Creating Worlds of Our Own
by Craig McNiven
Editor’s note: I’ve known Craig for several years now from the Worth1000 photography community and always loved his polarama planets. I’ve made a few myself, but they never looked anywhere as good or great as his and never had that sense of depth that truly made them look real. When I asked Craig about doing a tutorial on how to do this, I hadn’t realized that his actual .psd files were lost when thieves broke into his home and stole his computer, camera bag, lenses and several other items. He very kindly agreed to re-do one from scratch which to me shows his tenacity but also his grand generosity of spirit and sharing.
**note: Although anyone may attempt this tutorial, it is probably best suited for intermediate and advanced users as you should have a basic knowledge of how to crop, clone, sharpen, dodge, burn, and use layers and masks.
Polaramas are the polar coordinated conversions of panoramas and are easily generated in Photoshop. Many people often ask me how to get their polaramas to look more spherical and planet-like. This tutorial will attempt to explain how to get your creations looking less like a chaotic explosion of pixels and paint from the middle of space and more like the ordered plump, bloated and tranquil habitats floating serenely through space – the ideal worlds we imagined them to be.
For the uninitiated, a ‘polaramic’ image is a polar re-coordinated image created from a ‘panoramic’ image, which in turn is created from a set of individual shots stitched together to form a continuous 360 degree panorama – a long thin image in which the right side matches perfectly with the left.
In order to achieve this result, a series of overlapping shots are taken in succession from exactly the same point but at different angles . These individual images are then ‘stitched’ together either manually, in Photoshop using Photomerge, or via specialized software (e.g. Microsoft ICE, Autostitch, and others) to produce the final panorama. It is during this step that we should start creating our perfect worlds.
SHOOTING THE PANORAMA
Shooting the series of individual shot is a task in itself (and a most enjoyable one at that) and, like most activities – with each successive attempt the photographer learns new tricks to apply to the following session. Here is a list of handy tips that might make for an easier first time panoramic photographer:
- Stand in a Clearing
Don’t have any scenery elements too close to you; definitely none that will be disappearing out the bottom of the frame when fully zoomed out (e.g. a tree or a lamp post etc.) it will become clear later on why this is an important consideration in creating a good-looking planet.
- Shoot Portrait
Again, as we will later on see, the maximum size of your final image will be dependent on the height of your stitched image. In order to retain as much detail as possible, shooting portrait style is the obvious main solution
3. Compose First
Before starting the process, do a quick sweep around with the camera, not only to familiarize yourself with what sort of scenery to expect but to ascertain whether or not everything you want included is going to fit within the confines of the height of the shot, zoom out the desired distance at the point of highest features (note next point)
4. Zoom Out
If the scenery allows for it, zoom out as far as possible. It is important to keep some foreground in the shot and even preferable, but not imperative, to keep the top of the highest feature (e.g. tree, building, mountain etc.) within the frame. Should the main scenery be quite distant, with much ground between you and it, go ahead and zoom in a bit if you wish, just be aware that this will add to the image count that compiles the panorama.
It is important to overlap each successive shot with the next, overlap by at least 20-25% and err on the side of generosity. Don’t be tempted to include an item of interest in one shot when it will be well covered in the next. Start shooting centrally on a memorable landmark and finish centrally (or as close as possible) on the same item, it is surprising how easy it is to forget the starting point when you are mentally noting the end point of each individual shot also.
It goes without saying that the whole process should be done as quickly and smoothly as possible, especially if there are moving subjects such as people or vehicles in the scene, taking a smoke break after 180 degrees is unadvisable.
6. Stay Level
Don’t be swayed by the horizon – i.e. don’t keep the horizon line in the same position of the viewfinder for each shot, especially if the scenery is of vastly varying topography, this may cause the effective working height to be compromised – a situation which very easily goes unnoticed when shooting. Rather keep the camera plumb all the way around and allow the topography to vary within. A tripod with a spirit level will also solve all these problems, as well as remove any small height deviations between each successive shot, it will also keep each image square with the previous making the stitching process more concise.
