JEFF SUWAY| HUMAN FACTORS

Lou sits down with Jeff Suway of JS Forensics to discuss how he got into recon, the current state of his firm, Jeff's real love: retroreflective tape, quantifying luminance with a camera, video analysis, and much more.

You can also find an audio only version on your favorite podcast platform.

A rough transcript can be found below.

Links from the Show:

Timeline of Topics:

00:00:00 – Jeff’s background and introduction to forensics

00:19:40 – JS Forensics: Jeff’s vision for the company and his current team

00:31:10 - Jeff’s fascination with retroreflective tape

00:44:00 – Quantifying luminance with a consumer grade camera

00:58:25 – Presenting in the courtroom

01:01:18 – How to take contrast measurements

01:18:45 – Nitere applications

01:27:03 – Necessities for a recon trying to do nighttime analysis

01:38:49 – Methods and tools Jeff’s experimenting with

01:41:58 – How Jeff stays in touch with the latest tech in the industry

01:46:03  – Rapid fire: What Jeff uses point clouds for, best investment under $5,000, and most used papers

Rough Transcript:
Please find a rough transcript of the show below. This transcript has not been thoroughly reviewed or edited, so some errors may be present.

Lou (00:00:00):

Good day, esteemed colleagues. Welcome back to the show. My guest today is Jeff Suway, proprietor of JS Forensics and collision reconstruction/human factors polymath. Jeff is on my short list of people to call if I need help deep diving some topic, whether it's EDR, video analysis, or conspicuity.

(00:00:18):

In today's episode we'll discuss how Jeff got into recon, the current state of his firm, Jeff's real love, retro reflective tape, quantifying luminance with a camera, video analysis, and much more. Without further delay, please enjoy my conversation with Mr. Jeff Suway.

(00:00:35):

As I was prepping for this podcast, I realized you and I first met in 2014 at ARC-CSI. I don't know if you remember that. I think we were both presenting. It was a decade ago, and I remember the year, because that's the year I moved out to California. I met you and I was like, "This dude is a stud. Seems to have a fetish for retro reflective tape, but an absolute stud." Where did that retro reflective interest come from?

Jeff (00:01:19):

Yeah, absolutely. I do remember that. I remember the year also, because I had moved out to California in 2013, and I had just been doing this retro work. I remember working, meeting you. I think you presented on your motorcycle sliding deceleration papers, which was really cool.

(00:01:38):

Same thing, "This dude is cool. He likes to do this work and he likes to do the research." Instant bond of, "This will be cool." The retro work came from casework. I think the same thing with your motorcycle slide paper where you didn't have a good answer for your casework and you couldn't find it published elsewhere.

(00:02:01):

The same thing with the retro where I really couldn't find the answers and the quantification that I wanted, and so I really went out and started trying to understand it for myself. Those presentations back then at ARC-CSI in 2014 were pretty basic, and the first couple papers were pretty basic as I didn't understand it the way I understand today, and so kept learning more.

(00:02:26):

One of the things I ended up doing is joining the ASTM committee, where I couldn't find published information and I just wanted to sit in the room with the people that are designing this tape, and just started asking questions. That for me was a game-changer to start to understand the science behind it and how it's manufactured and how it's measured and how to quantify it.

(00:02:49):

I know you call it a retro fetish, but it was kind of my little niche of, "I can add some data here." There's a hole in our industry where clearly manufacturing or anyone in that space understands how this tape functions, but I wasn't seeing the information out there in the casework and I wanted to fill that gap for me so that I could do better. But also, the goal of publishing was to share that with people.

Lou (00:03:15):

Your educational background is traditional engineering. You're mechanical, bachelor's, master's, civil, and now I'd say you have a reputation that leans heavily towards human factors. Is that how that started to develop?

Jeff (00:03:33):

Absolutely. Yeah. Starting to work with that human factors element. It was funny. I was thinking about my background like I generally do before a depo as I prepped for this just to have some things in my mind, but one of the things that I always leave out in depo that I thought was interesting is before I went to school for mechanical engineering, the two programs I was looking at was an optics program and mechanical engineering.

(00:04:01):

I somehow found a way to marry those interests in what I'm doing now where I'm looking at a lot of the physics of light, and it is physics. It is physics based. Again, I'm a nerdy engineer coming from it from that point of view. Kind of trying to marry that traditional engineering background with some of the optics or how the visual system works.

Lou (00:04:26):

Yeah. I think there's a lot of power in that analytically and for careers in this industry as well when you take one discipline and then an interest in another discipline and start tying those two things together in a unique way that nobody has really done.

(00:04:41):

That's what you've done with retro stuff, and we'll get back into that a little bit. I just wanted to kick it off making fun of you a little bit. Maybe part of our bond straight out of the gate was the East Coast thing. You started your career on the East Coast. Did you grow up on the East Coast?

Jeff (00:04:58):

I grew up outside Philly. Yeah. Then like you said, went to school on the East Coast and started working also on the East Coast. Lived in Maryland and came out to California, like I said, 2013.

Lou (00:05:10):

What drew you to engineering straight out of the gate? What was your intent as you went to pursue your bachelor's in mechanical?

Jeff (00:05:17):

I liked the school. I chose Bucknell is where I did my undergrad. I was choosing the school and they have a strong engineering program. I really like working with cars, and one of the things I was always terrified of is getting stuck at a cubicle. That wasn't going to work well for me.

(00:05:35):

I was thinking engine development lab or something where I can do something hands on, and I randomly had an internship with an expert, Ken Brown. He used to work for ARCCA and he started his own firm decades ago now out of Philly area. He was doing a lot of seat back failure for bio cases and he knew a neighbor of mine. They were high school friends or elementary school friends.

(00:06:04):

I got an internship and I got to see this forensic industry and was like, "Oh, this is really cool. I get to apply my engineering background and what I'm learning in school, but also I'm not stuck in a cubicle and I can go wrench on cars a little bit. This would be really awesome." That's kind of how I started focusing more into the forensic field.

Lou (00:06:25):

Yeah. That's very similar to my story. I had no idea that it existed, and then in undergrad, my thermodynamics professor, he was a young guy, really intelligent. Knew how to get through to a group of 18 to 20 year olds and would just kick off by telling stories of tanks exploding and car crashes and drive-by shootings and talk about his work.

(00:06:45):

I was like, "Man, that sounds super cool." It seems like a lot of us get in that way. Then you got just by a random introduction. That was undergrad and I suspect that influenced your decision to go to grad school and what you did in grad school.

Jeff (00:07:05):

Yeah. Maybe more of the story was at the end of that internship I was like, "This is great. I like this work and I could see myself enjoying this. How do I do this?" His response was, "Go work for NHTSA." That became my goal was to go work for NHTSA. Ultimately, they weren't necessarily hiring, but through a contract, I ended up working as a contract engineer for their rulemaking division. One of the projects I was working on was they're always tasked with getting rid of regulations at some point or making sure they're not regulating unnecessarily. Through that, I started doing a lot of analysis of the NASS CDS system, the real world crash database that NHTSA has that's publicly available.

(00:07:52):

I'm starting to ask questions and trying to figure that out, and the person that I was working closely with, she says, "Hey, I'm lecturing at GW next week on this exact topic. Do you want to come and see my lecture?" "Oh, yeah. That sounds super cool." Then I'm sitting in class and I'm realizing there is a master's level class dealing with crash test analysis and statistical analysis of real world crash data.

