Where, When, and Who: the starting point of visualization
I took this photo on September 2nd, 2015 at 10:00 am in the morning. When I got back to my computer a few minutes later, I realized that there was a lot more to it than just a decent image of an almost complete project. With a little help from the internet, I could reverse engineer this image into a rendering. Since the photo has a time and date stamp, I can set the sun in ARCHICAD to the exact time and date of the photo. And I can set the Project Location to the exact latitude and longitude of the site; there are a ton of websites which will tell you the coordinates of any location on Earth—you can even just right-click on a location in Google Maps. Remember that the latitude and longitude in Project Location are the Project Origin. So build your site with an intelligent relationship to the Project Origin and make sure that the Project Origin is the point you find on whatever mapping website you use. While you’re in the Project Location dialog box, set the altitude above sea level as well. This is also something that’s easy to track down via the magic of the internet. Finally, don’t forget to set Project North either dynamically on the plan or in the Project Location dialog box. Don’t mess up your visualizations by forgetting to put north in the correct direction! For more about the Project Location dialog box, click here. If you aren’t exporting this data for others to use and you can’t be bothered to find the exact numbers, you can select the location from a number of pre-defined cities under Project Location and also make a guesstimate of the elevation. But knowing how to find and include the correct numbers for a project’s site is a good skill to have and adds value to your BIM. Beyond setting the location and time based on a photo, there’s something else the photo can tell us—or at least remind us of: the height and location of the viewer. I’m short, and held my camera such that it was roughly five feet above the ground. I can set my Camera Z in ARCHICAD to match that (remember to set the Z relative to the ground below the camera location, not Project Zero). I can also move the camera in plan to roughly where I remember standing. The photo gives me clues, and if I wanted to be really obsessive, I could have measured my location when I took the photograph. That actually isn’t a bad idea. Imagine you are taking a photo of an existing site, before design begins. If you document your photograph locations properly, setting up analogous cameras in ARCHICAD will be that much easier. Having digital approximations of your physical camera locations will add value throughout the life of the project. You can return to those spots throughout the design process and document the evolution. Treating photograph locations as an extra data stream for your BIM is one more example of documenting both visible and invisible forces affecting a site.My rendering is mimicking my camera, which is my POV. But really I should be photographing and rendering the site from my client’s POV. My clients for this project are taller than me. So I should be setting the camera at eye level for a 6′-0″ person, not a 5′-6″ person who rounds his height up tot he nearest inch. It’s a subtle shift, but an important one. We remember it when we’re designing a space for a taller or shorter client—or one with mobility issues. And we should make sure to represent our spaces that way too. Getting a camera in ARCHICAD to match an exact photo is tricky. It’s more art than science. There is the Align View feature, but I’ve personally never had much success with it. If you are interested in it, here’s an old BIM Engine article on it and the entry on the Help Center. Align View is definitely more accurate, but perfect accuracy is not always the goal. For my purposes, a view that is very close is good enough. When I plugged in all the information that I knew about the site and the viewer and did a quick rendering of my design, here’s what I got: Not bad. But it’s not going to win any awards. And it feels a little cold. We’ve set up the where, when, and who of the image, but it still lacks some magic. There is more basic context missing that we can add: trees and their shadows. Additionally the sky and grass are a bit too bright. So let’s fix all that (and also tweak the camera angle):What a difference! Now this image still isn’t the best rendering any of us have seen this year. And there are some things that bother me about it. But other than the siding material (which I’ll explain in a follow up post), everything in this image is out of the box and can be done by anyone with ARCHICAD 18 or newer. The basketball hoop was created by turning the Basketball Court Object into a Morph, deleting everything I didn’t need, and updating the Surfaces. The net is actually the “Metal – Perforated Squares 18” Surface from the default Surface Library with the color tweaked and the sizing of the mesh decreased. All the trees are out of the box default trees from ARCHICAD plus one that I found on BIMcomponents.com. The trees you see are all flat 2D image trees that always face the camera. The trees that you can’t see are less convincing 3D trees, but their shadows are beautiful (they are the ones from BIMcomponents). While technically not OOTB, I’m counting the BIMcomponents trees as defaults because that website is integrated into the search box of the Object Tool, and as such we should treat content from that site as fully accessible defaults. I placed all the entourage of this photo (specifically the trees) quickly to approximate the photo I’d taken. Their locations were all about feel; I placed them to look right. I moved them in 3D and plan until the image