I frequently receive emails inquiring about my photographic technique and equipment.
Here, I will attempt to explain these things, both in layman’s terms and in technical terms. (nerd stuff…)
But first, a quick run-down of functionality…
Double-click (or tap) any image to launch the interactive tour.
You can use the buttons on the bottom of the tour to enter fullscreen mode, open or close the map, share the tour,
flip the navigation mode from “drag” to “free” or to zoom in or out. You can also zoom in/out with the mouse.
There are different viewing projection modes available if you right-click. (Stereographic mode is fun…)
If you enjoy viewing these images, please help me achieve my goal of demonstrating to as many people as possible
how amazing it can be to stand in the places I’ve been fortunate enough to have stood by sharing the
page that brought you here, or by sharing individual panoramas (from within the tour…)
If you see this title stripe, it means that I shot this panorama with a special piece of equipment that
is fine-tuned to specific camera body/lens combinations to eliminate parallax.
This is the desired method of shooting a 360x180 equirectangular (getting into nerd territory…)
This means that the image will be as near to technically-perfect as possible.
By “technically-perfect”, I mean you’ll see no errors due to camera misalignment.
Blue = good.
See below for technical explanation...
If you see this title stripe, it means that I shot the panorama freehand, without a tripod or nodal head.
Perhaps I was perched precariously on a ridge at 14 thousand feet. Maybe I just felt lazy and didn’t want to drag
an extra 30lbs up a mountain in addition to my other seemingly omnipresent 30lbs of various gear.
Maybe I was on a technical climb where I really had little business even dragging a 5DMKIII up with me.
Whatever the case, no tripod/no nodal head means parallax errors. You will undoubtedly find some.
If you do, email me. I usually try to painstakingly fix them.
Maybe I just missed it. Maybe I was temporarily subscribing to the “It’s good enough.” school of thought…
Red = bad.
If you see this title stripe, it means that I shot a multi-frame panorama with a telephoto lens mounted to a tripod.
The end result of that process is a very high resolution image. Every time my camera takes a photo, it’s capturing an
image that is 5760x3840. 6K by 4K, basically. When you shoot a pano that is 10 shots across, you get a huge image.
(Though not 60K because there is overlap between images that serve as anchors for the stitch. Usually 20% or so…)
If the total number of pixels in the image is less than one billion, then you have a “megapan”.
If the total number of pixels is above one billion, you have a “gigapan”.
Another way to look at it is in megapixels. We all know what those are.
My camera shoots a photo that is 22.1 megapixels. (5760 multiplied by 3849 = 22,118,400).
My largest non-automated pano to date is 421 megapixels. The power of stitching…
If you see this title stripe, it means that I shot a multi-frame panorama, same as above, but without the use of a tripod.
Normally, this would not pose much of a problem at all, especially since I shoot many panos with an Image Stabilized lens.
Of course, I have a tendency to not do normal things, so I shoot 421 megapixel HDRi freehand megapanos sometimes.
An HDRi is basically one photo derived from many, pulling the best exposure values from each and combining them.
What THAT means is you may see what is known as "HDR ghosts". (boo!)
An HDR ghost occurs when anything in the frame moves between the constituent shots of an HDR set
and when the software used to process the image subsequently fails to solve for and correct.
Basically, you’ll see “smudges” and “smears” sometimes in some parts of the image.
I know…frustrating! (I guess I should have used a tripod.)
If you see this title stripe, it means that I shot a multi-frame HDRi panorama using an automated rig.
Basically a robot that allows me to program in a shooting protocol that the robot then follows.
Using one of these devices, you can quickly and accurately shoot a gigapan or larger.
None of these are on my site yet. I have the rig and have used it for other jobs, just not in the mountains.
I need to quit slacking!
(I was lucky enough to be able to shoot gigapans of the New York City skyline from the 102 floor
of the One World Trade Center before it open to the public, though…and that was really cool…)
Let's be clear, here; You’ll never see this title stripe.
Because…that would require a mix of patience and insanity that exceeds even my own abnormally-large portion of both…
I shoot 360x180 HDRs with either a Canon 5DMKII or 5DMKIII and either a Sigma 15mm f/2.8 EX DG or
a Canon EF 8-15mm f/4L. (usually at 15mm but sometimes at 8mm.)
My panohead of choice is currently a NodalNinja M1-s, and it’s mounted on either an InduroGear AT313 or an
InduroGear CT014, with either an InduroGear BD0 or an InduroGear BHL3.
The ballhead makes it easy to very quickly level the panohead, though, this is not critical when shooting a 360x180,
as any imperfections in leveling can be remedied in software. Segue... - - >
I currently use Adobe LightRoom CC to process my RAW HDR sets (mostly to correct for chromatic aberration and
whitebalance variance across frames), SNS-HDR Pro for HDRi tonemapping and Kolor Autopano Giga for stitching.
My 360x180 HDRi pipeline is as follows:
-Calibrate the NodalNinja M1-s:
Canon 5DMKII with Sigma 15mm f/2.8 EX DG = Upper rail 10, Lower rail 13.
Canon 5DMKIII with Canon EF 8-15mm f/4L = Upper rail 9.95, Lower rail 13.15
-Shoot in RAW with the camera “zeroed out”. (no in-camera adjustments to tone, NR, sharpness, etc…)
I shoot with a 15mm lens on a full-crop body every 45º, one zenith and three nadirs.
