As a photographer, or even a nature lover, you’ve probably stood staring at a wide vista wondering how you can possibly do that view justice in your photographs.  Maybe you’ve even wanted to be able to print that picture large to give a room in your home the same feel as actually standing there.  Have you ever thought about taking multiple images of a scene and stitching them together into a panorama?

Shooting “panos” has become more popular in the last few years with the advent of point-and-shoot digital cameras that stitch the image in-camera for you, but if you want to shoot a panorama with your SLR, there are some things to consider.  First, and this should be a no-brainer, pick something interesting to shoot.  It may look impressive as is, but if you pick out something with substance, it will look even better!

So you’ve got your vista picked out.  Now what?  In my experience, the key to producing quality panoramas is making sure your tripod is level.  If this is the case, you will have much better luck at the stitching process.  I have used two different methods to level my gear; the first is via a 2-stop process that’s super cheap.

My tripod does not have a built-in bubble level, so I start by removing the head, and placing a bubble level (~$3 at Ace Hardware) on the tripod.  Tinker with the lengths of your tripod legs until you level the tripod.  Now, put your head back on, and you’ll see that it can rotate on a level base.

By removing the head and using a spirit bubble level, I can make sure my tripod head is rotating on a level base.

The second, and more recent, method I’ve used is by putting a leveling base on my tripod.  I use the Manfrotto 438 leveling base, but other companies like Acratech also make them.  They’re more expensive than my first method, but they’re way faster and you don’t have to fiddle with your tripod legs.  To level your camera on your tripod use a hot shoe level like the Jobu Designs Dubble Bubble level.

My tripod head can be leveled easily and quickly using the Manfrotto 438 leveling base.

After leveling my tripod head, I level my camera on the tripod using a hotshoe level.

OK, now you’re level.  The rest of the pano practically shoots itself.  I usually work from left to right, making sure to overlap my images by anywhere from ~1/3-1/2.  This makes for a lot of images (many of mine are 12-15 images) but you’re almost sure to get a better stitch at the computer.

A couple of other things to consider before we discuss stitching:

• To shoot vertical or horizontal…that is the question.  In stitching, you’re almost guaranteed to lose a little of the top and bottom of your images because the software will correct for barrel distortion.  Because of this, I almost always shoot my panos in a vertical orientation.  That way I’m left with an image that is a little more proportional, instead of something VERY long and skinny!
• Just like with any shot, sharpness matters.  When I shoot panoramas, I almost always use a cable release and set my camera to shoot in mirror lock-up mode.  This reduces any possible vibration, hopefully ensuring a better stitch.
• Do not change your focus or any camera settings while shooting your panorama.  However, consider bracketing your shots.  That way you have more to work with at the computer, and you can even consider making an HDR panorama (let’s save that for another blog post, please).
• I haven’t addressed the problem of parallax yet.  Unless you are rotating on the nodal point of your lens (or using a pano slider), you will have parallax issues if anything in your pano is closer than ~10-15 feet.  The cheap solution is to make sure trees, rocks, etc are further than ~10-15 feet from my camera, and that seems to work to subdue parallax.  The obvious downside to my method is that you’re constrained as to your compositional choices.

So, now you’ve got your image files, and you’re back at the computer.  Can you simply align them in Photoshop manually, then crop the canvas down to its final size?  I think this works for most people ~1% of the time (I’ve managed to do it once, and I’m 99% sure it was dumb luck).  Every other time, the barrel distortion, as well as other factors, will necessitate the use of software to help you stitch your images together.

Dumb luck. I stitched this 4-image pano together manually--I've never gotten it to work since.

For stitching panoramas, I use the software package Panorama Factory.  I like it because it has a fully automatic mode, which should work really well if you leveled your rig properly and provided enough overlap between images.  Alternatively, Panorama Factory allows for various levels of manual stitching, letting you define stitching points, and for the occasional difficult to stitch image, I like that mode a lot as well.  If you are interested in shooing HDR panoramas, Panorama Factory will stitch different sets of exposure identically so they overlay perfectly.  I have to admit that I’ve tried this once, and it didn’t work quite as smoothly as I’d like.  I’ve found that using the auto-align feature in Photoshop CS4 works well before blending images.

So, that’s basically it.  This tutorial was meant to get you taking beautiful panoramas using a minimal amount of equipment.  I hope you enjoyed it.  Feel free to add to the comment section if there’s anything I missed, or needs corrected.  Hope you can find this useful!

An 11-image panorama of the Ritter Range, Sierra Nevada, California

Bookmark to:

I’ve decided to make Fridays the day for Marginalia; noting things from the week.  For this week:

• This weekend (Aug 15/16) is the U.S. National Parks free weekend!  Get out there and visit a national park for free.  For more info, click here.
• Last week, I mentioned that Jay Goodrich guest blogged on Darwin Wiggett’s blog.  In return, Darwin and Samantha Chrysanthou guest blogged for Jay, again on the acceptable uses of Photoshop in photography.  A very thoughtful commentary, you should check out.
• Speaking of Samantha, she recently blogged on the usefulness of social networking in photography.  So did Guy Tal.  So did a recent thread on Naturescapes.net.  The verdict?  Seems like resounding inconclusiveness.  I’m not sure if its worth it, but you can follow me on Twitter, Facebook, and Myspace.  In fact, please do.

