Friday, December 28, 2007
Why Edit # 1
Above is the end result of image editing, below are the steps I took to get here.
Having read all the helpful feedback about a possible future book, it seems that what is needed is not a how to book on editing - there are lots of books out there which explain the workings of Photoshop masked layers and other techniques for editing one's images. What does seem to be deficient though is an explanation of the process of looking at an image and deciding what changes are necessary to make it a better image. To that end I thought I'd go through the thought processes in an image I edited this morning.
Above is the image as it was brought into Camera Raw, with my standard defaults. Note the blown highlights in the water. Clearly this is of concern as the water is too important to accept pure white in significant areas. You could make the argument that I had overexposed the shot, however in-camera the highlights were just on the edge and from previous experience I felt they could be rescued. The reason I didn't reduce the exposure was to keep as much detail in the shadowed rocks on the left and also at the bottom right of the image.
This shows the result of a modest application of recovery to reign in those highlights without losing the shadows. The shadows warning is on but there is little to be seen in the way blocked shadows (they are dark blue).
Above shows the settings I used for sharpening. I confess it's a bit of an experiment in the last couple of months. Previously all sharpening was done in Photoshop but with v. 4 of Camera Raw, sharpening is more sophisticated than it had been and I have been taking advantage. Is it better than using Filter/Sharpen/Smart Sharpen in Photoshop - not sure, but I have it as my default in Camera Raw, so all images get this much sharpening, then I can add a little more in Photoshop if I need.
Above is the image as it arrives in Photoshop. It's time to start thinking about what I'd want to change in the image. Ideally I'd like to get a bit more detail in the white water and the shadows, though not blocked, could definitely use a bit of opening up. At the same time, the rock surface on the right of the image has some interesting detail but it doesn't really stand out. I was going to say it looks 'flat'. It occurs to me that I could improve things considerably through the use of Akvis Enhancer which should be able to help in each of these areas.
And above is the result of using Akvis Enhancer, default settings.
It's better, but seems to me the shadowed rock on the left could stand being opened up a bit and the rock on the right still seems a bit boring. I really like that small half filled puddle on the lower rock, but it needs to stand up as well, grounds for a bit of a contrast increase there.
Above shows the effects of opening up the shadows. Note that I have also lightend the mist in the canyon.
And these show the black mask into which I painted to apply the curve shown.
And this shows the effect of increasing contrast in the rock on the right - I'm much happier with it.
I haven't begun to deal with the cropping issue. Micheal Reichmann mentions in his Camera To Print Video series that he crops very first thing - but I have to tell you that there is nothing worse than deciding after two hours of working on an image that having a little more room on the left would be lovely - only it's long gone. Even if you did save the work before the crop, you are going to lose everything done since.
Photoshop does allow for cropping without losing the cropped area - it's just stored away in case of rain. If you can do your crop in a single swoop, this works perfectly. Unfortunately I often find myself refining the crop a step at a time so this won't work for me.
Bottom line is I prefer to to the basic work on the image before cropping. Personal choice, it works for me. Choose a system that works for you.
Anyway, getting back to decisions around cropping. There are two fundamental problems with the image, the small spruce branch in the bottom right and the jambed log in the bottom left. Either could be cloned out if I really need to but it gave me a chance to think about composition. There's absolutely nothing wrong with experimenting with composition via the cropping tool, until you come up with the best possible result.
On the one hand, I don't relish losing that interesting rock strata on the left, but I discover that if I crop out the log, I actually like the balance of the image better.
This shows the cropping tool in place. Note that I have elected to trim a little off the right too. It wasn't that the material wasn't good - in fact I regret trimming the curve of water at the top right, it just seems to balance better. I haven't removed all of the branch in the bottom right, but the rest can easily be removed with cloning.
I could, of course; have trimmed from the bottom of the image and removed both branch and log, but I really like that rock eddy on the bottom near right with it's rounded large pebbles and couldn't afford to lose any of it - in fact in an ideal world, I'd have kept a bit more of it so the second pebble was whole. Mind you, given that the shot was taken blind, with the camera 5 feet out on the end of my folded and hand held tripod, held almost horizontally out over the fall to eliminate foreground, I'm happy the framing worked as well as it did. I did manage to rest the tripod about a foot from the end on the safety railing to provide some steadying effect.
And this is the result of the crop and a little cloning to tidy the two bottom corners. I note that I forgot to clone out that stick lying on the rock. It could even be a fold in the rock, but if it looks like a stick, it has to go.
Other than that, I could call it quits - I like the image I have, there is good detail in shadows and highlights, overall a pleasant image.
