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

Jock Elliott

Hall of Famer
Location
Troy, NY
Check out this article on Luminous Landscape: http://www.luminous-landscape.com/essays/the_optimum_digital_exposure.shtml

Would somebody with post-processing skill get it a try and render their opinion?

I tried it, and didn't see any discernible benefit, but I will be the first to admit that my post-processing skills are slim to non-existant. (Some of the folks here remind me of Renaissance painters in their ability to render gorgeous post-processing effects. Me? I'm like a kindergarten fingerpainter.)

So how about it? Will somebody -- or perhaps plural somebodies -- give the techniques in the article a rip and show some comparative results?

Cheers, Jock
 
It looks interesting, but I wonder if it's worth it - if it reduces the fun factor of photography too much.
 
I've seen the basic idea expressed before and it seems to make sense - it's all about maximizing information. I think though it's always going to work best when you're dealing with higher ISO's on the basis that the more you crank up the shadows, the more your always cranking up the noise, whereas pulling back the highlights isn't going to add noise. The danger is always in overdoing it and losing everything in the highlights. I'm not sure I would be bothered going to that much bother - perhaps if you're shooting landscapes and have the camera on a tripod it's easier, but if you're doing that then presumably you wouldn't be using high ISO's.. One quick and dirty way to try it our would be to use spot metering and then meter for a darker part of the scene - obviously this would be a bit more hit and miss but I expect with time you could end up with a reasonable judgment about where in the scene you need to meter.
 
While it's technically correct and certainly maximizing the image quality, it's also adding an extra step of abstraction while you're shooting.
I guess that's fine if you're shooting with a tripod and technical camera (a setup that requires time in the field anyway).
But for me, I try to get as close to the final image "in-camera" as I can these days.
(Don't want to process every single image at home. Spend too much time in front of the computer anyway :) In my "Hells Canyon" set I'm happily using some out-of-camera-jpgs without any additional processing.)

In my opinion "expose to the right" was more important a few years ago, when sensors were noisier.
But then again, I'm usually not brightening shadows as much as other photographers.
If you like to do a lot of processing/tone-mapping/etc. on your photos, ETTR might still be the best way to get the most processing leeway.
 
I read the article, and understood somewhat, what he's talking about, without trying to absorb the finer detail, and committing then to memory, for instant recall. My take on it, is that it is on the level of, teaching "practical theory".

I have a Sunday School class, that is geared to "teaching to teach". It is not practical for the average person, just beginning to learn. Too many details, for average every day living needs.

I've been studying photography for many years. But, not so much from the hyper-detailed approach. Before digital, I shot and waited, to see my "hits" and "misses". Most of my roll was usually thrown away, for various reasons. Mostly blah composition, or subject matter. Over time and reading voraciously, I gradually gained an instinct, as to what would generally be an acceptable exposure. But, then with prints the processor would mess with it.And the results were not really my own.

Then came digital, and I was a rank beginner, all over, again! And again, I read voraciously, and practiced, practiced, and practiced some more. And, tried to learn PP. But with the instant replay afforded, by today's digital cameras, and near unlimited shots available, to "get it right". I have begun to be able to get it "close enough" in JPEG, to be able to tweak it slightly in PS, to satisfy ME. If I were working for clients, then I may be forced to be a bit more "clinical", in approach.

My conclusion to the article:
Some very good information there. But do you need that much technical information, to get good exposure,and final results? Maybe not.
 
At the camera club I used to attend, those who ran the infrequent tutorials were always saying expose to the right. I really could not see the point. I don't like overexposing, and thats effectively what you have to do. I'm not a post processing expert by any means, but I cannot take my photography as far as this, technically correct or not. I see something I want to photograph, and am happy to accept less than perfect, or less than optimum. Maybe its something people who go in competitions need to be considering, but for people like me... hobbyist photographers... nah... just wayyyyy too much fuss.
 
My own (admittedly limited) experience, I have found that ETTR results in rather unsatisfactory quality in the mid-tones. Such things are subjective of course, but I prefer to under-expose a little, and lift the mid-tones with a curve, as necessary.

You'll probably get a few different opinions on this, but that's the way I see it.:)
 
There is no such thing as one correct exposure.
The "correct" exposure for any scene will be the one that captures the range of shadow mid and highlight detail that you want so far as your sensitive material allows.
You can manage this by internalising complex calculations about dB , lumens, ettr, zones, h&d curves and whatnot, or you can manage this by noticing how your pictures come out and how you were metering and exposing, or you can do it by guesswork, or you can do it by letting the camera do it all for you. Or a combination of all of these.
Take your pick ...
 
The problem I have with ETTR is that it tends to only really matter much at higher ISO - with today's sensor you can raise shadows a ridiculous amount without noise anywhere near base ISO. And you tend to need higher ISOs when - in situations when there's not enough light to shoot otherwise. So when light is already in limited supply, you're asking your camera to gather more of it. If you're doing that by raising ISO, it's a rather pointless exercise because the noise you're trying to limit you're adding to by cranking the ISO. And if you're trying to do it with shutter speed or aperture, you may be in a situation where you're limiting your ability to get the shot - the aperture is probably already maxed out and then it comes down to how slow a shutter speed will work for the shot.

Having had this discussion with myself in the past, I've basically decided to just ignore it. Hell, I like noise anyway - I call it grain and me and it tend to be friends. So, no ETTR for me... I barely believe in diffraction either - I'm weird that way...

-Ray
 
Or you can use the EVF (if you have one) and adjust settings til it looks "right."
.

