Panoramic images date back to before photography, as there are many panoramic paintings. Panoramic photography is as old as photography itself. The first photographic panoramas were simply shots overlaid on top of one another. More sophisticated techniques were invented to create what appeared to be “seamless” photos. With modern digital photography, this is easy to accomplish with panoramic stitching software, which we will cover later. With this in mind, this tutorial will show, how to compose an HDR panorama, then generally cover the post-processing of such images.

First, what you will need:

• Tripod – preferably with a panhead. This is the sort of head that has independent handles for the pitch, yaw, and roll of the camera. “Ball” heads usually have some sort of ball-in-socket mechanism for moving the camera around. These aren’t really conducive to panoramas, unless the head also allows for panning too.
• Camera — Most any camera will work so long as it allows you to manually control the shutter speed.
• HDR software (Luminance HDR, Photomatix, or PhotoShop)
• Stitching Software (Hugin, Photoshop)

Helpful Items

• Lens Hood
• Polarizer

Taking the shots

• Determine the limits of what you want to photograph. This is important. Without this limit, you will either get too much or too little when photographing the subject. You can do this by choosing a landmark in the field of view.

Choose the Area

• Mount the camera in portrait mode for horizontal panoramas or in landscape mode for vertical panoramas.
• Choose an aperture and leave it there! You’ll want to use the same aperture for the entire composition. For general purpose panoramas, I think f/9 – f/11 works fairly well. You may want smaller apertures if you need long exposures or something like that.
• Set the camera on manual mode. With some cameras, you may be able to use aperture priority rather than manual mode because of AEB – Automatic Exposure Bracketing or similar programming. In any case, you want to leave the aperture consistent through all the shots used to compose the panorama to avoid variable depths of field.
• Take some samples from bright spots and dark spots in the field you want to capture. Note the shutter speeds of the various shots. If you’re camera does not have AEB, you’ll want to figure out the range of shutter speed you will need to capture the dark and light areas of the shot so they are not burned out or blacked out in the final image. That is, it may require a long exposure (say ¼ of a second) for the darks and a quick exposure (say 1/200 of a second) to get the details in the bright areas.

Metered Spots

• Determine the number of shots you will need to take. This is a function of the width (or height) of the panorama and contrast of between the light and the darks.
  • Conventional panoramas are usually composed of “slices”. I found overlapping images by 50% give pretty decent results for panoramas, as usually, the outer extremes are cropped out and I only use the center portion to the image appears more natural. This means that if 6 slices covers the area I want to capture, I will take 8 slices knowing I’m likely to crop out the half of the two edges, but need this information so the panorama software can correct the distortion.
  • Next, determine the number of images needed to cover the contrast (that is dynamic) range. (This is kinda technical…hold your breath) In the example earlier, I needed 1/4 of a second to 1/200 of a second to properly expose the content of the panorama. This is a rather large dynamic range. In theory, for any shot at a given shutter speed, doubling the shutter speed would mean that the sensor only receives half the light, which would correspond to one full stop on a camera. A digital camera shooting in RAW has a range of around 8-9 stops while JPEG is about 5, which means that a properly exposed image would have about +/- 4 stops each way for RAW and +/- 2 for JPEG. This means I need to set the camera at ¼ of a second, 1/32 of a second, and something around 1/256 of a second (1/250 would probably suffice) to capture the full dynamic range of the image. These are 3 stop intervals, but the camera could do it with 4 stop intervals. It is also good to have overlap here too, and with HDR, the more the merrier, I could do it in 2 stop intervals to get even better results. This would be necessary however for JPEG.
  • This means 3 shots per slice for 8 slices, a total of 24 exposures. Each color in this sample represents a relative slice, each overlapping the previous slice 50%.

Sample Slicing

• Start the pan by centering the camera on what you determined to be the edge of the field you want to capture and start taking shots at the lower shutter speed. Move up to the next interval, then take the shot. Repeat this for each determined shuttered speed. Then move the center so the edge of the frame is where the center was. This means 50% of the previous 3 shots will be in the next set, and repeat the process for the determined shutter speeds. You’ll want to avoid camera shake and want to lock the pan head with each exposure to make sure the camera does not move between each slice and HDR. Click the shots, then move to the next slice. You’ll want to do this as quickly as possible, (but not to get in a hurry!) because sometimes items in the panorama can change such as clouds or shadows. I cropped out the extremities on this shot, I was principally going for what is shown, so for this reason shot a lot of extra so I’d have lots to play with.

Start Here

Post Processing

HDR Panoramas are time intense on taking and in post processing. The example used in this tutorial took nearly an hour! The basic strategy here is to first, create HDR for each HDR slice, second create a panorama for these HDR slices, then 3 perform HDR post processing on the HDR Panorama.

