In the last century, we placed a very high degree of trust in eyewitness identification believing it to be nearly infallible. That was, of course, until DNA evidence provided us with significant insight that shattered the illusion of the infallibility of trusting our own visual perceptions. In this century, as a society, we placed a significant amount of trust in our perception of video evidence. We often believe that what we see in the video is what actually happened in the event that was recorded. After all, when the ‘eyewitness’ is a video camera, we do not have to worry about the human error in the evidence presented … or do we?
There are complicated technical issues involved in the capture and production of digital video, including body worn camera (BWC) video, that present significant concerns when it comes to the use of the BWC video in determining the reasonableness of the degree of force used by an LEO when detaining or arresting a suspect. The purpose of this article is to make the reader aware of these technical issues so they can be properly considered when making decisions based on personal observations of BWC video.
One of the biggest problems with the use of any digital video, including BWC video, centers around the methods used to capture and store that video. When you take a still image with a digital camera, the light passes through the lens and then is digitized at a certain predetermined resolution. The higher the resolution (i.e. the number of pixels) the more detailed the image. Many people erroneously believe that when a digital video recorder captures an event, it simply takes repeated still images one after another, at a certain frame rate (measured in frames per second or FPS) and then when you watch the video you see all of these still images captured on every frame. This is not the case!
The digital data storage limitations of raw video make it very impractical for a digital video recorder to operate like a still image camera. Ever since analog video gave way to digital video, many different companies have endeavored to find a way to “compress” video to reduce the file size so that it can be easily saved and shared. The more the compression, the smaller the file size which means the longer you can record. All digital video cameras, including BWCs, use compression to reduce file sizes to manageable sizes. Indeed, an entire industry has developed based on proprietary programming algorithms that seek to drastically reduce file size while still essentially conveying the recorded event.
It is very important to acknowledge and understand that video compression is a compromise! It is not practical to record raw uncompressed video of a significant duration. Compression makes digital video recording practical, especially on smaller pieces of equipment with limited storage capability like a body worn camera. The compromise occurs in the way compression takes place. In order to accomplish this “compression” the compression program will often compare one frame to one before it (often called a reference frame) and “borrow” information from the previous frame that it deems has not changed significantly and then reuse the previous information instead of storing the raw data in the current frame. This means you are not actually seeing what happened … you are seeing what the program “predicts” happened. Nationally-renown forensic video analyst Grant Fredericks explains that “compression actually employs techniques that are designed to fool the human eye.”[i] The compression compromise means that there are numerous predictive frames that borrow information from a previous reference frame and ignore the true data received for that particular frame. In other words, what you see may not be what actually happened!
Another significant issue with compression is that often frames are deleted entirely as the refresh rate is extended to maintain the desired level of compression. When frames are deleted it creates a perception by the viewer that events happened more rapidly than they really did in reality. For example, frames that are deleted as a part of the compression process can create the perception that an officer is slamming or shoving a person when in reality the officer moved the suspect much more slowly.
It is important that we all understand how compression works and the limitations that it places on the use of BWC video. The average person is simply not qualified to review a video and determine where these compression issues exist. A forensic video analyst may be needed in order to determine whether compression techniques have created a distorted or misleading video rendition of a specific event.
[i] Florida Association of Police Attorneys (FAPA) Conference, Delray Beach, Florida 2017.