unthinking
photography

Twenty-Six days after the shooting of Trayvon Martin by George Zimmerman on 26th February 2012, a 3D animated reconstruction of the crime was released by News Direct on YouTube. News Direct creates animated clips of news stories to “fill in the gaps when video footage is missing.” Distributed by Reuters, their videos are produced by Next Media Animation, themselves notorious for absurd and satiric animations of the news.

Leading up to George Zimmerman’s trial in 2013, the defence attorney Mark O’Mara commissioned a 3D animated reconstruction from the company Contrast Forensics. Contrast Forensics is owned and operated by Daniel Schumaker, a Graphic Artist who specialises in crime reconstruction animations. The California based company carries the slogan ‘seeing is believing’.

The two animations are similar in aesthetic, yet produced under different circumstances and intentions. News Direct produces their animations in a few days, an impressive turn around time to research, storyboard, model, motion capture and render a 3D animation. Their visuals are defined by their rapid workflow. An in-house created asset library of characters, accessories and environments allows them to assemble scenes at a fast pace. Contrast Forensics’ animation on the other hand has been created with a strive for accuracy. Their visual outcome is assumedly determined by the available data: witness accounts, scene measurements, police reports.

What will follow is a partial transcript of the 11th trial day. On July 9, 2013 Daniel Schumaker of Contrast Forensics is questioned for two hours by defense attorney Mark O’Mara as to the technology and processes used in creating the animated crime reconstruction. The questions are posed to prove the accuracy and authenticity of the animation in order to have it pass as evidence.

The questioning offers a rare insight into the language used to explain capture technology as objective tools of measurement. The exchange is an important insight on the assessment of accurate reproduction. Accuracy is reduced to the technical process of data collection and translation of such into imagery. The word ‘body’ is used only a few times while what is discussed is the death of an unarmed teenager.

The following partial transcript especially highlights Schumaker’s explanation of the motion capture technology used to create the animated reconstruction. Motion capture allows the registration of body movement as data points and was used to virtually recreate the fight between George Zimmerman and Trayvon Martin. The technology depends on a body wearing sensors and reenacting the movements. Used in the entertainment industry to surpass the human body and inhabit any fantastical creature, in this case the technology was employed to ‘accurately capture movement'. Is this then the accuracy of the technology capturing the wearer’s body movements, or the wearer’s accuracy in reenacting Zimmerman’s and Martin’s movements?

ZIMM_640x410_7.5_FINAL.mov
State of Florida vs George Zimmerman
Trial Day 11 — July 9, 2013

"From Design to Forensics"

Mark O’Mara (MO'M): Please state your name Sir.
Daniel Schumaker (DS): Daniel Schumaker
MO’M: The name of your company?
DS: Contract Forensics.
MO’M: What type of business is that?
DS: Crime Scene Accident reconstruction. I take a lot of different information, I take the coroners report, accident/crime scene reports, photographs from the scene, I take my own measurements and my own photos, Total Station data which is a survey of the scene, I take all that information and put it in context in the reconstruction of a crime scene.
[…]
MO’M: So how long have you been doing this type of work in California?
DS: It’s been the last about thirteen years.
MO’M: Have you been doing that under Contrast Forensics or has the name of your company or consulting group changed at all?
DS: It’s gone from Contrast Design to Contrast Forensics.

“Software Behind Those Suits”

MO’M: So within a context of the Aras 360, gathering if you will the environment digitally, how do you then make that relevant to your presentation, how do you put people, things, events, pieces of evidence in context to what you have now gathered?
DS: It creates the scene to scale and then what you do with that is you can combine it with other pieces of equipment. I have a motion capture suit for gathering, it’s the most accurate way of gathering a person’s movements. It’s wireless, it has accelerometers. You put on this suit and then in real-time you can capture a person’s movements. So then I capture a person’s movements accurately and then I can bring those into the scene, so you have the accurate movements of the person and you have the accuracy of the scene.
MO’M: Could you give the court a couple of examples when you talk about the motion capture suits, is that what they are called?
DS: Yes.
MO’M: If you would give the court some examples of how you utilised that technology in other cases.
DS: You can use it a couple different ways. I use it for…Well about a month ago I had to go down to San Jose and there was a shooting there that occurred in a vehicle. The advantage of my suit over other suits is I can take it out to a crime scene, I can get into a vehicle and it sees all of the movements. A typical motion capture suit that a lot of people see on TV has little balls all around it and then there are high-speed cameras that will capture your movements. The problem with that is, you can’t take it anywhere outside of those cameras and if you get into a vehicle or behind a wall those parts of the person disappear because you can’t see the balls anymore. So that’s why I bought my suit, I can take it, it has accelerometers that measure your movements, there’s 16 of them.

