a crucial aspect of my shooting style. Especially while I am photographing landscapes, I pay special attention to the image on the camera’s screen, and I adjust my composition according to how the objects in the photo look. If everything on the screen is green and has no contrast, I cannot judge an image’s success until I see it on the computer. If your method is like mine, using UniWB and a flat tone curve are probably not good options for you.
However, if you do want to implement a UniWB and a flat tone curve, be warned that your camera is fighting against you. All of the “background” JPEG settings that normally make no difference suddenly become important, since they affect how your histogram appears.
For starters, you must choose the lowest-contrast profile within the camera. This means using the Picture Control / Picture Style “Neutral” (or “Flat” on some of the newest Nikon cameras). Then, you need to lower the contrast and saturation as much as possible. Set sharpening to zero, turn off any automatic corrections (especially vignetting control), and set your color space to AdobeRGB. Although RAW files do not inherently have a color space, AdobeRGB is better than sRGB at showing a more accurate histogram. Also, turn off that silly Active D-Lighting, if you still have it on — not only does it mess with the histogram, but it also affects the exposure your meter recommends (yes, even in RAW mode). All of the other special effect modes that your camera has should also be turned off as well, as always. None of these settings will matter once you import the file into a RAW editing program, but they still affect how the histogram appears on the camera screen.
That is actually the point of UniWB in the first place — setting a white balance that does not falsely affect the appearance of the histogram. If the white balance were too warm, for example, a resulting histogram could show the red channel as blown-out when it actually still contained usable information. The “exposure-neutral” white balance, UniWB, just so happens to look extremely green.
Again, if camera manufacturers implemented a simple RAW histogram feature (or, if it’s not too crazy to ask, a different metering mode for RAW exposures), none of this would be necessary to begin with. Until that day, though, we have to work with the tools we have. In this case, that means that we are stuck with clumsy workarounds until the camera manufacturers decide to fix this unnecessary issue.
On that same note, if you want the UniWB file specific to your camera, you need to find it online from a kind individual who has taken the time to create one. It would be nice if a camera manufacturer provided these files on their websites, but I guess that would be admitting that some people need to use RAW histograms in the first place.
For the record, I do not use UniWB or a flat tone curve — instead, I use one of the two other ETTR methods that I outlined above (generally method two). They aren’t the most scientifically accurate ways to expose to the right, but at least they don’t turn my camera screen green.
9) When to Avoid ETTR
In theory, ETTR works with every possible scene. There will always be a “best” exposure for an image — one which is as bright as possible without blowing out any of the highlights. However, in practice, this is not always true.
If you aren’t shooting at your camera’s base ISO, ETTR is all but useless. For example, you wouldn’t want to shoot a scene at ISO 1600 and then decrease the exposure by one stop in Lightroom — it’s just as good to shoot the scene at ISO 800 in the first place, and that is less likely to blow out the highlights in your image anyway. The added noise from ISO 1600 would cancel out any benefits that come from darkening the photo in post-processing.
Similarly, ETTR is useless if it requires that you leave base ISO. For example, consider a windy landscape scene. Your metered exposure settings could be f/11, 1/2 second, and ISO 100 (base ISO), with no highlights in danger of blowing out to pure white. This scene may seem like the perfect candidate for increasing your exposure beyond the matrix meter’s recommendation, but that doesn’t work in this case. A longer shutter speed would blur moving objects, and a wider aperture would result in too shallow depth of field. So, to reach the suggested ETTR exposure, you would need to raise your ISO. Again, though, the increased noise from a higher ISO would cancel out the effects of darkening the exposure in Lightroom. In this case, it is better just to go with the metered exposure.
ETTR is also difficult to implement if you do not have much time to take the photo you want. For example, wildlife or wedding photographers may prefer a “safer” exposure — one that is darker than what ETTR recommends, but one that is less likely to blow out important highlights in the image. Sometimes this particular technique is referred to as “Exposing to the Left”, or “ETTL”. Although it is nice to have as much detail as possible in a photo, the small differences are not worth the possibility of missing a fleeting image in the first place.
Also, If you shoot JPEG, there is little benefit to using ETTR. Although you could see a decrease in the image’s noise, reducing a bright exposure in Lightroom has the potential to change colors around. Plus, JPEG files contain very little data compared to RAW by their nature (they are only 8-bit), so it is contradictory to use a technique like ETTR on an already-compromised image format.
It is also worth noting that you can sometimes go overboard on ETTR even if you do not blow any of the highlights. If you are shooting a very dark subject, for example, extreme brightness changes in Lightroom can shift the colors of the image somewhat. Still, these color changes are generally negligible and they should not deter you from using ETTR (from what I understand, this has more to do with the RAW converter than with the image data itself).
10) Is it Worth It?
With the extreme dynamic range capabilities of modern sensors, as well as a much lower amount of noise at base ISO, the value of ETTR is coming under scrutiny in the photographic community. Since it is so easy to recover shadows with modern cameras, isn’t it better to go with a “safer” exposure that is less likely to blow out the highlights? In many cases, yes. It is no longer an issue to increase an image’s brightness in post-processing by one or two stops, and few people print large enough to notice a substantial difference anyways.
Yet, as photographers, we strive to take the highest-quality images possible. We want our RAW files to contain as much data as they can, giving us more legroom to recover data in post-processing. If you are a landscape or studio photographer, especially, you should have enough time to bracket your photos in an attempt to get the ideal ETTR exposure. And if you have the time to implement ETTR, why not do so?
As a final note, it is worth mentioning that some photographers take the idea of ETTR too far, subconsciously putting its importance above the composition of the image itself. Exposing to the right can be a valuable tool for many photographers, but its benefits are only meaningful if the photo is already successful in every other way — composition, lighting, optimal focus, and so on. I am sure you have seen your share of photographers always taking 3 or more exposures (even when light is too flat or there is nothing interesting to photograph), ending up with a lot of wasted memory and too many images to even bother working with later.
As such, it is important to know when and how to use ETTR. If you get your shot set up and you have optimized everything else in your image, ETTR is valuable because it lets you squeeze the last little bit of image quality from a scene. For many photographers, that benefit alone outweighs the drawbacks of exposing to the right.
I’d be very interested to know other people’s thoughts about using ETTR when photographing birds. Indeed, how do others meter for bird photography (matrix or single point etc.)?
Personally, I would not do it. When shooting birds, you are typically always above base ISO and you mostly struggle with keeping high enough shutter speeds – ETTR would require longer shutter speeds, which is not always desirable. Most wildlife photographers do not bother with ETTR as a result.
Thanks for your response Nasim and much appreciated. Would you say that you judge exposure from experience and the type of bird photography that you are shooting (i.e. BIF, shooting at the sky etc)? Do you personally use Aperture Priority or Manual with Auto ISO?
I usually use Aperture Priority for most of my photography needs. However, for birds in particular, shooting in Manual mode with desired shutter speed + aperture and letting the camera manage ISO is my preferred method. For BIF, my shutter speed is set to 1/1250 – 1/1600 for fast birds. For slow birds like pelicans, I can drop it down a little to 1/1000 or so, obviously depending on what focal length I am shooting.
Spencer, a very absorbing and useful article. If I may add, for the benefit of more mature photographers who maybe came from using 35mm film cameras and moved to digita… This ETTR method shows the difference between the two media. With film I guess these days we would call it ETTL (to the left!) not that we had histograms to reference. Darkroom skills could rescue exposure errors up to a point, but the photographer was wise to underexpose a scene with wide contrasting values, and preserve detail in the highlights. This was never more important than with Kodachrome or Ectachrome positive film, where there was NO detail in overexposed areas, and no way to rescue the shot.
Your article, Spencer, alerts us all to the needs of the digital sensor being the opposite of those of film. A touch of overexposure allows post processing to carry out the image rescue that we had to do with film, and it is interesting that underexposed areas of contrasty shots are as lost as the overexposed areas in the days of film used to be.
Your example of the shot of trees also shows us how it is still very easy for the camera meter to be fooled, where you had to reduce exposure to -2/3, so it is not foolproof, but certainly has woken me up to something I was aware of, but not really giving any priority to when shooting.
My own technique, for what it is worth, when I encounter a trickily lit scene, is to try to find a mid tone within the scene (a stone, some grass, maybe a tree with light bark) and to spot meter that. Then see what the camera says when recomposing the image, and dialing in some compensation to get near the spot meter reading (if it varies by more than 1/2 stop). I have to say, after reading your article, I will be much more aware of exposure issues, and I think, using my camera’s bracketing system to secure, if possible, a perfect exposure among the group of 3 or 5 images. Coming from film cameras originally, I still have a hang-up about shooting more than one shot of a subject, as in those days, every shot cost good money… I have to keep reminding myself that these days, we have keepers, and deleters, and the card can hold 800 or more images instead of the 36 we had with film.
Great technical article though, I will be putting this into practice later today, off to shoot a waterfall in the woods so I hope to capture the flow of the water, and get the exposure correct, while using a strong ND filter, so wish me luck :)
Ross, wait until Iliah Borg publishes his series on digital camera sensors – you will need lots of coffee to absorb all that information :) I am not expecting to understand anything after the first read, although Iliah promised to make it simpler for all of us mortals to understand.
Digital cameras work completely different than film. Many modern cameras have near ISO-less sensors, so you can shoot at say ISO 100, then push images to ISO 400 in post and the results will be the same as if you were to shoot at ISO 400 in your camera. There is a lot to consider with digital. In fact, Iliah hates when we use the word “ISO sensitivity”, since there is actually no such thing in digital!
Awesome ! Looking forward to this series :)
I’ll start making coffee now. I’m thinking this will be a long read for me to understand.
Great! How many kg coffee need I buy? :-)
“Let’s get it on!” :)
When Iliah Borg writes an article, it’s usually required reading, resistance is futile!
Good article and response from Ross (as I’m very old Skool & have digital issues). I too do what Ross does but have instinctively been ETTRing as learnt behaviour as ONE method of obtaining the effect I want from using digital for many years now.
I still think photographers need to be guided by their creativity and expression rather than focusing (pun) just on post technique. Like spending hours correcting ‘errors’ and adding stuff to an image on Lightroom etc which destroy emotion in a scene. I know there’s an aguement that post production is ‘your art’!
