I wish somebody with competence would try this, post the results, and let us know what they think

If I look at the dynamic range graphs on dpreview the most sensors have a hard cut in dynamic in the highlight area . It is very simple to make a photographic with the Merrill cameras against the setting sun and get a lot of seperate rings around it and no smooth clorored sky.
This wouldn't happen with film (and the old sensors of a Fuji S3/S5).
I can imagine it is more difficult to hit a point before this cut and not to lose informations in the picture as to recover from the shadows where the slope of the dynamic curve of modern sensors is less high.
 
These have all been interesting responses, but since my own attempts to apply this technique have been far less than convincing, I'll think I'll keep doing what I have been doing.

Cheers, Jock
 
Hey All,

An acquaintance of mine and his wife shoot 30 to 35 weddings a year. They use ETTR all the time. if you think about it they are often in churches and such that aren't really well lit and so ISO is up and ETTR helps alot. They process on macbooks and then review the images on 55in 4k tv. Get them right highlights and shadows when viewing them that way keeps customers very happy.

-Ed-

It sounds as though they are a bit mixed up as to where using ETTR will provide benefits. As soon as you go above above a camera's base ISO, the only way to expose high is to raise the ISO since you are already presumably shutter speed limited. Raising the ISO is counterproductive to the goal of using ETTR to reduce noise.

If anyone has ever used a camera with an expanded Low ISO setting that is less than the camera's true base ISO, then you have already seen (in a simplified, uncontrolled manner) some of the potential benefits and also potential downsides of ETTR without having to do any of the processing yourself. Those Low ISO values promise less noise because the image has been overexposed by one stop and then pulled back, but also require you to be mindful of where your highlights are because of that same overexposure.
 
Hey All,

I guess the question is - is ETTR only a benefit at base ISO or is it a benefit at higher ISOs as well? I think what he was saying is he knows the depth of field (f-stop) he wants and he knows what speed will work so given those are fixed how do you ETTR and is there still a benefit? Also they use the newest full frame Canons.

-Ed-
 
Here's what confuses me. ETTR is better at high iso/low light. To do it you overexpose, which counteracts going to high iso in the first place, e.g., long shutter times. What am I missing?
 
Hey All,

I'll say the way I see it and if I am wrong somebody can correct me.

ETTR is all about putting more light into the dark areas of an image without overexposing (unless intentional) the bright parts of the image thus increasing the dynamic range (range of light captured) of the image. The basic point is that you can through post processing bring out more detail in the bright parts of the image by reducing exposure (light) in post processing than you can gain by adding exposure (light) to the dark areas of an image because of noise in the dark areas. I suspect that in theory this should be true at all ISOs but I doubt it is because of the properties of any particular sensor. By this I mean the "rate" at which high ISO noise is added as ISO goes up. As long as the benefit gained from adding light to the dark areas of an image outweighs the increase in high ISO noise in the image it is likely the image will benefit from ETTR. Now exactly where are those limits for any particular sensor combined with the nature of the scene being photographed combined with the photographer's intent - is up to each and every one of you to figure out for your photographic circumstances, intent and sensor.

How you "add" light (i.e. ETTR) depends as it can be done by increasing exposure, reducing f-stop, increasing ISO (which increases light sensitivity of the sensor as I understand it) or some combination. In the above I described it in terms of fixed f-stop and shutterspeed. Increasing light via F-stop and or shutterspeed has all the limitations inherent in f-stop and shutterspeed (e.g. changing depth of field, motion blurries, etc).

The other way to deal with images of scenes with a broad range of light in terms of dark areas to light areas is with multiple exposures and image blending - ala the proper use of HDR.

Now I may be wrong but this is my understanding of ETTR.

-Ed-
 
Hi EasyEd

I understand ETTR a bit differently.

Firstly ETTR has nothing to do with increasing dynamic range (dynamic range describes the ratio between the maximum and minimum measurable light intensities).

If your camera, say a D800, has 14 stops of dynamic range and you haven't clipped your histogram, then you've captured the entire dynamic range your sensor is capable of capturing from the scene.

By using ETTR you're not capturing greater dynamic range, you're just lifting your shadows (that is if you haven't clipped your histogram). If your camera can capture 12 stops of dynamic range, then that's what you'll get. (This is not strictly true, but will serve to illustrate this point)

ETTR will also not help you in capturing more dynamic range from the scene, unless by raising your exposure, you are recovering shadows that were getting clipped.

If you want to expand the dynamic range to something more than what your sensor can capture, use multiple exposures with HDR techniques to do so.

What ETTR has to do with is retaining more color information in your image. The lighter parts of your image retain more information on tonal value than the darker parts. This has to do with the way digital sensors are designed.

This example from LL might help clarify this -

A typical consumer DSLR recording 12 bits per sensel is able to record up to 4,098 separate tonal values.