This is exactly what I used to do with my digital cameras. Eventually I learned exactly how the EVF or LCD image would look for the sort of exposure I wanted. I combined this with using spot metering mode.
If you use a digital camera for long enough, you ought be able to get to know how the sensor is going to react to various lighting situations. This is one the reasons I think it's a mistake to change cameras every other week - how do you ever learn what it's capable of?
 
This is exactly what I used to do with my digital cameras. Eventually I learned exactly how the EVF or LCD image would look for the sort of exposure I wanted. I combined this with using spot metering mode.
If you use a digital camera for long enough, you ought be able to get to know how the sensor is going to react to various lighting situations. This is one the reasons I think it's a mistake to change cameras every other week - how do you ever learn what it's capable of?

I got the XT1 in, what, March? April? I am JUST now getting to where I can predict its output well, and get what i like pretty much right out of the card. And I shoot... too much.
 
Here are some photos from the OneZone™ book:

Book Photos – OneZone

With stuff like this, I look at the results and ask myself if I'd like to incorporate the look in my work. In this case, the answer is no as the results don't fit my "style" as it were. There's something off with them for me, an almost HDR quality where everything's been pushed to the middle of the histogram, but lacking the punch that rich blacks and strong local contrast can provide.

I was just working on a photo the other night that I struggled with until I stopped trying to lift the shadows. Once I let them be, the image came together. These photos remind me of that.

I do appreciate the tones preserved in the wedding gown as I've lately been pulling back the highlight slider to find subtle details in whites. But even there there's something off for me.
 
My own (admittedly limited) experience, I have found that ETTR results in rather unsatisfactory quality in the mid-tones. Such things are subjective of course, but I prefer to under-expose a little, and lift the mid-tones with a curve, as necessary.

You'll probably get a few different opinions on this, but that's the way I see it.:)


I'd imagine this is because you have to work in a completely linear color space to be able to retain the original midtone-saturation once the midtones are pushed into the highlights.

Most raw-converters apply an S-shaped filmcurve as a first step in the conversion.
These curves affect the saturation of colors in the shadows, midtones and highlights in different ways. Colors in the highlight range basically lose saturation by applying these curves.
Capture One allows you to work on linear data, so I might have a play with this on the weekend.

An illustration of the HLS-colorspace might help to visualize this problem (even though technically it's not an illustration of ettr)
You can see that midtones have higher saturation values than shadows or highlights, as the single RGB values can be further apart.
saturation_lightness.jpg


cheers,
Tilman
 
Here are some photos from the OneZone™ book:

Book Photos – OneZone

With stuff like this, I look at the results and ask myself if I'd like to incorporate the look in my work. In this case, the answer is no as the results don't fit my "style" as it were. There's something off with them for me, an almost HDR quality where everything's been pushed to the middle of the histogram, but lacking the punch that rich blacks and strong local contrast can provide.

I was just working on a photo the other night that I struggled with until I stopped trying to lift the shadows. Once I let them be, the image came together. These photos remind me of that.

I do appreciate the tones preserved in the wedding gown as I've lately been pulling back the highlight slider to find subtle details in whites. But even there there's something off for me.


I don't think the tonemapping/HDR-look of these images is inherent to ETTR or the one-zone thing.
But if someone is tonemapping his photos to a strong degree, he might get slightly less noisy results by applying these techniques.

At the end of the day, it's everybody's own decision how to work and what they find pleasing. There's always going to be some teachers promising you the "only way" to work in :)

Personally I admire more the imagery of a Rembrandt, a Gordon Willis or a Vittorio Storaro...
So in regards to this technique... I don't see why I should put in extra effort to preserve shadow detail, only to have even more effort later to get rid of it again. :)

Everybody's mileage will vary... and different techniques for different purposes...

cheers,
Tilman
 
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-
 
I use ETTR all the time, and I suffer for using it because of the way modern camera metering is set to recognize middle grey. It's mostly guesswork or peering at histograms, or making multiple exposures to get an ETTR image right.

I think of the image in camera as capturing "data". I don't think of it as the final image at all. The "final" image exists somewhere between the scene, the camera, the processing software and umm.. my head (or what little is in it).

From what I've read, modern digital sensors (unlike film) capture and retain more data (color information/tonal values) in the lighter parts of the image than in the shadows. This "data" falls of exponentially with every stop of light. Not so in film, where there was the same amount of "data" (information) in both light and dark areas. Some more info about it here (http://www.luminous-landscape.com/tutorials/optimizing_exposure.shtml)

I would simply like a metering option in all cameras that identifies the brightest white in an image, and helps me capture the image without losing data, regardless of how ridiculous it looks on the LCD at the back of the camera. I shouldn't need to jump through hoops to make an image like that if I so choose, with absolute certainty that I haven't lost any detail in the highlights.

It'll cost the camera makers nothing to do this, and will make people like me very happy.
 
I would simply like a metering option in all cameras that identify the brightest white in an image, and help me capture the image without losing data, regardless of how ridiculous it looked on the LCD at the back of the camera. I shouldn't need to jump through hoops to make an image like that if I so choose, with absolute certainty that I haven't lost any detail in the highlights.

It'll cost the camera makers nothing to do this, and will make people like me very happy.

Indeed - I have often wondered why no digital camera (AFAIK) has provided this. I wouldn't want a camera with this as a fixed characteristic, but I'd certainly welcome it as a setting that could be selected for particular circumstances.
 
Back
Top