This photograph was taken during a walk in the woods this weekend.
This is the as-shot photograph. This is a very typical family/friends/social snapshot in that it captures a personal moment that we want to keep, but It is hardly commercial nor polished. We may still want to make it better because we will be printing it for display.
Even if we will not be printing, we may want a stand-out social media post that doesn’t just look like a straight from camera photo, but equally, doesn’t look like you just slapped on an ‘Instagram color boost and light leak’.
This article will tell you everything you need to know to perform this workflow quickly. More importantly I will show how Lightroom works better if you assume its tools operate on light rather than pixels. This way of thinking is actually more comfortable for the photographer who doesn’t wish to think like a Photoshop retoucher.
The only thing that is common to both the as-shot and final photographs is the placement. The tree-hugger is completely off center and in an odd place, but the image seems to balance. Composition is probably not a part of the Lightroom workflow, but I want to integrate everything that was on my mind throughout the process of tree-hugger (and composition is often overlooked in other Lightroom tutorials), so we will get it out of the way first.
Composition: Visual Weight and Golden Ratio
There are two composition rules I am following: ‘visual weight’ and the ‘golden ratio’ grid.
A golden ratio grid is a much better option than rule of thirds: less regimented, less obvious and less overused
Most cameras allow you to superimpose a rule-of-thirds (ROT) grid on the live view. A few cameras (such as the Olympus Stylus 1, which I used for tree-hugger) give you the option of a golden ratio (GR) grid.
A golden ratio grid is a much better option than rule of thirds: less regimented, less obvious and less overused. It is more difficult to see the GR grid placements when you take the lines away, so the placements look right to the eye without there being an obvious geometry.
Here’s a perfect example this. Each of these dandelion shots look balanced, but there is no obvious grid.
You can display the golden ratio grid (and the related Golden Spiral) when using the Crop Overlay tool in Lightroom. Just Select the Crop tool and press ‘O’ to move between the different overlays.
Using a grid is a good backup system, but a poor replacement for something that takes into account the relationship between your elements. A better system is visual weighting. In this method, you
- Decide whether the composition is horizontal (reads in a line that goes left-right), vertical (top-bottom) or even an oblique (line between diagonal opposite corners).
- Then, assume that line is the arms of a weighing scale, and the primary visual elements are weights (with their ‘visual impact’ being taken to be their weight). You get balance simply by moving the camera such that the primary visual elements balance with the imaginary weighing scale’s arm lying in your chosen direction.
- If your composition has more than two subjects, pick out the two major subjects or group them. For example, if you are taking a picture of three children, assume them as one subject by grouping them, and look for a background element to balance them out.
In effect, you see hard to define ‘visual balance’ as explicit and easy to define physical balance.
My two visual elements are the two pairs of two branches to the left and right of the image. Not only do they balance, but before shooting, I realized that moving my camera slightly would cause the two sets of trunks to look like a mirror image of each other, and that’s what I did for the final composition.
One thing to be aware of is the visual weight of things you don’t notice. A background lamp-post in your scene may go un-noticed because you are concentrating on the main subject, but in terms of visual weight it may be throwing your scene completely! It is often a good idea to think for a second not in terms of people and things you want to be in your photo, but to get into the habit of blocking the scene out in terms of visual mass. This allows you to immediately identify seemingly extraneous and mundane background items that will otherwise break the composition.
NB – there is a relationship between visual weight and the Golden ratio grid when you only have one subject: one column of content visually balances two columns of white-space. You can see this in the first and third dandelion images.
Post-editing lighting with the Graduated and Overlay Filters.
I have changed the lighting in the scene using the Lightroom Graduated filter tool and Radial filter tool.
The Graduated filter is typically used in the same way as a physical graduated filter: to color the sky and to reduce its exposure, but it can be used for so much more. By seeing the graduated filter as a light source rather than a filter, you can use it to set light and shadow throughout your photo, and not just to control the sky.
