Sleeping Beauty Castle (little planet)

[inkista]

http://farm1.static.flickr.com/241/524664283_0a87f143d9.jpg (http://www.flickr.com/photos/inkista/524664283/)

Where Was it Taken?
Disneyland. Anaheim, California.

What settings did you use?
Manual mode. iso 100, f/8, 1/400s
Manual focus (focus set halfway between 3ft and infinity).
White balance set to "Daylight"

It's important to have all of these settings consistent for all the shots taken for seamless stitching.

What gear (camera, lens etc) did you use?
Rebel XT (350D)
Sigma 8mm f/3.5 EX DG Circular Fisheye lens
Velbon MAXi 343E tripod
Nodal Ninja 3 (http://www.nodalninja.com/) panorama head
Velbon QRA-3 Quick Release Adapter with QB-3 Plate

What software did you use?
Kekus PTMac (http://www.kekus.com/) (you could use Hugin or PTGui on the Windows side).
Adobe Photoshop CS2
Cubic Converter (http://www.clickheredesign.com.au/cubicconverter/)
Flaming Pear's Flexify plugin (http://www.flamingpear.com/flexify.html) for Photoshop.

How many images did you shoot?
Seven. Six images at 60-degree intervals, rotating around the no-parallax point of the lens, and one "zenith" shot straight up, again, rotating vertically around the no-parallax point of the lens.

Why did you compose the shot as you did?
I liked the vantage point. Inside the entrance, you couldn't see the facade. Just underneath it, there wasn't enough detail. Too much farther back, and the fisheye's tendency to exaggerate distances would have made the castle too small. The bridge gave me the right view of the castle, as well as offering a number of other interesting features for the VR view.

What post production work have you done on the shot (if any)?
A ton of post production work went into this.

First, the seven shots had to be stitched together in PTMac, to form what's known as an equirectangular mapping. A sphere is represented in an equirectangular mapping by simply translating the latitude of a point to the y-axis, and the longitude of a point to the x-axis (just like your usual map-of-the-world). You end up with a rectangular image that's twice as wide as it is tall that captures the full spherical view. (There was also some futzing with CubicConverter and Photoshop to patch the nadir. VR photography is a very deep subject. I'd recommend Panomundo (http://www.panomundo.com/panos/howto/index.html) as a first stop.)

Once the equirectangular (http://www.flickr.com/photos/inkista/524568764/) image was created, I then opened it in Photoshop CS2, and used the Flexify plugin to do a stereographic projection (http://mathworld.wolfram.com/StereographicProjection.html) and adjust rotation. (You can use Hugin (http://hugin.sourceforge.net/) to do stereographic projections as well, but Flexify gives you a very nice preview to work with).

What were you hoping to achieve with the shot?
Have something that was recognizably Disneyland, but weird enough to warrant a second look.

What did you do well?
Well, I got through the ton of post-processing involved. :) I'm particularly proud of the nadir (straight-down) patch job I did. A lot of masks and layers and cloning and patching were used.

How could you have improved it?
I could have planted the tripod in the middle of the bridge to get a more dead-on-centered geometry to the shot, but the heavy foot traffic made a wimp out of me. I should have corrected for chromatic aberration before stitching. And I could have shot RAW and done fake HDR to help offset the whole shooting-in-the-sun thing that's inherent in spherical coverage.

[Teewinot]

Wow! Very creative! I love it! It almost looks like a miniature until you see the people. Thanks for sharing this--it represents a lot of work and creativity and I appreciate the complexity of the whole process. :)

[inkista]

Thanks! I love VR photography and all the weird remappings you can get out of a spherical view of a place.

However, if I'm going for the miniature effect, I go old-school and use a tilt-shift lens (http://www.flickr.com/photos/inkista/514154581/). Gives me more time to dink about with VR photos in Photshop. ;)

[jiminyClickit]

inkista,

Outstanding work!

[xxpinballxx]

a really cool pic....
I think the top is amazing but the bottom seem a bit too distorted. Maybe try mirrorimaging the top hald to the bottom?
Cool idea though I like it and would work well with loads of city skylines.

[Darren Rowse]

I love it - very creative!

[JJthethird]

Really great idea, I look forward to the day when I'm confident enough with my ability to take on something like that. Excellent work and great information sharing.

Thanks,

John.:cool:

[Nieton21]

Really awesome shot, I love disneyland this image reminds me of the magic behind the scenes that bring it all to life. I was wondering if you could show a few of the images you used to compose the shot.

