HDR Cookbook – Taking Interior HDR Vertorama Shots
Among all the possible shooting situations, indoor HDR vertoramas are certainly one of the most demanding. The resulting photos are very rewarding as they offer an unusual perspective and can capture the greatness on many places. But how do you actually produce the source shots for such an image? What is the proper technology and an adequate technique? This recipe will give you valuable hints on how to proceed in such a situation.
Overview
Watch this interview on TWiT Photo to get a more tangible explanation of the principles, the technology and the techniques of HDR Vertorama photography.
In principle, a vertorama is nothing else than a panorama in vertical direction. Therefore, we have to take a series of shots (called “sections” in the following) from bottom to top and make sure there is enough overlap between them to be able to stitch them together on the computer. Combining this technique with HDR means that we have to take 3 or more shots with different exposures for every section of the vertorama – one normal, one underexposed and one overexposed. The challenges you will face when doing this are the following:
- Most interesting environments have low ambient light implying low shutter speeds. Blurry images may easily be the result of this.
- The composition of the final image is not obvious. It will not look like anything you see through the view finder
- You will have to be precise in taking the shots since being sloppy will prevent the software from working correctly.
You are going to learn how to deal with these challenges. In particular, I am going to explain the following:
- The gear you need (camera, lens, etc). Which camera functions are vital?
- How you compose a vertorama shot.
- How you need to configure your camera.
- The technique of taking all the source shots adequately.
Learn HDR Vertorama Photography
Everything about the technique in one ebook
7 Chapters - 183 pages - 112 illustrations for $19.95. Subscribers get 20% off!
Which gear do you need?
Camera Support
Most tutorials on panorama photography as well as those on HDR photography emphasize that for both you need a tripod – preferably a sturdy one such that camera movement and shake is avoided. Of course, it is true that using a good tripod enables you to produce highest-quality photos. The problem however, is that in many situations, you don’t have a tripod with you, you don’t have the time to set it up, or tripods are prohibited.
So, lesson number one is very simple: Forget about a tripod! We are going to shoot this hand-held!
Hold on! Don’t push the “Back” button on your browser just yet. With the right camera, the right technique, and a bit of practice you can do this.
In another recipe, I am explaining how to take HDR vertorama shots using a tripod.
The Camera
Which features does your camera need to have? I will go through these features in turn here.
An autobracketing function is a must! You need a camera that can automatically take 3 shots with varying shutter speeds that range from -2ev to +2ev. This enables you to click the shutter release once and make the required three exposures for each section of your vertorama. Each of these three-shot series covers the dynamic range more or less. If you don’t have an autobracketing function on your camera that can do three shots with +-2ev hand-held HDR shooting is virtually impossible!
Low-light performance is the key in this situation. You need a camera that produces useful images at ISO 1600 and above. You will have to set your camera to this high sensitivity to increase the shutter speeds such that you can shoot hand-held.
Speed is another important factor. Shooting a HDR vertorama may easily produce 15-20 individual photos that have to be shot in the shortest possible time. When I shoot 15 source images, this takes about 6 seconds. Having a high number of fps (frames per second) helps a lot here. Also the buffer size should be large enough to store these 15-20 RAW images. The buffer is a comparably small but fast memory built into in your camera. This buffer stores the images and from there, they are stored on your (comparably slow) memory card (SD, CF, etc). If your buffer is too small and can store, for example, only 5 RAW files, then your fps will drop after these ten shots to about 1 fps making it very difficult to complete the shoot. As a rule of thumb, if the buffer of your camera can store 10-15 RAW images, you are going to be fine in most situations.
From the statements above you can pretty quickly see that we are not talking about point-and-shoot type of cameras here. I am afraid, that you will need a DSLR here, and probably not one of the cheapest. I can recommend mid-range Nikon cameras like the D90 and the D7000 (anything above is even better, of course). These are great in high ISO settings and have sufficient buffer space (could be more though).
You have a Nikon D3x (just as an example)? Well, lucky you. You can’t do any better than this. BUT there is one caveat and one killer here.