I could shoot less around, 60º instead of 45º (six around rather than eight) put I prefer the extra overlap
and I like round numbers. It makes my two tripod nadir shots match the radial shots more visually intuitively,
(at 90º degree intervals) which makes for easier tripod paint-outs. (I wrote Photoshop macros to do this…)
I shoot the third nadir freehand, and this is usually the only time I do NOT shoot at ISO50 (see below).
I usually bump up my ISO by enough stops to make my slowest set-frame suitably fast enough to handhold.
Sometimes I reverse the head and shoot a proper third nadir (on-tripod), but usually I do not.
-When shooting outdoors (mountains) I typically shoot about a 10EV spread, more if necessary, but I’ve found 10
is usually sufficient to capture the illumination range in the environs I like to be in.
-10EVs can (obviously) be accomplished with a five-frame set at two full stops apart.
Sometimes I will shoot 1 1/3rd or 1 2/3rds, sometimes I will shoot 2 1/3rds or more.
It just depends on what’s happening in the environment and how much luma range I think I need.
This is always a payoff. I need ALL the range. Always. Getting it is usually not practical…
-I like to stay at five frames and adjust my stops wider rather than extending out to seven frames.
Seven frames takes a little longer, and quite frankly, five frames takes too long. (i.e., it’s not instantaneous…)
It can get pretty windy in the mountains, and clouds and trees have a horrible habit of moving,
as does my body when i’m shivering in 15F temps, being buffeted by winds, straining to maintain a
precarious position that would mean a very long fall if I did not, and cursing myself for taking my gloves off…
(“Take your gloves off.” They said… “It’ll make adjusting the camera easier.” They said… Yeaaahhhh…)
That aside, I have shot many seven-frame 10EV-15EV spread HDRi images in the moutnains.
It’s all part of the “payoff equation”.
-I shoot at ISO50. Almost always. Both for 360x180s and otherwise, tripod/head or no.
Lots of math goes down in the tonemapping process. Darks get pulled up and gamma/hue values tweaked.
All of that exacerbates noise, so if you start with some, you’ll end up with more than some…
I’ve read arguments that state that sensors work better at ISO100. I do not agree with those arguments.
ISO is not standardized across camera makes and models and chipsets. ISO100 on a 5DMKIII does not
look like ISO100 on a Rebel, for example. But ISO50 on a 5DMKIII is less noisy than ISO100 on the same cam.
Both the Canon 5DMKII and 5DMKIII process noise incredibly well. At ISO50, there is virtually no noise,
and that’s a great basecase to start from. I don’t like noise, in general. I think it detracts from what I’m
attempting to achieve. (a near-as-possibly-accurate-scientific portrayal of what my eyes saw at the time.)
-I process all RAW files through LR, first, for two reasons;
Equalizing the whitebalance across frames and eliminating chromatic aberration.
Very rarely, I will use the Noise Reduction capability in LR5 CC.
-It is here in the process that I leave the RAW workflow. I export all the constituent frames as 16bit/chan tif files.
I do this because the tonemapper I use just seems to prefer them. I’d love to feed SNS-HDR Pro CR2 files,
and I will probably keep trying, but I get different results when I do so. Bad results.
(For example, changes you make to a CR2 or any other RAW file in AdobeRAW are not saved in a way
that is accessible to non-Adobe products.) Any photogs out there can correct me if I’m wrong. I’d love to be…
I average 60 frames (twelve 5-frame HDR sets), and a 23mp 16bit/chan tif is 125mb on average.
That’s a 7.5GB disk footprint per (typical) 360x180. Ouch…
-I process all photography and video on a MacBookPro, so I need to leave OSX and bootcamp into Windows7
to use SNS-HDR Pro. It’s a bit of a pain, but i’ve done speed test vs. Parallels and bootcamping is much,
much faster. I’ve tried to convince Sebastian Nibisz to port it to OSX, but he aint haven’ it…
-I save all the tonemapped images as 16bit/chan sRGB tifs
I take these images into Photoshop CC to paint out dust specs, birds, people, etc.
I also compile a single nadir from the two tripod nadirs, leaving only the small area of the actual tripod head.
-These are the 11 images (or 10) I feed AutoPanoGiga
All of the shots taken from the tripod with the nodal head of course stitch flawlessly.
The additional freehand nadir is where some errors can occur, but you’re really only using that for a very small
area at the bottom, and AutopanoGiga has a few tools that facilitate the alignment and visual precedence
of this freehand nadir. You can pick the parts of that image you want to cover the tripod head in the other nadir.
I do this when there is a pattern in the nadir that needs to be maintained, but if there is no pattern (rock, snow)
I always do this in Photoshop by hand, thereby ending up with one clean nadir with zero stitch errors.
Because two of the nadirs (at 90º angles) were shot on the tripod, they match exactly with the radial frames.
The nadir patch that covers up the tripod head is simply that…a patch. I stay away from the edges when I
create this patch in Photoshop and thus maintain zero parallax.
(Shawn…there’s the answer to your question…”Where! Are! Your! Feet?!?!”)
-I save the final composited 360x180 as a 16bit/chan TIF or PSD file. Either works and there are pros/cons to both.
-The rest of the process involves boring HTML and XML coding.
As should be quite obvious from my site, I don’t like doing either, and I certainly don’t want to waste your time
discussing the very exciting ins-and-outs of CSS templates and JQuery Mobil Swatches!! (whatever those are…)
I use Kolor PanoTour Pro to do most of the heavy-lifting, which in turn depends heavily on the krpano engine.
Email me if you’re actually interested in that type of stuff…