Here’s how to reinterpret the graduated filter controls to assume a light source:
You see the graduated filter as either a large natural light source, or an infinite soft-box (a light that comes in from all directions and therefore casts no shadow, or what 3D modelers call a sky light).
- If you want to treat the graduated filter as a natural light source, then the Exposure slider is light intensity and the Color Temperature slider is the light color.
- If you want to treat the graduated filter as an artificial light such as a soft-box, then Exposure is gain intensity and Color is the soft-box gel.
Treating the graduated filter tool as a natural light source will give you more photo-realistic edits, but assuming is as an infinite soft-box allows emulation of more complex studio lighting effects. In either case, when you draw out a graduated filter, you can think of the filter size as the light throw: the light fades in relation to the size of the filter.
If we carry the analogy further to explain what our light source becomes when we use the graduated filter in its normal mode (i.e. like a physical graduated filter), it becomes clear that a filter is really just a negative light source. A filter takes away light whereas alight source adds it. The control that decides whether you have created a light source or filter is simply exposure: less than 0 gives you a filter or ‘negative light source’, and greater than 0 gives you a normal light source.
Consider the following photo
The steps and wall look a lot colder than the background, and this makes the photo: the steps look creepy. It’s done with two graduated filter tools, or two infinite natural light sources.
The first goes from bottom to top and has a negative blue light (exposure -1/2 stop, cool blue temperature). I have also increased shadow, clarity and sharpness for this tool, giving me a hard, contrasty and directional cold light. In short, this light is a shadow creator: it creates harsh blue shade. The second light is a negative yellowish light (-1 stop, warmish temperature). This darkens the sky and gives the background some warmth – a bit like an autumnal sunset. What these two graduated filters are doing is giving two opposing light sources: a cold and harsh blue light on the steps, and a warm yellow light on the background. If you are ever taking photos of people against an old building, the look we see here is fairly common: make the building cold and add more warmth as you move towards the people.
As you can see, like a 3D skylight, our two lights cast no shadows: they just change the ambient.
Back to the tree-hugger photo, you should now be able to see that there are in fact four infinite light sources being emulated, with the colors and directions as shown. All except the top filter are negative lights: they are creating shadows. The top yellow light is creating illuminating yellow ambient light.
A graduated filter gives us an infinite, shadow-less light (or infinite soft-box), but what if we want a directional light? We use the Radial filter tool.
The radial filter tool is most often used to create vignettes, but it can also be used to create circular or ‘beam’ areas of bright light. To use a radial filter in this mode, simply check the Invert mask checkbox, and select a positive exposure.. Consider the image below.
Again, the thing that makes this shot is the lighting. You should by now already recognize the negative light source that is making the bottom-right leaves darker.
What makes this shot though is the lighting in the center – I am using a radial filter as a spotlight aimed at the central grass area.
In tree-hugger, I take this concept to its extreme – rather than using the circle of the radial tool as the circle of a spotlight, I am using it as the light beam itself…
One thing that makes the tree-hugger shot is the subtle but almost volumetric golden yellow rays that seem to be coming in from the top and filling the background with yellow warmth. This beam of light does not exist in the as-shot: that lighting is totally synthetic. Here’s how I did it:
There is a radial filter running top to bottom as shown. This has increased exposure and reduced clarity settings, creating a white beam that causes its background to lighten and blur slightly. This creates our synthetic beam. It is enhanced further by the existing darkening gradients (negative lights) coming in from the left, right and bottom edges, which create the shadow areas that the light beam would be expected to create. If you look at the very top of the tree-hugger photo, you can almost see this light beam as it falls through the trees.
Using the radial tool as a ‘volumetric sunbeam’ in this way can make the most boring image look like a great naturally lit scene…
Simple! As you can see, the radial tool is much more useful than just a vignette tool: it’s a virtual spotlight tool!