Also endless talk of unnoticeable and unimportant vignetting etc or only buying lenses with wide apertures, as if subject isolation is the only way to convey your view of it!
ETTR seems to be one way of achieving what you want, and not used for the sake of it. Photography should be fun, and digital has given us freedom to play and express as well as the disposability of the instant image.
Keith,
I completely agree with you — at the end of the day, the technical aspects of a photograph make no difference, so long as it is a well-made image. ETTR is a great way to gain an extra bit of image quality, and I am glad to hear that you instinctively apply it to your photography, but it isn’t something worth losing creativity over : )
~Spencer
Benefits of ETTR are obvious, even if you use the technique over the base ISO of your camera. As soon as you darken your picture, the added noise will disappear (not completely, but you’ll probably see less noise than a shot taken at ISO800). BUT, it all depends on the scene. It does not work in all scenes and scenarios, especially if contrast at the scene is high.
I try to use ETTR as much as I can, even if I have to sacrifice some highlights. And if I have time, HDR is my preferred method.
Ertan,
Thank you for sharing your thoughts!
In theory, it is somewhat better to expose to the right even when doing so requires a higher ISO. However, during my research for this article, I took several images in an attempt to show this difference (for example, a metered exposure at ISO 800 and one exposed to the right at 1600), and there was truly no difference once I equalized the exposures. If you have the time, there is no harm to using ISO to expose to the right (assuming you don’t blow out any highlights), but the benefits are far smaller than at base ISO.
HDR is definitely the best of both worlds — you don’t need to worry about blowing out the highlights, and you get a good deal of shadow detail as well. The only issue is compiling a photo that looks natural — something that a lot of photographers need to improve : )
~Spencer
Instead of ETTR, which I have never found to bring me significant advantages, I use ERTFT (Expose Right The First Time). According to the circumstances (and there are many variables there), I will let the camera do its thing, or I will zone-meter or average-meter on a specific area then recompose, or spot-meter a couple of areas and take two photos to be later composited (when the range simply exceeds that of the sensor)… and when the best possible exposure is critical from get-go, I will use my Sekonic meter.
Thus, and given those habits of calculating exposure, I have never found that exposing to the right really brings anything of significance to my photography.
Dominique,
For many photographers, the benefits of ETTR are small to the point that they are negligible. It sounds like you have a good system for exposure, and I am glad that you found something that works for your photography : )
~Spencer
I currently own a Nikon D750 which has highlight weighted metering. Is this a reliable method? I have used it a few times and the results have worked out nicely so far.
Chris,
Highlight-weighted metering is useful, but it is not without its flaws (from what I understand, at least — I do not have a camera with this metering mode, so I cannot speak with complete certainty).
Read my reply to Bruce Crossan above — it should answer your question to some degree.
~Spencer
Nice article Spencer, a unique topic no one pays emphasis on. Nicely written :)
Thank you, Siddhant! ETTR is not as popular in the photographic community as it should be — we argue tirelessly over cameras with the tiniest bit better dynamic range, yet we ignore ways to improve the data from cameras we already have!
~Spencer
Spencer, your article is a good explanation, and it is very well illustrated.
I’ve never been sure what the term “base ISO” actually means: I think it means different things to different people rather than having an precise technical definition. In the case of Nikon cameras that have “Lo” ISO settings, the reason that they are labelled “Lo” rather than assigned numbers is because they are exposing to the right. E.g. if the lowest numbered ISO is 100 then “Lo 1.0” is ISO 100 ETTR by 1 stop. Although this is effectively ISO 50 it is not marked as such due to the 1 stop loss of highlight headroom being traded for a 1 stop increase in shadow resolution.
It’s easy to forget that the actual ISO of the sensor is fixed; the ISO dial alters the mapping of the electron counts onto the range of available output values — it’s a gain control. The reason it’s needed for RAW only output is because RAW doesn’t contain nearly enough bits-per-channel to encompass the maximum electron counts from the sensor. 12-bits has a maximum count of only 4,095; 14-bits has 16,383; whereas some sensors have a maximum electron count of over 100,000, which requires at least 17 bits per channel to represent and to be processed by external software (much of which uses only 16-bit processing).
For a perfect sensor, the per pixel signal-to-noise-ratio is the square root of the number of electrons counted. ETTR is for forcing the camera to record as near as possible to its maximum electron counts in the scene highlights, then map them into its limited 12- or 14-bit output range. This means that the deep shadow areas will have the lowest possible noise.
I cannot begin to understand why manufacturers make this so complicated for the users. All that would be required is to provide, say, a special lowest ISO setting that also causes the histogram to display the full range of RAW values. It would also be very nice to have the exposure meter switched to indicate the peak highlight level rather than the mid-tone level. The modern RGB multi-segment meters have enough information to do this, but it is not made available to the user. Instead, we have to perform a series of experiments, compensate the spot-meter then wave it around the scene highlights. This is a ridiculous waste of our time because the meter itself _is_ taking multiple spot readings of the scene, which could easily be displayed as a histogram; even as a low-res image with flashing clipped highlights!
Do the DSLR manufacturers believe that we are not smart enough to use such features, or are the manufacturers not smart enough to properly understand the theoretical and applied science of digital photography? My guess is the latter.
Pete,
Thank you for your detailed comment, and for clarifying some info about ISO values. When I referred to base ISO in the article, I did not mean any of the “Lo” options, but I see how that can be confusing.
Camera manufacturers can be contradictory at times, especially when it comes to optimum data. They spend a tremendous amount of money and energy designing new jpeg styles (which are inherently compressed, of course), yet they don’t give photographers an easy way to extract the best possible image quality from a scene. Your suggestions would not be particularly hard to implement (Magic Lantern has done so to some degree with Canon cameras already), so it amazes me that no manufacturer has caught on yet.
~Spencer
Dear Spencer,
When I “coined” :) the term UniWB, Uni part was standing for “unity”, that is white balance coefficients all equal to 1. It is often possible to use b/w mode in the camera to avoid green tints.
I do not watch the in-camera histograms, only blinkies. That helps with the scenes with light sources and specular highlights. Those highlights need to be blown out most of the cases, as to keep them an “underexposure” of about 3EV is necessary (which is the case when I shoot welding).
One of the reasons I introduced UniWB is to allow for magenta filters in front of the lenses or studio flashes. Using filters like CC30M to CC50M completes the optimization of the per-channel exposure.
ISO setting does not add noise. If anything, increasing ISO decreases noise. What increases the noise in shadows is the low light conditions. Increasing ISO while decreasing the exposure is exactly that, decreasing the amount of light accumulated in the shadows, hence more noise. However, ETTR with ISO setting still makes sense if one is hitting the limit on decreasing the shutter speed and opening the aperture wider – that is, if the exposure is already maximized and there is still some unnecessary room at the left.
The method I use most is spot-metering for the important highlights where I want to keep some texture and adding positive exposure compensation. For most cameras it is +2.5 EV to +3EV. Only few cameras require +3.5EV. I often use hand-held spot meter for that, as I found 2° to work very good for lenses shorter than 105mm. It is also easier for me to scan the scene for the highlights with a handheld meter when the camera is on a tripod; especially when shooting panoramas.
Not all cameras are optimized for ETTR. For now strong ETTR can’t be recommended if the camera is in lossy raw compression mode as such ETTR causes image degradation – the compression is designed to work the best when the midtone is placed where it belongs. Case in point – current Sony cameras. ETTR on such cameras result in better shadows, but the resolution is lost in midtones.
Thank you for writing such a good article.
Iliah,
You have made a very important technical point: ETTR without the application of colour-balancing filters is often a far less than optimal solution when using either digital sensors or film. Adjusting white balance post-recording always results in a loss of maximum achievable image quality from the recording medium — hence having films that are optimised for tungsten lighting rather than for direct sunlight and flash.
Digital white balance is achieved by manipulating the values of the RGB sensel electron counts. This is why scenes illuminated by candle light, and captured at high ISO, exhibit disturbing chroma noise in the blue channel. Digital sensors for general-purpose photography are optimised for a colour temperature in the range 5000 to 6000 kelvin; obviously, the sensors used in space telescopes are optimised for vastly different ranges.
Thank you for sharing your experiences.
Hi Illiah, clearly I still have much to learn with regards to maximising
digital files and dSLR metering. I am hoping you can lend me some
insight via a comment or PM, concerning the comment I made earlier with photos
attached, since you seem to have a solid grasp on such things. Hoping
you can find the time to do so, because I struggle with this all the
time with outdoor portrait shooting.
Best,
Brian
Iliah,
I was hoping you would join in on the comments on this article : )
Thank you for clarifying this information — I didn’t know that you coined the term UniWB! I fixed the acronym in the article, and I linked to RawDigger.
I do have two questions. First, do you find that a magenta filter allows leads to a noticeable improvement in noise performance? I have always been curious about how significant its improvement really is.
Second, how does monochromatic mode work with an accurate RAW histogram? I had thought that Nikon’s black and white jpeg picture control was far more contrasty than the Neutral or Flat controls, even with contrast at its lowest. Is it close enough that it still works for ETTR, or am I missing something? Perhaps a custom monochromatic picture style?
I know that, in theory (and perhaps in practice under very strict conditions), ETTR still is preferable even if it requires a non-base ISO. However, I have never been able to produce two test images (one at a higher ISO, then reduced in post-processing) that show any sort of difference, no matter how much I enlarge it. For me, the theoretical gain at this point is just so small that I don’t bother implementing ETTR at anything other than base ISO.
Clearly, you are the ETTR expert — I couldn’t even write an article on the subject without mentioning your name — so I appreciate that you have taken the time to clear up some of the confusion with your comments.
~Spencer
Dear Spencer,
The improvement due to balancing filters is two-folded.
First, the exposure for the shadows is raised about +0.7 to +1 EV. This results in less noise, more resolution, and less issues with flare skewing the white balance in shadows.
Second, demosaicking works better when there is more “meat”, and when the values in all channels are balanced. Blotchiness in shadows is reduced, as one of the results.