(00:08:18):

"This is a whole master's program? This is really cool." Then that's when the goal was, "Okay, I'm going to go to GW in their transportation safety master's." It was under civil, because it started as roadway furniture and DOD barriers and things like that, but there's a lot of mechanical in there.

(00:08:40):

In my Civil masters I did LS-DYNA work, we did MADYMO work. Like I said, Finite Element, and we're running real world testing of a vehicle into a wall for DOD. There's a lot of mechanical aspects to that. Through GW, again, I had a master's level class in accident reconstruction. There's not very many programs in the country. Maybe Northwestern. It's like, "This is the program I want to go to. No question, I want to go to GW," was at the time.

Lou (00:09:21):

We were probably looking around the same time. I started grad school in '07, and one of the big reasons that I chose the school I did, Worcester Polytechnic Institute, was because they had an NHTSA contract that was headed up by Malcolm Ray. He was in the civil engineering department.

(00:09:39):

I managed to still get a mechanical engineering degree, but a lot of work was with the civil department at WPI, and they let me tailor my master's toward reconstruction type stuff. Roadside safety, impact mechanics materials, fire dynamics. But I could never find a school that had traditional forensic engineering. I think there was one in England at the time, but I never caught an eye to that recon class. Who was that taught by and do you remember what the topics were?

Jeff (00:10:12):

I would have to go back. The professor, and I forget his name and that's terrible, but he did a lot of work on rollovers, like the Jordan Rollover System, the JRS out in California. He did a lot of development work in that and a lot of testing.

(00:10:33):

I knew him from him coming to NHTSA and saying, "Hey, your rollover tests aren't great." Trying literally a little scale model that he brought, rolling a vehicle over dynamically literally on the desk at NHTSA. The funny story with that class is it was advertised, but it would never go.

Lou (00:10:56):

They couldn't fill it up?

Jeff (00:10:57):

They couldn't fill it up. Everyone except for me was there as a research student on a stipend. You get paid and your classes are free to do research through the school on NHTSA contracts. I basically said, "Hey, look, if you take the class, they won't charge you and it won't count against you if you drop the class in the second week. You can't drop the class first week, because then they will kick me out also, but I'm begging you." I was driving from SEA at the time an hour and a half to go to school.

Lou (00:11:34):

Oh, you already had the SEA gig?

Jeff (00:11:36):

Yeah. At this time I was working at SEA and I didn't have time to get dinner, and so I would stop at Chipotle every class and I would literally be eating my burrito in class. Eventually everyone would be giving me money. I'd pick burritos up for everyone. I said, "Look, I will pick up burritos free all semester if you sign up for this class." Everyone signed up, we got enough people. They got the professor to go. We're going to go and they all tried to drop it and the school was pissed. The school, "No way." Then everyone got forced to take this recon class with me. Sean Haight, which I know you know him. He was also in that class. He wanted that recon class, but you can't have a class for two. I worked the system to get that class is what happened, and hopefully free burritos made up for it for a couple of the other students.

Lou (00:12:34):

Yeah. Man, I'd go to that class obviously willingly. It's funny that it was hard to fill, because in my physics class, I remember physics 101 in undergrad. When the professor would talk about using momentum to analyze a car crash, everybody would perk up a little bit and found it interesting. But I guess at that stage it's a little different. I never knew that you went to school with Sean. That's cool. Man, it's crazy how quickly you got into recon and that's a bit of a rarity and that set you up to have a lot of early success.

Jeff (00:13:08):

Yeah. Meeting Sean was a big deal for me coming out to California, too. When I started working at SEA and I was going to GW, and Sean was one of the only people in that program that had some recon background already. Literally Rusty has been showing pictures of Sean in diapers at crash tests at every one of his classes I've been to.

(00:13:36):

Sean and I instantly hit it off, too, of like, "Oh, we both crashing cars. This is cool." He told me. He's like, "You got to come to ARC-CSI. You got to come to ARC-CSI." I had no idea what it was. But the same thing, just knowing Sean and being able to meet the people that work out in California like Judd Welcher and the BRT people, that really helped me move out to California pretty easily and work with a great firm. That was super cool.

Lou (00:14:06):

What year did you first go to ARC-CSI? Are we talking '05, '07 area?

Jeff (00:14:10):

No. It would've been maybe 2010, 2011 kind of timeframe, and then maybe 2012 was when I started doing more with the BRT crew and Judd Welcher and all those guys, and really kind of met them and knew them better. Then like I said, August of 2013, I came out to California.

Lou (00:14:33):

The NHTSA contract work that you were doing, was that simultaneous to your master's?

Jeff (00:14:39):

Yeah. I was working at NHTSA doing the master's, and then at some point I switched and I was working at SEA and still doing the master's. One class at a time, it takes a lot longer when you're working 40, 50, 60 hours a week. Doing one class is a lot especially. It kind of spanned two jobs.

Lou (00:15:02):

It makes it easier to come out without debt, but that's got to be pretty grueling.

Jeff (00:15:06):

Right. But worked out, right?

Lou (00:15:09):

Yeah, exactly. It worked out. That's how you met the BRT guys. You're at SEA. I imagine you get that job. Did you just knock on their door or did they have a posting?

Jeff (00:15:19):

They had a posting. Initially, I didn't meet any of the qualifications. They wanted someone already with a master's, already with a PE. I called and said, "Hey, I'm not eligible for a PE yet. I'm in my master's. Can we make this work?" They basically said no, and I just kept calling until they finally gave me an interview.

(00:15:38):

I know more than I should, but basically the person they gave the offer to didn't work out and I was their second choice. It again worked out, and that's how I met Tony Cornetto. I was working in his office in that office with him and a bunch of other recons.

(00:15:57):

Tony and I do a lot of collaborations still, which has been really cool. I know that you talked about this in your last podcast series and I think your WREX keynote was about just collaborating in the industry and how valuable that is. I think you're kind of hearing that from what I'm saying today, too, where just collaborating with people like Sean and Tony and yourself has been really cool. Really cool and really helpful. Instrumental in my career and ability to do what I'm doing.

Lou (00:16:32):

Yeah. Same here. I was going to mention when we were talking about Sean, he's now what, head of crash worthiness at Tesla or something ridiculous like that.

Jeff (00:16:39):

Yeah. I don't know his exact title, but I know he does a lot of their crash worthiness testing. I know he's at CARCO a bunch, and now in Germany doing stuff out there, too. Running crash tests and looking at their safety systems. It's really, really cool.

Lou (00:16:54):

Yeah. That's impressive. You met the BRT guys through ARC-CSI. Obviously, they're very easy to get along with. You got probably back then, Isaac Ikram, Tom Szabo, Judd Welcher, and they just kind of probably instantly through their personality, and the work that they've produced over the prior two to three decades is very enticing. What was it like meeting those guys and how did you end up deciding to come out and start working with them?

Jeff (00:17:27):

Yeah. That was super cool. Like you said, super easy to get along with those guys. Brian Randall. Chris Furbish, I think, had been hired by those guys maybe six months or so before I started. He just kind of started there when I was coming in. If you looked at any SEA, maybe not any, but at least coming out of the Baltimore office, any low speed rear end kind of paper that SEA was writing or report, nine out of the 10 references was Welcher, Szabo or Szabo, Welcher. Those were the low speed bio guys. I incorrectly thought, "Well, with my background at NHTSA and understanding injury criteria and those things that I can come out here and learn from the BRT guys and work in bio." I quickly came and realized, "No. This is not the world I need to play in. I'm not qualified for this. I would need to go back to school to really do something in the bio world."