If we assume a 10 stop dynamic range this is how this data is distributed...
The brightest stop = 2048 tonal values
The next brightest stop = 1024 tonal values
The next brightest stop = 512 tonal values
The next brightest stop = 256 tonal values
The next brightest stop = 128 tonal values
The next brightest stop = 64 tonal values
The darkest stop = 32 tonal values


Now at any given ISO, your sensor will capture more color information in the lighter parts of the image, than in the shadows.

It's a function of how all sensors are designed and has nothing to do with how one shoots an image.

ETTR is just a tool to maximize the capture of color information by your sensor in any given situation.

How you achieve the exposure is totally up to you. You can widen your aperture, lower shutter speed, increase ISO all with the intent to push your exposure to the right.

Should one increase ISO to achieve ETTR. Well no. Should you raise your ISO to negate shake in an image. Certainly.

Currently the way the metering in your camera is set up, the priority is not to capture the most color information your sensor is capable of capturing.

Given this information, it is up to the photographer to decide how to use it. Should one raise the ISO to achieve ETTR? Or reduce shutter speed? Or go for a larger aperture?

If it's a fast moving action scene, I would not use ETTR at the expense of raising my ISO or lowering my shutter speed to introduce motion blur to my image. If it's a landscape/architecture shot and I'm on a tripod and can lower my shutter speed, I would definitely use ETTR.

With the way our cameras are metered to middle greys currently, ETTR is just not a very convenient option. If that changes, I would see myself using it a lot more often.
 
Hey All,

I guess the question is - is ETTR only a benefit at base ISO or is it a benefit at higher ISOs as well? I think what he was saying is he knows the depth of field (f-stop) he wants and he knows what speed will work so given those are fixed how do you ETTR and is there still a benefit? Also they use the newest full frame Canons.

-Ed-

At it's heart, photography is about capturing light, and there are two ways to increase the amount of light being captured by your sensor. The first is to increase the intensity of light by opening up the lens aperture, and the second is to increase the duration by lengthening the exposure. Of course the third point of the exposure triangle is ISO, but this only increases the sensitivity to light and doesn't increase the quantity of light.

The practise of ETTR exists solely to improve the technical quality of your images, and ideally means using a higher exposure than the camera's meter might indicate whilst being careful not to clip any highlights where you want to retain colour information. So, to take advantage of ETTR you wouldn't increase the brightness of your image by increasing the ISO speed, you wouldn't select a larger aperture than is ideal to give you the right depth-of-field and the least lens abberrations, and you wouldn't select a shutter speed slower than your lowest limit for subject movement or camera shake.

All that means that the only time where it is potentially beneficial to apply ETTR is in good light at base ISO and where you have room to lower the shutter speed without causing issues with subject movement or camera shake.
 
With the way our cameras are metered to middle greys currently, ETTR is just not a very convenient option. If that changes, I would see myself using it a lot more often.

Offering a metering mode tailored to ETTR is one way that might it easier to achieve, but the other is to use a camera with a) live view, b) a reliable and accurate representation of exposure, c) a jpeg mode that provides a good representation of the dynamic range of the sensor, and d) exposure warning tools like histograms and "blinkies" that also accurately reflect the dynamic range of the sensor.

Based on those criteria, the best cameras I have used for ETTR are Olympus m4/3 bodies and the Canon G1X. The worst are any cameras that rely on an optical viewfinder.
 
Hey All,

Very informative Boid! I understand the issues of inherent sensor dynamic range and of clipping and I assumed that ETTR was all about not clipping on the left thus increasing the dynamic range captured by the sensor. I think it can be about not clipping on the left but what you are saying is different as I would have assumed a linear response in information capture (voltages) across stops of light since I assumed all you are doing is recording voltages on a uniform medium across different wavelengths of light - but apparently not. The number of tonal values captured changes with stops of light - wow - I've no idea what is going on at the sensor/software level to cause this but apparently a sensor can output a greater number of voltage options at brighter wavelengths than darker - whood a thunk it! I am not arguing with you at all. From a practical perspective this would suggest to me that over exposure by some amount - as long as you are not clipping on the right - when you are committed to post processing is going to routinely give you more tonal range in your final image. If you are not committed to post processing and live with OOC jpegs it is different. Maybe overexposure by a stop is best as a rule of thumb unless you accept what you get OOC where people commonly underexpose by a stop.

Thinking further maybe over exposure by some amount is good as if I want what I get OOC I can reprocess the image/voltages in camera via push pull menus so having the extra information for an image and then if desired pulling it back for OOC is not bad - provided you are not clipping what shouldn't be clipped on the right in your over exposure. You can alwaya subtract information but you can never add it if it was never captured to begin with.

Fascinating - I've more to learn. Thank You!

