Films
Sin City and
300, and the work of stills photographers such as
Andrzej Dragan and
Jim Fiscus have recently helped popularize a dramatic, dark, muted colour palette style of produced photography. Obviously lighting and post production are the keys to these looks. I'm not going to speculate about how other photographers produce their images other than to state that contrary to many peoples suggestions I do not think
LucisArt filters are used at all. Each photographer develops their own unique techniques and then applies them differently to each image to best suit subject and brief.
My processes are no different in this respect. They're not used too often but on occasion they're the right way to go. I'll take you through one image's production step-by-step...
LightingA lighting arrangement that results in strong highlight and shadow lays down the right foundation for the post production process. Lighting that clearly delineates the subject from the environment is also a target. There are techniques similar to those outlined below that work with flat lighting. This results in a different, watercolour-like look.
In the case of my image below the background was lit with two tall white V boards placed at each end of the screen.
The model was heavily rim lit with two reflector dishes at 2 stops over placed at 45 degrees behind, left & right, and at 35 degrees above eye level.
Just to the right of the camera position and a little behind it was one light fitted with a 80cm circular dish diffused with tracing paper. This was elevated 25 degrees above eye line and was 1/3 under. It's at a distance that makes it a fair bit smaller than usual light source for female portraiture...
Raw ConversionUsing ACR the image was correctly WB'd from a grey card reading. Then it was adjusted from defaults. I'm not quoting exact settings because they're different for every shoot. Aim for an accurately coloured, low contrast file that holds all important tones...
Photoshop core process
Once the subject was extracted from the Chromakey background the edges were tidied up using a mask. The layer was duplicated. A Channel Mix layer was placed above this. The red channel was reduced (usually by 50-100) and the blue increased by the same amount. In this case each was shifted by 90. Sometimes the red channel reduction is replaced by some increase in green as well as blue. This results in a subtler look and slightly a cleaner skin tone. The channel mix was set to monochrome and then the layer set to Luminosity. Often this is then set to less than 100% opacity. In this case it was left at 100%. The Channel Mixer layer was then merged with the duplicated subject layer.
Keeping an eye on the histogram an S-curve with both input points close to centre (127) was then applied. Only a slight shift from input to output at each point is required, just 2-3 is usually enough. For example a typical adjustment would be an input 123, output 121 for the bottom curve and input 133, output 135 for the top. Both ends of the curve were held tight with two additional points to minimise shadow & highlight clipping (eg. 25,25 & 230,230).
Next was a full skin retouch. Use which ever technique(s) you prefer. I'll detail my usual method on another post. For female subjects especially, a full paint over is essential as the blue channel boost used to set Luminosity reveals a lot of latent noise.
After the retouch I decided drop in a new background made with the Vignette setting in the Lens Correction filter. I also cropped the image a little to tighten up framing. Here's how things looked at this stage...
Photoshop finishing touches
It was important to unify tones in this image so several transparent layers were made over the image layer. Each one was set to colour mode, and each then 'painted' with a brush in colour mode to change the colour of the jewellery, makeup, areas of skin and hair. Of course you could do this on one layer but if you put each colour on its own layer then adjusting it is simpler. I usually use the Hue/Saturation menu for any tweaks. Once all the colour changes were finalised the colour layers were then merged to the subject layer.
Next was a touch of detail recovery. Notice on the image above how the hair and knicker line/trousers are blocked-up. The original raw conversion (the base layer) was duplicated and placed as a layer over the image and set to Luminosity mode. Using a mask painting these dark areas have their detail restored from this layer.
The visible layers were merged. The subject layer was then duplicted, desaturated, and then this B&W layer was duplicated. One of these B&W layers was set to Multiply mode, the other to Screen. Each was 'Hide All' masked and then a soft and careful paint was used to subjectively enhance highlight (Screen layer) and shadow (Multiply layer) areas. A light hand with a pressure sensitive stylus makes this job much simpler than a low opacity paint with a mouse. Once this was done and the masks were applied, the B&W layers merged, this merged layer set to Luminosity mode, and its opacity set to 33%. The opacity is another image-by-image variable, typically I use 20-40%. This was then merged with the subject layer. It's pretty much job done at this point. All I usually do at this stage is a final small retouch for skin and fine detail (on a transparent layer) as was the case with this image. I also shifted the overall Hue towards red by 9 and reduced the overall Saturation by 25 on this one.
Here's the finished image...

Labels: Post Production, Technique