Perfect (Using a Tripod):
Shooting without a tripod is challenging but not impossible, there are bound to be some shots which don’t line up too well and it should be kept in mind that any discrepancies dictate the inner limits of the shot determining the effective working height. That is why it is always nice to get excess portions of foreground and sky in the composition, better too much than too little.
Using a 55mm lens, fully zoomed out and overlapping at 20% (shooting portrait) usually results in a total of about 12 to 14 shots for every 360 degrees. The more we zoom in, the higher the count becomes.
Turning clockwise or anti-clockwise is a matter of choice, although clockwise will result in each successive image falling to the right of the previous which is handy when the intention is to manually stitch your images together
It should be noted that the pivot point of rotation should theoretically be the camera, and not the feet of the photographer, however, most scenery cases and stitching software seem to be quite forgiving of this and also, sometimes, you may be shooting around obstacles such as your vehicle or a tree etc.
Take an extra picture of the ground near your feet. This might sound odd, but we will later see how this might come in handy.
(note: See how, in the above example, the hikers were duplicated as they walked into each shot, such situations should be avoided or capitalized upon, depending on the will of the creator). The option now is to trim out any unwanted area of the image falling outside the effective work area:
Sometimes one might want to leave these chunks in as it could provide some quite valuable information later when the cloning process begins, but that is entirely optional.
Now is a great time to finalize all the colour adjustments and your preferred settings of the image as these images serve more purposes than just converting into planets – they look great printed out and even more impressive when viewed in panoramic viewing software (such as Pixaround.com).
Now we get to the crux of the matter, here we will learn how to covert our panoramic image into a customized planetoid.
Before we begin with the creation of our world, let’s take a quick look at why it’s not such a good idea to have scene elements intersecting with the bottom edge of the image. In the example below, we can see that not enough grass has been included between the road and the photographer, who, had he but stepped a couple of paces backward, would have gotten a nice seamless walkway included in his polarama.
Instead, he now has a cloning nightmare on his hands. The same applies to tree trunks and lampposts, although – if unavoidable – it is often quite feasible to simply remove the vertical element and rejoin either side manually. Elements such as trees or tall buildings intersecting with the top edge of the image are less of a problem as they can simply be cropped out of the final polarama, as demonstrated in the park scene below.
The conversion process is a relatively simple one, but if we wish to add depth and character (which we most definitely do) it will require a fair amount of trial and error. First we need to ascertain the height of the image (in pixels) by clicking on Image > Image Size and consulting the height in the Pixel Dimensions portion of the dialog. Memorize the figure – the more precise you have been – the closer it should be to the native width resolution of your camera. For our working example, let’s assume the height of the image is 1200 pixels.
Ordinarily, one might go ahead and create their polarama at this point, which would be done via the following procedure:
Image > Image Size…
We are presented with the Image Size dialog. Make the following adjustments :
Uncheck the ‘Constrain Proportions’ field – this allows the height or width to be adjusted independently.
Enter in the ‘Width’ field the same value as that of the ‘Height’ field – this turns the image into a perfect square with the image horizontally distorted.
3. Close the ‘Image Size’ dialog and goto: Image > Image Rotation > 180deg – This flips the image upside down, a step which is necessary to create a polarama around the earth rather than the sky – which would produce and invert our intentions (also an interesting concept if ever you are in the mood to experiment) .
4. Goto : Filter > Distort > Polar Coordinates… if it’s not already active, activate the ‘Rectangular to Polar’ button in the dialog and click OK. The images coordinates are converted and the resultant globular world is presented.
The first question we must ask is: is this what we wanted? In all but a few cases, the center of the image will have an awkward pinched feel about it with obvious lines radiating outward much like a dartboard (this is usually the stitching lines from the panorama) The 4 outer corners of the image where the difference between the square and the circle are, are filled with a consistent spread of radiating lines which originally were the top row of pixels from our initial image. Depending on our original position relative to the scenery, the items on our planet may look outrageously exaggerated and comical – trees could be huge bushy behemoths atop spindly little trunks, nearby people may be so distorted to the point of being unidentifiable, buildings illogically shaped and structurally unsound. Apart from having a flat, splat-like look about it, our planet also may appear to be in the process of exploding.