• To combine each HDR slice, open your HDR software and work through the wizard for each slice. This typically means selecting each image that composes the slice and letting the software align merge the images into an HDR. If you are using the same software that combines images to HDR that you will need to merge the images, you may not need to save the images, although this is recommended because HDR and Panoramas by themselves use tons of RAM, so you may want to have only one slice loaded in memory at a time to conserve system resources. Luminance HDR (a.k.a. qtpfsgui, a free HDR Workflow) and Photomatix are two independent programs that perform HDR. Photoshop since CS2 performs HDRs and will do panoramas too. Here’s a tutorial on composing HDR’s in Photoshop. Luminance HDR (aka qtpfsgui) does pretty well to. Check out post processing in my HDR Waterfall tutorial for a how to on this. If you use a separate software, you’ll want to save the images out in some format that supports 16 bits per channel (or more), such as TIFF files. JPEGS, PNGs, and BMP use 8 bit. Saving in these formats will hinder any post processing after the panorama is created. WHATEVER YOU DO, DO NOT TONE-MAP THE IMAGE! This will be done after the panorama has been created. Save each HDR slice in a separate file for combining later in the panorama stitching software. Here are some samples for one of my slices I used in making the HDR Panorama in this guide:

• Next, open your panorama software and created the panorama from the HDR’s you created from each slice. Photoshop’s “Photomerge” wizard (File –> Automate –> Photomerge) does a phenomenal job at combining images and creating panoramas. There are some free tools such as Hugin will create panoramas quite nicely too, albeit it is more technical to use. These programs will walk you through a wizard like interface and combine the images. Photoshop does not “flatten” the image (that is combine the layers it creates into a single image) when you are done with it, so you’ll have to do this manually. You’ll want to crop the image too afterwards. Hugin and others will do this for you, so use the software accordingly. As before, you’ll want to save the image in a 16 bits-per-channel or 32 bits-per-channel when finished.

Photoshop Photomerge

• After creating the panorama, take the panoramic image back into your HDR software and perform tone mapping on the image as you would any traditional HDR photo. The tone mapping will attempt to compress the high contrast image into a dynamic range suitable for onscreen viewing such as a JPEG. Once you have tone-mapped the image you may want to perform some other adjustments such as levels, saturation, and balances, but this is purely discretionary.

Panoramic HDR

Here are some examples I did using these techniques.

I hope you have found this tutorial helpful. Happy shooting!

The first thing you’ll want to do is select or take a good candidate photo for processing.

• A good candidate photo should be well exposed—that is there are not a lot of blacked out or burned out spots on the image. HDR images attempt to prevent this from happening by merging images together. Just about any image that is well exposed can be used, although you may be able to get away with a poorly exposed image.
  • If you look at the tone curve below, you see two peaks and a valley. If the tone curve contains peaks or valleys that are flat, then the image is probably blacked out or burned out.
  • Most cameras nowadays have a way to see the tone curve while you are shooting the image, or right after you’ve shot the image.
• If possible, shoot the photo in a RAW format so that the camera does not do any processing to the image.
• You’ll want a low ISO too.

I shot this image at as a RAW file at ISO 100. This is the image with no processing save a resize and conversion to a JPEG, but even conversion to a JPEG does  alter it some.

Pseudo HDR Original Image

Flat Tone Curve

The next step you’ll want to do is to “flatten” the image. An image that is flat is an image that seems to lack contrast such that the colors look washed out. The human eye has a tendency to gravitate towards localized contrast, and images that don’t have such are typically described as flat. RAW images right off a camera almost always look flat without some sort of post processing. Images converted to JPEG usually go through some sort of algorithm to make the image look more natural, typically done by adding more contrast. The white line represents the bias by which the photo processing software is altering the photo. This line is straight (i.e. flat) while the human eye sees things in more of an “S” shape with the line flattening out at the top and bottom.

• If your image is flat already, you may not need to do any post processing.
• If the image has light and dark areas you may want do some post processing to flatten the image more.
• Add fill light if there are dark spots. You may want to decrease the overall contrast of the image and tweak the brightness some.
• If the image was shot at a high ISO, you’ll probably want to run a noise reduction algorithm to smooth this out some. Some HDR processes can really expose noise even at low ISOs.

Pseudo HDR Flattened Image

Once the image is flat, you’ll want to save it as a 16 bit TIFF or something that supports a higher bit range than 8 bit because tone mapping attempts to compress higher dynamic range images into lower range images, typically by reducing the color information. 16 bit TIFF files store enough information for most modern cameras.

After saving the image, you can process it like any other image that was produced using HDR, so fire up you’re HDR program and start tweaking the image with tone mapping!

Pseudo HDR Image

After I ran this image through Qtpfsgui (a free HDR workflow), I rotated and cropped it some to come up with this image.

Pseudo HDR Final Image