clicked (thank you Nudge command!). I didn’t worry if they were too tall or too close or the wrong type of tree. They are approximations. They are sketches whose goal is to match the photo, which means their locations are fairly correct, and could be more exact if I took a few measurements at the site. I could also have grabbed a satellite image and used that to better approximate the tree locations. But matching the photo is good enough. It’s worth noting that I matched the photo in two ways: the height and relationship of the trees to each other and the garage, and also how their shadows fall on the site. I knew that this photo was taken on September 2nd at 10:00 am. So if that’s the time in my model, then the shadows should match. I kept moving and reshaping the trees (those that I see and don’t see) until the shadows looked correct. Here’s the plan showing the camera and the loosely placed trees. For this exercise I didn’t care what the plan looked like. But if I was doing this on a live project, I’d unify the graphics and put them on a hidden tree layer.Had I taken a good picture of the existing garage, I could have used those parameters to set up the rendering on day one of the project. And if I wanted better trees, I could have tried harder to work with the existing Objects or used custom images in them. It wouldn’t be hard to find tree images to use to improve the trees. I could even use images of the trees from the site but that’s overkill. Too much effort. Too much perfection (most of my decisions in this process were done to balance effort and value). The default tree Objects in ARCHICAD offer a nice range of visualization. When you change their proportions and mix from all the available images between the different Objects, you can create a pleasant diversity. Foliage changes and plants grow. The goal isn’t photo realism but contextual essence. By focusing on the feel more than the details I am approximating a sketch. No one would complain if a hand drawn tree didn’t match a photo perfectly. Instead of using a composite image, which isn’t hard to do with photoshop, I prefer this solution of semi-generic ARCHICAD trees. It keeps everything in the image consistent. It’s all a little cartoony—but in a good way. I’m not sure I hit the right sweet spot on the graph of the Uncanny Valley for this image, but it’s definitely more than adequate. Plus not only does all this effort apply to the rendering that mimics the photo, by using placed elements in ARCHICAD, all views work. Or most. And that’s pretty amazing. The rendering I’ve been discussing is good enough to approximate this one view. And that’s fine. But we need to remember that this effort gives us all views. I spent the time to set up the trees, the location, and the sun. And I got a nice rendering. Great. But I actually got an infinite amount of renderings. And videos. And sun studies and BIMx pro models and much more. This is the power of BIM. Here’s a quick example. The original rendering took some time to go from the first image to the second one. But the rendering below only took me the time required to move to a new spot in the 3D window and click render (plus the few minutes of rendering in the background while I continued to work).I had the benefit of doing these renderings post construction. But it didn’t need to be done that way. I could just as well have taken the photo of the existing garage and matched that imagine, setting up the trees and sun during the existing documentation phase. How would the design have evolved differently? Would I have noticed how the site reacts to the seasons? Working in BIM we claim more than two dimensions. We talk about 3D, 4D, 5D, 6D ad nauseam. When we look at our sites and document existing conditions, let’s remember to look beyond the basics. Let’s remember time. Let’s remember Winter, Spring, Summer, and Fall. We can document so much from being on site and gather much more from the internet. The power of BIM grows exponentially when we move form replicating one thing (a rendering that mimics a photo) to using one thing to inspire many (a photo as the starting point for a plethora of visualizations and experiential explorations). It took a few hours to render while I was sleeping, but here’s one more thing that came from that original photo and some extra data:
One final note. Another side benefit of creating renderings that mimic photographs is that they become easy comparison tools. Scroll back to the original photo. Now look at one of my renderings. Notice the differences in the architecture? There’s actually quiet a few. I wasn’t involved during CA, so the contractor made some judgement calls (that’s the polite way to say it, right?). Looking at these images it quickly becomes apparent what’s off. Imagine taking photos during construction and running renderings to match those shots. Even if you don’t get the view perfect, you’ll be close enough to make it very easy to catch deviations from the documentation. Remember, the renderings are the construction documents. The same data that got permitted is what we are seeing rendered in still images or videos. There is no interpretation, just ARCHICAD processing the information. The model does not lie. When I look at the construction documents, the photo, and the rendering, and there’s a discrepancy, I know which one is obviously incorrect. Using BIM makes before and after visualization mindbogglingly easy. We aren’t being fooled by good artistry or masking flourishes. We are seeing what is and what could be. We are judging based on information, not distraction.
Are you following GRAPHISOFT North America on Twitter? Click Here to keep track of all the latest ARCHICAD News in North America (and beyond).