Tuesday, 19 January 2016

The Making of an Image

One of my favourite outings from last year would have to be the Greymouth Street Races. It was the shoot that convinced me that I had made the right decision moving to mirrorless. Up until then, I had been nervous about using the Olympus OM-D E-M5 MkII to photograph fast-moving subjects. But rather than being disappointed, I was blown away with the images that I got that day. It’s been a few months since then, and I have lived with the images for a while before deciding how to process them (there were over 2000 images!).

In the Lead. OM-D E-M5 MkII with 40-150mm f4/5.6 lens. 1/4000th sec @ f5.6, ISO 1600
As I said, I was thrilled with the quality of the images I captured – especially with the ‘cheap’ Olympus 40-150mm f4/5.6 consumer zoom. Cheap and plasticky it may be, but boy-oh-boy is it one sharp lens for the price. Even wide open, it’s as sharp, if not even sharper, than the Canon 400mm f5.6 and 70-200mm f2.8 pro zooms. And that’s not just anecdotal. That’s from actually comparing images taken with these lenses from previous years capturing the same event.

Cutting it Close. Olympus E-M5 MkII with 40-150mm f4/5.6 lens. 1/2000th sec @ f5.4, ISO 800
Yet as sharp as it may be, the one slightly limiting factor for the lens, and to a certain extent mirrorless itself, is the larger depth-of-field exhibited due to the smaller sensor size. F5.6 on the EM-5 MkII is equivalent to F11 on a full frame sensor (and probably around f8 on an APS-C sized DLSR). This is great for landscapes when you want everything in focus. Not so great for sports where you want a decent separation from the action and the background.

So, with the editing of the images, I set out to give more background blur, for a greater 3D quality and subject ‘pop’. I also wanted to simplify the backgrounds as much as possible, to really focus attention on the bikes themselves. I had tried to achieve this on the day by positioning myself somewhere on the track with fairly clean backgrounds. But even then, I wanted to make it cleaner still by cloning out a few problem areas.

Catch Me If You Can. E-M5 MkII with 40-150mm f4/5.6 lens. 1/2000th @ f5.6. ISO 640
What follows is my post-processing steps to create the final images included in this post. I will cover what and why I did what I did at each stage of the process, without going in to too great a detail about the actual how-to’s of the individual processes themselves. I’m very happy with the results, although that doesn’t mean I wouldn't do it slightly differently six months or a year from now. That’s the great thing about post-processing an image. You can take it in many different directions – as long as you keep the original!

First step is cropping. I don’t crop all that often, preferring to ‘crop’ in-camera as I shoot. But sometimes the action is fast and frenetic and you don’t have time to crop perfectly. So in this instance, and again, to help simplify the composition and distractions, I cropped afterwards.

Complete Concentration. E-M5 MkII with 40-150mm f4/5.6. 1/2000th @ f5.6. ISO 1000
Second step was actually sharpening. Yes, the images were already sharp, but I wanted to give them even more crispy sharpness. I achieved this with ‘high pass’ sharpening (my favourite way to sharpen), and then added a layer mask to mask out the areas I didn’t want sharp (like the asphalt on the road and the background).
 
Third step, cloning out the unwanted distractions. Cropping got rid of some of them, but many of the plastic barriers had dark holes and branding on them that were very distracting – even out of focus. So I cloned them out using a combination of 'content aware' fill and the clone brush.

Next was some dodging & burning and levels adjustments to get the exposure looking exactly how I wanted it. This usually meant some light dodging on the faces, and some opening up of the shadows with the highlight/shadow setting on +10 for the shadows. Then a simple levels ‘S’ curve finished things off.

Layers Palete in Photoshop for Blur
Having achieved all that, I was then ready to create the background separation. I wanted the final images to look like they were shot with an f2.8 aperture rather than the f5.6 (f11) they were actually taken with. To do this, I cut around the bike and rider using the pen tool (make sure you use a good graphics tablet if you’re going to be doing a lot of pen tool work) and selected this with a 0.5 feather on the selection. I pasted this on a new layer (Ctrl + J), and then went back to the original layer and duplicated this as well (Ctrl + J again). I re-selected the selection (found in the Path palate), this time with no feather, and hit Delete (Backspace) on the keyboard to cut it out of the layer. To this layer I applied a 15% Gaussian Blur, then added a layer mask. Using the gradient tool I masked out the lower third of the image, graduating towards the top where I wanted the most blur. And hey-presto, an f2.8 background!

This is the most time-consuming element to the image, but still only took me about 10 minutes to achieve. And I’m more than happy with the results. I wanted the bikes to really ‘pop’ out from the background, without making it look too unnatural. Hopefully I’ve achieved that?

All the changes made in post process to get the images looking how I wanted.
To finish it all off, I flatten the layers into one final image, and then darkened the top and bottom a little, before adding a 10% black feathered vignette to the image. Subtle, but just something to again help draw the eye to the action in the image.

It sounds like a lot, bit it really isn't when you sit down and get into the flow of processing. Each image probably took about 15 minutes, and it certainly helps if you have a good graphics tablet (Wacom Intuos).

And the finished image after all that processing.
Admittedly, I could have saved myself a few hours of post processing work by just shooting with an f2.8 lens in the first place. Agreed. Except, I don't have an f2.8 150mm lens on me, and am probably never likely to. So you make do with what you have, and just know you're up for some computer work afterwards. I don't think that 10 to 15 minutes is overly excessive in terms of post production anyway. I certainly wouldn't want to spend that time working on every shot I take - and I don't. My average is probably only a minute or two. But for images that I want to take the time to elevate to a higher standard, then 10 minutes is totally worth it (IMHO).

Out in Front. E-M5 MkII with 40-150mm f4/5.6 lens. 1/4000th sec @ f5.6. ISO 2000
There may be some photographers out there who still believe any use of Photoshop in an image is a crutch. Hard to believe maybe, but I'm old enough to remember back to Photoshop version 1 and the hot debates held by 'purist' photographers and those crazy early digital adopters. For a long time, Photoshop was a dirty word in camera club circles and while that may have changed, I know that many photographers still believe that you shouldn't spend longer than 1 minute 'working' on a photo on the computer.

I can sympathise. But sometimes, I also think that you've gotta do what you gotta do to help create your vision. You can't use Photoshop to make a silk purse out of a sow's ear, but you can (and should) use Photoshop to help you achieve the end result you envisaged when you took the photo. And that's exactly what I've done with the images in this post. Hope you like them?