Monochromatic mode histogram is pretty accurate when it comes to displaying what I’m actually interested to see – the overall levels distribution, including highlight clipping and shadows plugging. YMMV of course. What I do is modifying the curve in Picture Control to load a linear one – if for Nikon. You can do it using free Nikon Control utility, it will save the controls to the memory card, and they will appear in the camera – http://imaging.nikon.com/lineup/microsite/picturecontrol/adjustment/custom.htm , “2) Creating with Picture Control Utility 2, a dedicated Picture Control management and adjustment software”. Enable the curve and set it to linear. You will be able to load the necessary Picture control through camera shooting menu, “Manage…” item and activate it through “Set…” item. It is possible to exchange the controls between cameras and users, but it is not guaranteed – that is why I’m not posting mine.
The gain from ISO-based ETTR strongly depends on the camera. It is not much with Sony sensors (Pentax, many Nikons, Fujifilm, Sony) but very important with Canon cameras and Nikon sensors (D2H, D3/s, D700, D4/s, Df). THe further the camera is from “ISO-less”, the more is the gain.
As to UniWB, I posted on DPR what it is, how to use, and why the filters when D2X hit the shelves ;) At that time UniWB was easily created with just Nikon Capture. They took off that option pretty soon.
Iliah,
Thanks for the helpful information. For what it’s worth, I did my ISO ETTR experiments using the D800e, which (being a newer Sony sensor) seems as close to ISO-less as they come.
I guess that my next purchase will be a magenta filter : )
~Spencer
Dear Spencer,
Stay with gels or plastic filters – they are cheap, easy to replace, and do not introduce the spikes in spectrum like the glass ones.
Thank you, Iliah. So, would something like this be ideal for a Lee holder?
http://www.bhphotovideo.com/c/product/102489-REG/LEE_Filters_M40_4x4_Magenta.html/BI/5562/KBID/6400
Depends on the holder, but if you are going to use them with foundation kit or full compendiums, you may want to add http://www.bhphotovideo.com/c/product/87154-REG/LEE_Filters_CM100_Filter_Frame_4x4_Plastic.html
Iliah,
thanks a lot for your detailed explanations (and Spencer for initiating them).
I have two more questions:
1. The JPEGs of images captured with the Flat Picture Control added to the newer Nikon model look really awfull in most cases, but I assume the Flat Picture Control shows a histogram more closely resembling the RAW histogram than the other Picture Controls. Is this correct? Does it make sense to prefer the Flat Picture Control over the others in order to assess the blinkies or the histograms?
2. If I understood it correclty there are two things needed for doing / using UniWB: First, a linear picture control, as you described above. Second, an image with a special white balance which can be copied to a WB preset. These images are supposed to be camera specific. Can you recommend any source for images with a UniWB white balance or a method for generating them?
Interesting. Probably more than I’ll ever get into (by way of the technical aspects), but shooting to the right is very important, as I see it, with something like a wedding dress. You really want brightness and yet all the detail you can get. The rest of the tones will fall into place.
Thank you for sharing your thoughts, Keith!
Wedding dresses, or any other bright objects, are prime candidates for ETTR. The issue is that the camera’s matrix meter may underexpose the dress, thinking that it should be a neutral gray color in the image. Although it can be hard to implement ETTR with subjects in changing light, it is definitely a good idea to use some positive exposure compensation in situations like that.
~Spencer
Spencer, thanks for the article. I completely agree with your (and many other people’s) opinion that the camera manufacturers finally need to give us a RAW histogram!
To really be useful the camera should measure the correct ETTR exposure (with an “intelligent” way of recognising specular highlights like street lamps or the sun that will always blow out) and then automatically darken the image for the JPEG preview. If the image was not darkened in-camera it would make a good judgement of the final look very difficult (similar to the way UniWB does).
An additional (extremely useful) exposure mode would be an option for ISO-less shooting. Basically the camera should show you everything the way it would be needed for “correct” exposure. But when you actually press the shutter it underexposes by using the base ISO. Then – again – it pushes the exposure for the JPEG preview so you can properly judge your images. With today’s ISO-less sensors that mode would potentially save tons of highlights :-)
One more way to shoot ETTR would be to find out how much headroom the JPEG “blinky” warnings have. Basically find out the difference between JPEG and RAW highlight clipping with your specific camera. Then always leave your picture control settings set. Find out when your first highlights clip according to your camera. Then lower your shutter speed by a 2/3 stop (or whatever you found out works for your camera) to expose as far to the right as possible without RAW clipping. I much prefer the blinking highlight warning because it tells me what is blown out. It’s very important to know whether it’s the street lamps or the red channel on your subject’s face!
Jano,
Thanks for sharing your experiences! The method you described would be wonderful — I hope that some day (especially once more people discover the ETTR technique), enough people will ask camera manufacturers for this feature, and they’ll have to give in to consumers.
Your method is another great way to implement ETTR, and one that makes it easier to differentiate between your subject and useless background highlights.
Some day, if we’re lucky, camera manufacturers will get smarter about their implementation of exposures. Their current system works well if you shoot jpeg, but it falls short if you are shooting in RAW.
~Spencer
Hey Spenser, thanks so much for such a great article explaining the details of an ETTR technique. I have often struggled with how effective this technique is though when shooting portraiture, especially in high-contrast situations, whether using a flash or not. I have one such example below (I think it will post below; I’ve never uploaded photos to a PL thread before). I am hoping either you or any other PL reader can give me some pointers, because I struggle with this a lot… how to expose correctly on a fast-paced portrait shoot with high contrast and where the sun becomes my enemy.
PLEASE NOTE: this photo is a failure for many reasons, and was not included in what I delivered to my client, so please, let’s just focus on the exposure and recovery I performed, and not on my composition, my models’ body position, the colour grading I tried out, etc.
Some background, as I’m sure many other photographers here have been in similar situations. All photos were taken at 11:30am: a wretched time to shoot portraits, but that was out of my control. When time allows, I prefer to use a hand-held light meter and meter from my model back to the camera/flash, but on lots of environmental portrait sessions, this is impossible, because of a fast-paced or more candid nature to the shoot, such as the examples below (I had one and a half hours to get in three outfit changes and four locations for the client).
As for metadata: all shots were obtained using a Nikon Df set to spot metering (my preferred for portraiture), captured in RAW (NEF), sRGB colour space, at ISO 400 – a relatively low ISO for the Df – and a Nikon 70-200mm VR1, at 120mm. All shots were also captured using a diffused, on-camera flash, which was recharging during some shots, and didn’t fire at full power – which brings us back to recovery later in post, when either my exposure or my gear failed to help me obtain a great shot.
In the first shot (1/500 sec. at f/4.5), the model’s skin is fairly close to a good (workable) exposure, but the highlights have been completely blown not only in the grass (secondary to my model’s exposure), but also along her right arm and across her hair. Additionally, there is a notable loss of contrast due to the back-lit (more above and back) scene. There were many shots from this session that were much worse concerning loss of detail on my model’s face, and which I opted to use a shot that obtained with ETTL rather than ETTR, but I cannot post those photos because of the terms of my client’s session contract, so let’s just focus on these shots as a stand-in.
For this first shot, I would later try to develop it in post, only to find it didn’t process as well as the second, “under-exposed” shot, largely, because it was far easier to correct for my model’s skin than it was to correct everything else going on.
The second shot, taken moments later at 1/1000 sec. at f/5.6 – a reduction of one stop from the shutter and another 2/3’s stop from the aperture, was technically under-exposed for my model’s skin. Worse still, it was obtained when the flash was recharging (TTL, set to High Speed Sync). Needless to say, the result on my model’s skin is a sub-optimal result. However, this was a stronger photo for me to work with in Lightroom and Photoshop in the end, in that my settings retained much more in terms of detail in my model’s arm and hair; in that the “environment” retained important context; and in that I didn’t lose too much in the face when recovering shadows.
These situations arise all the time for me when I’m out doing portraiture, whether from side-lighting from the sun, or from a recharging flash. I don’t think I’m alone in this, either. This almost always results in either losing too much detail in the side of my model’s face, or blowing highlights in other areas, such as a model’s attire. A lot of the time, ETTR scenarios make it difficult to get a good range of tonal variation and shadow in the face. So, to state it again, I worked the “under-exposed,” second RAW file as well, and obtained the final photo’s result.
Although I never delivered this particular photo to my client, there were many more like it that I did deliver, with very similar lighting failures both on my part and on my gear’s part. Still, I was able to deliver very beautiful results to my client after some post work recovering shadows from back-lit scenes, and I never needed to torture the RAW file in post to obtain those results. Surely though, there is a better approach I can take? Or if not, then maybe there’s a case for ETTL, especially at lower ISOs, when dealing with a person that is one’s subject?
To me, losing detail in skin due to over-exposure is about the worst thing I can do during a shoot. I realise no one here will dispute that “over-exposing” is a no-no, especially when details are lost in critical areas such as skin. Sometimes, I can regain detail through frequency separation techniques, but these are arduous tasks in Photoshop. Crushing the highlights in Lightroom will never make skin look good – even if it recovers some detail. The results are flat skin, in which all the contrast, burning or clarity in the world can’t seem to restore and still retain a natural look.
Spenser, I welcome your or anyone else’s feedback on obtaining the best possible exposure for portraiture in particular, given high-contrast, back-lit scenes where ETTR seems to fail me more often than not. Nasim, you in particular might be able to lend some advice, since you and your wife often shoot portraits and weddings together.
Thanks again for the great article, Spenser!
Best,
Brian
Readers, please note: the order in which the photos are displayed are not in sync with my comment. The first photo is the ETTR photo; the last photo is the ETTL photo, and is the one I more successfully worked, to produce the final, second photo. All images should be at a web resolution sufficient for zooming in.
Looking forward to any suggestions!
Brian
A number of things strike me in the scenario you present.
First, the light. In some situations, the light is so adverse that it is better not to shoot as the results will almost inevitably be sub optimal. Sometimes, even with a good subject I will not lift the camera to my eye simply because the light is such that I know I will not be able to achieve the particular ‘look’ I seek in my work no matter what I do either during shooting or in post process. I see no point in shooting to feed my trash bin. However, I accept that having taken on a paid assignment, it behoves a professional photographer to come up with a satisfactory result.