-Ed-
 
Based on those criteria, the best cameras I have used for ETTR are Olympus m4/3 bodies and the Canon G1X. The worst are any cameras that rely on an optical viewfinder.

The two evfs I have used were on the X100 (original) and the very old Sony DSCR1. They weren't good enough to make critical judgements of ETTR. The rear LCDs are pretty much useless. So that leaves studying histograms and "ain't nobody got time fo' dat"
 
I've played around with it a bit. Problem I had was that getting the color right in the mid tones proved difficult. Medium sky blue turned to a lighter cyan and getting it back to blue was difficult.
 
Hey All,

I think I got it Boid although I may not have said some of it right.

Likely the sensor which just generates voltages when light hits it is uniform in it's response. I'm wondering if the separation into more or less tonal values is the "divi up" of the light spectrum by software converting the voltages - hence more of a software issue than a sensor issue. Irrespective it is interesting and good to know.

In terms of practical use it leads to a lot of questions. For example on my XE-1 I can have a histogram show while framing a shot and rotate the exposure dial to move the histogram left or right. But what would I really be doing - "changing" f-stop, shutter speed, iso or just software push/pull of what the camera wants to do anyway. If it is the last am I gaining anything by rotating the dial? So it seems to me I need to know the exact workings of the camera. jeez I'm not sure the manual even always says.

-Ed-
 
I've played around with it a bit. Problem I had was that getting the color right in the mid tones proved difficult. Medium sky blue turned to a lighter cyan and getting it back to blue was difficult.

I don't think that the tendency to turn blues into cyan is related to exposure, but rather it is symptomatic of digital sensors in general and particularly so with Sony sensors in my experience. This has, in part, led to my general dislike of that brand of sensor. Even with darker skies, cyan will still show through and give them an unnatural, cartoonish look.

The unfortunate thing is that with Adobe software it is now more difficult to remove cyan. The current HSL panel within Adobe Lightroom no longer has cyan as a colour channel that can be manipulated separately...

HSLLightroom_zpsaf5d67da.jpg
Join to see EXIF info for this image (if available)



...whereas my near ten year old copy of Photoshop CS3 does, and makes it quite easy to remove cyan because it has it's own colour channel.

HSLPhotoshop_zps88f59dd3.jpg



My usual workflow is to do raw conversions in Lightroom and them export as a tiff file to finish off in Photoshop since it has some tools that I prefer to use over Lightroom, and it is at this stage during processing where I will attempt to remove cyan from any affected images.
 
Be aware that ETTR only works for raw files, not for jpegs. I personally find it too much effort (I find myself hardly ever using raw in the first place these days), but it is a good way of minimizing noise and maximizing color depth. Once you have the ettr exposure, you can do anything you want with it, so porchard, the midtones issues you've experienced should be fixable. In the end it's just 1's and 0's, and the more different values you have available, the more control you have over where you want them to end up.
 
I don't think that the tendency to turn blues into cyan is related to exposure, but rather it is symptomatic of digital sensors in general and particularly so with Sony sensors in my experience. This has, in part, led to my general dislike of that brand of sensor. Even with darker skies, cyan will still show through and give them an unnatural, cartoonish look. The unfortunate thing is that with Adobe software it is now more difficult to remove cyan. The current HSL panel within Adobe Lightroom no longer has cyan as a colour channel that can be manipulated separately......whereas my near ten year old copy of Photoshop CS3 does, and makes it quite easy to remove cyan because it has it's own colour channel. My usual workflow is to do raw conversions in Lightroom and them export as a tiff file to finish off in Photoshop since it has some tools that I prefer to use over Lightroom, and it is at this stage during processing where I will attempt to remove cyan from any affected images.

It never ceases to amaze me how major software development teams will remove important functionality from their product with no more thought than to issue a disclaimer, if anything at all! The rules were laid down at the beginning of time (circa 1981-84) -- never remove functionality from a product unless you absolutely have to, and when you do make sure the prior version is available indefinitely, even if unsupported. That last part gets sticky. Shouldn't there at least be a plug-in, or are plug-ins a no-no with the subscription model?
 
I don't think that the tendency to turn blues into cyan is related to exposure, but rather it is symptomatic of digital sensors in general and particularly so with Sony sensors in my experience. This has, in part, led to my general dislike of that brand of sensor. Even with darker skies, cyan will still show through and give them an unnatural, cartoonish look.

The unfortunate thing is that with Adobe software it is now more difficult to remove cyan. The current HSL panel within Adobe Lightroom no longer has cyan as a colour channel that can be manipulated separately...

HSLLightroom_zpsaf5d67da.jpg
Join to see EXIF info for this image (if available)

Hey Nic, shouldn't the Aqua work in the same parameter as the Cyan did? I've noticed with my X-T1 that it wants to put more of the Cyan/Aqua in the sky so I usually push that over in the hue section closer to blue and that usually helps out.
 
Back
Top