The ability to see what is happening during an experiment is often critical to human understanding. High and ultra-high speed cameras have for decades allowed scientists to see these extremely short time-scale events; starting with film cameras and now with digital versions of these cameras. The move to digital cameras has invited the use of computer analysis of the images for obtaining quantitative information well beyond the qualitative usefulness of merely being able to see the event. Digital image correlation (DIC) is one of these powerful and popular quantitative techniques, but by no means the only possible image analysis method. All of these analysis techniques ask more of the camera technology than simply providing images. They require high-quality images that are amenable to analysis and do not introduce error sources that compromise the data. Possible error sources include image noise, image distortions, synchronization and spatial sampling issues. As a minimal starting point, the introduced errors must be well understood in order to put error bounds on the results. This is because in many experiments some result is better than no result; with the caveat that the error sources and the relative confidence of the data are understood. The concepts will be framed in relation to ongoing ultra-high speed work being done at Sandia. A call and challenge will be given to begin thinking in more detail about how to successfully turn these cameras into diagnostic instruments.