Second, it seems to me that to overcome lighting conditions as difficult as these, something more than a single flash is needed. Reflectors, both using flash and sunlight, are invaluable in these situations as being large, they are more effective than even a diffused flash in producing softer highlights and softer, more flattering shadows. Also, difficult situations often demand more complex solutions viz. more than one flash. However, the use of multiple flash and reflectors of course also demands a higher level of knowledge and skill not to mention the likelihood of needing the services of an assistant. Iliah and Mike are dead right, you would be better off looking at ND filters and light modifiers not trying to blast your way out of trouble with a single flash gun. As professionals, we should be asking ourselves, on taking on assignments, whether or not we are properly equipped to deliver to the brief, come what may.
Third, the choice of background and wardrobe seems to me to have some flaws. Photographing a dark haired girl in a pure white dress, standing against a white painted gate at midday in blazing sun does not in my mind add up to good planning. Rather than stacking the odds in favour of achieving a good result, you seem to have made a rod for your own back and are subsequently trying to find technical fixes both in camera and in post for the inevitable shortfalls. However, given that sometimes things are what they are, why did you not simply stand the model under a tree or on the shady side of a hedge? Shadows are so much more forgiving and easier to cope with, so that all that would have been required would be a little fill flash to add catchlights to the eyes and some sparkle to the teeth and hair? And why did you shoot while the flash was recharging? It makes no sense. The result will be underexposure of some or all of the image and why would you want that?
With regard to exposure and the attempts in post process to rectify matters, I would first like to say that ETTR is not a ‘technique’ suited to one or other photographic genre, it is a universally applicable method of achieving optimal exposure in digital photography. Conversely, ETTL is almost universally wrong in achieving the same aim and if you are universally applying it to portraiture then your results can never be other than sub optimal – even if only slightly, depending on circumstance.
Similarly, underexposing (and so losing data and increasing noise) and then trying to recover shadows in post process is not the pathway to a perfect output file. You could have recovered highlights in the arm and hair with a curves adjustment+/- tonal masking+/_selective dodging and burning, etc. and at the same time opened up the (more detailed) shadows if necessary given the high dynamic range). This is only possible of course if ETTR has been executed properly and the there is highlight detail to recover. Proper exposure (ETTR) will never damage important highlights. However, the answer to overexposure of highlights is not ETTL underexposure combined with boosting the shadows! Two wrongs have never made a right. Also,you seem confused about whether you are using flash or daylight as your primary light source or whether you are using daylight supplemented with flash to control contrast. I am at a loss to understand why you would be using a shutter speed of 1/2000sec for portraiture ( I know she’s a bird – but this kind don’t fly) and combining it with really weak, auto FP flash. It seems to me that is choosing the worst of all worlds – basically a daylight exposure in high contrast conditions combined with flash which is too weak to make a useful contribution. That’s why you seem, in these two examples, to be stuck between overexposure and underexposure and locked in a struggle with excessively high contrast.
It seems to me that the poor choice of clothing, venue and time of day and lack of preparation, combined with insufficient equipment and insufficient knowledge of technique can only be a recipe for disappointment.
But also sometimes there are no technical fixes that are practical and no matter what we do, the result will never be great. Sometimes it’s just better to go home and put your feet up.
Hi Betty,
First, thank you very much for a thorough and helpful reply. Before I get too far along in relying, I just want to reiterate things I stated earlier, in my original post(s), before addressing your great suggestions. First, in case you didn’t read it (and there was a lot to read), I started by asking PL readers to focus on my question about ETTL in such lighting situations, and, as I mentioned to Mike above, to bear in mind what I premised from the start:
>>”PLEASE NOTE: this photo is a failure for many reasons, and was not
included in what I delivered to my client, so please, let’s just focus
on the exposure and recovery I performed, and not on my composition, my
models’ body position, the colour grading I tried out, etc.”<>”As
professionals, we should be asking ourselves, on taking on assignments,
whether or not we are properly equipped to deliver to the brief, come
what may.”<<
That's certainly true. The clients who hire me are getting results that exceed their expectations, which are based either on their past experiences with photographers or with what they saw from my previous work, and I attribute the latter to my desire to always do the next shoot better. However, the clients who hire me are also not big-name firms or agencies or models, and I think they intrinsically expect less of me as a result. Still, that is not to imply I settle for anything less than what I am capable of delivering. I, like you and so many other photographers on here, am keen to learn and improve. I hope my comments relay that wish of mine; otherwise, why would I ask for suggestions? Thank you sincerely for giving me yours!
Thanks also for touching on the actual question I posed regarding ETTL vs. ETTR and post work. I again wish to clear up one small thing, though: I don't not ALWAYS prefer ETTL in portrait shoots. If I didn't make it clear before, I should state now that I have come to integrate ETTL only under strongly backlit scenes, because it tends to offer me better results later in post, with regards to a model's skin. I mentioned that in my original post, but again, there was a lot to read and you may have missed that. Per your suggestion that I could have recovered highlights in that first photo example, I actually tried (with the first photo I showed), and could not recover what I deemed were important details. That's why I chose to develop the latter, under-developed photo, which, as I explained, still was a dud in my mind. Which is to say that I never obtained a good ETTR file during that series of shots — that I did not properly execute ETTR.
I wanted so much to have it come clear to everyone that the whole point of my comment was to express that getting a properly executed ETTR file was not working for me because of the compromise between background and model, and the harsh sun overhead. The suggestions you and others have made with regards to ND filters and reflectors is certainly something that I will try out next time, as this seems to offer possibilities. But my chosen shutter speed and ISO should be obvious, given my comments about focal length, high winds, large apertures, no tripod, fast-paced shoot, etc, and that I was attempting to get my model's skin under-exposed a tad on purpose (ETTL). None of which excuses my poor use of FP flash to compensate, and all of which go back to my question about ETTL versus ETTR, which I think you've all done a great job answering me, and so I thank you (all) again. In a perfect world, I would have had the best light ever. I dealt with the conditions I had to the best of my ability at the time, as imperfect as they all were. Which is why I brought these examples up. I've learned a lot from everyone's comments. Maybe someone else will learn from my post as well. That's the idea to this forum, no?
On a more personal note, thanks for taking the time to phrase your comments in a manner that did not come across as an attack. I noted a difference immediately, and that left me wishing to engage you in a thoughtful rather than defensive manner. I know you and I have a history of misunderstandings. Thank you sincerely for taking the time to reply to my comment. I'm glad you did.
Best,
Brian
Thanks Brian, I am trying to behave myself and act more decorously as befits a lady.
Brian, no offense intended with my comment, but it appears to me you were concentrating more on ETTR and ETTL than on the shoot itself. In trying to achieve technique in camera it would appear you didn’t pay enough attention to your environment. In the first photo of the young lady sitting in the field, you might have just turned here a bit more to the left which would have put the back light from the mid day sun more to her back decreasing the amount of gross highlights on her arm and leg and acting more like a key light. This would have enabled you to use a reflector to fill in shadows to the front. When working alone as you most likely were, I would use my 70-200 mounted on a tripod, set back from the subject. Focus and tell my subject not to move, (if they are paying you they will do anything you tell them to do, trust me on this). Using a remote shutter release I will hand hold the reflector myself. This too, will take practice but by having a long focus (150-200mm), back far enough from the subject you could stop down incase she would move a bit and still have a sharp photograph. Plus, It appears to me that the wind was blowing from right to left and she was trying to keep a shock of hair from blowing to her left. Using the wind to keep the hair in place will become second nature to you in time. Personally, I don’t like the angle of her right arm in this picture. Regarding the photo by the beach, in order for that picture to be successful you would have needed two people holding reflectors on both sides of the subject. Seems to me you didn’t have that facility. Therefore, making one of Betty’s comments below something to keep in mind. Sometimes it just isn’t worth taking the shot because you will not be able to improve it. Hope something here helps.
Hi Mike,
Thanks for taking the time to comment. I realise there have been a lot of comments in this thread and you might not have had the time to read them all, but I just wanted to point out that I began my initial comment/request with stating the following:
>>”PLEASE NOTE: this photo is a failure for many reasons, and was not
included in what I delivered to my client, so please, let’s just focus
on the exposure and recovery I performed, and not on my composition, my
models’ body position, the colour grading I tried out, etc.”<<
I did not post any photos I ended up delivering to my client — per the contract I had with my client. Each of the photos I posted were all "failures" in my mind. I completely agree with a lot regarding what you wrote. For those reasons and others, these photos I posted failed to make the cut, as it were. I really wanted to avoid having my posting them turn into a critique of my photos or poor choices I made, and rather, focus on my question about whether using ETTL and then open back up shadows in the face later in post was a viable option. Many people have graciously given me ideas about how to avoid that all together, and for that, I thank you all.
I haven't read Betty's reply below yet (about to, though), but for your input, the use of the reflector/tripod combo is something I'd like to try out in the future. However, as of now and in the past, I shoot alone on my assignments. Thanks again for the valuable feedback, Mike!
Best,
Brian
Brian, sorry, you are correct. I have been very busy the past two weeks with work and have only read snippets of several PL threads. I didn’t pick up on your asking others regarding only ETTL or ETTR. I am also sure you would only deliver, to your client, those images you felt best represented the subject and your work. I have no doubt about that. In any photo shoot we all wind up with duds and keepers; and for the most part, we all try to make more keepers. Keep experimenting, keep working and learning and it will all come together for you because you love what you do. All the best Brian.
Dear Brian,
Shots 1/500 and 1/1000 – were those with flash? Because you are going over the flash sync speed 2+ stops. Flash will not actually help lighting the photos at such shutter speeds.
Iliah, the clue to the use of flash was in Brian’s comment “(TTL, set to High Speed Sync)”!
No good, too much power drop. And with a diffuser it will not help at all at such distance and contre-jour light.
An interesting response. Now please write a blog post that uses solid evidence to clearly inform Nikon, Canon, and every user, that high-speed flash sync is nothing other than marketing bullshit.
