Image Autopsy - Tree Reflection

I am about to start a new series. In it I am going to pick one of my images that means a lot to me and explain the circumstances of shooting it, what I was trying to achieve and what I did to the image(s) to get there. I'm going to talk about the elements of the image which are important to me. It will be up to you to decide if they work but I'm not looking for feedback on whether it's a nice image. Rather, what I hope is that you may get something from my own struggles, sucessful or not. I definitely want feedback on whether the whole autopsy idea is worth while.

Some time ago I did a series on dissecting classic images but difficulties in getting permission to reproduce the images meant only that I could provide a link and any comments I made about the images were largely speculation about the intents of the photographer. Time has shown that such analysis may explain why an image is good but may be completely off the mark in explaining what the photographer was doing and why. This time I have an intimate understanding of the photographer and the only limitation is my memory, faulty as it can definitely be.

Circumstances - I was looking for interesting patterns in the thawing ice of the Elbow River near downtown Calgary (pop. 1 million). My back was to the bus barn and across the river was Inglewood, an old and now artsy part of town with quaint shops, live theatre, galleries and restaurants. It was January during a warm spell and about 10 degrees C. (50 F.), brilliantly sunny with a strong blue sky typical of Alberta. There were a couple of clouds hanging around.

The Search - I didn't find any really strong patterns in the thawing ice. A number of shots were taken, none have seen the light of day. As I worked my way along a bicycle path next to the river, I came across a very slight inlet in the river edge. Itmade a sweeping curve which could be accentuated by use of a wide angle lens. At this time of year, multiple freeze thaw cycles had given the ice at the edge of the river a large crystaline structure, easily visible. I looked both upstream and down for intersting patterns in the ice leading away from this foreground curve but nothing struck me. Then I realized that water running along over the ice in mid stream was reflecting the few clouds. Further, by positioning myself I could get the water to reflect the upper part of a tree on the opposite bank.

The Setup

I set up my tripod and moved it around to fine tune the image. I was already into stitching and it was a simple matter to use my zoom to frame the image vertically, then simply create however many images I needed to set the horizontal boundaries of the image. As it happened, this meant only two images and a square format, largely determined by the limits of the curving inlet and overhanging bushes to left and right of me. As I like square pictures, this was no problem.

As I was setting up the shots, the sun went behind one of the few clouds. This resulted in better control of contrast, avoided harsh shadows and unbeknownst to me at the time, gave the image a blue cast since the scene was now lit entirely by a deep blue sky.

The Equipment

At that time I was shooting with a Canon 10D, using my new 17-40 mm. lens, set at 27 mm. (don't forget the 1.6 X factor for the small sensor, more or less cancelled out by stitching).

A similar image before any manipulation looks like this:



You will note the really subtle colours, also some telephone wires needing cloned out. Still, you can see the blue tones of the blue skylight.

At that time I was using a wooden Berlebach tripod with built in ball joint centre post combination. This tripod served me very well for several years for both 4X5 and digital. I was already using a Really Right Stuff L Bracket and with a purchased RRS lever clamp and a piece of oak, I fashioned myself a nodal point arm which would let the camera body sit behind the rotation point of the ball head. The ball swivel/centre column fitting on the tripod meant that I could level the centre post and rotate a real ball head (Arca Swiss B1) perfectly horizontally, even if the camera aimed up or down. In this case the camera was aimed down - the horizon is slightly above the top of the image.

It's important when stitching with the camera angled down this much to note that your carefully found nodal points are no longer accurate - the rotation point moves back to the camera body as you tilt the whole mechanism down. The resulting images have angled edges and you need to be absolutely sure you have enough image at either end to end up with a proper rectangular image (unless you are into trapezoidal prints - a limited market).

I subsequently lost that home made nodal point slider and have since replaced it with a RRS one.

To take the picture, I used mirror lock and cable release. Exposure was 1/15 at f11, ei. 100.

Post Processing - Camera Raw, whatever version was extant at the time. A number of images I have reprocessed with more recent versions. The first big step was going to 16 bit and subsequently better sharpening algorithms. I'm a great fan of smart sharpen but I have never reprocessed this image so suspect the original camera raw output was 8 bit.