MO’M: Could you explain to the court just a little bit more of the technology of what that is when you talk about an accelerometer and how that is better than just the visual balls that are sitting on a suit.
DS: It’s perfect for court because it gives you a running log of all the movements. For example, somebody hit something and causes a certain amount of damage. With this suit, I can do that same action and from the accelerometers you can see how much force it took to cause that amount of damage.
MO’M: To the extent that you can then advise the court, what software exists behind those suits that interprets the movements of the suits, that make that information relevant to your presentation?
DS: It’s just software that came with the suit. I don’t know the physics behind it but the engineers that have developed it, they not only use it for motion capture like this but it’s used in biometrics for physical rehabilitation and physical therapy and it monitors the person’s movement from day to day so they can see if there is any improvement.
MO’M: Let’s continue with that and advise the court if you would then what you were able to recreate and what data you used to recreate those movements firstly, and then we talk about possible changes to the movements later. So what were you able to do in that case as far as recreating it and what data did you use, what information did you use, to recreate it.
DS: In the truck…?
MO’M: Yes, in the car case.
DS: I had accurate measurements of the truck so I had recreated the truck and since I had the motion capture suit on interacting with the truck, my movements fit well within the truck. So then I take my movements, it exports a .bvh file which is a motion file, and then I bring that into a 3D program that I also have the truck in. I can just place that person…What you do is you map that person’s movements onto a figure that matches the person that was involved. You map it onto the person’s skeleton. So then his movements are all accurate, the size of the person is accurate.
MO’M: How do you get the size of the person to be accurate?
DS: From police reports or coroner’s reports or…
MO’M: So you would use height and weight and the software, or any function of the software, does it take that data put into the software and then it recreates a body in proper proportion to height and weight and what not?
DS: It doesn’t do any simulation like recreating like that. I just have a figure that I size to the size that I need and when I bring in the motion capture file you can specify what height the person is.
MO’M: And can the software do weight for example?
DS: Weight is more the appearance from photos.

“The Computer May Not Know It’s a Body”

MO’M: So in that animation, can you explain to the court, is there a name for this motion capture suit, this…
DS: It’s an Xsens, X-s-e-n-s, inertial motion capture suit.
MO’M: And you use that in the shooting … explain to the court how it benefitted you and in the presentation of the animation.
DS: What I did is I put on the suit and got into a vehicle that matched the vehicle that was used and recorded my movements getting out of the vehicle. And then I put it in the computer and in a figure that matched the victim, I put in a rod to match the bullet trajectory and then looked at his positions getting out of the car to see what positions the bullet trajectories would be and if it was consistent with the height and the position of the shooter.
MO’M: Now let’s talk about, you know, a bit more context. So when you get in the car and have the motion suit on and it’s communicating with a computer, correct?
DS: Correct.
MO’M: So the computer knows exactly where the body is in that it knows exactly where all 16 accelerometers are, correct?
DS: Yes. You calibrate it at the beginning of your recording.
MO’M: You calibrate for accuracy?
DS: Yes.
MO’M: So at that point you are doing movements but the computer is actually gathering data on 16 discreet points in space, that are moving throughout space, is that correct?
DS: Yes.
MO’M: The computer at that point may not know it’s a body, it’s keeping track of 16 data points, would that be accurate?
DS: Well in real-time on the laptop you can see a manikin. So if I move my arms up the manikin will follow all of my movements.
MO’M: Ok. Explain to the court the frailty of an animation because it is not the exact video of in this case the driver moving his arm one particular way but rather you deciding how the arms are to be moved and how the body moves. So how do you address that frailty of the animation?
DS: You look at certain things that you know from the discovery, like the bullet trajectory the opening of the door, and also the physical characteristics of the inside of the truck. And you look at different scenarios to find which one best matches the scene and the evidence and the discovery that you have.