Lou (00:18:27):

Yeah. They're playing in a big sandbox, too, where you're going to get beat up if you don't have exactly the right pedigree.

Jeff (00:18:35):

Right. Again, that was a quick learning for me of like, "Oh, I'm not going to do bio. That is not in my future." But this night visibility or retro reflective tape, my retro fetish as you called it, that was something I could pursue. Like I said, there was a hole. Judd was great letting me pursue that and supporting me through that, which was really cool.

Lou (00:19:00):

Yeah. Ned and you started JS Forensics. You were at Mechanica for a little stint and started JS Forensics, which now you're in the building. We were joking before recording. You're in your new office. It still feels new to you even though you've been there for several months, which just speaks to how busy you are.

(00:19:18):

I was commiserating, because I've been in this office for a year, and this is the first podcast I've done, because this is the first time my studio has actually been ready to record. It's tough to find the time to start hammering nails into walls and hanging pictures when you're as busy as we are, but I guess we should hire help, but we're not great at that either.

(00:19:41):

What was your vision for JS when you started that? First of all, I guess, you're in Long Beach at that point, or were you already living in Carlsbad? How was the move and why did you decide to set up where you did, and then what was your longterm vision for the firm?

Jeff (00:19:57):

When I moved in 2013, I moved to Carlsbad. At the time my wife was going to school in San Diego, and that was kind of the reason I was down here, but I was commuting to Long Beach, which is about 160 mile round trip commute every day.

Lou (00:20:16):

Brutal.

Jeff (00:20:16):

Right. Again, I wanted to do that. I wanted to work with Judd, I wanted to work with BRT, and I don't regret that at all. That was fantastic time and I really liked those guys. I had a kid and he starts turning about one and having some health issues, and commuting didn't really work anymore. Three hours in the car was not on the table for me.

Lou (00:20:42):

Yeah. That's a surefire way to get into some trouble on many levels.

Jeff (00:20:46):

Right. Even again, my wife at the time in school or seeing patients. She's a clinical psychologist. Get a phone call from daycare. "Hey, you got to pick up Jacob." "Okay. I'll be there in two and a half hours." It doesn't really work so well.

Lou (00:21:06):

"If SoCal traffic cooperates, I'll be there in two hours."

Jeff (00:21:10):

Right. Exactly. That was more of the push of staying in Carlsbad. I don't know. I like this area. The LA area seems too busy. I like the slower beach towns. I know we're kind of fast paced in our industry, but it feels like a slower beach town down here, and that was kind of the goal of living in this area.

Lou (00:21:33):

Yeah. I'm with you. I was in West LA, as you know, with Dial Engineering. Very similar thing where I was just like, "I love Eric Deyerl. I love what Dial Engineering does, but I need to get out of LA proper and start to live a slower life." That's why I started my firm. What was the vision for JS? Multi-disciplinarian, or were you trying to focus on one thing?

Jeff (00:22:01):

I always have done recon and the human factors type work. I wasn't really aiming to focus on one or the other. I think we have a pretty good split between the two still. I need to learn from Lou Peck and learn to say no to cases. I'm not as good as you with that. I need to get better.

Lou (00:22:24):

It was a muscle that took a lot of developing. Years and years.

Jeff (00:22:28):

The vision was just running it the way I wanted to run it and not worry about commuting. It just so happened that we were able to grow. Like I said, I've had trouble saying no, and if I'm not going to say no, there's only so many hours, and just started hiring people that could help and fill that time that I'm not available.

Lou (00:22:54):

Yeah. If you're not going to say no to cases, you have to say yes to hiring. You can only have it one way.

Jeff (00:23:00):

Right. Exactly. Luckily have some really good people across the board that are working with us, and that's been really just really, really nice.

Lou (00:23:11):

Yeah. What's your team look like right now?

Jeff (00:23:13):

We're now in Encinitas. We moved. We're now in Encinitas, which is right down the street from Carlsbad. In actually January of this year, Eric Hunter came down from Seattle, Washington. He does a lot with teaching at HV Forum if you've done that. You've probably seen him teach there.

Lou (00:23:31):

That's a great get. I've always loved Eric.

Jeff (00:23:34):

Yeah. A super nice guy, great recon, super great in HV, obviously. He's fantastic. Hit the ground running. Initially I started trying to email people like, "Hey, we've got this new guy." He's too busy from cases in Seattle still. I don't know. Doesn't really seem to make sense to try and get him busier. He's just busy.

(00:23:57):

We also last year got a guy from [inaudible 00:24:01] from Drexel, Ziad. He's a mechanical engineer and just learning this and getting up to speed on everything. He's been really helpful and fantastic. Then I know you know Justin Ngo. He was at Mecanica for a while before. Same thing, mechanical engineer. He's kind of picking up his own caseload, which has been really cool. Still helping me a lot, which is great. I see in the next couple of years, he's not going to have time for me anymore.

Lou (00:24:30):

Yeah. That doesn't shock me. I don't know Ziad. Is that how you pronounce it?

Jeff (00:24:35):

Yeah, Ziad.

Lou (00:24:36):

I don't know Ziad very well, just things that you've told me, but I've seen Justin around for years and he took one of my photogrammetry classes and I've had some conversations with him. I'm just like, "That guy is going places." There's no doubt he's going to be one of the superstars in this industry over the next 10 to 20 years.

Jeff (00:24:52):

Yeah. Absolutely, absolutely. Then you also know Henry Vega. He's out in Chicago. That was really cool. I wasn't really looking for an office somewhere else, but really looking for someone that knows what they're doing and cares and can help, and I didn't care where they were.

(00:25:11):

He was looking for something similar, and it just happened to work out where he can come on board. I know he's kind of out there on his own, but hopefully with all the Google Meets or whatever that we do to kind of interact with him daily and weekly, it feels like he's still part of us here. He's a good guy.

Lou (00:25:32):

He seems like the type that is probably well suited though for being out there on his own, because he's doing all of this kind of unique exploratory work that I suspect solitude is a benefit for. We'll talk about that later. I don't know how much you want to give away, but Henry has been developing some crazy techniques with audio that I think is going to blow the whole industry away and change a lot of the way we do some stuff. Are you growing still? It sounds like, I thought that Eric was a hire to help you deal with your work, but I'm just finding out right now that that's not the case. Are you still on the hunt for additional engineers?

Jeff (00:26:17):

It's always about the right person. Well, that's not true. There's been times where it's like, "Oh, we need help. We need help," and specifically targeted people for help. Eric was just, he wanted to move down here, and so I think it was kind of beneficial for both of us where he wanted to move down here.

(00:26:37):

He wanted to be part of a group where he was kind of on his own. It wasn't that we needed the help or needed the work, but there was certainly the room for that. Again, he's kind of coming with a full caseload, so that was a really easy yes. Like you said, he's an awesome guy and a good recon.

(00:26:59):

That wasn't me looking. It was more of mutually discussing like, "Hey, would this work out?" I wouldn't say we're actively looking. Again, if the right person comes along, I wouldn't say no to that at all. There's a lot of work in Southern California. I think everyone out here is kind of just getting slammed. Again, you're better at saying no. I'm just working harder is my solution, which doesn't work super well long term.

Lou (00:27:31):

I was going to ask you about that. Getting here, my ass in this seat for 10:00 AM this morning was a serious challenge, just because I was doing testing down in San Diego yesterday, way out in the mountains. I'm schlepping this motorcycle around and unloading it in the wee hours of the night last night, and still trying to get a workout in this morning and talk with my kids and see how their day is going, and back to school night last night, getting all that together.