Once in Photoshop, I applied a series of curves layers to increase contrast, get some dark tones and most importantly reveal the ice crystal pattern. it is in the nature of Photoshop to increase colour saturation at the same time as increasing contrast with an S shaped curve. This gave the image it's dramatic blue colour. While the original blue was real, it wasn't seen by the eye which rapidly cancels strong colour hues (try wearing orange coloured sunglasses for a while). In the end I probably had more than 20 such curves, gradually increasing the contrast, in each case locally only through the use of black masks and painting white into the mask with assorted levels of opacity.

The near ice received the greatest changes and in fact I had to tone down the colour saturation several times as I worked on the image, because of the increased contrast. At that time I didn't know about the luminosity setting in the layers palette which means that saturation doesn't increase when you increase contrast.

The image isn't in fact super sharp and it took some work with creative sharpening with Photokit Sharpener to ge the ice crystals sharp. In hind site, f11 wasn't enough depth of field for such a near far composition. Nowadays I'd shoot multiple images and blend with Helicon Focus, having determined that f16 is the smallest f stop on my current 1Ds2 which actually increeases sharpness in the out of focus areas. A smaller stop just makes the in focus bits fuzzier while not sharpening the out of focus bits any more (diffraction).

This is, however; one of those images which doesn't need to be super sharp and I actually like a 3 foot square image I have at my office.

The telephone wires were cloned out as was a very sharp straight line of bubbles in the water which looked more like a negative defect than real. A single dark coloured leaf had started to melt its way into the ice and was cloned out.

The End Image - this is one of my favourite colour images, and that the colour is exagereated doesn't bother me one whit. I like the centering of the tree while the curve of the near ice on the bottom isn't centred (enough is enough). I love the clouds reflected in the water and the brush stroke like effect of the melting ice on the far side of the river. Note how I darkened the two bottom corners to keep the eye within the image - too much? - perhaps, but it doesn't bother me.

I'm not one for rules but certainly this image is very much divided into thirds vertically.

There is a consistency of colour throughout the image - given that the reflected sky is part of the image and the rest is lit by sameself sky, this isn't too surprising, though I did have to do some subtle colour balance work for the real image to look this way.

It's really important to have a profiled monitor for this image as subtle changes can make the blue look purple or turquoise, neither of which works.

If you go back to the original image file shown near the top of the article, you will get a sense of the amount of work done on the image. This was done over half a dozen different sessions complete with about three total restarts from scratch. Total time working on the image was probably about 20 hours. Who says digital is easy?

To many people the image is very disturbing because of it's upside down nature. At the time I was completely unaware of this aspect of the image and it has never looked upside down to me. It gets a lot of comments and some sales, but not as many as my more conventional images. People tend to 'get it' or not. Those who have to work to see it right way up typically don't buy it, but a small number of people absolutely fall in love with it.

About 5% of the public see it as a reflection immediately and about 30% never get it, even after having it fully explained to them.

Sales - I'm a little uncomfortable tying sales hype into what is supposed to be an information and discussion blog, but what the hey - I'm going to sell prints of each image discussed in this series (if it's popular). Prints on 8.5X 11 (approx. 6.5 inches square) will be $40 Can., regular postage included, anywhere in the world. Prints on 13X19 (approx. 10.5 inches square) will be $69 Canadian. Both will be shipped flat and well protected and guaranteed to arrive safely. Prints will be on whichever archival paper and made on whichever archival printer/ink which seems best suited.

The series are open, prints are signed.

Payment can be made by Paypal using my email address george dot barr at shaw dot ca and specifying which print it is.

You can also send me your visa or mastercard number and expiry date. I would suggest spelling it out - eg. two four nine... just in case someone is trolling for entries with the word visa and a 16 digit number in the same email. Alernatively you can fax your request to (403) 256-4852 and address it to George Barr Fine Art Photography, so my secretary doesn't throw it out as a junk fax.