Installment 1: Photometry and Driver's Eye Video
Nov 14, 2016
Why/ what are the challenges in presenting this type of evidence in a valid manner?
If performing the recreation with sufficient fidelity to the real-world event was as easy as pressing “record” on the camera we have available, this discussion would have been closed permanently and 24 years ago I would have been involved in something entirely different. Instead the innovations and refinements continue unabated from various areas of science and forensics analysis, from psychologists, accident reconstructionist and vision sciences to color theorists, electrical engineers and signal processing professionals.
The challenges inherent in this task fall into a few major areas:
- Properly controlling for critical factors during the recreation; properly memorializing the recreation with video
- Properly calibrating the result to match the real-world data and presenting the data at trial in a manner that maintains this fidelity and does not introduce information not available to the original driver.
Note: I am using ”driver” as a convenience as well – in some cases the party whose view we are recreating is someone other than the driver, such as the pedestrian who did not see the oncoming vehicle.
Each of these areas involves many sub-factors that must be properly dealt with throughout the process in order to have the presentation shown to the jury and to survive cross-examination - such that the jury believes the presentation is valid. If this is done successfully, the impact on the jury in many cases is greater than any other type of demonstration they are likely to see at trial. This last point is important as being successful in showing the jury the presentation but failing to defend the foundation of your work during cross can be devastating. I have worked with many attorneys who have enough damaging facts gleaned from a deposition of the authoring expert to exclude the video but instead choose to score points during cross.
Throughout the history of this type of visualization, there have been various methods to capture and display an image for the jury that is substantially similar to what was seen during the reconstruction. These fall into two basic categories: those methods that attempt to match the final output to the “real world” view “by eye”; and those that attempt to use measurements taken at the scene to calibrate the final product. Done properly (and thoroughly) the method utilizing measurements of the actual scene should always trump any attempt to match the two “by eye”. In practice however, neither method guarantees either a false result or a valid one – they both can be used to good effect or result in a visualization before the jury that that is misleading and obfuscates as opposed to illuminate.
Given my background and preference for methods that can be mathematically validated, I advocate using real world measurements to calibrate this type of video evidence and will be focusing on this practice throughout the article. First, I will provide a little bit of background as to how each method is employed generally. The “by eye” method involves experts at the recreation, (typically the accident reconstructionist, video expert and human factors expert) making their own observations of the recreated scene, noting relative luminance between the subject and its background, areas of darker/lighter illumination, areas of high/low reflectance on the subject and other subjective values. These viewers then view the recreated scene through the lens/monitor of the recording device and adjust the recording parameters (aperture, shutter speed, gamma settings, ISO and similar) until the view on the monitor appears to be similar to the real-world scene.
There are several drawbacks to this method in general, including (but not limited to):
- The subjective nature of such observations; the different (and potentially sub-standard) visual capabilities of these observers.
- The small image size of the screen resulting in image scaling errors compared to the size of the real-world objects.
- The ambient lighting of the observers’ immediate environment.
- The adaptation that takes place in the eye/brain system while viewing a bright, backlit scene.
- The fact that the monitor that is being used at the scene will not likely be the one used in the lab to calibrate the video or the one used to display the result to the jury.
- The differences between this monitor and the one used in calibration/final display (often unknown differences).
- The fact that the monitor being used and the values being recorded are not the same, as the monitor has its own color and luminance curve separate from what the camera is recording.
- The lack of any true, objective memorialization that can reliably be used after the fact when it is time to edit the recording in the lab.
In essence, the by eye method asks the final viewer to trust that the expert’s recollection of what he/she saw at the recreation is valid and is maintained in the final presentation. These limitations on the demonstrable validity of this method have spurred many experts and forensic videographers to search for a methodology based more on real world measurements and less on subjective perception.
In the next installment, I will introduce the general characteristics of those methods that attempt to use specific photometric measurements of the recreation to memorialize and present a calibrated presentation. This type of mathematically validated methodology is the future of driver eye video presentations.
Look for our next installment in a couple of weeks.
Craig Fries is the Founder and President of Precision Simulations, Inc., the foremost forensic analysis and animation firm in the US. In this role he has created or directed over 1,000 3D forensic animations for use in criminal and civil litigation and has maintained a 100% record of non-exclusion at trial. As a pioneer in the field of forensic laser scanning, Craig created the first 3D animation based upon laser scan data ever admitted at trial in the US, and has been at the forefront of the use of the technology in homicide and accident investigation since.