There is a second cool way you can use the radial tool, and this has nothing to do with lighting but is something Photoshop users often wish Lightroom had: a reverse mask.
Reverse masking with the Radial Filter tool
The Adjustment Brush tool in Lightroom is cool. It allows you to select specific areas and apply changes to only that area. We call this a selection tool: you select what you want to apply the change on, then make the change knowing that nothing else is affected by it.
It is sometimes easier to select what you don’t want to change. For example, in tree-hugger, I wanted to increase clarity on everything except the face (because increasing clarity on skin tends to age it). In Photoshop that is easy: select the face and then reverse the mask with Select > Inverse. In Lightroom, you have to take the long way around and use the Adjustment brush to explicitly select the ‘everything else’.
To use the radial tool as a reverse mask, simply place the radial filter around the thing you don’t want to edit and make sure the Invert Filter checkbox is unchecked. In tree-hugger I am using the Radial tool as two reverse masks, as shown below.
Okay, that’s the fun ‘how to hack Lightroom tools to do something completely different to their normal workflow’ stuff over with. There is one more tool I used in tree-hugger and it is perhaps the most important and least understood tool in all of Lightroom…
RGB Curves In Lightroom
Keep your changes small when using curves, because the effects of even a small change are significant.
The final piece of the puzzle is RGB curves and the Contrast ‘S’ curve. I’ve saved this till last because I know many readers will turn off at the very mention of Curves, but bear with it. If the tools we looked at in the last section provide our virtual lights, then Curves provide the other important kind of lighting: they define our global ambient light.
Let’s see why:
Have a look at this photograph. The colors and tone look very ‘current’, and are very similar to those used in a lot of commercial photography today (especially fashion). The darks look colder, the whites are warmer than they should be, and the skin-tones have a healthy warmth to them.
You will probably already know about the Tone Curve in Lightroom, because it sits so prominently in the user interface:
The gradient of a tone curve represents the contrast. By lowering the gradient at the very bottom and top, we create a large area in the middle with a steeper gradient (and therefore greater contrast), and the resulting curve gives us the shallow ‘S’ shape. Out-of camera photographs generally lack contrast, and adding a slight ‘S’ curve is one of the first things you typically do.
The more important RGB curves are hiding in the Tone panel – click the icon bottom right (circled red) to see them.
You will see the RGB curve to start off with. This is the same as the tone curve, except that clicking on the curve will create a point on the curve, and dragging that point allows you to alter the shape of the curve. Its essentially a more advanced version of the tone curve because you can add many custom points to give more control.
The magic starts when you click the Channel control below the RGB curve and select Red, Green or Blue. This splits the tone curve into the three channels, allowing you to work on each color in isolation: you are no longer just controlling tone (black/white) balance, but color balance too.
At this point, most discussions go into a long description of what the tone curve and RGB curves actually are, but I don’t think that’s important. All you really need to know is what they do. Here’s the most succinct image of what they do I could come up.
If you are new to Curves, here’s a big tip. Print out the above image, laminate it (or put clear sticky tape on the printout front and back, then cut it out) and stick it on your keyboard or monitor. Not only does it look pretty, but this image is the Rosetta Stone to understanding what RGB curves do: you don’t need to know anything else about curves but that image.
The image consists of four color gradients. These correspond to the RGB, R, G and B curves. If you look closely, you will see that the diagonal from bottom left to top right are tones only (pure black to pure white). That diagonal line corresponds to the initial curves (a diagonal line, bottom left to top right).
The top left corner of the RGB, R, G and B curves are White, Red, Green and Blue. The bottom right corners are the complementary colors: black, cyan, magenta and yellow. If you know about print process, you can remember the complementary pairs because the complements to RGB and white are the print based channels, CMY and K (‘Key’ or Black).
