Align View

[Emoke Csikos]

Align View is an option to set your ARCHICAD model rendering to an imported background photo. It enables the architect to automatically reconstruct the position of a camera in the interior or on the street from a picture taken of it. This feature can be used both for inserting buildings into an existing street or designing an interior.

How to align the camera view

Let's say you have a photograph of a building and its plan as an ARCHICAD file.

Step 1: Place the selected photo

Place the original photo on the Floor Plan of the building with Figure Tool. If the picture is placed without any rotation then the horizon is calculated automatically as horizontal. You can, of course, override this horizon calculation manually, and enter the horizon angle yourself; this is done by rotating the picture on the floor plan. When the picture is rotated with an angle other than 90º (or 180º or 270º) on the floor plan, then Align View calculates the horizon by itself, otherwise the horizon is assumed to be horizontal.

Step 2: Placing the Hotspots

You need to know the height values of two vertical lines (four points in all) on the building and their locations on the photo. On the screenshot below you can see two black hotspots on the Floor Plan and the heights of two distinct points of the building below each. Zoom in on the photo and place four hotspots at the locations of the four points with known height values.

After you have your hotspots on the picture, select the photo and choose Document/Creative Imaging/Align View... command.

Step 3: Linking the Hotspots

(Follow the instructions of ARCHICAD given in the bottom left corner.) Click with the cursor first on the point on the Floor Plan that corresponds to hotspots on the photo. (In the above example, the side of the window on the left where you see the black hotspot with two heights, +2.64 and -2.21.) Now click on the respective pair of hotspots on the photo, one after the other. Repeat this operation with the other point on the plan and the corresponding pair of hotspots. It does not matter which point on the Floor Plan you start with. In addition, after marking a point on the Floor Plan, you can select its two corresponding points on the photo in either order.

When correctly done, the result will look as shown above, and the Align View dialog will appear.

Step 4: The Align View Dialog

Type the height of the hotspots in the appropriate fields in the dialog box. The upper points are located along the same vertical line as the lower points. In this example, these will be 2.64 and -2.21 meters (left upper and lower) and 3.08 and -3.40 meters (right upper and lower). After you click OK, a new camera will be placed on the Floor Plan. Select the original scanned photo to be the background picture in the PhotoRendering Settings dialog. Make sure the width-height ratio of the resulting PhotoRendering is the same as that of the original photo. You can also modify the size of the photo to suit the size of your PhotoRendering window.

NOTE: you cannot crop the photo. If you still have to do so, do it in a way that the original proportions of the photo stay the same (i.e. cover a part of the photo with a solid white fill) because Align View needs the original proportions of the picture to locate the original target point of the photo. The view will not be aligned precisely if the proportions of the photo were changed. For example, if you have a scanned picture of 1200x900 and you want a 3D window or PhotoRendering window image of 600x450, you would reduce the size of the background photo to 50 percent in the PhotoRendering Settings dialog. You can choose this same background in the 3D Window Settings dialog. By selecting the created camera and generating a 3D View/PhotoRendering View you will see the result - the building is accurately placed in the existing environment.

Please not the following: You may have difficulties to reach an exact result if your image of the environment was taken with a device using a wideangle camera, or GoPro or it originates from Google street view or if it was taken inside of a building. As a rule of thumb please consider that the closer the image was taken to the building the higher the risk of inaccuracy is, because a small difference/inaccuracy generates a large difference in the result.

[ape2000]

Hello,

I have been trying this but I get the following message all the time:

'Unable to place Camera!' Check entered data or try with different points!

I have tried many different points and it keep showing the same thing

Would someone be able to help?

Regards

J

[Karl Ottenstein]

Yes, there are a bunch of old (very old) threads about this back from the days when we were so excited when this new feature was introduced. 🙂

The root of this and other problems with Align VIew comes from the fact that the Archicad "camera" is perfect - there is no camera lens in the world as perfect as a 3D CAD projected view.  So, the photo has to be made as close to perfect as possible....

The typical problem is the camera lens with which the photograph was taken. Lens distortion is most common (and can be extreme) with wide angle lenses.   You have a few choices.  The best is to take the photo with a good camera/lens combo - which does not include most cell phones.  Ideally about a 50mm to 85mm lens, but a telephoto can be OK.  Lens distortion is the least in the center of the frame, so take your photo in a way that you can crop all edges ... or that the model aligns with bits that are in the center portion of the image.

If you do not have the option of taking a better photograph, then take the image into Lightroom, Photoshop, or some other photo editor that has a lens distortion removal filter, distortion/alignment tools, etc to make walls perfectly vertical, etc.

In the end, I stopped using Align View because of the difficulties of getting a perfect alignment ... combined with the more important issue of getting the rendered model to have a believable atmospheric rendering for the tone of the model and image to make sense together.  Additionally, resolution/grain/pixelizaiton of the model and photo will differ greatly.   Thus, many of us chose for static presentation context images to get the closest alignment possible, but then to merge the model image with the photo in Photoshop where one or the other can be distorted (free transform tool) to fit perfectly, and with layer control, the model can be toned and feathered into the photo.

[Karl Ottenstein]

@ape2000 My comments above are almost identical to ones from 18 years ago here:

https://community.graphisoft.com/t5/Design-forum/align-view/m-p/84731?p=18095#18095

Use the search function of this Community and just enter your error phrase: unable to place camera

and yo