In order to give it some spherical appearance or ‘bloat’ we need to gain control over the perspective aspect of the image, i.e. we need to either bring the background closer, or move the foreground farther from us, ideally resting on the perfect compromise for our needs. The way to do this is to add perspective in the form of space to the top and bottom of the image in varying degrees depending on our requirements – this is done through trial and error.
As a starting point we can add the same amount to the top and bottom as to gauge how best to proceed. Let’s add 25% of the height to the top and bottom of the image, re-run the process and look at the result. Return (or undo your way back) to your panoramic image and call up the ‘Canvas Size’ dialog (Image > Canvas Size…). Change the measurement units to pixels if necessary and add 50% to the figure in the height field, make sure the middle block of the anchor point (default) is selected – this will ensure that the added 50% will increase equally to the top and bottom i.e. 25% each. Also note the canvas extension color which is the colour that will be inserted – this is unimportant for our needs. Click OK.
Now lets go back to the ‘Image Size’ dialog and repeat steps 1 to 4 again, this time using our new height value for the width setting in step 2.
We may instantly notice a difference from our initial attempt. The ground appears more rounded, the sky clearer (so to speak) and more coherent, the radiating stitch lines are less apparent and – despite a large gaping hole in the center of the image, giving it a CD-face look – the scene does have a more spherical feel about it. The elements within the scene also appear more geometrically correct and appealing.
Once again though, we should ask ourselves if this is what we want? – are there any key elements such as people or vehicles in the scene which are still distorted? Is there too much foreground with your scenery consisting of a thin strip on the horizon? Is there too little foreground with the scenery being bigger than the planet itself? These are all subjective values which differ for each individual scene and each individual creator, the best we can do is have a look at the perspective extremes and find our perfect Utopia somewhere within.
At this point it might be wise to save a copy of what we have, then go back again to your panoramic image, and once again call up the ‘Canvas Size’ dialog. This time, again using a figure of 50%, add a generous portion of that space to the bottom and a minor amount to the top. This time the operation will have to be done in two steps, each time using an opposite anchor point.
This procedure will produce a result looking something like this:
Once again, construct your polarama from the image as per steps 1 to 4 and save a copy of it. Return back to your panorama and repeat the unequal canvas size exercise, this time reversing the top and bottom values, to get a perspective of the opposite extreme.
Now, again, construct a polarama of this version, our third – which will give a fair indication of the extremes and also of the average. Let’s take a look at what each image produces.
Comparing the two images, we see that relatively, the one extreme (A) produces a larger ‘land mass’ while the other (B) has more ‘sky’ area. The elements in (A) are exaggerated more horizontally while (B) are exaggerated more vertically, (A) has a large expanse of grass with the scenery sitting further on the horizon, while (B) has more defined scenery looking a little top-heavy on a smaller sphere. The decision now lies with the creator as to what proportions to divide the top and bottom perspectives into, in order to achieve a result they are happy with. In our example here, I choose to make the top space double that of the bottom as I feel it provides a nice balance for this scene.
A Final Point
One final point to polarization : Should your scene have no defining clearing in which to stand (like on a road, atop a mountain etc.) then there is always an option to shoot for an ‘inverted polarama’ – where the image is warped around the sky rather than the earth.
The process and theory are all the same, save for step 3 in the process – rotating the image 180 degrees – which is omitted. In this instance, for best results, there should be no elements intersecting the top frame of the image.
The results are debatable; less entertaining than it’s plump cousins, but still quite surreal.