(00:27:58):

I am saying no a tremendous amount to cases. The vast majority of case calls come in, I don't accept. How are you doing, I guess, is the best way to say it. How's it working out, and how much are you working, and do you plan on making any changes or are you just going to kind of keep trudging along and hire when the opportunity arises?

Jeff (00:28:25):

I have been actively pushing cases more towards Justin. He's more than ready. He's more than ready to take the case and run. Quite frankly, he's doing that already on my cases. If I can't convince him, I'm getting more aggressive about, "I'm not your guy. If you want us to work on it, I'm more than happy to overlook it if you're nervous, but I'm not your guy. I'm not going to be the one sitting in the hot seat on this one."

(00:28:59):

Trying to just be firmer on instead of like, "Hey, I think this would be better for." It's like, "Nope. It's not me. If you want us in general, it's going to be someone else." That's been the change I've been making more recently, and hopefully Justin is okay with that. At least that's my understanding is he's good with that.

Lou (00:29:23):

That is what I did, too, when Ben was here and I was actively trying to grow. They would call for me and I would just say, "I can't take it, but I've got a guy here who's great, and I will be able to, like you said, look over the case and everything." That worked really well.

(00:29:40):

Most people want to keep it with your firm, because they trust while they don't get your presence on the case, they get your standards and they get your oversight, so there's a lot of benefit to that. How many hours a week do you find yourself working right now? Are you staying sane?

Jeff (00:29:56):

Yeah. I've also been better at picking my kid up, dropping him off, going to the gym in the morning. I'm doing better. I'm doing better.

(00:30:03):

Dropped them off, going to the gym in the morning. I'm doing better. I'm doing better at that. A vacation here or there, at least a day off. But yeah, I mean it's probably, I'm not billing, but it's probably somewhere in the 40 to 60 hour range. Still trying to make it closer to the 40-hour range, especially my kid, he just turned eight, he's in second grade. At some point he's going to leave the house and trying to capitalize on that while he's still here and still wants to hang out with me, right? He's only got a couple more years before he is like, get out of here. I got friends or whatever.

Lou (00:30:38):

Yeah, we're in the same boat there. My kids are 11, so I've got, I mean probably they have a middle school dance tonight, so I'm not really going to see them today at all. So it's starting and they're going to want to hang out with their friends a lot more. So I figure I got a couple more years where I'm hanging out with them a ton and then it's going to start to diminish from there. So I'm trying to be very conscious of that while also growing Lightpoint. I'm not actively trying to grow Axiom, but well, growing Lightpoint, which takes a lot.

(00:31:14):

Okay, so let's start getting into the human factors stuff. I meant to mention this at the top for those that aren't intimately familiar with you and your work, I love having you in my Rolodex because if I have a question about really anything like high techie and state-of-the-art, I can call you up and you will have an idea of what's going on. I'm like, okay, how can you know about HVEDR, CDR, Kia, Hyundai, human factors, general recon, simulation, renderings, animations, I put you and Tony in that same category where you know an unbelievable amount about an unbelievable amount. So I don't want to take away from the fact when I start talking about all this human factors stuff, I don't want anybody to start thinking that you're just a human factors guy and you don't do recon too. You are a badass recon and also a badass human factors guy.

(00:32:14):

So looking at your CV though, I counted 41 references to Retroreflective tape. Okay, so this is where the fetish idea came up with. I was like, this guy loves retro tape, but I think we've all had those cases, which got you started down that path, like tractor trailer stopped on the shoulder and gets rear ended and they're trying to sue the tractor trailer company. So all of us eventually get exposed to that retro tape. How should we be looking at... What are the first steps essentially in documenting the retro tape and making sure that we have everything accounted for if it gets to the point where that's going to be attacked.

Jeff (00:32:54):

And that's one of the things that I spent a lot of time trying to figure out and learn about was everyone can go up to the trailer and look at it and some tape looks real bad and worn and some tape looks real good and new and the tape that looks new is going to be just terrible at night and the tape that looks worn actually performs really well and that's not always the case. Some of the terrible looking tape is really terrible at night and some of the good tape is really good, but I've gone on Amazon, you buy whatever cheap tape you can find and straight out of the box, it's not even meeting new standards. Normally there's a factor of safety in this stuff and some of this stuff instead of white is supposed to be 250, let's just say, some of it's in the 800, 900 range, huge factors of safety, and I've bought brand new tape where it's like 150, not 250, which is really scary and terrible that you could try and buy new tape and it looks good.

(00:34:02):

And so trying to understand what this means and how to use it at night. Same thing with color. Sometimes the night color is different. One, it's a different light spectrum that we're looking at. You're using headlights, which is different than the sun, but two, it's the way that that tape is functioning and it's reflecting the light back that's going to change that color. And so there's just so many nuances in the tape. Basically the end result or the real answer there is there's equipment you can buy that'll measure it. And that's what I like to do when we can is actually inspect the trailer whenever we can and physically measure it and the equipment's going to shine light on the tape measure how much comes back at different angles that we care about and some of the equipment will also give you a color and tell you, is this red really red?

(00:34:55):

Is it going to look red at night? 'Cause it looks red in the daytime, but that doesn't mean it's red at night. So measuring that and understanding that is important. There are other ways to do it. Obviously you can put that same tractor trailer at night with your car on the roadway our exemplar car and headlights shining at it, quantify it that way, but just looking at it in the daytime is really not sufficient. At least at a minimum, I like to take a daytime photo with a flash on 'cause you'll at least see if it's reflecting. You don't know how well it's reflecting, but if nothing comes back, you know it's not. You know you're in trouble. If something comes back, oh, maybe it's worthwhile measuring and actually getting a value. But yeah, there's Road Vista makes it some equipment, Delta makes some equipment. It's all basically for sign sheeting, but that's essentially what the retro reflective tape like that DOTC to tape and white tape on your trailer. It's essentially sign sheeting that's applied to the trailer.

Lou (00:36:01):

So what do those tools look like? Is this a test that has to be performed at night or does it create its own shade by just applying it directly to the tape? What's that process?

Jeff (00:36:10):

You don't need your own shade. You don't need to do it at night. I'm laughing now because I'm remembering a story of an attorney that hired me that refused to let me inspect the trailer at daytime 'cause I wanted to measure the tape. That was the big issue in the case. And I'm trying to explain to him, and I just kept sending him emails. I was like, "I don't care if the other side sees it. I don't need to see this trailer at night just to measure the tape." Anyway, part of it is the way that the tape is functioning. The entrance angle is the term that would match it is the angle that the light is entering the tape in relation to the light that you're viewing the tape from.

(00:36:56):

And so, sorry, observation angle. I don't know what I said. Observation angle. And so if that observation angle gets much greater than even at one degree, basically the tape doesn't do anything. And so if you have the sun up in the sky and you're measuring the tape perpendicular, that observation angle is so huge it's not going to change. And so the way that I've shown that is measure the tape at night and then take a flashlight and just put the flashlight on the tape, which is clearly a different scenario, and then measure it again. You're going to get the same answer. That observation angle is so great, it's not going to make a difference.

Lou (00:37:36):

So when you're taking a photograph, you have the flash right above your head, so what the lens, what the camera sensor sees and the origin of the light source, your observation angle is essentially zero. They're both about the same.