[inkista]

a really cool pic....
I think the top is amazing but the bottom seem a bit too distorted. Maybe try mirrorimaging the top hald to the bottom?
That's an interesting idea--to do a mirror-image, but it would sort of defeat the whole "little planet" feeling where you can still recognize the actual space. I agree about the oddness of the distortion--it's caused by the fact that I planted my tripod at the side of the bridge, rather than in the middle of it. I tried to mitigate it a little by rotating it to the bottom of the image, where you won't notice it as much.

Here's an example where I planted myself squarely in front of the target instead of off to one side:
http://farm2.static.flickr.com/1091/525836336_0ac59f1e04.jpg (http://www.flickr.com/photos/inkista/525836336/)

Cool idea though I like it and would work well with loads of city skylines.You'd think that, wouldn't you? Actually, if you get a traditional city skyline landscape shot, you'll get a very boring circular little planet. In the shot above, I was only about 25 feet away from the sun-sculpture. The fisheye exaggerates the distance and makes it seem much farther away. Rather than standing on the heights and getting the wide view as you would for a traditional cityscape; for VR photography, you'd do best standing in an urban canyon. This is what's counter-intuitive. VR shooting is almost the opposite of traditional landscape photography. You can turn your bathroom into a vista, while huge vistas turn into very boring squashed-to-the-horizon shots. You want things overhead.
Really great idea, I look forward to the day when I'm confident enough with my ability to take on something like that. Excellent work and great information sharing.Grin. It's my party trick. :D Like everybody else here, I'm still learning and struggling my way to better images. I've learned so much from everyone here, I thought I'd share a little of what I know about this one weird thing.

[Murtasma]

I always wondered how these were done now I know. I may try one sometime when I feel ambitious. Seems like a lot of work but, would be satifiyng when complete. Great write up this should go on the front page.
:D

[inkista]

Really awesome shot, I love disneyland this image reminds me of the magic behind the scenes that bring it all to life. I was wondering if you could show a few of the images you used to compose the shot.
Sure. I'll turn over the rocks and show you the seamy underbelly of VR photography. :)

Here are the eight images I took that day for this shot:
http://farm2.static.flickr.com/1173/531889542_d21a59dae8.jpg (http://www.flickr.com/photos/inkista/531889542/)

I use an 8mm fisheye lens because it gives you a 180º field of view across the frame. Unfortunately, it was designed for 35mm film, so on a crop-body, I don't (as you can see) get the entire image circle on the frame. To maximize vertical coverage you turn the camera into portrait view. You cover the horizontal 360º by spinning around the no-parallax point.

Because you don't get perfect 180º coverage, shooting a zenith and a nadir can come in handy for patching tiny holes, or simply for using the center of the lens (where image quality is best) where the stitched image would otherwise rely on the edges of the frame from the other images. The problem with the nadir shot is that you have to take away the tripod, and the chances that you've found exactly the same spot in space and absolutely the right amount of rotation is unlikely. And in low light, you have the usual problems with handheld. Still, it can sometimes be handy for patching that little hole in the middle, but unfortunately in this case, my handheld nadir was unusable because I didn't watch where my shadow was falling.

If you're feeling lucky/dangerous, you can shoot four images handheld at 90º intervals, and probably still get something that stitches, but you'll probably have a bit of work to do in Photoshop to hide the inevitable stitching errors from parallax error.

And here's the final equirectangular:
http://farm1.static.flickr.com/209/524568764_c82f10201b.jpg (http://www.flickr.com/photos/inkista/524568764/)

If you want the interactive VR view (and have broadband, shockwave, and patience), here's the SPi-V viewer link (http://fieldofview.com/flickr/?page=photos/58939979@N00/524568764&tags=equirectangular) for the image. Dragging around will change the POV, shift will zoom in, ctrl will zoom out. As a warning, it's been bombing out my browser of late (sigh). Someday, maybe Flickr will let us upload QuicktimeVR files...

[inkista]

I always wondered how these were done now I know. I may try one sometime when I feel ambitious. Seems like a lot of work but, would be satifiyng when complete.
You know, that's just what I said when I first saw the cubic VRs at panoramas.dk (http://www.panoramas.dk). I'm so sorry. :D

Great write up this should go on the front page.
:DAs a professional tech writer, I try to use my powers for good, not evil... ;)

[xxpinballxx]

That's an interesting idea--to do a mirror-image, but it would sort of defeat the whole "little planet" feeling where you can still recognize the actual space. I agree about the oddness of the distortion--it's caused by the fact that I planted my tripod at the side of the bridge, rather than in the middle of it. I tried to mitigate it a little by rotating it to the bottom of the image, where you won't notice it as much.
thing.

really interesting. I see what you mean about the mirror image.....
Wonder about merging two different halfs such as a city and a country side or mountain range and crating a image that looks like two totally different sides of the world so to speak. a farm and a skyscraper or something. Great post now I got tons of ideas.....Thanks for posting this.

