This is a guest post by Dan Villeneuve
Creating HDR panoramic photographs requires a lot of attention to detail and a ton of effort, but when all is said and done the result pays for the work. After receiving a lot of great feedback and tons of questions on Phogropathy’s Forum, John recently asked me to share my process for creating these HDR panoramas in a guest post here on Phogropathy. As I am always trying to help people understand the process I thought this would be a great opportunity to do just that – so sit back and grab a coffee we’re in for one epic post on creating HDR panoramic photographs – first is the set up!
Quick note: Before we get too far into this post if you are completely new to HDR you might want to check out this introduction to HDR post that I wrote on my own site.
Today I’m going to use an example to help illustrate the workflow that works for me.
And now for the practical example
That first part from yesterday is pretty straight forward. So let’s use an example to help illustrate the workflow that works for me. Recently I was invited to photograph the inside of a B-29 Superfortress Bomber at the New England Air Museum here in Windsor Locks, CT. It was an amazing experience which I will write about in a separate blog. This is the end result that we’ll be striving for (minus the watermark):
The final tone mapped panoramic image of the B-29′s cockpit. Click the image to see the larger version.
In order to achieve this shot, I realistically want to do a panorama using three portrait oriented compositions (tilting the camera 90° on its side). You’ll notice that I deliberately did not say “exposure” or “shot”. That’s because we are going to actually do multiple exposures per frame (That’s where the HDR comes into play, which I will detail later).
So why not just take it in one shot and not go through the extra effort of stitching together 3 separate panned shots? Simple. Detail. My Nikon D80 only has a 10 Megapixel (MP) sensor, so the amount of light data captured will be limited. And if I were to crop this view from a single frame, then there will be less detail than if I captured the same area with 3 frames worth of 10MP. In other words, 30MP worth of data is going to give oodles more detail than <10MP. If you want to get technical, when taking HDR into consideration, we’re actually looking at 3 bracket groups of 12 exposures each, totaling 36 exposures… so 36 x 10MP = 360MP worth of detail! Not to mention that I would need to have a wider lens than my 18mm focal length if I were to attempt this with a single frame composition.
First thing is to set up your tripod. I like to do my panorama compositions from left to right, then top to bottom if I need more than one row of height. Here’s the camera setup during the left-most frame:
Camera setup for 1st frame in the 3 part panorama. You can see the histogram mode active in the LCD.
When you’re composing for a panorama, I find it’s best to make sure your tripod head is as true to level as possible. I will look through the viewfinder while panning back and forth to ensure I’m going to end up with a well leveled capture. That’s one of the reasons why I adore this beast of a tripod/head combination. Manfrotto generously provides spirit levels for just about every axis you could need a level for. I’ll leave the pan lock (leftmost handle on the head, in the setup picture) loose so when it comes time to pan to the right, I’ll minimize any unwanted alignment variations. I also try to position the tripod so that the panning axis is actually along the sensor mid-line, and not the tripod head central axis. For large outdoors landscapes it’s not as critical, but in a tight space, like this aircraft cockpit, it’ll make a big difference in the final rendering as far as visual balance goes.
As I mentioned earlier, the example image is comprised of 36 total exposures (3 bracket groups of 12 exposures per group).
I prefer to use the term “Bracket Group” to describe the total number of exposures per composition. Typically the term “Bracket” is used to refer to the 3 exposure group that most DSLR cameras will take automatically (The Nikon D7100 and Df now allows you to take 5 exposures per bracket, the D4 allows for up to 9 exposures). So for the purpose of my workflow, I’ll use “Bracket Group” to describe the entire exposure group taken to make one framed composition.
I triple check my camera’s settings:
- Aperture Priority Mode
- VR turned off
- Bracket shooting turn on and set to 3 exposures at +/- 1EV (stop). I see some people do +/- 2EV, and I used to do that, but about 6 months ago, I discovered that this is too far of a gap in data between exposures and causes weird (mainly grainy/noisy) artifacts when Photomatix merges the exposures. Photomatix has to interpolate the data in between the given sources. Ideally it’s best to shoot more exposures at a tighter stop difference.
- Shutter actuation in remote mode (wireless remote)
- Focus triple checked
- Set the camera LCD to display the histogram when the image review kicks in after each exposure is taken.
- Expose for -4EV to begin the bracket group.
So why am I electing to expose for -4EV?
Being the super organized person that I am to a fault, I like to start my bracket group at the underexposed extent (exposing for the brightest areas of the composition) and work my way up to the overexposed extent (exposing for the darkest areas of the composition). Most DSLRs have a tick mark exposure meter in their display that range from +2EV to -2EV and are usually graduated in 1/3EV increments with the larger tick marks indicating a full stop. Nikons have the + on the left and the – on the right. Canons are the opposite.
Nikon Viewfinder layout
I have my D80 configured to increment in 1/3EV steps. So every click of my rear control wheel equals a change of 1/3EV. So when starting out my bracket group, I’ll adjust the exposure to where the exposure meter displays -2EV, then manually click off 6 clicks (an additional -2EV) to the right. This will put me at -4EV. I’ll fire off the first bracket (-5EV, -4EV, -3EV), paying attention to the histogram. After the first three exposures, I’ll ratchet the exposure back up to the -2EV meter and kick off the next bracket (-3EV, -2EV, -1EV). Repeat and rinse until we get to the topmost part of the meter range, +5EV. Now for those of you paying attention, you’re probably saying, “Hey! Wait a second! You’re making duplicate exposures!” And you know what? You’re absolutely right! I’m not worried about duplicates because Photomatix rather conveniently will recognize this and offer options to vary the duplicates by 1/3EV or 1EV.
After I capture the first bracket group, I carefully pan the head to the right so that there is between 1/3 and 1/2 frame overlap between the two compositions. The more overlap you give, the more usable area you’ll have left once Photoshop finishes its stitching process. I’ll then reset the meter back to -4EV and shoot the next bracket group. Repeat and rinse until you finish the rightmost side of the total composition. The number of composition frames you capture is really going to depend on the subject. I have some panoramas from that B-29 shoot that were 4 and 5 frames total.
When I finished the 3 bracket groups, this is what I was left with:
Bracket Group 2. Under exposed to overexposed in 1EV increments. Click the image to view larger.
Bracket Group 3. Under exposed to overexposed in 1EV increments. Click the image to view larger.
You can see how each of these bracket groups have a compositional overlap as I panned to the right, this will become important as we go into the processing phase where Photoshop will be matching the different grouped images with one another to create the final photograph.
We’re Not Done Yet!
Back over on my site I put all these images together and explain the software side of things – continue to that post now - and don’t forget to share this series with people who you think might enjoy it.