Surveillance video does not always have a frame rate that allows for continuous or near continuous recording of an event. Frame rates may be reduced to account for multiple camera feeds, rotating camera activation, reducing the amount of data that is captured, etc. When the frame rate is reduced, the viewer must be particularly careful when attempting to interpret the images. All we know for certain with a reduced frame rate is that the camera has recorded what it saw the moment the image was captured. We do not know what occurred between the captured images, so while each image may be accurate, the images collectively may portray a misleading view of what actually happened.
I will use this example to illustrate the point. When I teach for LEVA (Law Enforcement and Emergency Services Video Association International Inc.), and I talk about frame rate, I give an example of a frame rate of one image every five seconds. I stand at the doorway to the classroom and that is one image. The next image is of me at the lecture podium. We know for sure that I was at the doorway and then made it to the podium. During the interim walking period, I may have tripped, thrown something at a student (which I would never do just to be clear!), made a face or simply walked to the podium uneventfully. The point I make is that we do not know what happened between the images that were captured and that it is dangerous to draw conclusions with so much data missing.
It is not necessarily the accuracy of each individual image that is at issue but rather how the images are collectively interpreted (or misinterpreted). An untrained viewer may draw unsupported conclusions from video with a reduced frame rate and without expert assistance, the trier of fact may reach an unjust conclusion based on erroneous interpretation of the evidence. Many a prosecution has been premised on video evidence that the key parties do not properly understand.
An example of this problem is illustrated in State of Florida v. Muro, 29 October 2004, Broward County, Fla. No. 03-016981CF10A (17th Cir. 2004), wherein the defendant nanny was accused of child abuse. Suspecting abuse, the father of the child installed a “nanny camera,” consisting of a DVR with four cameras, for the purpose of monitoring the defendant’s interaction with his daughter. The cameras were motion activated and were capable of recording from all four cameras simultaneously, depending on where movement was detected. If only one camera was activated, the frame rate was 30 frames per second (fps). If two cameras were activated, the frame rate was 15 fps. If all four cameras were activated, the frame rate was 7.5 fps. State and defense experts testified that at a frame rate of less than 30 fps, the video would still be recorded in real-time but the appearance of the playback would be more “jumpy.”
Several months after the defendant was hired, the parents discovered a bruise on their daughter’s face. Suspecting that the defendant was responsible, they watched that day’s video and claimed that it showed the defendant shaking their daughter. When the investigator came to the family’s home to obtain the hard drive, the father refused to turn it over to the police claiming that it would show intimate activity between him and his wife. As a result, the investigator connected a video recorder to the DVR and copied various portions of the recorded images as pointed out by the parents, amounting to two hours out of a possible eighty hours. Since the parents continued to use the system, it recorded over the entire hard drive so that the only evidence remaining was that copied by the investigator.
In the suppression hearing, the defendant sought to suppress the video evidence for several reasons, the most cogent being a claim that her due process rights were violated as a result of the destruction of the original recording. It then fell to the State to prove that there was no prejudice to the defendant as a result of the loss of evidence and to the defendant to show that the original recording would have been more than potentially useful.
The Court held that the original recording would have been useful to the defense because it could have shone some light on the interpretation of the images given the reduced frame rate. It may have shown scenes involving the parents that were similar to the impugned scenes involving the defendant. Given the frame rate, the video may make it appear that the impugned conduct was abusive when in fact it was not. At the same frame rate, video of the parents with their daughter may also have appeared to be abusive conduct. If so, this would clearly have raised a reasonable doubt as to the defendant’s conduct being criminal in nature. Further, by destroying the original recording, the defense was prevented from determining the exact frame rate.
Since the video evidence was the only potential evidence of child abuse, the Court ruled that the State did not meet its burden in showing that there would be no prejudice to the defendant based on the destruction of the original images. The suppression motion was accordingly granted.
This ruling was appealed by the State to the District Court of Appeal, 24 August 2005, No. 4D04-4302 (Florida District Court of Appeals, Fourth District). On the issue of whether the original recording would have assisted the defense in determining the frame rate of the captured video images, the court ruled this to be too speculative. The Court ruled that at trial, the defense could call expert evidence regarding the interpretive value of the video images and that the jury would be entitled to consider all of the possible technical explanations for the impugned video that appears to show the defendant shaking the baby. Accordingly, the Court held that there was no due process violation and the trial court’s suppression order was reversed. It would then remain for counsel at trial to litigate the merits of the video images and for the jury to make a final decision. However, owing largely to the unreliable video images intended to be used by the prosecution in this case and the absence of other compelling evidence, there never was a trial as the State stayed the proceedings against the defendant in 2006. As a result, no court ruling was ever made on the reliability of the images. However, the defendant spent two and half years in pre-trial custody in part on the basis of key parties drawing premature conclusions of what occurred based on reduced frame rate video without the benefit of effective expert assistance.
The impact of frame rate on the admissibility of surveillance video evidence was at issue in the more recent Canadian case of R. v. Zoraik, 2012 CarswellBC 1858 (British Columbia Court of Appeal). The defendant, a lawyer, was alleged to have deposited an incriminating envelope on a certain desk in the courthouse. The relevant courthouse security camera made a 74 second revolution and recorded at 6 frames per second. In one image, the camera recorded a counter free of objects and then in the next image there appeared a white object on the counter. The defendant was standing near the counter in both images but no image showed him touching the white object. The trial judge ruled that while the rotation of the camera and the reduced frame rate limited the number of images of the room in question, that did not render the video inadmissible. Rather, this was a question of weight for the trier of fact to assess. The Court of Appeal agreed with the trial judge’s ruling in this regard, meaning that it was proper to admit the video evidence at trial, despite the reduced frame rate.
Concerns regarding the completeness or production of individual images affected by frame rate generally go to weight rather than admissibility absent an overriding FRE Rule 403 violation. The digital image evidence presented must be fair and accurate and its probative value must outweigh the prejudicial effect of the evidence. See for example Commonwealth v. Rogers, 459 Mass. 249, 945 N.E.2d 295 (2011) (Supreme Judicial Court of Massachusetts, Suffolk), wherein the defendant argued that reduced frame rate images were misleading. The appellate court ruled as follows on this point:
27. “When, as here, the demonstrative photograph is generated as a digital image or video image, the judge must determine whether the image fairly and accurately presents what it purports to be, whether it is relevant, and whether its probative value outweighs any prejudice to the other party.” Renzi v. Paredes, 452 Mass. 38, 52, 890 N.E.2d 806 (2008). “Concerns regarding the completeness or production of the image go to its weight and not its admissibility.”
It is reasonably safe to conclude that, properly argued, reduced frame rate video is indeed admissible subject to an assignment of weight to be afforded the evidence. It is imperative that those who make decisions on liability based on video images and those who tender video images in court seek out and have the benefit of expert assistance when interpreting what conclusions can fairly be drawn from the video images. The goal – always – is to find the truth.