Good so far. Here’s the cool bit. If you add a point to the curve, then drag it anywhere, the color shift you have created is the color under the dragged point on our gradient. So if we dragged the midpoint of the R curve up, we are changing a mid grey to a mid-red. Dragging down would create a mid-cyan shift.
It is important to realize the difference here between RGB curves and Lightroom’s HSL/Color panels. If you change the reds in HSL/Color panels, you are selectively changing the saturation, hue or brightness of reds only. If you change the red channel in RGB curves, you are introducing a red color cast that affects all colors. How do you know how our red shift will affect the greens and blues? Simple: try it! Right click on the ‘What the RGB curves do’ image, save it and import it into Lightroom then apply the curves change. You will see something like this:
Practice with the bitmap for a while later. Curves feel odd to start with (much less controllable than the Hue/HSL panels), but after a while the curves and my 4 gradients will start to link up in your mind, and you will know what your changes will do in real photos.
And that’s it. That’s all RGB curves do! They produce color casts. You have some control over them because you can decide which tonal range (lows, mids and highs) they affect, and by careful editing of the RGB curves, you can actually be very specific over where they occur. In essence, curves edit the ambient lighting in your scene, and you can control the color and tone of the low, mid and high key lighting separately. You typically use RGB curves to change the lighting to affect
- Global ambience (e.g. make an image colder) or
- Emphasize a particular color range that is already in the photo. Using the Hue/HSL panels to selectively change color can result in a very color obvious edit. Adding a tonally controlled color cast via curves is harder, but the result is always more subtle.
- Color cast removal. This is not a usual use case in Lightroom, as you have the White Balance selection tool.
- Separate the main subject from the background. In general, your main subject has one tonal range (typically mid or highs, but not always) and the background has an opposite tonal range (e.g. if the main subject is the lightest thing, then the background will be the darkest). When this tonal separation occurs, you can enhance it by adding a color separation via curves. This is most often seen in portraits, where the skin-tones are given warmth, and the background is made cooler to make the eye go naturally to the warm skin-tones. This is used all the time in fashion photography, and also in the Hollywood ‘blockbuster’ look for film. Doing the same thing in your photography is the single quickest way to take away the snapshot look and add a more mainstream aesthetic.
Finally, curves are useful to change the time of day through lighting. You can make a cold winter sunrise look like a warm summer sunset because almost all the differences between the two (except for the sun size and placement, which you can alter via lens focal length and post editing) are caused by ambient light color.
NB – Adobe Premiere has RGB curves, so much of this discussion also applies to video (in fact more so than stills – curves rule in video because you almost always need to change ambient light between shots to maintain continuity and keep to the impression of a common time of day even though the shoot occurred over the best part of the day and in variable lighting).
There are a few gotchas and things-to-know:
- Curves make drastic changes. Although you can move a point a significant distance for certain special effects, for most photo-editing you will drag points by a very small distance. By small I mean 2-3 pixels on the graphs… the RGB curves are for subtle changes!
- The gradient of a curve is the contrast.
- Don’t move the black and white point unless there is a specific reason to do so. Moving the low left point or the high right point drastically changes your black and white. If you move either, you lose true black and true white. You should either move the endpoints very slightly or not at all
- Making big changes on the curves will always lose color range. When you move a point you are actually pushing the colors in a direction. Think of a grid of coins completely covering a square table: if you push the central penny up, some coins are pushed off the table. Pushing a curve too far will always lose you some color range, and this can result in banding. Another reason to follow rule 1!
- There are two ways to think of the RGB curves: as RGB or as their complements (CMY). Moving the red and green curves up to create a blue-green cast is the same as thinking of cyan being increased, which equates to moving the red curve down. This means that if you are moving two RGB color curves in the same direction, there is always a simpler way to do the same thing involving one CMY color. This is an advanced concept, and many professionals don’t use it (its so damn complex once you have multiple point changes across curves!) but it i always worth bearing in mind.