Jeff (00:37:50):

Yeah, it's not going to be zero, but it should be close enough that you'll get at least to see if it's going to reflect or not. Again, it's not really a quantification, but if that's the best you're going to get at least to say, "Hey, look, it's reflecting some light." Body worn camera from officers at the scene is sometimes nice too 'cause they put a flashlight on it and if the flashlight is going on the retro tape at the right angles, you can at least see if it's going to function as retro tape. Again, it's not a quantification, I don't know how bright, but it's at least something that I know it's better than duct tape 'cause I've also had that opinion too where not only did I measure it, but look at this body worn camera, the officer at the scene and he has a flashlight on it, and it's the same increase in brightness as the trailer side. That tape is not retro, that's just essentially a red piece of tape.

Lou (00:38:46):

It's not doing anything. And we should, I guess, define for people who don't know the difference between reflectivity and retro reflectivity.

Jeff (00:38:55):

So everything's reflective, right? White is obviously more reflective than black and matte is less reflective than shiny. Retro reflective is different in that it's physically designed or it's manufactured to reflect the light back towards the light source where your shirt that you're wearing is reflecting the light essentially equally over the whole area, retro reflective material is reflecting light only towards the light source.

Lou (00:39:29):

So that device then, is it similar to a camera in that it emits light of a certain intensity and then has an image sensor and it's measuring how much of that light is coming back to the image sensor?

Jeff (00:39:42):

So I think it's more of a light sensor like an illuminance meter inside. It does have a sensor that's measuring the light coming back, but yes, it's essentially there's a light inside. You're shining that light. It's going to change the intensity over its life, which is why you calibrate it at right before you measure is how this equipment works. You take a calibration plate that you get NIST certified calibration on that plate, and the procedure is you calibrate to that plate at the scene with the ambient temperature and humidity and illumination and everything else. And so you calibrate it right at the scene or right at your inspection site and then you go and measure and it's like you said, it's shining a light at known angles and then measuring what comes back.

(00:40:32):

And so realistically, anything over one is officially retro. It doesn't mean it's going to help anyone if that's so low, but anything over one would be different than again, duct tape. It's doing something that's being retro reflected, but one's not the number you want. That's not really going to help anyone on the roadway. Again, like NIST regulations for new on that tape is like 250 for white, and that's like coefficient of retro reflection. There's a lot of math in there.

Lou (00:41:13):

I remember I went to one of your presentations at IPTM and I was like, holy cow, there's a lot going on there. So I guess for people who aren't going to be trying to dive so far down that rabbit hole that they can intelligently testify about it, but hopefully they can document something, get somebody like you involved if it becomes a big part of the case. A lot of the times, as you know, I've brought you in on retro cases before and been happy to have you there. Sometimes you don't get hired for a year after the reconstructionist. So what I have historically done, like you said, I will be inspecting the truck or the trailer, I'll crank down my exposure setting and then fire the flash so everything's dark except for hopefully, not super dark but darker, I don't want to blow out exposure on the rest of the image.

(00:42:09):

So I try to keep that low, fire the flash and then see what comes back. And then I've done, like you said, in cases where you just get the tractor out there, you get an exemplar car out there, you get the headlights going and take some images, but is that the best thing that recons can do that aren't going to invest in the equipment, just fire off a photograph with a flash and document it that way?

Jeff (00:42:33):

That I would say would be the minimum I'd want 'cause then you at least get something. You get to see something. It might not be able to be quantified, but it's better than nothing. The other options for that would be, and I think Delta does this too, but I'm more familiar with Road Vista. You can call them, they'll rent you the equipment. I think it's like six or 800 bucks and they'll overnight it to you, assuming it's available. We do the same thing. If you need the equipment for something, we'll rent it to you if you know. A lot of times I know people across the country get a phone call at 4 A.M. "Hey, a truck just was in an accident, go now", it's not going to work for that. But surely if you have a little bit of lead time renting it either from the manufacturer directly or for us or from another recon out there that has equipment, that would be my preference is to be able to actually quantify it. I think the range in the equipment is somewhere between 9 and 15 grand depending on what you get exactly. And so I get that's a lot of money, especially if you're only going to do this once or twice a year, but renting doesn't sound so bad, and generally you pass that on to the client and there is a benefit to the case obviously to understand, especially if it's a, like you said, tractor trailer across the roadway at night, car goes underneath it. We're trying to understand what happened and why that happened from all aspects of it, right.

Lou (00:44:00):

Which brings me to my next line of questioning, which I've been living with luminance for the past couple of days reading some of your papers, and man, I love what you've been doing there and the long and short of it is you're ultimately using a consumer grade camera, granted a high end one, but to quantify luminance, so the best way to probably start there is to define for the audience what luminance is and then what drove you to that project.

Jeff (00:44:32):

So the depo answer, right, luminance is perceived brightness. And so it's really the amount of light that's reflected back towards your eye. That's the amount of light that we care for vision. And so I know that there's been lots of discussions and people generally understand the concept of contrast. And you can go by contrast gradient boards to help calibrate your photos and things like that. Basically what you're looking at is quantifying that, measuring it. Let's see how much light is coming back towards your eye for different objects. And so if we have, and again, the depo example or trial example I like to give is one of the reasons I like luminance is the amount of illumination, the amount of light hitting the surface can be correlated to visibility, but it only is correlated to visibility because of contrast. And so if you look at the papers and you look at Mustard's response software, if you're looking at a nighttime equation method, that's on an unlit roadway with a black background.

(00:45:43):

And so if you need less light on a pedestrian dressed in all white than you do it on a pedestrian dressed in all black, because there's going to be more contrast there. And so that trial answer is always if you can put as much light as you want on your piece of paper and no one will be able to read the words if you used white ink because there's no contrast between the white ink and the white page. It doesn't matter how much light you put on it. And so you can measure and you can quantify that contrast. And again, there's lots of published and peer-reviewed research on what level of contrast is needed for different detection. And there's different detection models for that.

(00:46:22):

And so I tend to use that a lot. I use the response software and I use the nighttime equation a lot also or headlight method. I use all that stuff in combination, but especially if the background's unique or there's other things going on, I want to look a little bit closer at contrast. And to do that, you need to understand the luminance of the object or the background. Again, that light coming back towards your eye is what you want to measure and what you want to quantify. That's what set me off into that contrast world.

Lou (00:46:53):

So you really want, if you say have a pedestrian walking across a roadway with a white shirt and it's nighttime, so the background is presumably black, to really analyze the ability of a driver to detect that pedestrian, you would get the luminance of the shirt and then the luminance of the background and compare those two. And that's the contrast you're looking for?

Jeff (00:47:17):

Exactly right. And so there's a road in California that people kept crossing the road at night to go to the beach and getting killed and the city kept getting sued. And the attorney hired me on maybe one of them, and I was like, "This case is terrible." And then finally he calls and he's like, "Oh, I got a good one. I got a really good one. The guy just bought this brand new all-white jumpsuit. He's in a white top and white pants and he's going to be able to be seen, right?" The whole roadway is dark. The problem with that was the facts in the case where the guy is trying to cross and he gets to the center and he gets hit by one car and just knocked into the oncoming car.

(00:48:08):

And so the background is like the headlight bloom. And so you have a pedestrian dressed literally in a white tracksuit and the background is white. And I placed Jeff Schmelter who was at BRT or at BRT now in a white painter suit, and I'm on the roadway and I know he's there and I'm on the phone with him and I'm going 10 miles an hour and I can't see him. I literally couldn't see him. It's like there's nothing I can do here. I can't even measure, I can't see anything to measure. And so that background does matter, and that's obviously a glare issue also in that case, but white on white and black on black are hard to see, right? I don't think that's a hard concept to understand.