The next step is to finish off our world. Firstly, let’s rotate it to an orientation which appeals to the image. There are many factors that may come into play in this process, for instance one might want to orientate the focal point of their shot (a human/vehicle/building etc.) to the top of the shot so that it draws all the attention, or one might want to keep the sunlight point at the top, or any body of water which would look a little more natural sitting at the top, or any other reasons which might present itself to the creator. In our working example, I choose to orientate the city of Sydney to the top of my world so as to command some attention (also that’s where the sun is, which is another valid reason) despite the water of Sydney Harbour ending up on the right side of the planet and looking like it wants to fall off.
The next step is to crop. Remember that the corners will need some cloning attention which should be considered when choosing a crop, so cutting out some sky or clouds is not always a bad thing. Also, the crop needn’t always be square, cropping rectangles often help in presentation. Here is my crop option, also with some out-of-bound trees just to lessen the volume of cloud cloning in the corners.
Once the rotation and crop have been fixed, the next step is pretty obvious – getting rid of the open spaces in the center and corners of the image, let’s start with the corners. Clouds are usually not too difficult – using a combination of the cloning tool and patch tool as well as a varying and random degree of opacity usually allow for a great quality clone without too much obvious repetition, any unique cloud information which fell outside the crop area can also be copied and blended in if necessary. Of course, clear blue skies are an absolute cinch.
The middle circle is not as easy, especially if the photographer was standing on a paved area. Grass or beach sand are way more forgiving but still quite tricky – not only is the perspective pointing the grass blades toward the middle of the image, but the brightness of the grass on each individual photo usually varies as one goes around the perimeter. If the photographer had taken a shot of their feet at the time of shooting as suggested in the tips, this would be the time to use it. Otherwise I would suggest simply cloning inward around the perimeter using a fairly softish brush of about half the diameter of the hole. Once the hole has been filled, it’s a matter of blending in the varying directions and tones of the grass with an opaque cloning brush or patch tool. After a fair bit of practice and experience, this step will become quicker and more natural looking. In the case of grass, sand or snow, the sharpen tool is also quite useful at disguising the repetitiveness of the cloning. Finally our world will start looking presentable.
An optional last step is to add some shadow to your image. Unless it is a night shot, the position of the sun will usually be quite obvious to the viewer even if it is a midday shot, and adding some corresponding lighting and shadows will contribute a fair amount to highlighting the illusion of roundness. When looking at the shading of a sphere it becomes quite obvious where and how our shadows should work in relation to the position of our sun. Mentally break your planet into 5 or 6 ‘zones’ of increasing darkness to get a feel for the shading. How to achieve the final effect is up to you, one method is to perhaps burn in the shading manually using an increasing levels of exposure according to your mental diagram as you approach the ‘dark side’ of the planet – apart from being a lot of fun, this also allows for some freedom relating to individual elements that might occupy your scene, i.e. you may want to shade the sides of tree trunks or topography relatively in accordance with the sunlight and degree of detail you hope to attain.
Another more simplified method is to use this free downloadable shading mask (simply rotate the mask to the wanted direction) similar to the one above to globally shade your… er… globe. Remember this will apply constantly to all the elements in your scene and will not differentiate between ground and vegetation. I feel the object, however , is not to cast one side in total darkness as to try to emulate day and night, but, to subtly add some shadows to the ground in order to achieve a more spherical appearance. The shading of scenery and skies on the ‘dark side’ of the planet is once again totally up to the creator and desired effect of the image, here I have to chosen to discard it.
And that about rounds it up folks
There are few other activities, I can think of which makes one feel more god-like than creating one’s own planets, especially when doing so on a lazy Sunday afternoon. All the best and enjoy every minute of it. ~Craig
Click on images to view full-size
Craig lives with his wife and 2 young sons in Sandton, Johannesburg where he works in the structural engineering field. In his spare time his favorite hobbies include outdoor photography and photo manipulation, which he has been enjoying for about 7 years now. He hopes one day to catch a shot fabulous enough to hang on his living room wall! (editor’s note: He is much too humble! )
All photographs © Craig McNiven
Check out Craig’s Photos & Cool Manipulations HERE
Perfect Worlds with Craig McNiven – Creating Worlds of our Own by Craig McNiven is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.