The caveat: The higher the resolution of your source images, the more power you need for the post processing. With 18 21 mega pixel RAW files, producing the HDRs and stitching them is not going to be fun, and maybe your 3-year-old computer is even going to give up in the attempt for a lack of memory. In his case, you should reduce the resolution of the images before you proceed.
The killer: The D3x does up to 9 shots in autobracketing mode. Wow! But only with a max step of +-1ev. Mhhh.. Probably not enough to cover the dynamic range. This means that you would need to do 5 shots (+2, +1, 0, -1, -2). At the full-resolution frame rate of 5fps this takes a whole second and will produce considerable offset in the images and increased problems with moving objects. If you leave your D3x at home and take your D7000 (at a 1/4 of the price) instead, you will get three shots (+2, 0, -2) at 6fps in 0.5s. Perfect!
So here, bigger is not necessarily better!
The Lens
The prefered type of lens in the typical shooting situation (inside a church) is an ultra-wide-angle lens. There are two reasons for this:
- It is easier to shoot at low shutter speeds, and
- you can get close and still fit everything into the frame. Getting close is important as it minimizes the number of people between you and the subject, and you may simply not have the necessary room to go back another 20 meters.
A 10-20mm lens at 10mm focal length is my preferred weapon here. Using an 18mm lens you may have difficulties to get the right framing. Anything above that is not going to produce the desired result in most situations.
The Memory Card
Your intuition may already tell you that this kind of photography produces a hell of a lot of photos that have to be stored somewhere. Get a big memory card or even two! With a 12 mega pixel camera set to store highest-quality JPEGs and RAW images, it is no problem to fill a 16GB card in one day if you are shooting for HDR and HDR vertoramas/panoramas. Get a second one, just in case!
The Technique
Ok, you’ve got adequate gear. So, how do you use it in this situation? There are three main things you need to achieve:
- Avoiding camera shake: You need a steady hand to avoid blurred images at low shutter speeds and to produce images with minimal offset for the HDR production.
- Avoiding people in your shot: Having moving people (or objects) in your source shots can be a real problem for both, the HDR production and the stitching.
- Get the composition right: The composition of your final image exists only in your head. You need a bit of experience to know how the final result will look like when you take the source images in a particular way.
The Detailed Process
So, here is the recipe describing the workflow that I use most of the time. I assume that you are at the location and you have spotted something that could make a great vertorama:
Plan the final composition:
Think of how you would like the final image to look like. Which features would you like to emphasize? What is the best spot for the shoot? In many situations, symmetry is a very important factor. Shooting down the central aisle of a church with the altar area and the ceiling as the most interesting features is a typical example of this. Such a shot needs to be absolutely symmetrical. So, you need to stand in the center of the aisle as precisely as you can. Especially with an ultra wide-angle lens (e.g. 10mm) small deviations from this center are very hard or impossible to fix on the computer. Look through the view finder and make sure that all the central elements (the grooves on the floor, the altar, the central points of the arches in the ceiling etc.) are exactly aligned. There are, of course, also asymmetrical shots (e.g. the sample image at the start of this section). But if you decide for a symmetrical composition, make sure that it is precise. If you are sloppy here, the shot will be ruined.
Another really important point is that you need to leave enough space at the edges of the images for cropping. Why? Because the stitching process has to distort each image such that they fit together. As a result the stitched image will have uneven edges. As a rule of thumb, assume that about 10% on either side of the frame will be lost in cropping. It might be less in some situations, but you want to be on the safe side here. If you do not account for the cropping while shooting the source images, the shot will be ruined!
Preparing the shoot:
When you have a good idea of the final image and you found the right spot, you need to prepare your camera and yourself for the shooting.
- Put your camera in A-mode (aperture mode) and set it to its lowest aperture value. For most ultra wide-angle lenses this will be something like “3.5”. You need that to get the highest shutter speed possible. At 10mm, DoF should not be an issue as most of the image should be sufficiently sharp even at the widest aperture.