Pete A, I think my above reply (and photos) to Illiah shows that it’s not a bunch of bull at all. I use high-speed sync all the time, and it works… unless the flash is recharging when I take a given shot it obtained. Each shot, as I mentioned above, was acquired with 1/2000th of a second. And each shot under-exposed my model’s skin just enough to have the flash fill the details enough, whilst not allowing the strong side-light from the sun blow out all the details in other parts of her skin.
Best,
Brian
Brian, I’m really pleased that Betty has responded to your question. What irritated me was being swiftly dismissed by lliah’s reply: “No good, too much power drop. And with a diffuser it will not help at all at such distance and contre-jour light.”
Clearly, I’m not the one who was suggesting that high-speed sync is a bunch of bull. I’m glad that Mike Banks challenged lliah on his reply.
Brian, something you might add to the mix, which helps the flash recycle a bit faster, not much but sometimes just enough would be the Nikon SD-9. I use this all the time and if that unit is too expensive for you, at this time, Bolt makes a model for Nikon flash which are much cheaper. I have both and the Bolt works just as well as the Nikon. When shooting certain types of fast paced events, I too found the SB-910 would not fire at full power. However, when adding the SD-9 or Bold equivalent, I stopped missing shots because of recycle time.
Do you know how FP is implemented?
No, not at all. But I’d love for you to explain it, which based on all your other posts, you could — you’re extremely knowledgeable about such things — because I use it often and want to know it’s strengths and limitations. Thanks!
I was actually asking Pete :)
Look at your flash manual, see GNs for Auto FP. You may also be interested in reading the classic article at http://www.scantips.com/lights/flashbasics2b.html
Many thanks for the links and advice, Illiah. I will read it as soon as possible!
Sorry, Iliah, I didn’t know that my name had been reduced to a dot. Yes, thanks, I do know the technicalities of Nikon’s various flash systems, including high-speed sync. I eagerly await your own article dismissing it.
What is it that I’m dismissing, exactly? Please read my responses with some attention. What I’m saying that for the particular situation and one flash unit the FP mode is suboptimal due to reduction of power. Did you overlook that? Now, I do not want to participate in reducing Nasim’s site or Spencer’s article to a show. So please refrain from your sarcasm, especially because it is unfounded. I will not be answering anything but technical questions stated in technical terms.
Ilianh, the best way to avoid the typical/annoying ‘British sarcasm’ is to do what we do with it here in Australia and that is to simply ignore it and converse with those who stick to the pertinent aspects of the discussion.
Thanks for your replies, Illiah and (Pete A). I really appreciate it. However, is there any other way to accomplish my goals (stated in my comment above), though, aside from using 250 shutter sync and dropping ISO? That wouldn’t have worked for my goals in so many of this session’s shots. Using my flash on this particualr shoot (an SB700 with a Gary Fong Diffuser, because it’s portable), with my camera set to High Speed Sync (as Pete A noted), allowed me to use larger apertures and very fast shutter speeds in bright daylight, throughout the shoot, which took place on a very windy beach, with white water from waves splashing about, and my model’s hair blowing all over the place. My first examples above (the under-exposed one) show a result where the flash was recharging (as I stated), but shots such as these other photos below were also shot through ETTL and my model’s skin was what anyone would call “under-exposed” before post work.
Both shots were taken at 1/2000th sec, f/3.5 and f/3.2, respectively, and again High Speed Sync, TTL), getting all else in the environment workable, and wherein I brought back the detail in the shadows of my model’s face in post. Flash fired each time, which should be obvious, given the direction the sun is coming from. The ETTL method I used in these shots really made it easier in post for me to get better results on my model’s skin. Hoping you or anyone else can give additional feedback for lighting such as this, and whether ETTL or ETTR is the better bet. Thanks ahead of time for the replies. I have much to learn! The photos:
Brian, learning how to use large reflectors might do full justice to your model; learning how to master flash will definitely not!
I can assure you that my comment is intended only to be helpful; it contains no malice nor disrespect.
Sincerest best wishes,
Pete
Pete A, I hope I didn’t show ANY disrespect in my previous comments to you (or Illiah), and I took no offence from your feedback at all. It’s wanted and appreciated! I’m serious: I am asking for help here, because I struggle with this scenario all the time: super strong daylight/side-light/back-light, flash, etc. I have a 5-in-1 reflector set that I use when I can afford to have an assistant, which is rarely! My problem with reflectors and light stands in the types of shoots I do, is they blow over or blow away in the wind without someone to mind them. I recently recruited a father/mother combo to hold my light stand and umbrella during a family session in the mountains, so I could have my flash off camera. That worked… until I needed the whole family together for a series of shots. I tried to anchor my pack to the light stand, to no avail. I am still (and always will be) learning!
Pete, what are your thoughts though on my above examples (all examples), where each was acquired with ETTL and wherein I brought back details from the shadows in post, whilst retaining *most* of the highlights? I realise the above shots are hard to discern without the original RAWs, as I posted before. I’m glad to post the originals, if needed, but I don’t think that’s necessary since I’m trying to figure out if ETTL is the better bet for situations such as these.
I’ll take your advice, for sure, and try to learn my reflector setup better!
Best,
Brian
Brian, my typically British subtle (and often acidic) sarcasm wasn’t aimed at you at all. I know that high-speed sync has its uses therefore when a certain person persists in quickly dismissing my comments and those of others (on this and other PL articles), I initially do my best to remain polite, but I’ve become so sick and tired of his belligerence that I now have zero tolerance for it.
I would be more than happy to provide a little guidance, but I shall backdown and let the self-proclaimed ‘expert’ take over the discussion.
Pete A,
If you could take the time to provide me with guidance, I would really appreciate it. You or anyone else! If you feel more comfortable doing so privately, please feel free to email me: brian@briangaschler.com
I sincerely hope you can make the time to. I’m eager to learn, and since I am a portrait photographer by and large, and since I am confronted with these lighting scenarios most of the time during my *environmental* portrait sessions, I can use all the help I can get!
As for the British humour, Betty (yes, that Betty) knows all too well that even I can confuse sarcasm with witty, yet benign humour. So much is hard to convey through text, anyway… one of the many downside of our modern, digitised lifestyles.
Betty was, and still is, my teacher :-) I have much yet to learn…
Gosh, I have a fan.
More than you realize!!!
Brian,
Pete’s and Iliah’s recommendations to use off-camera flash (possibly more than just one) with reflectors are certainly valid since they allow for more control in modelling the light. (I understand this might be difficult without some helping hands, as you described above;-)
Whatever flash setup you use, on-camera or off-camera, however, if you want more control about the exposure, you might consider the following: you need to control two different exposures—one for the ambient light in the background (and maybe some illuminated skin parts in case of side-lit scenes) and one for the flash light.
How do you control exposure for the ambient light? Set the camera to manual mode and dial in aperture, shutter speed, and ISO to get a correct exposure in the background. You can use ETTR for this part of the scene if you want. If you need a lot of light from you flash (see below), prefer shutter speeds equal or lower than the flash sync speed (e.g. 1/250s). In a very bright scene in addition to reducing ISO you can make use of an ND filter in order to get the shutter speed down to the flash sync speed.
How do you control exposure for the flash light? Set the metering mode of the camera (Nikon) to Spot Metering in order to use Standard TTL flash mode instead of i-TTL Balanced Fill-Flash mode used when in Matrix Metering mode. (The problem with i-TTL Balanced Fill-Flash mode is that the camera tries to balance between background and foreground exposure, which is not what you want. What you want is to control both exposures independently from each other on your own. In Standard TTL flash mode the camera just controls the flash exposure on the foreground model.) Meter your model with the AF focus point.
If the shutter speed is above the flash sync speed and the flash doesn’t provide enough light, try to reduce the shutter speed to the flash sync speed (as described above).
You can also use Manual flash mode by setting the flash to Manual and dialing in the appropriate amount of light on the flash. You can even use ETTR for the foreground model if you want.
In the essence, this is what Iliah already recommended above, with some added details about exposure, metering, and flash modes.
Hi Carsten,
Thanks for taking the time to reply. Lots of good info you’ve provided me with, and I look forward to integrating it into my shooting! For clarification, when I use off-camera lights, I almost always use a handheld light meter to balance out my flashes, which are set to manual output at this point. I control them wirelessly via Phottix Odins, and have knowledge of but am still understanding the result of controlling flash output via aperture settings and ambient light via shutter speed settings. I always have (and probably always will) shoot in manual mode, and always have used spot metering in the camera for other occasions. I can’t tell you whether it comes from my film days or not (some of my film cameras had spot functionality whereas others only centre-weighted), but nevertheless, I am now too used to metering off bright areas with the spot meter and adjusting exposure accordingly.
Many thanks again for the input, especially on setting my flash up outside of normal TTL mode. I need to learn more about GNs and the application of using an ND filter while doing portraiture, that’s for sure. Thanks to you all for this valuable feedback.
Brian
Dear Brian,
Can you please explain why you want to maintain ISO 400?
You can use and ND filter (usually 3 stops) with flash set to normal sync.
The trick to have high speed sync (FP) is to switch the flash to stroboscopic mode, to make it shine the light through the duration the slit travels. This is much less efficient use of flash power compared to one burst.
Thanks for the tips, Illiah. I still have a lot to learn with regards to artificial light on-location. Lots!
Per your question, I don’t *want* to maintain ISO 400 per se; I want to maintain as low an ISO as possible, given that I’ve come to use ETTL more frequently in my portrait shoots, so I can hopefully obtain a good end result after post, and not have to try too hard to recover highlights in parts of my models’ skin, which seems to happen to me a lot when the light is strong from the side/back and I use ETTR and get most of my models’ skin workable, but lose too much elsewhere. ISO 400 just happened to be what these shots were obtained with. Looking at Lightroom metadata from my past 15 shoots, I actually ranged from ISO 200 to ISO 800 on a most of those shoots, wherein I use high speed sync, large apertures and fast shutter speeds. I like to use the fast shutter speeds when I need to control wildly blowing hair, or to make up for shooting rapidly at 150mm-200mm, where I will inevitably shake the lens around.
In any event, clearly I need to better understand the nuances of my speedlight — its strengths and limitations. I still would like to know your (or anyone else’s) thoughts on my ETTL propensity when it comes to shooting portraiture under extremely back-lit or side-lit scenes.