- When you change the RGB curve, you also implicitly change the R, G and B curves, but Lightroom doesn’t show this. The RGB curve (tone curve) is really just an amalgamation of the separate R, G and B curves: the tone curve itself doesn’t really exist. So you would think that changing the tone curve will change the R G and B curves, but it doesn’t (they are added together behind the scenes to produce the fnal R, G and B curves). Instead, think of the tone curve as changing brightness and the R/G/B curves as changing color. Again, this will only really be true if you make small changes to the RGB curves and save any big changes for the tone curve: yet another reason to follow rule 1!
Although there are a lot of rules here, there’s really only one you need to remember: Rule 1. Keep your changes small when using curves, because the effects of even a small change are significant.
Here’s exaggerated as-used tone and RGB curves for tree-hugger:
The RGB curve is a typical S curve and gives us a better contrast to the out-of-camera photo.
Next, I worked on the Blue curve. First, I raised the Blue shadows (1) and increased yellow via points (2) and (3). This makes the dark tones bluish, and the mid and high tones more yellow. The effect of this changes is the increase the color contrast between the shadows and mid-highs, by making them colder and warmer respectively. Doing this is a common use-case for the blue curve.
Note that moving points (1) and (3) on the blue channel change the black and white points of the entire photo. Professional retouchers avoid this by adding more points near the extreme ends of the curves to make sure the black and white point never changes as per (1a)
I then worked on the Green channel. I use this channel to modify the Blue channel. Pure blue is far too unnatural a color, so I tone it down by adding a little bit of green to the shadows, giving us a blue-green for our shadow color. Points (1), (2) and (3) don’t actually do anything: they are just there to stop the mid and high tones from moving.
Finally, I worked on the red channel. Here, I am lifting the mid-point via (1) and moving the black point right a little. This adds warmth to the mids and loses it from the shadows.
Point (1) on the red curve and point (2) on the red curve work together: they are warming an area of the tonal range. If you are just warning the mid-tones as I am it will be near the mid-point of the curve as shown. Often, you are warming skintones, and the exact place the points go will be defined by the lightness of the skin tone itself: the lighter the skin, the higher up the point is on both curves.
Point (1) on the blue curve, point (4) on the green curve and point (2) on the red curve are working together to give us our final shadow color. This is normal: your shadow color is often the most important and most complex because it does two crucial things: it sets the color of the darks (and this goes a long way in setting the ambient color of the scene) and it acts as a counterpoint to the warmer mid-tones. Adding too much warmth to skintones can often look unrealistic, so instead you create most of the warmth by taking it away from the shadows. In fact, some photography has almost white skintones with no intrinsic warmth. The skintones still look warm overall, because the shadows have been made still colder with the blue curve.
Finally, its worth seeing the actual curves for tree-hugger. As you can see, the actual changes are much smaller than the graphs above imply! I’ve also used the Tone curve rather than the RGB curve to enter my contrast ‘S’ curve (it is easier to use).
The changes above are not specific to tree-hugger but are very close to a typical curves workflow for most photos. You typically use curves to create opposing tints between the background and main subject, and this works to separate the two visually, as well as giving the main subject a clear prominence.
In photos including skin-tones, the main subject will be the skin-tones themselves, so you will typically use exactly the same curve edits. In fact, almost all of the photos in this post use small variations of the same curve changes (if there is such thing as a text-book ‘how to make your images look more commercial’ curves change, the ones I have used are probably it!).
Before raising your camera, think about the composition you will end up creating, and if nothing else, think about the visual weight of the elements in your scene.
By seeing some Lightroom tools as ‘light editing systems’ rather than pixel based tools, they can make more sense, and allow quick and easy editing despite complex visual changes.
Lightroom curves are a crucial yet well hidden color editing tool within Lightroom. By thinking in terms of the complementary pairs of colors Red-Cyan, Green-Magenta and Blue-Yellow instead of just Red, Green and Blue, RGB curves become infinitely more usable.