- You need to have a good idea about the lowest shutter speed you are able to hand-hold (and still produce sharp images). With a bit of practice, you should be able to hand-hold a 1/15s shot with a 10mm lens. Thus, your longest exposure should be above 1/15s. Point your camera at the section of your vertorama that has a medium brightness. If you are using a 3-shot (+2, 0, -2) setup, the shutter speed in your view finder after pressing the shutter release half-way through should read “60”(1/60s). If it’s less than 60, increase the ISO speed accordingly. Try to keep it at ISO1600 or below to avoid overly noisy images, unless your camera is a low-light monster.
- Activate the autobracketing function and set it to 3 shots with 2ev step width.
- Activate the RAW mode to make your camera produce RAW images. There are a number of reasons for doing this:
- It gives you more dynamic range than JPEGs.
- It allows you to fix chromatic aberration in post-processing more effectively.
- It allows you to set the white balance of all your shots to the same value in post-processing.
- It helps you produce the highest possible quality in all respects.
- Pick a spot and focus your camera. This spot should be at a medium distance (not the furthest point and not the closest one). This ensures that everything will be in focus.
- Take an autobracketed sample shot of the afore-mentioned section with medium brightness. Check the shots on the camera display.
- Are they blurry?
- Is the noise level acceptable?
- Are they well-exposed (not too much on the bright or on the dark side)?
- Ok, now you’re set!
Final preparation of your camera:
As the very last step before taking the shots, you need to prepare you camera:
- Point your camera again at the medium-brightness section and press the shutter button half-way through to focus.
- Without moving your camera, press the AE/AF-lock button. Any decent DSLR should have such a button. It locks the autofocus and the exposure settings. If you press and hold this button for the entire duration of your shoot, each of your shots will have the same focus and the same exposure values (shutter speeds). This is vital! If you don’t do that, you will have problems in post-processing. If your camera does not have such a button, you can also use manual mode and manual focus to achieve the same thing much less conveniently.
- Now your camera is ready and you are in the right spot! Wait until your view is clear of people (if at all possible).
Taking the shots:
Now it’s crunch time! It’s time to relax and take the shots.
Always shoot the source images from bottom to top. If you do it the other way around, you may find that people are moving through the frame in the very last shots of the entire series, and you will need to redo everything.
Take all the shots fluently without interruption in a fast series of moves from bottom to top (click-click-click – move/breathe – click-click-click – move/breathe – …). If you need to stop, wait, or take the camera off your eye, start over again! Otherwise, the conditions may change between two sections of the vertorama or you may produce an offset that becomes a problem in stitching.
Start with the bottom shot by pointing your camera straight down at your feet. End with a shot of the ceiling behind you. You may not need all of this in your final image, but cropping can be really cruel in taking parts of your image (also at the top and the bottom) that you may not want to crop. So it is wise to produce more material than you need in order to retain maximum flexibility later. The final composition will eventually be made on the computer screen.
Especially those last shots when you point your camera at the ceiling behind you can be tough. You are bent over backwards contracting muscles that have not been used before. In order not to produce too much camera shake here, it is important that you relax and do the shoot fluently not wasting any time.
Make sure there is enough overlap between the sections. At least a third of each section should overlap with the section before. A good technique to achieve this is the following: When you completed one section, do not move the camera yet. Fix a point in the frame with your eye that is about one third away from the upper edge of the frame. Now move the camera up while you keep your eye on that point until that very point reaches a position in the frame that is about one third away from the bottom edge. Take the shots for this new section and move to the next section in the same way.
After you have finished the entire series, don’t move your feet! Check the images on your camera’s display immediately. Here’s a check list:
- Any camera shake or blur especially in the longest exposures?
- Any moving people or objects that may become a problem? (sometimes you don’t see them when you shoot!)
- Is the noise level acceptable?
- In the entire series, is the exposure chosen adequately (brightest overall shot not too bright, darkest overall shot not too dark)?
- Is the overlap between sections sufficient?
