There is a type of image where HDR is an essential tool: Panoramas and vertoramas (vertical panoramas- for brevity I will refer to both simply as “panoramas”) are usually stitched from multiple exposures and depict a much wider angle than you can capture with a single shot (except maybe for shots taken with fisheye lenses). Due to this larger angle, the tonal range usually gets much higher, for example, because the sun is in the frame. This is a problem common to many panoramas and the logical choice is to use HDR to solve it. However, it is much harder to produce a proper HDR panorama than it is to produce a normal HDR image or a normal (non-HDR) panorama image. Again, I will not go into the details of taking the shots for a panorama (called “source images” in the following). There are numerous guides out there for doing this. Google is your friend here. I will explain how you can produce the final panorama based on a series of bracketed source images that are suitable for a panorama (e.g. have enough overlap). We call each set of three bracketed shots for a specific section of the panorama a “source image set”.
My assumptions are the following:
- You took your source images handheld. This is usually not recommended and it produces problems for both HDR and panorama processing. However, with the right tools, you can produce decent panoramas from handheld source images. Of course, the workflow also works for source images taken with a tripod, but this is not strictly necessary.
- The source images are in RAW format for maximum quality.
- You want to produce a single-row panorama – a a panorama made from a single series of shots taken e.g. from left to right).
- You have enough hardware resources! Creating an HDR panorama in full resolution requires a lot of memory and CPU power. While a lack of CPU power simply prolongs the process, a lack of memory can void your chances of finishing this process. Both Photomatix and Photoshop may require well beyond 2 GB of main memory for this task. So, my advice is that you get a computer with at least 8 GB RAM and 64-bit Windows. If that is not possible, you have to reduce the resolution of your source shots before you start the process. You have to experiment to find the highest resolution that your computer can handle.
Some Advice on Shooting the Source Images
Please check out my recipe for Taking Interior HDR Vertorama Shots to get a very detailed description of this process.
As I said, I am not going to write a tutorial on shooting panoramas. However, there are a few things that may make your live easier:
- Get the exposure right: When you identified what should be in your panorama, use the exposure meter of your camera to measure the shutter speed for the entire panorama (the aperture should, of course, be kept constant): Look through the view finder and scan the area you would like the appear in the finished panorama. Watch the shutter speed suggested by your camera (usually displayed in the view finder). Choose an area with a medium shutter speed (middle between the highest and the lowest shot). Focus on that area. Your camera should have an exposure lock button. After focusing, press that button and keep it pressed while shooting the series of source images. This will fix the exposure and the focus such that they are the same for each source image set, making the stitching easier and leading to better results. Note that this may lead to the situation that the darkest source image is almost black while the brightest source image is almost white due to the increased tonal range. We have to live with that! If you let the camera choose the best exposure for each source image set, then each section will be optimally exposed, however, in the final panorama, you will have unnatural transitions between the stitched shots – it will just look wrong.
- Leave enough space at the top and at the bottom as well as on the left and right margin: After the stitching process, the edges of your panorama will look like waves. You will have to crop the image to remove these uneven edges and produce a clean rectangular image. Consequently, some portion of the image at the edges will get lost. You have to take that into account while shooting.
General Workflow – Overview
The general workflow will be as follows:
- Produce a 32-bit HDR (not a tone mapped LDR!!!) from each of the source image sets (the 3 or more bracketed shots for each single section of the panorama) with Photomatix
- Stitch the 32-bit images with Photomerge (a tool present in Photoshop)
- Tone map the stitched 32-bit image with Photomatix
- Post-process the tone mapped image in Photoshop
So you see, that this is not a straight forward workflow as you have to change between Photoshop and Photomatix multiple times. It is very important that you create the 32-bit HDRs first. In principle, you could also stitch a panorama for each exposure setting (one for 0ev, one for -2ev and one for +2ev if you use standard bracketing). However, the alignment of these three panoramas in Photomatix will not work in most cases. So we let Photomatix do the alignment job for each of the source images individually, and afterwards, Photomerge does the panorama alignment between the overlapping images. This work very well, even for handheld panorama shots.
- If you want to create a pseudo HDR panorama (using a single RAW image for each section of the panorama instead of a set of three genuine exposures), then you should load each of these single RAW images into Adobe Camera RAW (ACR) and create a set of three TIF images from each of them, each with a different exposure. In the simplest case, you set the “Exposure” slider in ACR to -2 and save, to 0 and save and finally to +2 and save. However, you should also try to get back as much details in the highlights and shadows as possible using e.g. the “Recover” and “Fill” sliders. After you have saved the three TIF files for each RAW, you can just continue with these files in the same way as you would with original RAW files (or JPEGs).
- Put all the source images of your panorama into a separate folder. This folder should only contain these images, no other image files. If you have multiple source image sets for different panoramas, create a source folder for each of them and copy (or move) the source images to the respective folder.
- Start Photomatix and go to “Automate > Batch Processing”
- In the dialog that comes up, tick “Generate HDR Image” and click on the “Setting…” button behind this option. Set the HDR options as desired. I recommend that you tick the “Reduce chromatic aberration” and the “Reduce noise” boxes. Also, if you have not set the White Balance to a fixed value while shooting the source images, then define a white balance value in this dialog (only possible for RAW images). Note that we only create the 32-bit HDRs in this step. There is no tone mapping involved yet. We will do the tone mapping after stitching the panorama.
