Behind the Curtain | Advanced projection techniques downstage axis along with rotation of the scenic objects within the x, y, z coordinates. Note that the limitation here is not in the image processing electronics but in the stage design and corresponding scenic automation system, which may limit the range and form of movement. As stated earlier, the object to be tracked can be fl at or dimensional, although the dimensional objects will pose a greater challenge for the image processing system than a fl at surface. For clarity purposes, the following process is based on using a single fl at plane as the tracked projection surface. With a selected point on the scenic object designated as the image origin point (e.g. the lower stage right corner of our box) and the window in the projected raster sized correctly to match the size of the surface, the projection system is ready to receive positioning data from the moving scenery. This can happen in a couple of different ways as described in the following sections. accurate and timely, however, it may be possible to utilize the information for tracking purposes. including laser-based distance measuring devices, which can be adapted for data output. These devices have the advantage of being single components rather than a reader/marker pair and we have used them for some unusual applications, including a project that required tracking the movement of a building elevator. Tracking with external data Another method used to track scenery is to add devices to the objects that will provide exactly the kind of location data the image processing system requires. If the scenery and the design allows for the addition of the devices, this is an excellent option as the device can be tailored or programmed for the specifi c object and movement. One tried and true method is to place a shaft encoding system on the mechanism or motor that is moving the object, see Figure 3 . The small, patterned wheel is read by a sensor and calibrated for the speed and length of travel. The resulting data, which is output instantaneously, provides accurate and timely location information for image processing system. A similar system utilizing a marked tape strip, which is applied to the tracking surface along the path of travel, can also work but is not quite as reliable as the shaft encoding method. There are other more exotic methods, Tracking with video image data Another method for tracking movement, which does not require mounting equipment on the scenic object, is supported by using video image data derived from a video camera trained on the scenic object. Specialized image recognition software, typically developed for the specifi c application, is used to interpret the video image, recognize the object being tracked, Tracking with automation data Assuming the moving scenic object is not being pulled by a stagehand with a block and tackle, but is being actuated by a computer controlled winch system, it’s likely that the computer system has an idea of where this object is in the x, y landscape. If that’s the case, it may be possible to extract that location information and create a data stream that the image processing system can interpret. While seemingly convenient, this method presents some pitfalls. One is a timing issue that can occur if the automation system outputs data that translates as “I just moved the object from 100, 150 to 200, 250,” versus “I am moving the object from 100, 150 to 200, 250.” In the case of the former statement, the image processing system is receiving old information for an action that has already taken place. This means that the image window will follow the object but will lag behind, which will be very noticeable. If the automation data stream is Figure 3 – Image of scenic automation scenery with shaft encoder highlighted FALL 2011 33 PROTOCOL “ “ What makes this technique so interesting is that it requires a com-prehensive bridge between two very separate disciplines—scenic fabrica-tion and projection design . . .