If you are not sure about the quality, repeat the series (that’s why you should not move), possible with slightly adjusted settings. I always take more than one series of shots, sometimes three or even four.
Examples
You can find a number of examples where these techniques where applied in my Vertorama gallery.
Concluding Remarks
I know that sounds like an awful lot of information. But believe me, once you have done this a couple of times it becomes a natural thing.
Before you get to a location where it really counts and that you will not visit again in your life, go out and practice this whole thing. Go to your local church and take a few series of shots. It may not be the most beautiful location, but getting experience with shooting vertoramas and the processing later on your computer is vital. Check how different ways of shooting influence the final result. Get a feel for how much of the image gets lost by cropping.
Make sure you check out my recipe for Creating HDR Panoramas and Vertoramas in which I explain how you proceed at home on your computer once the source shots have been taken.
Checklist
To make your life a bit easier, here is a checklist for you to easily remember the most important stuff explained above.
The gear
- Camera
- autobracketing with 3 shots and a step width of +-2ev
- Good high-ISO performance
- High speed (fps)
- Sufficient buffer space for 10-15 RAW images
- Not too much resolution
- Lens
- The shorter the focal length the better (but no fisheye!)
- Get a big memory card
The preparation
- Plan the composition
- Beware of symmetry
- Leave sufficient space on the edges for cropping
- Prepare the camera
- Put in A-mode
- Check exposure to make sure that you do not fall below the lowest possible shutter speed
- Set ISO accordingly
- Produce RAW images as output
- Focus at mid-range
- Take sample shot and check
- Focus again, press AF/AE-lock button
- Wait for people to leave
Taking the shots
- Relax! Don’t cramp!
- From bottom (shot of your feet) to top (shot of ceiling behind you)
- Do a fast click-click-click – move/breathe – click-click-click – move/breathe – kind of motion
- Use the eye-technique to create enough overlap (et least one third)
- No interruptions! No waiting! No pausing! If interrupted, start over again!
- Stay where you are and check the shots!
- Do it again, just in case!
And now, go out there and have fun shooting HDR vertoramas!!!
Very informative, as is the rest of your cookbook. I have just one question though, how much overlap do you use between shots?
Hi,
thanks for the question. I actually mentioned that under “Taking the shots”. It should be at least one thrid of a frame. More can be better, but don’t go beyond 50%.
Cheers
Thank you very much for your tutorials.
You’re welcome, Jorge!
Just posted my first cathedral interior hdr vertorama using as much as I could remember from your excellent cookbook. Looks pretty cr@p compared to your work but I’m quite happy with it for a first attempt.
Klaus … I read the Vertorama Article at least 4 times. Excellent work … and again thanks so much for sharing you knowledge and expertise to use. It is very gracious of you to do so.
I have a few general questions:
1. Do you “always” shoot wide angle inside (like major buildings, like churches, halls, etc.)? BTW, I have a wide angle lens, but “never” use it. I certainly am going to try it.
2. Is there a time when “Non” Symmetrical vertorama pictures would work? If so … can you explain.
3. I am sure that you do not re-meter when you take the second, third, or fourth series of bracketed photo’s bottom to top).
The reason I ask is that after the first series of bracketed shots, I have to take my finger off the shutter button. When I do that, I loose the metering for the second and consecutive shots. Or … to you lock in the meter setting from the first series of bracketed shots on the “bottom” shots.
4. I am a acronym person. Its a way of me retaining tips like the checklist you mention in “taking the shots” section.
I like NOPE…S which means if I don’t follow it and someone ask me if I got the shot, I would say NOPE…S (if I missed one of the checklist_)
N = NOISE level acceptable?
O = OVERLAP sufficient?
P = PEOPLE is the picture?
E = EXPOSURE capture the scene
S = SHAKE of the camera – long exposures.
Just had to share … it may not mean much to anyone, but it works for me.
Hi Dan,
oh that was you viewing my blog! 😉
To your questions:
1. Yes, my 10-20mm Sigma is glued at 10mm as soon as I enter a room. In these occasions nothing beats a super wide angle. Even with an 18mm lens (27mm on my crop camera) which is already wide angle there are shots that you simply cannot make.