[muddy250]

Now that is a beauty!
Great shot.

[inkista]

Thanks. Your method's still a hell of a lot easier (and lighter on the gear bag). ;)

OTOH, I can fold mine into an origami balloon (http://www.origami-instructions.com/origami-water-balloon.html):

<a href="http://www.flickr.com/photos/inkista/2521697391/" title="Fold the Sleeping Beauty Castle by inkista, on Flickr"><img src="http://farm3.static.flickr.com/2183/2521697391_c164dc821e.jpg" width="500" height="500" alt="Fold the Sleeping Beauty Castle" /></a>

Or turn it around to get a little sky instead of a little planet:

<a href="http://www.flickr.com/photos/inkista/2521739771/" title="Tunnel Sleeping Beauty Castle by inkista, on Flickr"><img src="http://farm3.static.flickr.com/2373/2521739771_034dcb11a2.jpg" width="500" height="436" alt="Tunnel Sleeping Beauty Castle" /></a>

Or just give you 180º opposite views of the bridge from where I shot:

<a href="http://www.flickr.com/photos/inkista/2521742077/" title="Castle (coming and going) by inkista, on Flickr"><img src="http://farm3.static.flickr.com/2295/2521742077_1b4b6b8109.jpg" width="500" height="250" alt="Castle (coming and going)" /></a>

I know. It's a sickness.

[dlaf]

Very cool - thank you for sharing your expertise!

I see you took the photograph at DisneyLAND and I'm wondering how many steps could be removed if you try the same photograph at DisneyWORLD...

Hmmmmm. ;)

[HockeyFan]

Thanks for providing so much info on the how-to. Great results.

[DEBBIE BROCK]

Amazing!!!! so creative!!

[MWerner]

But it caught my eye! Do you know of any websites or books that give a step-by-step on how to do the little planet process? I only have PSE4... Will that work for me?

[inkista]

Don't think PSE4 is going to work for you. You're going to need something that can do layers and masks for the pano editing, but PSE4, afaik, can't do the stereographic remapping, although it can probably do a polar little planet (you stretch the equirectangular into a square and then do a Cartesian -> Polar coordinate remap).

Remember that you have to create the equirectangular pano first, though. This is not a trivial task.

I put most of the links I know in my original post at the top of this thread. If you have $100+ to burn, get PTGui (http://www.ptgui.com). If you have $0 to burn, go get Hugin (http://hugin.sourceforge.net). They'll be able to do the stitching for you, and probably can remap to stereographic as well. I've done stereographic little planets using nothing but Hugin (http://digital-photography-school.com/forum/showpost.php?p=78972&postcount=10), although using the Flexify plugin in Photoshop makes it easier to see what you're doing. SebPrzd (http://www.flickr.com/photos/sbprzd) used the Gimp and the Mathmap (http://www.complang.tuwien.ac.at/schani/mathmap/) plugin to do his (then again, he knows the math for stereographic projection and has mad coding skillz).

This is not a simple step-by-step process, really, not yet. I'm still using a handful of different software tools to get the final result. So, be prepared for a steep learning curve. The guys at the panoguide (http://www.panoguide.com/) forum can also help you out when you get stuck (particularly with nadir patching and shooting tricks and tips). Their how to section (http://www.panoguide.com/howto/;jsessionid=FF7C354ECB78B1F6DEAED9356EE150AA) is pretty comprehensive, as are the Hugin tutorials (http://hugin.sourceforge.net/tutorials/index.shtml). I got started with Panomundo's tutorials (http://www.panomundo.com/panos/howto/index.html), but they assume you're on a mac, and are a little outdated, now.

Also don't forget Flickr groups. I only learned about stereographic little planets from seeing what Seb Przd was doing. I'd recommend checking out the equirectangular (http://flickr.com/groups/equirectangular/) group.

[MOM2SNOX]

Whoa!! That's radical and so cool! I'm totally impressed!

[bebekuririn]

Simply beautiful!

[candleman]

phwoar..! that sounds like hard work.. but a fantastic result.
good work

[inkista]

grin. After a while, it just turns into something like HDR--just another tool in the arsenal.

<a href="http://www.flickr.com/photos/inkista/2507449698/" title="Beach Ball by inkista, on Flickr"><img src="http://farm4.static.flickr.com/3254/2507449698_2156445a27.jpg" width="500" height="500" alt="Beach Ball" /></a>

XT. Sigma 8mm f/3.5 Circular Fisheye.
Four shots, handheld, @90º intervals (up to my ankles in the water--had a hard time timing the waves, so the nadir patch sucks). Stitched in PTMac to equirectangular; nadir patched with Photoshop and CubicConverted, remapped to stereographic little planet in Photoshop with Flexify.