The Digital Darkroom

The Camera is a Tool, But The Photographer Doesn't Have to Be

As technology has advanced, and digital cameras have become more affordable, the barrier to entry that digital photography once enjoyed has been smashed apart. For a few hundred pounds you can now pick up a digital camera that even as recently as five years ago would have been many thousands.

What hasn't advanced as quickly, however, is the ability of the camera to take Pulitzer-prize winning images without any help from the person behind the lens; and - truth be told, and contrary to the myth perpetuated by the cameras' manufacturers - that won't ever happen. Just as some knowledge is required to ride a bike or play the piano, so you need to learn a little about photography to take great pictures.

If you do decide to study photography, or simply ask any enthusiastic amateur, one thing they will always tell you is to turn off the camera's automatic settings, and do everything manually. For real control over what's in front of the lens, the bit behind the lens (that's you!) has to be in total control: exposure, lighting, focus, shutter-speed, and all the other widgets that high end cameras have buttons for.

Sadly, you can't do this on most of the consumer-level cameras. They are what you'd call point-and-click.

So where does that leave you?

Don't Worry, We'll Fix It In Post

Sure, there are a few things you can do before you press the shutter to improve your photos, but you might be surprised to find out what you can do in the digital darkroom once the lens cap has gone back on.

To show you what I mean I recently asked for some of evolt.org's members to send me some of their... erm... less successful photos. That way I could be sure I was dealing with some real-world issues, and that I wasn't taking staged photos simply to prove a point. I picked a few of the images sent to illustrate some different techniques for this series of tutorials.

But first, lets have a look at one thing you can do that will really help you create acceptable raw material for our experiments...

Please Don't Flash

The first thing you should do - if at all possible - is turn off the camera's in-built flash. We've all seen the startled, demonic red eyes in family portraits; that's not because Aunty Nelly is a hellspawn, but a side-effect of the camera flash hitting the blood vessels in the eyes.

More modern cameras do this unnerving, double-flash thing to reduce the effect, but a far better solution is to turn the camera flash off (or cover it up with a well-placed thumb) and get the required light from elsewhere. If budget will stretch to it, a removable flash can be used. If you can get one with an adjustable head, use it, and 'bounce' the flash off a white wall or ceiling. If not, consider a table lamp pointed at a piece of white card held out of shot. If you are careful - and don't hold it close to the bulb - you can even use tracing paper in front of a lamp to soften the light, and avoid those heavy shadows and bleached hilites that are the trademarks of flash photography.

Remember, in nature the only direct source of light is the sun, and that's 93 million miles away; there is simply no way that a flash some six feet away is going to be as flattering to your complexion as our fiery friend in the sky. For a good photo, indirect, or bounced light is always the way to go.

Giving Back Some Snap

There's not a great deal wrong with this photo, submitted by Javier Velasco (and copyright mantruc.com 2002); as Javier himself says, "In general terms it's a nice picture, it's got a decent composition and is well focused, but the cloudy day makes it look quite dull."

So, the question becomes what can we do to it to give it back some 'snap'. Obviously, I wasn't there when Javier took the picture, so I have to assume that the flower is the right color. For now I'll just concentrate on how we make it even more colorful, but - and this is key - without altering the color.

Which nicely brings us round to the most obscure of color spaces, LAB. For those of you unfamiliar with LAB, it's very different to RGB and CMYK color, in that it separates the color elements of the pictures from their contrast. Whereas in RGB and CMYK, altering a single channel will alter contrast and color at the same time, with LAB you can choose to alter them individually; when you've got a picture like this, where the color is fine, this is a huge advantage.

The L in Lab stands for Lightness, which is just a technical term for what most people think of as contrast. (As an aside, the A and B don't stand for anything in particular, and carry the image's color information in a startlingly non-intuitive way; suffice to say, even the smallest alteration of the A and B channels can have massive effects.)

So, stage one is to take our image into LAB mode, using Image->Mode->Lab Color. There shouldn't be any visual change, but if you examine the channels now you'll see what this has done. This image shows what the channels look like before and after.

Now we have to actually make our color correction.

I use Photoshop as my image editing program, so I'm going to use that for the walkthough, but most other programs have similar color controls. The key thing here is the ability to control the numbers in the different channels of the image; to alter the L, A and B, or the R, G and B independently.

I'll start with the Lightness (L) in this image.

Open the curves dialog in Photoshop (Image->Adjustments->Curves or Command-M) and select Lightness from the pop-up at the top (Command-1).

Make sure your dialog matches mine, in that shadow values are on the right, and hilites on the left; if this isn't the case, click the light-to-dark bar at the bottom of the panel to flip them. You'll notice if you flip the bar, the values toggle between a scale of 0-255 and 1-100%; you can still correct the image in the same way, but everything will be reversed, and you'll have to calculate your target values accordingly. If you are a hardcore web-geek you can even use #FFFFFF-style hex, but I'll leave that as an exercise for the reader.

This dialog simply shows how the current Lightness values across the image - mapped along the horizontal axis, left to right - are translated into to our intended, corrected values - along the vertical axis, bottom to top. When you click the okay button, the color values in the image are re-calculated according to how you have adjusted the curves in this dialog.

You'll notice if you run your cursor over the picture while this dialog is active you get an eyedropper tool, along with a readout of the current lightness value and a circle on the line indicating where that is on the correction curve. If you run this dropper over the area of the image you are interested in (the focus or main subject of the image) you can see where the curve should be corrected. In this case, the flower's lightness runs between 50% and 80%. As a general rule of thumb, to get the most detail in an area of the picture, the curve for that area should be the steepest part. So that's where we want to concentrate our efforts.

To correct the curve, then, I click once on it at around the 50% mark, and move that down to around 25%. I create a second point by clicking on the curve again at the 80% point, and move that down to about 88%. This 'steepens' the curve around the 50% to 80% region, as well as increasing the lightness of this area. This immediately gives our image more 'snap', without altering the color.

Now we need to increase the intensity of color.

You'll recall that I said that the A and B channels hold the color information in a very non-obvious fashion; I'll not go into that in this simple tutorial, but I can show you how a simple move in these two channels can increase the 'saturation' in an image fairly painlessly, greatly helping the image's washed-out look.

With the curves dialog still open, change the pop-up to the A channel (Command-2). Now bring the bottom-left point in 10%, and the top right in 10%, so that it reads 90%. By bringing in both ends by exactly the same amount, we don't change the neutrality (or color cast) in the image.

You can do the same with the B channel, bringing in the two points 12%. Of course, the exact numbers are a matter of taste and fairly subjective. Without seeing the actual bloom, I've just done a move that appeals to me; you might like a stronger or weaker move.

Now if you apply the color correction, you should be left with a much brighter, more colorful and 'snappier' picture. All that's left now is to convert back to RGB mode and save it.

Any picture where you want to keep the color the same, but change the relative brightness, will benefit from this technique.

Next time I'll show you how to color-correct images with bad color casts, and how 'doing it by the numbers', and knowing a few simple rules, can help you determine exactly what it is you need to do to an image. Alternatively, if you'd like me to cover any particular problems you have, which you feel might benefit other evolt.org readers, let me know.

In the meantime, if you want to explore these and other techniques in more detail, I highly recommend the book 'Professional Photoshop' by Dan Margulis, which goes into far more detail than I ever could.