2. Asymmetrical compositions works very rarely as a vertorama. Nothing is perfectly symmetrical of course, but I am talking about e.g. a chruch being photographed as a vertorama while you are not standing dead center in the aisle. Try it! You’ll see that your brain (and mine) cannot handle it if it is not symmetrical.
3. Do you have an AF/AE-lock button on your camera? Most Nikons have one. Configure it in the menu such that it locks the autofocus and the metering. Measure the scene, then press the AF/AE-lock button and hold it through the entire series of shots. Then, it does not matter if you take your finger off the shutter release. The other way is to put your camera in manual mode.
4. Nice way of memorizing it. But you should look for acronyms that sound more positive. 😉
Cheers
Hey man, nice tutorials, highly appreciated!!!
I was wondering, I am the happy owner of a Nikon D300S, but it can only take autobracketing exposure at 1 EV intervals, so I would need 5 exposures of each frame to get a decent dinamyc range, so wanted to ask you which is the highest ISO value on this camera that could be acceptable to get nice results on print, and also do you find JPGs are OK in this case, or would I need to work with RAW files every time?
Thanks in advance, and kind regards pal!
Hi Francisco,
the D300s is a fine camera. It should be sufficient in most shooting situations. It should produce even cleaner images than the D90 (which I have experience with). Thus, I would say that you can go as high as ISO1600. With a bit of post-processing magic, it is possible to produce acceptable results for a print size up to A4. Of course all of this heavily depends on the shooting conditions and your hand-held shooting ability.
As for the discussion about JPG vs RAW: There is no reason not to shoot in RAW. Buy a big memory card or two and you’re fine. Configure your camera such that is produced JPG and RAW images. Then you can work with the JPGs if that’s sufficient and work with the RAW files if you need the extra quality and dynamic range.
Cheers
I just came back from a little field trip today and took a lot of bracketed -2,0+2 photo’s for HDR processing.
Something hit me today – something I didn’t think of.
I’ve always heard that RAW photo’s are flat and need processing to make them POP.
If, as I have been, take my OOC Raw photo’s into Photomatix and start processing them.
Shouldn’t I be making the RAW photos the best they could be “before” sending them to Photomatix.
I see things like black and white point – white balance as examples that need to be adjusted.
If I take them un-adjusted into Photomatix, won’t I be enhancing a flat image that will need these adjustments later.
I need some clarification.
Dan,
indeed, you should convert your RAW files using a RAW converter like Adobe Camera Raw and save them as TIFs before feeding those into Photomatix. If I remember correctly, this is also the advice of the guys at hdrsoft.com. Photomatix would convert them automatically, but this is not the best choice. Doing this manually gives you the most control over the process. Take a look at my recipe for “Correcting Chromatic Aberration” (https://farbspiel-photo.com/hdr-cookbook/correcting-chromatic-aberration) for more details.
Cheers
Excellent post. I have been shooting panoramas for quite a while now and I also wrote a post about panoramic photography – equirectangulars to be more precise. If you are interested you can find it here. I’m also a Nikon D7000 owner and shoot my panos with a Sigma 8mm fisheye.
http://experimentaldigitalphotography.wordpress.com/2011/02/06/equirectangular-panoramas/
I have some questions. First of all, which software are you using to stitch your panoramas? Secondly, do you shoot a full 360 panorama and then crop it or do you stitch the final image from several pictures.
As for HDR, I take my shots in RAW (usually 8) convert to 16bit TIFFS, stitch the panorama from them and then generate brackets in Lightroom and stack them in Photomatix instead of stacking every picture individually.
Anyway, thanks for sharing
Hi Grzegorz,
nice site! I like your panos.
Now to your questions:
I use Photoshop’s Photomerge. It does an amazing job of handling even problematic shots almost automatically. I have looked at the other programs (Hugin, PTGui, etc) but, honestly speaking, they are too complicated for my little brain. Moreover, PTGui has this paranoiac approach of stamping 1000 little watermarks into your picture making the preview completely useless. I used it for about 10 minutes before I deinstalled it again. At least that is how it was a year ago.