There is, however; a lot more that can be done to polish the image. For example, though the highlights are there, it might be possible to give them a bit more punch by increasing contrast from pure white to light gray - easy to accomplish with another black masked curves layer.
The first of the three images shows the effect of the curve, applied across all the image. Note that the effect is way too strong in most places but remember that since I am painting into the mask I can control the amount of the effect and I want a few areas to receive the maximum effect shown here so this much will be just fine, once tamed.
You can see the curve used. Note the second point almost at white. That was essential to keep the curve from muddying the white water. If you doubt it, try clicking in the before image to show it full size, then bring it into Photoshop yourself and play with curves to help the water.
The third image shows the effect applied where and to what degree I want. There is a lot more detail to the water, but I see that in some places the water has become too dark in the highlights. As this is very easy to fix later, I'm not concerned.
Above is the result of a little fine tuning. I decided to increase the contrast in the puddle and on the flat rock, though I used separate curves so I could get things just right. It turned out that in both cases, the colour saturation increased with the increase in luminosity contrast. In the case of the puddle, it was really objectionable so I set the layer blend to luminosity. This worked perfectly for the puddle but left the rock looking dull again. Setting it at normal has left the colours a bit unrealistic but I can tone that down later with a hue/saturation layer.
In increasing the contrast in the puddle, I liked the effect in the part that is reflecting the rock, but the bright part looked unrealistic and had to be toned back.
Time to stop and reflect, what further changes would help the image?
OK, so what have I done this time. You will remember I thought the colour of the flat rock was a bit too much but what really bothers me is the yellow green tinge to the rocks on the left. What if I applied a subtle change in colour here to better match the hue if not the saturation of the flat rock? There are a number of ways to do this. I could use the separate colour curves in the curves adjustment layer but that's more control and hassle than I want to deal with for a very modest change. I could use colour balance but it tends to change everything. Oddly, I have found that using the hue setting in hue/saturation adjustment layer is often quite helpful. A small shift to the left (-6) nicely takes the yellow green out of the rock without fundamentally changing it. Applied with a black mask painted into as usual it better blends with the flat rock. I then did the opposite with the flat rock, moving the hue +4 to the right and now I see that I don't need to desaturate the colour and there is an overall consistency to the image without going too far. I see that the small triangle rock in the bottom left is too bright so I do desaturate the colour in it - but the hue is correct so no other changes.
I could go on making more and more changes, but the image is very close to what I want at this point and there's a real risk I'll adjust it too far - just one more change needed - the water needs cleaned. I could use another curves layer but experience tells me that subtly whitening water isn't easy with curves. Better is using the dodging tool set to highlights and the amount set to around 5% so I don't over do it. Even better is if I have some way to tell how close to pure white I have come since staring at the monitor, the difference between a brightness of say, 245 and 250 is subtle, and the difference between 250 and 255 impossible to see.
Can't remember where I read about it but what I use is a threshold layer, set to around 250. Applied on top of the layer I am dodging, this tells me when I 'break through'. This is so handy I set it up as an action. I also have one for shadows, warning me when I drive the tones below around 10. Depending on the printer and paper, these numbers may need to be changed but the latest printers and semi gloss papers can actually show the difference between zero and five, between 253 and 255 - heady stuff.
One more precaution, I want to do the dodging on a duplicate of the image, so it's time to flatten the image (you can save beforehand if you are inclined - I don't an do occasinally regret it but the files are so large...
This shows the application of the threshold layer on top of my image. Note that only a tiny area of water in the upper left is even as bright as 250. All of the bright water on the right hasn't come close to paper white - it looks fine on screen but I'm losing out on a lot of punch here so I'm definitely on the right track.
I start applying the dodging highlights at 5%, gradually building up the effect I want, occasionally flipping back and forth between the image and the threshold layer by turning it on and off (but all the work being done on the image of course).
And now you see that I have broken though 250 in a few more areas, not a lot, but it took quite a bit of lightening the water to get there so I know it will make a more impressive print. The second image shows the results of the dodging.
I still wasn't 100% satisfied though and decided to darken the water a bit further now that the whites are a lot closer to white, two more curves helped with that and voila you have the final, for now, image.
It's time to walk away from the computer. I might well make further adjustments in the future, I already have some ideas, but at this point I am not sure if they will really improve it and experience tells me I'll do a better job coming back to the image in the future, if at all. Don't forget you can click on any of the images to bring up larger versions. I did an especially large version of the final result at the top of the blog entry.
This has been a test run for a future book. I'd appreciate feedback on whether exercises like this are helpful and if so for what level of photographer (hey, market research). If you don't find it useful, that's ok to report too. Thanks.