Best,
Brian
If your flash is well-balanced you may find that the dynamic range of the resulting scene does not call for ETTL. Use low ISO, use ND filter when necessary (it is well-described how it works on many sites), use off-camera flash when possible, and it will all come together. Auto FP is mostly for sports, when one needs fast shutter speed to freeze the action, and usually a serious shooter will use 2 or 4 flash units to compensate for the power drop.
Thanks for the reply, Illiah. I will research some articles, grab one of my ND filters and some speedlights and experiment
Best,
Brian
If you don’t want ND-filters, using a stronger (but more expensive) strobe (600Ws) gives you all the possibilities. I recommend Jinbei for a cheap but powerful solution: http://www.foto-morgen.de/Studioausstattung/Studioblitze-portabel/Jinbei-New-HD-II-600-V-HSS-Akku-Studioblitz-mit-600Ws-Fernsteuerung.html.
lliah, I would not ordinarily contradict anything you say but I have been using TTL high speed sync for a long time with SB-910’s and PocketWizzard radio triggers with none of the problems displayed here. OH yes, I do use modifiers on the lighting as well.
Well, the point I’m making is for the particular scenario of a strong backlit situation – here the exposure is not balanced between the subject and background because the flash is not powerful enough in FP mode to compensate for 3 EV+ difference. At 105mm GN for SB910 is 49 in standard mode, and 18 in Auto FP. With a strong diffuser GN of 18 makes for too little, given the distance.
True, but when moved closer to the subject it does work. One only need to add about 20-30% light to that situation.
No disagreement.
Brian,
The loss of contrast and ‘veiling’ in the first (overexposed) photo looks to me like it was caused by flare in the lens. It has all the typical characteristics. The reason it looks better in the second (underexposed) shot can probably be attributed to the fact that you stopped the lens down for the second image and additionally the darkness hides the veiling to some extent. Flare is always on the cards when shooting against the light with the sun just off axis and may explain why a number of other shots were similar in appearance. However, in this frame there, are in my view, two things going on at the same time – flare AND overexposure.
However, neither the flare nor the overexposure have anything to do with ETTR.
Brian, if you use ETTR correctly you cannot lose anything in the highlights as not losing highlights (clipping data) is the whole point of the technique.
I wouldn’t even call ETTR a ‘technique’ as if it were something extraordinary or different, it’s just simply the best way to get an optimal exposure with a digital camera.
Similarly, unless the dynamic range of the scene is absolutely outrageous, you will not lose shadow detail either – in fact you will preserve the maximum detail and the greatest range of tonal values possible as you double the amount of data for every extra step (stop) you push the histogram to the right. More data=more detail. Fully 50% of ALL the data in the file is contained in that first step in the histogram so the more you can push your exposure in that direction the better. You can then normalise those highlights, skin tones and shadows in post process to whatever appearance suit your fancy. You just need to learn the controls in the software.
I don’t know how much you ‘shake around’ but if you need 1/2000 sec to control a lens at 120mm, you should perhaps be consulting a doctor or joining AA. Same with the hair. I presume there wasn’t a tornado approaching on the day? Unless you are in a wind tunnel, 1/250sec – 1/500sec seems ample to me.
If you were shooting at, as you say, a fast pace, why use spot metering? This is best suited to contemplative landscape and studio photography, not fast moving location work. It is the most demanding and difficult method and takes a great deal of care and experience to get right. The slightest deviation from the chosen tonal value will throw exposure out. What were you metering from? If you were metering off anything other than a grey card, then your exposure will have had to be tempered with a compensation factor in manual brain mode. Compensating from a non-standard tone ‘on the hoof’ is notoriously difficult and without the benefit of great experience, prone to serious error. If you were using spot metering in auto mode and not locking the exposure, it is little wonder that exposure is all over the place. This seems to be borne out by the fact the two exposures, “taken moments apart”, are different by nearly two stops. And the fact that both shutter speed and aperture changed together, implies that you may even have been in programme mode. And if you were trying to use ETTR, then spot metering is pretty much an irrelevance because the exposure is ultimately determined by the histogram. It seems to me that you are still applying film exposure thinking to digital exposure and getting into a terrible muddle. Finally, not waiting for the flash to fully recharge and shooting regardless, negates any exposure measurement, ‘spot’ or otherwise, and makes the entire exercise completely random and haphazard.
It is clear that you are struggling with many aspects of your photography and as you say, have much to learn, but please, above all, take on board the fact that there is no “best possible exposure for portraiture”, there is only ‘best possible exposure for image making’ – and that is best achieved through ETTR.
ETTR is not failing you, you are failing you because you have not grasped the principles of correct digital exposure or the appropriate use of flash to balance foreground and background exposure. I apologise if that sounds hard but sometimes being hit on the head results in a change of perception.
What would I do?
Well, as the Irish say, if I was setting out to get to there, I wouldn’t be starting out from here…
No matter, given the situation, (very bright with high contrast) and no access to reflectors, additional flash guns or shade…..
I would use matrix (or even spot!) metering to get a ball park ambient light exposure to set an appropriate aperture and hand-holdable shutter speed combination, adjusting ISO (auto or manual) to suit.
If the light was too bright to allow sufficiently wide aperture without running out of shutter speeds, I would use a ND filter to get these parameters under control.
In this situation, if you have no neutral density filter you are IMHO basically screwed and you can pack up and go home. As you have found, using a suitably (longish) lens for full face and full length portraits, the flash to subject distance involved is usually too great for the reduced power of high speed synch flash to be effective even for fill light. And the faster the shutter speed the worse it gets. And putting a diffuser on the flash is truly shooting yourself in the foot.
I would use a (fill) flash in standard TTL mode (NOT daylight balanced iTTL) to lower the contrast ratio, taking a number of stepped test shots to determine the optimum level of compensation to achieve a nice balance.
I would check my histogram for blinkies on any highlights I considered important and reduce the overall exposure until they disappeared.
If the light was constant I would do all of the above in manual mode.
If the light was changing a lot I would engage auto ISO to take care of any variance.
In both cases I would periodically check the histogram and/or the LCD screen for blinkies to make sure nothing was going off track.
Hope you find this helpful and constructive.
>>”Hope you find this helpful and constructive.”<<
I did. Thank you. I've replied to your earlier comment as well, so there's no sense in saying again here what I stated there. Thanks for taking the time to reply, Betty. Sincerely.
You are welcome.
I have always exposed to the left, i.e. for the highlights, and used the sliders to bring up the shadow detail. Nikon even conveniently added a “highlight priority mode” to cater for my exposure preferences, which I use almost exclusively now on my D750.
But I have found this article very thought provoking so I will experiment with this technique for a while now. One option could be to use the highlight priority mode and set EV Comp to +2.3. . . what do you think?
Thanks for this article. . . I thought myself once a “good” photographer who then realised he was not really that good. . . now questioning everything I do, tearing it down and starting again. My recent switch from Aperture to Lightroom started that process.
Bruce,
Thank you for your interesting comment. It actually sounds like your method of exposure really was ETTR in disguise — as I mentioned briefly in the article, a high-contrast subject will probably need negative exposure compensation (in matrix metering mode) if you want to preserve the highlights. After all, preserving the highlights is the most important part of ETTR.
The highlight priority mode was Nikon’s opportunity to implement a proper RAW exposure option, but as far as I can understand (not having a camera with highlight priority mode) it doesn’t quite succeed. I think that its issue is that, with scenes that do not have bright highlights, the camera effectively reverts back to Matrix metering. In other words, the highlight-priority feature only activates when the image has noticeable highlights in the first place.
This means that you could still overexpose a shot that doesn’t have highlights, since you would just be applying exposure compensation to a normal, matrix-metered exposure.
Again, I do not have direct experience with highlight-priority metering, so I recommend that you do some tests yourself. It could very well be an accurate way to expose to the right, which would be quite a nice feature.
If anyone has some experience with this metering mode, I would love to hear your experiences.
~Spencer
My experience with the Highlight-Weighted Metering on my D810 is that it tries to preserve highlight detail to the max which in many cases results in an underexposed overal image, basically making it more ETTL then ETTR.
Excellent article. Very comprehensive and well explained. Thanks for taking the time to post this.
Thank you, Mac! I am glad that you enjoyed the article.
~Spencer
Hi Spencer
Great article
Here is what I think is the math of +2.3 stops, to add when reading pure white.
A grey card reflects to the camera 18% of the light. Plus 1 stop would reflect 36% (double), while +2 stops is refecting 72%, and +3 stops would reflect +144%, The pure white color beeing 100%, and while +2 stops reflect 72%, 28% is still missing + 0.388 stop ( 28/(144-72) stop) is to be added . So metering with the camera pure white the camera metering sensor will try to see it as a grey card (18%), Because its pure white we metered and not a middle grey (18% reflectance) color requires we add +2.38 stops to the reading, to make the reading as a reading from a grey card (18%).
Hey Luc,
Your math is quite interesting, and it makes a lot of sense. I had never thought about ETTR that way before, so thank you for showing me a new perspective.
There are just a couple things to point out. First, of course, the values in the real world are never quite so nice. For example, I have been able to recover detail in a spot-metered area even with an exposure compensation of +3.0 on a D800e. Further down in the comments, Iliah Borg mentions that some cameras require up to a +3.5 EV compensation for spot metering.
Also, although it is a common myth, most spot meters actually do not meter for 18% gray, but rather 12-15% gray. This could be the reason for the difference — i.e., 12% -> 24% -> 48% -> 96%, a three-stop boost that hovers on the edge of pure white.
Thank you for sharing your perspective.
~Spencer
Spencer
If we do it the other way from pure white 100% , one stop down is 50% two stops down is 25%, three stops down is 12.5%, while four stops is 6.25%. So 3.5 stops down needs to substract to 12.5% -0,5 X (12.5-6.25) for a value of 9.375% for their grey card reading, that seems to me way off from the 18%, I suspect the camera sensor get fooled (overloaded) by the brightness level on some cameras light meter…….