Lou (00:48:52):

No, yeah, exactly. And you wouldn't expect nighttime white suit that it's going to be tough to detect, but like you said, you got to take that background into account. We were talking this week, and I was mentioned in that example, there was a video going around that got viral where there was a Detroit Lions defense dude in the end zone and the opposing quarterback threw it to his guy, had no idea the Lions guy was there because he's in blue and silver and the background is blue and silver. I don't know if that was intentional on whoever designed the stadium, but that's camouflage. That is the definition of camouflage.

(00:49:27):

So I was looking at the papers and you get these beautiful graphics where you're essentially mapping out luminance and you can see contrast really quickly. So can you describe the process? Say I have a Sony Alpha 7, and I want to go out to a nighttime scene and try to figure out what the drivers was presented with, what kind of challenge they were presented with. Well, not me. Let's say, how do you go about doing that and what are the underlying concepts?

Jeff (00:50:03):

So it always starts with a scene inspection. And so in general, I want to be on the roadway with a substantially similar lighting, whatever street lights are on or out, whatever we can do to control that, which sometimes is nothing, but whatever we can do. And then I generally do this from an exemplar vehicle. Does the exact headlight make a difference if you're in the range? Probably not. Does the exact eye height in a passenger car matter plus or minus a couple inches? Probably not. I want to get rid of the stupid questions in depo and trial, and so I get the exemplar car, whatever make model matches with the right headlights. Again, there are some cars where they have an HID version and a regular halogen version or an LED version or a laser version. I would want the right one. Those are all going to be different.

(00:50:58):

But again, if you have a halogen light, 9,006 bulb in one car versus another passenger car, probably the same. Anyway, the goal is to get the exemplar car. And in general, what I do is I drive through with video, and so there's lots of suction cut mounts and other mounts you can use. I tend to use a Monopod. I used to spend what felt like hours adjusting these suction cup mounts and whatever I was doing, some of them had three or four suction cups and I was trying to mount it to the sunroof and the window and the windshield, and I would spend all this time to do that and then the video was still shaky.

Lou (00:51:41):

That I was going to say, because you're at these lower shutter speeds with the more modern cameras, you can jack that up a little bit, but you get a lot of shake. You don't have stabilization on most of these. You're not using a GoPro or something with this digital stabilization. So yeah, I'm curious about the mounting and how you stabilize that, whether that's done in pre or post.

Jeff (00:52:03):

Most of the time it's post. And so that's why, unless there's a better version that I don't know about, I also, I kept trying, they make stabilizer plates or vibration mounts, and in my mind, I just haven't spent enough money on one yet 'cause they have to work. You see these videos shot from moving cars in Hollywood and the video's just perfect. I imagine that there's something out there that I can buy if I were willing to spend enough money.

Lou (00:52:33):

I saw Stein. I went up to Stein Husher's shop probably a year ago, and he had this giant setup that goes on the hood of the car. It's super magnetic. It's got these crazy looking stability things on each of the... It has three magnetic mounts for the hood, and then each one has a stabilization thing. And he showed me some of the videos. First of all, it's very expensive, but it's beautiful, but that's still on the hood. It's not at the driver's eye. So that presents its own unique challenge.

Jeff (00:53:00):

Well, one, it's not the driver's eyes, and two, I want it to be on the other side of the windshield. And for me, that matters because if we're going to talk about luminance and detection of an object by DOT regulations, that front windshield can be tinted and it's going to be. There's going to be some light loss going through a windshield. And so I want to be on the driver's side of that. I can account for it and I could measure it, but now I have to explain that and I'd rather just say, "Hey, the camera is in the car and you're looking through the windshield." The other thing is I'm on a plane a lot and it's hard to travel with a lot of gear. Whatever doesn't fit in my Pelican case and a backpack doesn't really go.

(00:53:46):

And so a Monopod solves that. And essentially what I do is I know in general you want about a 50 millimeter focal length to look at what the driver's actually able to focus on. What I do is I shoot at 24, and then when you stabilize it, you're essentially cropping in and matching to stabilize that video. And so we're cropping in to about that 50 millimeter mark in the stabilization while we're stabilizing at the same time. And so that to me seems to be more reasonable. I guess the other thing that I've been playing around with Tony with this a lot is I'll take two of those cameras with a fisheye lens and I'll mount them back to back and I can get 180 degrees and then you can again crop in and stabilize. And if you need to have the driver turn their head or look somewhere specific, you can do that now that you have 180 degree video.

Lou (00:54:46):

Dang. So what cameras are you using there and how are you stitching them together or is it a camera that automatically stitches them?

Jeff (00:54:53):

We're using two Sony Alpha 7S IIs that we're just mounting on the same tripod, like I said, back to back and then stitching them together. And so it's not as an automatic process, but it allows us to get that really nice night video. Again, using a fisheye lens, you're going to lose some of the clarity and some of the resolution, but it's still better than like you said, a GoPro.

Lou (00:55:20):

Yeah, dude, I got so many questions there. I'm like, okay, what ISO, what frame rate, what lenses, how are you accounting for distortion? I'll just let you talk a little bit more about that. I won't actually ping you with all those questions, but I'm super curious.

Jeff (00:55:37):

So again, the settings I'm using are pretty similar for what I'd be doing, just regular video. And so, okay, I'm driving through. What I've done is, and so I know I've talked about this in other classes or presentations, but really the goal for me is to be able to drive through the scene taking video from approximately the driver's eye perspective. And I'm capturing a couple of things here. I'm capturing a video of the drive-through that hopefully we can use later. And I'm also capturing, because using specific settings in the camera that I've calibrated in the Terra that I can now take luminance estimates from. And so that gives me a lot of information. It gives me a lot of information to be able to do a visibility analysis with, but it also gives me a lot of information to calibrate the photos and the video later as well as a lighting simulation.

(00:56:40):

Because if we know the real-world luminance values from the scene, when you bring it into your simulation software, like Tony will use Blender, Henry likes to use 3D Studio Max, you can use whatever and Tony's validated Blender render engine, and Henry did validation for Max. As long as you're ensuring that you're matching real world values, if you take the video in, you can add your element in. So I drive through the scene at night. Let's say the vehicle hits a pedestrian. Well, now we can take that video, add the pedestrian, and add the motion based on the recon and add the lighting based on mapped headlight file, which Tony and I also did some work on and published on, or an IES file if you have the headlight file some other way. And then you can actually ensure that what your output is is fair and accurate to the real world luminance values.

(00:57:35):

Now you have to display it properly. You can take your really nice calibrated video that we spent all this time to ensure the real world values or the luminance values in the render are right. And then you put it on a monitor that's cranked all the way up in brightness and it's going to be blown out. Same thing, cranked all the way down in brightness and you can't see anything, and so you have to display it properly, but as long as it's displayed properly, the input is video you've taken at the scene with known luminance values 'cause we're using that video to estimate that. And then your output of that video is the same thing, known luminance values that you're matching. And so again, the goal is to try and be as fair and accurate as we can and make sure that the end product is appropriate.

Lou (00:58:26):

Yeah, that sounds like a huge challenge. Getting the proper presentation come, I mean, depo not that big of a deal, but once you start presenting it to the trier of fact, now that's a huge deal. Are you calibrating the courtroom systems, bringing your own, how do you handle that?

Jeff (00:58:42):

I'm laughing because I've been doing this for some time now, and I always make a big deal. And maybe over the years I've maybe calmed down of how big a deal I make in Depo, but I generally make a big deal out of how you have to display the videos properly and what the papers say and what the research says. And I basically lay out everything the opposing counsel needs to get my videos excluded if I don't do these things at the time of trial. And then I always tell my client, we need to do these things. And they almost always ignore me. And then I get into the courtroom and they're like, "Okay, Mr. Suway, this is the video you took?" Yep. And then they play it, and then I'm sitting there, soul crushed.