- Further down in the automate dialog, tick “Align Source Images” and “By matching features”. Tell Photomatix that each HDR should be created from 3 images and give it the name of the source folder.
- Under “Save 32-bit HDR images as”, choose “Open EXR”. Photoshop can read this file format.
- Click start and let Photomatix create all the HDR images. Note you can also create every single HDR image manually, but the batch processing option saves a lot of time and work.
- After the HDRs are created, they are stored on your disk (usually in a subfolder of the original source folder created).
- Now, start Photoshop for the stitching: Go to “File > Automate > Photomerge…”.
- Click on “Cylindrical” and select the HDRs you just created. Choose “Blend Images Together”, “Vignette Removal” and “Geometric Distortion Correction” and click “Ok”. Photoshop now merges your panorama. This may take a while.
- When the merging is done, save the resulting image in the OpenEXR format. This is an HDR format that Photomatix can read.
- If you are stitching a vertorama (a panorama in vertical direction – source images not taken from left to right but from bottom to top) you should now rotate the whole image 90° counter-clock wise: Choose “Image > Image Rotation > 90°CCW”.
- Go to Photomatix again, load the newly created 32-bit panorama, and click on “Tone Mapping”
- Set the tone mapping settings as appropriate and click “Process”. After a few seconds (or minutes, depending on your computer) you will see the tone mapped image. Note that I usually set the white and black points to 0 in Photomatix since Photoshop does a much better job at correcting this. Hence, the result you see below looks very dark with muted colors. All of this will be corrected in Photoshop in the next step.
- Save the resulting image as 16-bit TIF (this gives you the highest possible quality for the post-processing).
- Go to Photoshop again and open the saved TIF.
- Crop the image to remove the black parts on the edges.
- Sometimes, especially if have not got a lot of experience, you will find that the necessary cropping also removes some parts of the image that you would rather like to preserve. For instance simple cropping may remove too much of the sky, affecting the composition of the image. In this case, you can do the following:
- Crop the image such that sky has the right proportions (possibly leaving some of these ugly black areas around the image.
- Use the clone stamp tool: sample parts of the sky and stamp them into those black areas.
- Use the healing brush to fix any unnatural transitions between the stamped areas. However, be sure not to get to close to the edge of the image with the healing brush.
- Apply any post-processing magic you like
- For the normal post-processing in Photoshop, the same principles apply as for any other image. Correcting the contrast and the brightness as well as the saturation globally is the least you should do. For this particular image, I have treated the sky and the foreground differently, using a layer mask. To get a coarse impression, click on the image below. Also go to the photo’s flicker page to get some more details on the post-processing done here.
Special Considerations for Vertoramas
The workflow for vertoramas is very similar to the panorama workflow detailed above. However, there is one step that is particularly important in this case: distortion correction
Vertoramas look particularly spectacular for certain subjects like churches. The interior of a church lends itself to this technique since it can bring out all the impressive details and create very unusual perspectives. Below, you see some examples of interior vertorama shots:
For this kind of shot, symmetry is very important. To achieve this, you have to adjust the distortion of the stitched image. Photomerge normally does not create a symmetric result as it tries to distort each source image such that they fit together.
Learn HDR Vertorama Photography
Everything about the technique in one ebook
7 Chapters - 183 pages - 112 illustrations for $19.95. Subscribers get 20% off!
- After stitching, go to “Image > Canvas size…” and add 1000 pixels to each dimension. Leave the anchor in the middle such that 500 pixels will be added on each of the four edges. This gives you the room required to correct the distortion. The excessive space around the corrected image will be cropped again later.
- Press Ctrl-H. This will bring up the grid that can help you in the correction.
- Go to “Edit > Transformation > Distort”. A bounding box will appear around the image area. You can drag any of the corners of this box to distort the image. Use the grid to judge whether everything is straight and whether, and make sure that symmetrical element of your image are really symmetrical (at the same spot to the left and right side of an imaginary vertical middle line). You can hold the Ctrl key while dragging the corners of the bounding box. This keeps the corners from snapping to the grid and allows you a very precise positioning.
- When you are done, press Enter on your keyboard.
- Crop the image by activating the crop tool in the Tools palette. Make sure your crop is symmetrical too. You can use the reference point at the center of the crop box. Drag the corners of the crop box such that this reference point is on the vertical center line of the image.
I have created a little time-lapse making-of movie for the Valletta church shot. This video shows the whole process.
Taking the Source Shots
- As you can see in the examples above, there is usually some architectural element that defines the focal point of the composition. For most church shots, it is the dome. This element has to be in the frame of the final image. Therefore, it is vital that you take the source shots accordingly:Take more shots than you think you need. Distortions and cropping will take away a comparably large part of the image on the edges. I start with the bottom shot with the camera almost pointing straight down (yes, my feet are on that shot). For my last shot I am bending backwards shooting the top part of the wall behind me (if there is any). This covers an angle of more that 180° in vertical direction! It’s easy to crop the image, but it’s impossible to add anything in post-processing.
- When you shoot, the center of your lens needs to be on the vertical axis of symmetry. For the church shots above this is the imaginary line running straight from the center point of the dome down to the floor. If you are not on that line, your shot will not be symmetrical or you need to crop too much from one side later on.
- Keep your camera square: Imagine there would be a plain wall at the end of the church. When you point your camera directly at this wall, the plane of your camera’s sensor needs to be absolutely parallel to this imaginary wall. Otherwise, one side of the shot will be further away than the other, making it impossible to achieve symmetry.