3 EV is the equivalent of 12.7%, Canon 1Dx has 3.4 EV headroom in highlights, which is the equivalent of about 9% log2(0.09) = 3.47EV
> A grey card reflects to the camera 18% of the light
Dear Luc,
Things are not exactly as they seem to be. Average scene reflectance is considered to be between 11.5% and 13%, that is below 18% middle grey. That is why we tend to use dome diffusers on exposure meters (unless doing flat repro work), and that is why the grey card needs to be tilted midway between the sun and lens axis if used to determine exposure.
I agree. I have always thought it was 12%!
Anyway, I remember in my professional days when shooting medical/dental subjects for clinical records, checking my Nikon (D2x) cameras against an 18% grey card and finding the histogram, which in theory should be sitting straddling the centre, was in fact offset by about 1/3 stop. However, as my subjects (teeth and bones) were predominantly whit(ish), I figured the 1/3 stop underexposure would stand me in good stead as a safety buffer – and in general, it did, especially as highlight recovery in post process was not as good then as it is now while the tiny bit of extra noise was undetectable.
Excellent article! So good that it must be deemed archive and file worthy. Pros, Cons, practicality, and practice is discussed. Thank you Spence. :)
Thank you, Mon! I am glad that you found the article useful.
Technology can do a lot, with compensation values, histograms and alike and I know, I bought it all. I own the top of the line gear, cameras and lenses but none of it can make the picture. It can only take it. Knowledge of light and composition exceeds all investments. For landscape photographers it means getting up in the dark, often arriving in the dark, hiking to the place in minimal light in often freezing temperatures to shoot the light. Yes, I am shooting the light and it its interaction with the subject matter. While we frustrate ourselves in tweaking the technology, in good light, it becomes less important if important at all. I chuckle at the comments below especially the one that states the grey card meters 12-15% grey as oppose to 18%. If you don’t have good light and a mastery of compensation, leading lines, juxtaposition, etc. A 3% difference in the gray card isn’t going to mean much. If you don’t have good light no matter how you expose, left or right will not render the desired image. Knowledge of the technology is important, don’t get me wrong but it can not displace nor will it displace light and composition.
Xrulan,
Thank you for putting things into perspective. Sure, it’s nice to have the best image quality possible, but no one really cares so long as your images are strong. Articles like these are for people who are vying for that last bit of extra image quality, or for people who are simply curious how all this works.
~Spencer
Great article, Spencer. I think the key to ETTR is “important” highlights. There are many highlights that are unimportant or have no data anyway so they confuse the issue (and the jpeg histogram).
The resistance to ETTR that I have observed from many people is usually one of gear-level origin. When you are trying to capture wildlife or wiggly things like kids and your 300mm lens only brightens up to f/5.6 and your camera really doesn’t have that great of an ISO performance if you pump up ISO too high, it is hard to be excited about jacking that exposure comp to the plus side to move the histogram to the right. If you don’t have a high end camera and lenses you are deeply relying on squeezing out that little bit more shutter speed but you can’t let in any more light and you are afraid to boost the ISO too high and get muddy pictures. It is more tempting to go negative with the exposure comp to get more shutter speed and then boost exposure in post (even though many times you are shooing yourself in the foot doing that when you could have just added the extra noise with the higher ISO rather than the minus exposure — which starts the whole ETTR argument over again lol).
Joshua,
It’s a wonderful cycle : )
The funny thing is that most people who claim to shoot ETTL (expose to the left) claim that they are trying to retain highlight information. In reality, ETTR (done properly) saves just as much highlight information — and it makes the shadows cleaner.
Thanks for sharing your position.
~Spencer
Yes, exactly. There is so much data on the right and, in my experience, so much noise to the left when you try to move it right in post. :)
Important note on this technique, it’s only effective in superbly bright scenes. Take the last example from above, “ISO 100, f/7.1 at 1/320”, that’s pretty fast for a mid range aperture. That’s a pretty bright scene, if that photo is the by product of a pull.
I find that in normal to darker light it’s actually more effective to expose to the left and push a photo, as it will allow an equivalent exposure while allowing less shake, sharper edges and allow for deeper, more richly colored shadows. This will generate some noise, but the noise is easy enough to rectify by flattening the blacks slightly in curves – literally a 2 second fix in PS, in curves you drop a dot at your darkest shadow, your brightest black, and then drag the darkest black up to flatten the black end of the curve, effectively evening out the noise.
Something else to consider, when pulling a bright lightly colored object, you will introduce color noise. This can be a fairly noticeable problem if you’re using say the BnW filter on Luminosity to darken an overcast or cloudy sky. It’ll get splotchy on you quick as blues begin to separate into purples, aquas and yellows.
facebook.com/shepshot
John,
So long as you are at base ISO, even if the light is low (so you would probably be using a tripod), ETTR works better than pushing exposure after the fact. The ETTR dandelion image was shot at 1/15 second, f/3.2, and ISO 100, so not exactly a bright scene. Still, as the comparison shows, ETTR is noticeably better than the metered exposure.
~Spencer
Right…. But you’re still shooting at 1/15. If you had pushed instead, you would have added crispness, improved the depth of your shadows and improved the vibrancy of the flowers. All which would combine to improve on what is otherwise a fairly flat image. Just saying this is a good technique, but the applications are more limited than you’re presenting which might be a little misleading to some of the new guys. Do you have skype? I’d be happy to show you how this works in PS
John,
Increasing exposure will always introduce noise, including color noise, which means that the photo has less color information (hence, it is less vibrant once you reduce color noise). Likewise, pushing an underexposed image does not improve detail or crispness over a properly-shot ETTR exposure reduced to the same brightness level — it actually decreases detail thanks to the added noise. The amount of light in the environment doesn’t affect this.
Every sensor has an innate level of read noise. If you expose to the right, you are improving the ratio of signal (actual image detail) to the read noise. The higher your signal-to-noise ratio, the more detail your image will have (especially in dark areas, making for “deeper” shadows). Obviously, exposing too far to the right (ETFTTR?) will blow out important highlights, and it is never preferable.
Perhaps the increase in noise/grain you get from pushing the exposure is giving the appearance of added sharpness? And maybe the lack of shadow detail (leading to “crunched,” darker shadows) is making them look richer to you? I may also be misunderstanding what you’re saying.
I think it would be most helpful if you could upload images directly that demonstrate what you are seeing, since it would make it easier for other readers to understand what you mean.
Thank you,
~Spencer
The sharpness is not a by product of editing but shooting. It allows the the ability to use a faster shutter speed and narrower aperture, giving more opportunities to reduce camera or mirror shake, and allow you to get closer to the sweet spot on your lens.
Also, crunchy blacks are the opposite of what I’m talking about here. Think of it as reducing the contrast of just your blacks. So instead of a sharp cut off, you get an even gradient, where it’s hard to tell where the shadows end and the blacks begin.
This is one of those things that seperates film and digital senors. What you’re saying used to be correct. Noise was essentially a false positive, registering luminosity where there shouldn’t be, a by product of over sensitivity. But it doesn’t work that way anymore. Rather what you’re seeing is contrast where it shouldn’t be. If you manage the contrast – ie flatten that minuscule section of curve – you manage the noise.
The same holds true for the brights. If you lower the luminosity of a bright lightly colored object you are essentially raising the contrast between what would normally be fairly similar rgb values, with the color differences separating (read: producing color noise) at an expodential rate as the values decrease.
I’m mobile at the moment. But let me see if I can put together some examples when I have a minute.
If I’m exposing anywhere at all, it’s going to be to be to the left. Tories are not to be trusted.
Benefits from ETTR are not in theoretical but documented facts and it has to do with how digital photography works. In digital, the more you expose (before blowing out) the more detail/levels you allocate in blacks and shadows, which will effectively boost your IQ. This will not work in analogue photography.
Who would have thunk it.
I think in a general sense it’s pretty close to the technique for exposing color negative film. In both cases one is aiming for good shadow tonality (low noise/grain) and tonal separation without blowing-out (digital) or blocking-up (neg film) highlights. The key difference is how one meters the subject — with digital it’s the highlights that are critical, with neg film it’s the shadows. For studied shots (other than studio flash), I contend there’s no better tool than a handheld 1˚ spotmeter with its own precision viewfinder. Testing and studying the resulting negs or Raw histograms is how I learn the camera’s response in order to “place” certain brightness values in the subject.
Rather than making colors more vibrant, ETTR makes colors look more washed out, which can clearly be seen in your dandelion and is also explained here: https://howgreenisyourgarden.wordpress.com/tag/expose-right/ Hence, in my book, ETTR is a useless technique in most circumstances. Also, if ETTR were such a clever idea, camera manufacturers would long have built in an ETTR button, as the camera software would be able to decide much more accurately how much to overexpose than a user can from glancing at the histogram.
There are 3 conditions to use ETTR successfully:
– sensor linearity
– lossless raw
– raw converter linearity (that includes the linearity of colour transforms, white balance, “exposure adjustment”
If the above conditions are met, the noise in shadows is reduced without anything bad happening to colour or resolution. And even with lossy raw ETTR still makes good sense if the DR of the scene is more than 7 stops – compare http://www.fastrawviewer.com to http://www.fastrawviewer.com
Not sure what the above two pics are supposed to show me…. The problem is that sensor linearity and raw converter linearity are not be found in the real world. Hence, ETTR produces washed out colors, of which I’m not a fan…
> The problem is that sensor linearity and raw converter linearity are not be found in the real world.
You are so wrong…
Then show me two pics, one with normal camera exposure and one with ETTR and then “down-exposed” with a RAW converter, where the colors are exactly the same. You won’t be able to do that. In the picture with the dandelion above, it is obvious that the greens and also the yellows are washed out in the ETTR pic.
ETTR was maybe a useful technique 5-10 years ago when sensors produced much more noise in the shadows than they do today. Take a pic from a D600-D810 and you will never have any issues with noisy shadows at sensible ISO levels…
Nuff said!
> you will never have any issues with noisy shadows at sensible ISO levels…
> Nuff said!
Very true, you said just enough to demonstrate you do not realize how things work. High ISO settings are not the cause of noise.
> show me two pics…
Well, take those two shots with your camera, and send me the raws.
Iliah, I know very well that you are highly knowledgeable in some specific areas of photography, but will you kindly refrain from writing your highly inept comments on PL.