“Other People in Suits”

MO’M: So what information did you then gather or was given to you to assist you in the work that you have done.
DS: The discovery from the district attorney’s office, photographs, coroner’s reports, police reports, audio from the 911 calls. There was also questioning at the police department of witnesses.
MO’M: Depositions?
DS: Depositions, yes. Also photographs they took of the scene and then also Total Station data and their measurements.
MO’M: Was there any information that you needed in order to create the animation or the graphics that were not available to you.
DS: Use of the motion capture suit at the scene. That wasn’t available to me. So I made two visits to the scene.
MO’M: So in effect you came back once you had the information and the foundation of the animation or recreation, you then came here to utilize the motion station, motion capture suits?
DS: Yes.
MO’M: Ok. Do you recall when that was done approximately.
DS: Second week of May.
MO’M: Explain to the court then what you do with them, who did it and how you assure accuracy in that regard.
DS: To ensure accuracy, that’s where it’s good that I can interact with the suits at the scene because if there’s a slope or other physical characteristics of the scene you catch that movement of the person that’s walking over the surface.
[…]
MO’M: So then what did you do with that information? We were talking now about the motion capture suits. Who did them and what frailties exist with having other people use the motion capture suits.
DS: There were a couple volunteers from the law office that put the suits on. Then as far as the timing for the second part of the scenario where they were on the ground, I used the audio from the 911 call for the timing of it, to time it with the shot of the gun.
MO’M: Did you also have available to you John Good’s police statements and I guess at that point deposition testimony regarding what he said he saw in what happened?
DS: Yes, that was used for their position, for the positioning.
MO’M: Did you also have information available to you from a Miss Mora, a woman who said she saw George Zimmerman on top of Trayvon Martin some seconds after the shot and did that assist you in the creation of your animation, in putting the bodies where other witnesses saw them.
DS: Yes.

Selective Resolution

The following day Judge Debra Nelson ruled to not allow the animation as evidence. It was however allowed as ‘demonstrative exhibit’ and the jury was able to watch it during the closing arguments. Dismissing the animation as evidence prevented the jury from being able to watch it repeatedly during their deliberation.

Judge Nelson questioned the animation’s accuracy due to the lack of witnesses and Trayvon Martin being depicted as right handed even though he had used his left hand to punch Zimmerman. Assistant State Attorney Richard Mantei referred to the animation as a ‘supposition inside a riddle wrapped in an enigma,’ pointing to the inaccurate light and the deliberate omission of the murder weapon.

Reading the transcript, it appears that the capture technology used for the recreation performed accurately. But what or who is performing the technology? Whose body is embodying Trayvon Martin? Whose body is embodying George Zimmerman? The use of motion capture technology is discussed largely by example of another case. After almost 1.5 hours of questioning O’Mara finally specifically asks how the technology was used in recreating the shooting on Trayvon Martin. When Schumaker mentions two volunteers wearing the motion capture suits, O’Mara does not inquire further. Questioning if a sensor is accurately capturing data is only relevant when there is a legitimate context and body. Why was it not questioned that a white volunteer embodied a Black teenager?

Technology as a tool of accurate measurement for reconstruction must be challenged beyond functionality. Specifically motion capture technology can be subjectively performed while still registering ‘accurate’ data. Captured data does not simply transform into an image. This was not the case with celluloid photography and it is not the case with today’s capture technology. Just as chemical processes affected what was visible in the resulting image, today calibration software, render algorithms and virtual illumination impact the visual result. Use of capture technology involves a level of selective resolution; an emphasis of certain elements over others. As the image production process isn’t a linear hermetic process, decisions are taken in it’s production that will be visible in the rendered outcome.

Further, as is the case with motion capture technology, who wears the suit and where the sensors are placed are crucial to the outcome. When I spoke to Xsens in 2016, they mentioned the ability to remove one of the sensors and attach it to someone or something else. The software still registers all 16 sensors, regardless of where they are attached. They mentioned this as a testament to their product as well as a creative possibility. The software will still read the data as part of ‘the body on screen’ but now suddenly one foot is across the room.

*This text is part of the research project Parametric Truth on the use of 3D animation technology in evidence production.

(The commission of this essay was generously supported by Molior London)