Lou (00:59:25):

Yes.

Jeff (00:59:25):

I've done all this work and no one's cared, including opposing counsel. I was hoping opposing counsel would make some big deal out of this. I say that, and then I just went to trial. Jason Ars was the recon, and I did the night visibility stuff, and the other side was making a big deal out of it. And so we ended up getting our own monitor in. I had to come to the court, the judge allowed me in before the jury came in earlier in the day, and I could calibrate that monitor for the courtroom. And so we went through that whole process again.

(01:00:03):

And so we went through that whole process again. The other side made a huge deal out of it and they were like, "Our experts are going to come and they're going to watch you and they're going to be able to cross you on this and it's going to be a big deal." And then I show up and plaintiff's counselors wasn't even there.

Lou (01:00:14):

None. Exactly.

Jeff (01:00:15):

It's literally like me in the corner of the courtroom by myself. The judge is on phone calls ignoring me, and I got to talk about it now. But no one cared. Still no one cared.

Lou (01:00:25):

Yeah. Lots of bark, little bite.

Jeff (01:00:27):

Right. And I think by the time they figured out they called our bluff or they didn't call our bluff, we were actually going to do this, they figured they weren't going to fight it anymore. But like I said, one case, one trial.

Lou (01:00:41):

Oh man, I'm in the same boat where I'm so often disappointed in the AV, we do so much work to get our animations, presentations, simulations, all that jazz, just perfect. I'm getting the line thickness right and the dash length right and the color right. And then who knows what you're actually going to get come trial.

Jeff (01:01:04):

It's a printout on an Elmo or something.

Lou (01:01:07):

Yeah, exactly.

Jeff (01:01:09):

Okay, great.

Lou (01:01:10):

I think that the presentation in front of the jury is super-duper important, but you can only take your clients so far. They have to believe the same thing. This episode is brought to you by Lightpoint of which I am the principal engineer. Lightpoint provides the collision reconstruction community with data and education to facilitate and elevate analysis. Our most popular product is our Exemplar vehicle point clouds. If you've ever needed to track down an Exemplar, it takes hours of searching for the perfect model, awkward conversations with dealers, and usually a little cash to grease the wheels then back at the office. It takes a couple more hours to stitch and clean the data. Time is the most valuable thing a person can spend. Don't waste yours doing work that's already been done. Lightpoint has a huge database of Exemplar vehicles all measured with a top of the line scanner like as RTC360. So no one in the community has to do it again. The Exemplar Point Cloud is delivered in PTS format, includes the interior and is fully cleaned and ready to drop in. Your favorite programs such as CloudCompare, 3DS Max, Rhino, Virtual Crash, PC-Crash among others. Head over to lightpointdata.coM/datadriven to check out the database and receive 15% off your first order lightpointdata.coM/datadriven. Sounds like there's a couple ways to do it, correct me if I'm wrong. It sounds like most of the time when you drive through with a video camera, you're capturing all these frames. You'll pull key frames, you'll quantify the luminance based on your algorithm that you've developed in these 2017 and 2022 SAE papers, which will link to in the show notes. And then you take that into a lighting simulation in Blender or 3DS Max. Are there times when you will just take contrast measurements straight, something like your software Nitere? Is that how you pronounce it?

Jeff (01:03:05):

Yeah. Nitere. Yeah.

Lou (01:03:07):

Okay, so will you take contrast reading straight from Nitere and I've only read it's like one of those words that you've read for five years and you have no idea how to say it.

(01:03:18):

But that I imagine would a lot of the times be sufficient. So if you have a ped walking across the road and you've got a surrogate and they're walking across the road during your video capture, then you can just say, Hey, give me the luminance of this shirt, give me the luminance of the background. The difference is this, I have these papers over here. Yay or nay. Is that a potential workflow as well?

Jeff (01:03:43):

Yeah, absolutely. What I was describing was the workflow to get this lighting simulation in some of what I was saying, I'm assuming you're not going to be to close the roadway down or you want that video. The other way I've done it is absolutely that where you have a pedestrian dressed in whatever the appropriate clothing is, hopefully the right height and things like that. And then we place them on the roadway and we'll take still images or video. I try not to drive it live pedestrians, but-

Lou (01:04:12):

Good idea.

Jeff (01:04:14):

So what I have done though is for tractor trailers but also for pedestrians, I'll just have the car idle and I'll record video with the car idling forward at whatever it is with GPS on it. So we're going five miles an hour. Well now we can speed the video up and I'm not worried about running into the tractor trailer that stopped on the roadway or hurting my pedestrian that I've placed there. And so you can do that in post where you speed it up and you get a lot of data. I'm capturing 24 frames a second, we're going five miles an hour for a thousand feet or so. You have a ton of data to measure from. But yeah, absolutely. And I know that you had a poster or you're to the point email with an image from Nitere and that's exactly what I did in that case where there was a pedestrian in the roadway or having them walk across the roadway and then you can draw a polygon.

(01:05:09):

And so one of the reasons that I like this method besides I'm not law enforcement, I can't close the roadway down always is I can, instead of just getting stuck with, if you have a spot meter, a luminance spot meter, you get stuck with a 0.1 or a one degree cone, so you can't get the whole torso. You kind of just get an average of the pedestrian. And so I can more precisely draw a polygon and say, "Hey, what's the shirt? Luminance of the shirt? What's the luminance of the pants and what's the background near the pants and what's the background near the shirt?" You can get more fine tune your luminance measurements. It becomes really important for retro tape because that retro tape, it's only two inches thick or two inches in width, and if you're at a distance away, that's important, a one degree cone or even a 0.1 degree cone from your spot meter, you're not going to get the retro tape. You're going to get the retro tape and some underneath the trailer and some above the trailer. It's kind of an average.

(01:06:11):

And so to be able to kind of pick just that point is one of the reasons I like this software. There are other commercially available like luminance cameras you can buy. Cost goes way up.

Lou (01:06:24):

I was looking at the costs and it's what, seven to 10K essentially?

Jeff (01:06:28):

Yeah, you can spend 30 grand if you really want to between the software and the hardware for some of these luminance cameras. The big push for me was those luminance cameras are cameras and sometimes for night work, low luminance, it needs to have a 32nd exposure and that doesn't work very well if I need to do a video and I'm not going to sit on the I5, like a major highway in California and take a 32nd exposure from the travel lane, I'm going to get hit. And again, I'm not law enforcement, I can't close that road. And so to be able to take video as I drive through feel safer, I'm not playing Frogger and I'm not in the middle of the street for long. The other thing with those luminance cameras is they have a lot of power requirements. And so I blew up a bunch of inverters when I was using them and trying to test them.

(01:07:25):

Literally it smells like it's about to catch fire quickly unplug it. So again, there was a bunch of reasons that kind of pushed me towards what we're using now. It's really the same science, it's just applied slightly differently.

Lou (01:07:42):

And from the traditional methodology, like I said with the spot meter, that is super time-consuming, super tedious I, Think I've done it a couple of times with Mutar when I was on the East Coast for unquote fun. But you have to measure a million different things to get a proper map of what you're trying to analyze. Whereas if you get the video you can export those key frames and then in something like Nitere go in and draw the polys and be like, all right, the shirt is this, that light? Is this the retro tape? Is this, the sign on the store is this, and you end up with a beautiful graphic so that you can then present it to the jury assuming you have some AV equipment there.