Your far too frequent use of a terse authoritarian style is far from serving the best purposes of PL, which is to both inform and encourage non-expert readers via using non-intimidating and highly-informative rhetoric backed by empirical evidence and easy-to-understand illustrations.
Your knowledge of photography does not in any way compensate for your lack of understanding in other areas, especially:
1. In science, there are no authorities; at best, there are just a few experts.
2. As Dr. Phil Plait requested in 2010 at TAM 8: “Don’t be a Dick.” (It’s available on YouTube.)
You know that most of the information is in the right part of the histogram? I don’t see “washed out” colours. If you have more information you can make use of it and have more leeway in post.
Hans, I’m not seeing what you are seeing in the three dandelion images. In the first image, which hasn’t been compensated for the ETTR, obviously the colours look washed out.
In the second image, which has been compensated “AFTER ETTR (DARKENED)”, I’m unable to see any washed-out colours. I can see only a tiny difference that I think is caused by the compensation being very slightly below the exact amount required — by slightly below, I’m talking a tenth of an f-stop or less i.e. a few percent less than optimal.
Bearing in mind that tiny exposure difference, please explain how Spencer has not shown you “two pics, one with normal camera exposure and one with ETTR and then ‘down-exposed’ with a RAW converter, where the colors are exactly the same. You won’t be able to do that. In the picture with the dandelion above, it is obvious that the greens and also the yellows are washed out in the ETTR pic.”
The third dandelion image (100% crop) also has no washed-out colours and it clearly shows the reduction in noise resulting from ETTR. I’d say that the camera used to take the shots and the RAW converter used in post-processing both exhibit remarkable levels of linearity.
I’m not claiming that you are wrong in a general sense, Hans, because I’m sure that there must be a plethora of camera plus RAW converter combinations that do not exhibit adequate levels of linearity to endorse the use of ETTR to extract better image quality from those combinations.
As always with photography, it is each individual photographer who must experiment with their particular hardware and post-processing in order to get the best images for their intended purposes.
How about hue twists? “changing the exposure sliders in Adobe Photoshop Lightroom or Adobe Camera RAW will also change those colors on the fly” source: https://www.hansvaneijsden.com/colorchecker-perfect-skin-colors/
Hans, the photo’s in the article are not washed out, the ETTR image just shows less contrast. As you might see from the accompanied histograms, the shadows in the histogram are leaning to the right (because of ETTR). Pulling the shadows in post to the left (Blacks slider in ACR) will increase the contrast and improve the colors.
The link you give is an interesting and thought-provoking read, so thanks.
Hans,
Perhaps my terminology was wrong — rather than increasing the vibrancy per se, ETTR increases the color information. That is why the comparison shows less color noise with ETTR than with a metered exposure.
Any differences in the dandelion color (and to be honest, I see very few) are likely due to changing lighting conditions — I shot this image right as the sun was setting.
If you ETTR too far and blow out a color channel, you would see a decrease in vibrancy and color information, although that is not what happened here.
I think it is a dangerous game to predict that manufacturers would have implemented an ETTR feature by now if it were useful — see how long it has taken them to implement an electronic first curtain shutter, and even then one that does not work perfectly.
~Spencer
Nice article, I think auto ettr in canon cameras when used with magic lantern is also a good option as camera does all the calculations for me.
Mustafa,
Yes, all of us Nikon users are jealous of Magic Lantern : )
~Spencer
But we need ETTR less then Canon users (at native ISO) because we have several stops more dynamic range to push shadow detail in post ;)
Not really. If we fail to use ETTR to the full, we lose data or, more accurately, we fail to gather all the data available to us. Pushing in post is never the best way – it’s a partial correction of a mistake.
Sure I know it is true for every sensor that ETTR gives the best data I don’t dispute that at all.
With my reaction, I was actually reacting to the Magic Lantern mention by OP with intended pun, that’s why I finished with a ;)
Hey Marco, I got the pun without the explanation. Just saying, sometimes read between the lines – especially when you see a smiley
I just have observations because I really do not understand the technical aspects well enough to “know” why things happen the way they happen. Here is my general observation about ETTR – it works well in circumstances like the one I am about to describe. A week ago I was shooting action pictures at a dog show. It was held in a big exhibit building that doesn’t really have any windows (a few doors were open to let in a little light) with high ceilings and fluorescent lights. I was taking pictures of dog handlers running their dogs around the ring. With a D7100 and fast glass, I still needed high ISO’s to get the shutter speeds (1/500 or above, in general, although a little slower is possible with the action coming toward you) I needed. This leads to too much noise. By exposing to the right, I am able to get good results (after PP) at ISO 3200 and usable results at ISO 6400. If I didn’t expose to the right, I would have had too much noise. Fortunately, I knew this from past experience. Regardless, this is how I have used ETTR – to get less noise.
Spencer, if I may quibble with a fine article…. ETTR is generally MORE useful when shooting JPEG. That’s because its chief advantage historically has not been to lessen noise but to avoid banding produced during editing, especially in darker tones — and this happens primarily with 8-bit images with their fewer tone levels. In Photoshop’s histogram, you can actually see the gaps created when editing such images. But once you get up to 12- or 14-bit, banding is much harder to produce in normal editing, so the benefits of ETTR are more subtle, and as you say, in many cases even questionable.
When somebody has any intention to edit their photo’s and want to prevent posterization (banding) they should be shooting raw in the first place! Shooting ETTR is risky in JPG as blowing out (overexposing) the highlights will be unrecoverable.
The article states correctly that one of the benefits is that ETTR produces cleaner (more detailed) images, this is both in raw as well as JPG mode.
Nice article. However my own experience is that ETTR gets more important the higher the ISO gets (using my D810).
It seems like lifting shadow details from low ISO images during raw development is no problem, but the higher the ISO get, the more noise appears when lifting the shadows. To illustrate this, I have attached an image that I shot at ISO12800 and is exposed far to the right as the overexposed towel in the mans neck shows.
Just wanted to say whether we choose to use this technique or not, this article helped me learn many concepts and reinforce some concepts.
Very well written and was easy to read.
Hey Spencer,
I’ve just started reading your article on BTTR, but I’m afraid I’ve got stuck on the first two histograms, and I wonder if you can help. The first shows the ‘correct’ exposure for the columns photo, and the second the ‘ideal’ exposure. What I don’t understand is why the two histograms have a completely different shape, as opposed to the second one having the same shape, just shifted to the right, as I would have expected. Could you ‘shed some light’ on this?
You’re correct — in general, increasing an image’s exposure will create a histogram with the same shape, just shifted farther to the right of the histogram. However, if you increase the exposure enough (more than a stop or so), the new histogram will barely resemble the old, which is what I did above. This happens because of the process of increasing exposure: mid-tones bunch into highlights, but some of the deepest shadows stay behind.
Take this scenario as an example: if your original exposure had a histogram shaped like a perfect bell curve, increasing exposure would cause the histogram become skewed — most of the curve will bunch to the right, as shown in the picture below, but some would remain on the left-hand side. With a more complex histogram (like that of any photograph), this same effect occurs, just in a way that isn’t as neat and tidy.
If you have Lightroom, you can test this out on your own images by moving around the exposure slider, and then studying the resulting histogram.
…
Image By Asitgoes (Own work) [CC BY-SA 3.0 , via Wikimedia Commons
Thanks Spencer, I see what you mean. OK, I can get on with the rest of the article now!
♪♪♪♪♪♪♪Best of list s I RECEIVED FIRST DRAFT OF $13000!@ak27 by photo:>
➨➨➨➨ Read Full Article
—————————
< col Hiiiiiii Friends….uptil I saw the receipt for $6027 , I be certain …that…my friend woz like truly erning money part time from there labtop. . there sisters roommate has done this for only twenty one months and by now repaid the debts on their mini mansion and purchased a top of the range Citroën DS . More Info SEE FULL DETAIL
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i
How Manual Mode and Spot metering can tell me the exposure setting???
“Set the camera to manual mode and spot metering. Point the metering
point on the brightest part of the scene, and record the exposure that
the camera tells you to set”
Can I do it in Program Mode?
If you point your spot meter at the brightest part of the scene and record the exposure reading, you will get severe underexposure.
You cannot apply ETTR in Program or any Auto mode without looking at the histogram and applying compensation until your histogram is just short of the clipping point on the right hand side.
To use ETTR you must first have a good understanding of how exposure works.
To further Betty’s instruction, I would add that you need to make a study of dynamic range of any scene you want to capture. Finding the middle tones and then exposing correctly will get you closer to a good exposure which then can be adjusted by reading the histogram. When taking a photo of a person and filling the frame with your subject, were you to just meter the small amount of light from the background sky, as Betty stated, your picture would be underexposed. Equally, if you meter for a white blouse on your subject or other light colored object in the frame the same would happen. You need to find a mid tone to take your exposure. There are many very good books that would help you learn how to read the dynamic range of your subject.
Hi Spencer. Give credit to Canon. You can get raw histogram with Magic Lantern.
Hi Spencer,
just one short note (I didn’t find anything about this in the comments). I am an avid astronomer and like to shoot astrophotos. Noise is the biggest enemy here. That is why astrophotographers like to use a technique called ‘stacking’ where 10, 20 or even hundred shots are composed to one final picture. Noise is greatly diminished this way (=1/squareroot(N)). ETTR is a simple technique for us doing this the simple way. If you expose with +1EV it’s like adding two photos and this way reducing noise to 1/squareroot(2) or if you expose with +4EV it’s like reducing noise to 1/4 which can help a lot.
One renownded astro photographer is Justin NG (http://www.justinngphoto.com) who uses this technique (to its limit!) to take photos of the Milky Way from the extremely light polluted city of Singapore.
Regards
Peter
I unintentionally did this while taking some sunset photos, and I really noticed a difference! Thanks for explaining so that we know what it is! It really did have a large impact on the final result.
Thanks, Spencer, glad you found it useful!
From Method 2, you stated to set your camera to manual mode and use exposure compensation. I always thought that exposure compensation doesn’t work in manual mode.
Thanks for this very useful, in-depth article. I just stumbled across your site and will be around here often from now on.