I do not produce 360° panos mainly for one reason: To really show their beauty, you need a player. I like images that unleash their whole beauty when printed. Old-fashioned? Maybe! I shoot 3-6 source shots and then stitch them.
Concerning your workflow: I think you lose a lot of tonal range first by taking only one RAW and second by converting it to a TIFF and then making the different exposures. This conversion impacts on the tonal range and on the quality. I always take at least three RAWs, convert them to TIFFs and create a 32-bit HDR from these.
Cheers and keep up the great work
Hi Klaus,
I follow your blog daily as well as Flickr, and now facebook. 🙂
I have just started processing my first Vertorama. I have processed each of the 5 EV individually (with the same settings) and now have 5 processed TIFFs. When I put them into photoshop (Automerge) I get a very short vertorama and the horizon is not straight.
My question is, do you just then transform that final image to stretch it and straighten the image and then crop?
I have thought about rotating the images 90 degrees and then stiching, assuming Photoshop is happier trying to make a Panorama instead of a Vertorama, and then retating back the final image.
Regards
Bruce
Hi Bruce,
good to have you on board! 🙂
In Photomerge, make sure you select “Cylindrical” and check the “lens distortion” and “vignetting options” (not sure if these are the exact labels but you should recognize them). Photomerge does not care if your source images are landscape or portrait. It should stitch them either way. You should rotate the image (if necessary) after the stitching is done. Then you should indeed apply the transformation tools to straighten and align the image. Make sure you have the grid turned on to help your with the exact alignment.
This should do it! I hope this helps. Otherwise, feel free to ask!
Cheers
___________________________
Visit me at facebook.com/farbspiel
Thanks for that Klaus, much appreciated.
Here is my final image, I hope you, the master, likes?
http://www.flickr.com/photos/benny450d/5837540242/
Regards
Bruce
hi,
Thanks for the tutorials. I’m bit disappointed as my nikon d3000 does not have any autobracketing. What technique i can do if i use manual bracketing ???
Hi Krishna,
you can bracket manually:
1. Put your camera on a sturdy tripod (mandatory for manual bracketing).
2. Put your camera in A-mode and take a shot. This is your 0EV shot.
3. Set the exposure compensation to +2EV and take a shot.
4. Set the exposure compensation to -2EV and take a shot.
This is the simplest way of doing it manually. Consult your manual to check how “exposure compensation” works on your camera.
Cheers
___________________________
Visit me at Google+ gplus.to/farbspiel
and facebook.com/farbspiel
Hi Klaus
You have said to take picture from bottom to top. You choose A-mode for autobracketing shot. Say, if you took 5 sets picture, exposure date of each set picture would not be same. Is it ok during stitching? My experience to take Panorama shot was manual exposure, f & shutter speed were same during shooting, then go to stitching.
Cheers
Sam
Really knowledgeable
Would the Tamron 11-16mm, f2.8 lens be better? You loose 1mm but gain in Aperture. Thanks Bill
sorry, I meant TOKINA 11-16mm,f2.8 lens not TAMRON. BILL
Hi Bill,
the Tamron will be fine. 1mm does make a difference, but not a huge one. Using something like a 16-85mm kit lens may not produce the desired results though.
Hi Klaus,
Can you please tell me what metering mode should be used? I shoot with Canon 450D and metering set to Spot. What would you suggest?
Thanks
Usman
Hi Usman,
for a HDR Vertorama shot, I use the matrix metering mode to measure the entire scene. If a test shot reveals that the photos are too dark or too bright, I compensate that with the exposure compensation. Then I use the same exposure throughout the entire set of photos by using the AE/AF-Lock button on my Nikon.
I hops that helps!
Cheers
Klaus
While I’m experienced with HDR and panoramas, I’ve never thought about trying a Vertorama, even if it’s such a simple idea! I’m going to try soon, your guide is quite essential but detailed at the same time