That’s all I know this week.  See you where the light is.

Bookmark to:

Yesterday, I outlined one of the biggest problems with HDR photography I’ve encountered since venturing down that road a couple of months ago: the dreaded halo.  The “fix” of getting around it is surprisingly simple, and I’ll walk you through it with an image I recently edited:

Sunrise, Joshua Tree National Park, California, January 2009

This is a 3-shot HDR that I processed just last week, after getting the courage to try HDR.  You might be asking why I’m not following through on the image I showed you yesterday.  Well, in the interest of full disclosure, I still haven’t processed it.  However, I will follow the same workflow on that image as for this one. 

I won’t post the source images, but I will post the tone mapped images.  For my workflow, I generally follow Royce Howland’s suggested settings  for both the DE and TC methods of tone mapping.  With that said, here is the image tone mapped using the DE algorithm:

There is some lens flare that I’ll later clone out, but do you see the halo around the joshua tree?  That light purple line all the way around the whole thing?  That’s what we want to get rid of.  Other than that, this image doesn’t look too bad; its lacking some midtone contrast, which is also characteristic of the DE tone mapping algorithm in Photomatix.

Now let’s look at the other tone mapped image, the one done using the TC algorithm:

What do you see?  First you see that its very contrasty–almost too contrasty–especially compared to the other tone mapped image.  However, what you don’t see is the halo!  The TC tone mapping algorithm does not give you a halo.  Good.  Now, we can open these two images in Photoshop and pick apart the best of both tone mapping methods.

I like to use the DE tone mapped image as my base image, and I copy and paste the TC image over the top of it.  The next step–the crux of this workflow–is to choose a blending mode for the TC layer that–at least mostly–eliminated the halo in the DC image.  I’ve had pretty good luck with Overlay, Color Burn, and Linear Burn, but probably the most success in Overlay.  Overlay multiplies the dark areas and screens the light areas, and logically would be good for eliminating the halo.  

Once you’ve settled upon a blending mode that looks more or less “natural”, you will need to adjust the opacity slider to make it look even better.  On this image, I settled for 35% opacity.  

I hate to be anticlimactic, but once you’re done blending these two layers, you’re pretty much done and can edit the image like normal.  There was a little halo still left at the top of the joshua tree, so I selected the inverse of the tree and just closed the sky to look less halo-ish.   In addition, I cloned out some lens flare, and applied noise reduction (I use Imagenomic’s Noiseware).  Other than sharpening and levels/curves I didn’t do much to it.  One thing I’ve found is that I don’t have to apply much saturation to HDR images–they look pretty good as is.

I hope that this has been helpful, and that you find it useful.  I’d love to hear feedback as you try this method with your own images!

Bookmark to:

Although its been popular for a while, I’ve recently begun working with multiple images to produce high dynamic range (HDR) photographs.  What is HDR?  HDR is a series of methods developed to produce a high dynamic range image from a set of photographs taken with a range of exposures.  There are multiple methods for producing an HDR image, from blending multiple shots in Photoshop to processing a single RAW image multiple times and blending those in Photoshop to using HDR software like Photomatix.

My goal here isn’t to tell you how to make an HDR photograph, because there are already fantastic tutorials out there, written by people much more talented than I am.  Probably the most complete tutorial was written by Royce Howland over at Naturescapes.net.

Because you can probably already produce an HDR image on your own, I want to share some nuances of the process I’ve learned to make your images look more realistic.  These aren’t new techniques, and I didn’t come up with all of them myself–to give credit where credit is due, Royce Howland, E.J. Peiker and Paul Stoczylas are the masters.  I’m but a student.

Part I: Getting rid of the halo

Photomatix is a powerful program and has a very user-friendly interface that pretty much does the work of producing an HDR image for you.  Because there are no monitors that can visualize that kind of dynamic range in an image, you must first tone map the image before being able to display it.  You have two options: Details Enhancer (DE) and Tone Compressor (TC).  DE is the default, and although it lacks some mid-tone contrast, it looks a lot better than a straight TC image which is way over-saturated.

The problem with tone mapping with only one algorithm is that you’re limited by each one’s constraints.  For instance, halos are a big constraint of the DE algorithm:

Towers of the Virgin, Zion National Park, Utah, June 2009

This shot looks pretty good, but you notice at the junction between the sky and the cliffs that the sky turns noticeably light blue–that’s because I only tone mapped with the DE algorithm.

This halo can be especially problematic in images where you have trees against the sky.  In this series of posts, I’ll cover a couple of methods I’ve learned to make the HDR image look a little more  realistic and squelch that halo!

Stay tuned…

Bookmark to: