I am a hoobyist photographer. I like to do landscapes when I can so this article interested me. I want to know how other people feel about this. I would also like to know how I would go about taking a picture that is overexposed but not to the point of clipping. I personally think it would be better to do bracket the picture then layer them in photoshop.
ETTR is very common for landscape photographers, most pros will do it to some degree (although some don't realise that's what they're doing and do it to reduce shadow noise, which eventually boils down to the same basic concept).
Easiest way to do it is to enable 'blinkies' on your camera and either use exposure compensation if you're shooting in a priority mode, or adjust your exposure in manual mode, until you are just at the point of having no blinkies (or you have specular highlights/point light sources, they are the only parts blinking).
Note that it doesn't increase dynamic range per-se. If your image has a wider dynamic range than your sensor, then you are only ever going to capture at most the range of your camera in a single shot. If you follow the mantra that underexposed shadows are less obtrusive than overexposed highlights though, it can be used to get you to the best settings given the camera limitations. If the scene has more dynamic range than the camera and you want to capture it all, you will either need to modify the light (e.g. using ND grad filters) or take multiple images.
Where ETTR is most useful is when your scene doesn't exceed the dynamic range of your camera, either because that's how it is, or because you've modified the light using filters. In this case you can capture all of the range of your image with a variety of exposure settings, all the way from capturing the blackest part of your image as black and the whitest part of the image as grey, to capturing the blackest part of your image as grey and the whitest part as white (which is what ETTR does).
In this situation, where the image dynamic range is less than your sensor's, you will almost certainly want to modify it in post-processing so that both the blacks are black and the whites are white. If you do this by raising the whites, you will introduce more noise and have fewer colour tones than if you do this by lowering the blacks (as explained in the article).
The benefits are rather subtle -- if you're not at the level where you are shooting RAW, using ND grads to control light, and doing lots of selective dodging and burning in your processing, it's questionable whether you'll benefit from it (and it's generally considered to be a bad idea when shooting JPEG). But if you are making every pixel of your image count and want the ultimate quality, it's a useful technique to understand and learn.
Alexis is a major expert in this, heed his advice. Landscapes particularly benefit from pushing to the right (portraits, not so much).
My addition is, live by your histogram. If you can't push it right with more exposure time, push it up. (Night time landscape shooters will know what I'm talking about.)
Old school (analog/chemical) photography dictated the opposite: push to the left. Digital dictates the opposite: push to the right. It's *way* easier to recover highlight detail than shadow detail. If the light's not there, no editor can bring it back.
Shooting RAW is essential. Lightroom makes RAW as easy as JPGs in any other software, so I use LR as my primiary photo editor. Occasionally, my images make their way to Photoshop.
Imagine Adobe Camera Raw on steroids, and that's Lightroom.
@abirkill@cameronknowlton Thank you for your clarifying comments on ETTR. The DPS artical that has fueled this discussion is good, but lacking some practical comments as you are making.
ETTR is just a way of saying expose for the highlights. back in film days we would say expose for the shadows and develop for the highlights. You had a chart listing chemical temp and based on the temp and type of film you could develop longer and recover more highlight detail. This was called pushing your film. There was the famous zone system that you used by taking meter readings at what you wanted to be pure black and what you wanted to be pure white and figure out the dynamic range basically and you could figure out your exposure and if you needed to push the film and by how many stops. Thank god we dont have to do that anymore. It sucked terribly bad lol. We had a whole class just on the zone system. Now you do the opposite. You expose for the highlights which is what is on the right of your histogram.
@soia@cameronknowlton@abirkill i am guessing that in the digital world, "semi-blown" highlights don't blow all three RGB channels typically and detail can be recovered on one of the channels. Whereas "lost in the black" is just that. If it is not there then no channel has the detail. Would that be the sound byte at the bottom line of this discussion?
@frankhymus Yes...ish, but even blowing the highlights on a single colour channel results in a rather unpleasant effect. Here's a shot from a looonnnnng time ago with a camera that only displayed a histogram/blinkies based on the green colour channel:
The white sections (e.g. just above the observation deck) are where the blue light was bright enough to blow out all three channels, and the solid blue sections are where the blue light was only bright enough to blow out the blue channel.
Write a Reply
Sign up for a free account or Sign in to post a comment.
Tweets ahoy!
Join the Group
Pin With Me
Easiest way to do it is to enable 'blinkies' on your camera and either use exposure compensation if you're shooting in a priority mode, or adjust your exposure in manual mode, until you are just at the point of having no blinkies (or you have specular highlights/point light sources, they are the only parts blinking).
Note that it doesn't increase dynamic range per-se. If your image has a wider dynamic range than your sensor, then you are only ever going to capture at most the range of your camera in a single shot. If you follow the mantra that underexposed shadows are less obtrusive than overexposed highlights though, it can be used to get you to the best settings given the camera limitations. If the scene has more dynamic range than the camera and you want to capture it all, you will either need to modify the light (e.g. using ND grad filters) or take multiple images.
Where ETTR is most useful is when your scene doesn't exceed the dynamic range of your camera, either because that's how it is, or because you've modified the light using filters. In this case you can capture all of the range of your image with a variety of exposure settings, all the way from capturing the blackest part of your image as black and the whitest part of the image as grey, to capturing the blackest part of your image as grey and the whitest part as white (which is what ETTR does).
In this situation, where the image dynamic range is less than your sensor's, you will almost certainly want to modify it in post-processing so that both the blacks are black and the whites are white. If you do this by raising the whites, you will introduce more noise and have fewer colour tones than if you do this by lowering the blacks (as explained in the article).
The benefits are rather subtle -- if you're not at the level where you are shooting RAW, using ND grads to control light, and doing lots of selective dodging and burning in your processing, it's questionable whether you'll benefit from it (and it's generally considered to be a bad idea when shooting JPEG). But if you are making every pixel of your image count and want the ultimate quality, it's a useful technique to understand and learn.
My addition is, live by your histogram. If you can't push it right with more exposure time, push it up. (Night time landscape shooters will know what I'm talking about.)
Old school (analog/chemical) photography dictated the opposite: push to the left. Digital dictates the opposite: push to the right. It's *way* easier to recover highlight detail than shadow detail. If the light's not there, no editor can bring it back.
Shooting RAW is essential. Lightroom makes RAW as easy as JPGs in any other software, so I use LR as my primiary photo editor. Occasionally, my images make their way to Photoshop.
Imagine Adobe Camera Raw on steroids, and that's Lightroom.
The white sections (e.g. just above the observation deck) are where the blue light was bright enough to blow out all three channels, and the solid blue sections are where the blue light was only bright enough to blow out the blue channel.