Jeff (01:08:22):

Right.

Lou (01:08:25):

Okay. So what does the calibration process look like and how low can you go with a camera? Sony Alpha 7S II, not that expensive nowadays. You could probably grab one on eBay for what, 800 bucks tops. What cameras do you recommend and what does the calibration process look like?

Jeff (01:08:44):

I really like the Sony Alpha 7S IIs. The 7S III is nicer. I haven't upgraded yet. It's really a pretty similar camera. I don't think it's a huge upgrade, but it's the newest, right? The calibration process is really, we need to correlate the amount of light coming into that sensor with a real-world luminance value. And so I have a spot meter, kind of Conica Minolta spot meter, and I actually use Mutar's contrast gradient board that he sells because it's a really nice target. It's great for a lot of things, but I use it as a really nice target for this where it's 10 different C's, like literally the letter C across, five on the top, five on the bottom, and I can take my spot meter and measure the center of the C, and it just gives me a nice reference point. And it's different.

(01:09:39):

We have different white, we have white, we have gray, we have black. And so for a lighting, and we talked about me moving to this new office, the one requirement that the architect or whoever was helping get all the plans together didn't understand is that our conference room. She's like, "What do you mean you don't want windows?" And it's like, no, no, no, we don't want any windows. We want a door that goes as far to the floor. We want no light. And so our conference room acts as our calibration room two where I have enough distance and we have lights on dimmer switches, and so we're not having external lighting. And then I can measure at different light levels, the amount of luminance in each one of those C's. And I take a picture. I think one of the differences in what we're doing versus a luminance camera that you'd spend seven to 30 grand on is the frequency that you're going to take what we call a dark image.

(01:10:40):

And so it's a digital sensor on this camera, and when the lens cap is based on temperature, there's just an ambient voltage going through each pixel. And so we want to understand that because that's clearly zero. We want zero to be zero, and so we have to account for that. The luminance cameras, they account for that essentially once or maybe one time at the calibration phase or even one time before you take measurements because those sensors are liquid cooled, which is really cool. But obviously our Sony cameras aren't. And so we're taking dark images more frequently as that temperature changes of the sensor as it heats up as you're using it, that steady state, like zero voltage changes also.

Lou (01:11:29):

But you still don't have to do it that frequently, from what I read in your papers, like every five minutes or so grab of a dark image.

Jeff (01:11:35):

And so I've also been playing with, there's a really big range in temperature that doesn't make a big difference. If you go to the extremes, I've put the camera in the middle of the summer out in Bakersfield in the sun and let it sit there and bake for a while and get it real hot and then take dark images. That's different than if I put it in my refrigerator for a while and then take a dark image. At extremes, it matters. But realistically in the middle, I don't think it makes it, at least from what we're testing, it doesn't make a big difference. But I'd rather take more than less and not have to worry about it or the person on the other side.

Lou (01:12:14):

I think actually you just might've figured out why they call Bakersfield Bakersfield. By the way, that was the first time I made that connection.

Jeff (01:12:24):

Yeah, right. Me too. Yeah, you can bake bread out in the middle of the summer.

Lou (01:12:28):

Exactly. Fry an egg on the sidewalk.

Jeff (01:12:30):

We're taking these dark images and then subtracting out that dark current or that negative current and doing that for different light levels. And then you can get... We're putting it into the software. Essentially it's taking each pixel and at each pixel level we're then taking that amount of light that pixel is captured and subtracting the amount of light when they have the dark light, your lens cap on, and using that to correlate that with illuminance value. That again, based on a conical Minolta spot meter is what we're using. Trying to go back to a NIST standard. And so that's NIST certified or calibrated to a NIST standard. And then like I said, it's a pretty simple, we're just fitting the data kind of thought. We've done it with Sony's and I like the Sony Alpha 7S because it does 4K video on board. It has a nice big sensor that's not crazy megapixel. I think it's only 12.1 megapixel.

Lou (01:13:40):

Yeah, it's 12.

Jeff (01:13:40):

And so it's the same sensor size, and if you only have 12.1 megapixels instead of 48, each pixel sensor is going to be bigger, and so you can get more light in per unit time for a bigger pixel. So it ends up having better night performance. Anyway, you can do that with other cameras too. And the price point is tough. Like you said, you can get a used one really cheap. I look back and the first one I bought and I like Black Friday sales, but the first one I bought, I got the camera and a nice lens and 128 gig high speed SD card for like 4,800 bucks shipped to my house after tax. Yeah.

Lou (01:14:26):

Yeah. I think I was about in the same ballpark. I have a 7S II as well, and then I got the 50 millimeter Zeiss lens, which is phenomenal.

(01:14:36):

And for daytime stuff, that's great, but to your point, and I'm realizing now I do have I think a 20 millimeter fixed as well, and maybe I could use that and then crop in so that I have a room to stabilize, but I hadn't thought about that before. That's really smart. And then like we were saying, if you're going to be doing nighttime stuff, you really need that sensitive sensor. And the Alpha 7S II, that's its whole thing. It's like take video, very sensitive sensor, otherwise you got to stop down. Your shutter speed has to be so low that there's just going to be camera shake everywhere. So you need... I don't know if you have found out where that threshold is when you're recording from the driver's eye, how fast does your shutter speed need to be and then how does that generally relate to how high you need to crank the ISO and you don't want to crank so high that you get noise, and that's another benefit of that Sony. So what have you found through your experience with regard to those settings?

Jeff (01:15:35):

The general rule of thumb for video is you want to record double then the frame rate. And so in this, I'm recording 24 frames a second, so really the minimum you'd want would be a one over 50 frame rate shutter speed.

Lou (01:15:50):

Which is pretty quick. Man at night, that's quick.

Jeff (01:15:53):

And so with this camera, I've found that one over 80 actually gets me nicer images, I think one over 50. I still start to see some, they call it shutter roll is what you get with some of these digital cameras. One over 80 is kind of the minimum I want to do for video. That gets me a good number and everything has a positive and a negative with the camera settings. You want to have a lot of light, so you want a slower shutter speed, but you don't want it to be blurred and you don't want it to have shutter roll. And then the higher you go, the darker the image and it becomes just black. The same with F-stop. And so initially I was buying these really crazy low F-stop expensive lenses. And the problem for me is when I'm trying to focus manually focus at night, the lower the F-stop, the depth of field becomes super narrow.

(01:16:53):

And so I might focus on the target when I'm a thousand feet away, but by the time I get to 400 feet away, it's out of focus and blurry. It's like that doesn't really help or in the distances you care about, it becomes blurry. And so F-four is about the lowest I really want to go to make sure that that depth of field is big enough. And so again, that's for what I feel gives me a good balance of depth of field, but also low F-stop for a lot of light. And then at least on this Sony camera, I think anything up to a 20,000 ISO, which is a really absurd ISO, gives you a pretty good image.

(01:17:32):

I mean, 20,000 ISO I would be using more like... I do some premise work where someone trips on a parking stop at night, there's no headlights involved. 20,000 ISO might be more appropriate for that. When you have headlights involved. It's more in the 10,000 range is kind of what those settings on that camera. And if there's a lot of overhead light at the roadway, it might be more in the 5,000 range ISO. The same kind of thing, the higher the ISO, the more light you get in and the higher the F-stop can be, and the faster your shutter can be, but the more noise you get. And so at least again, on the Sony camera, anything under 20 seems to be pretty reasonable.

Lou (01:18:16):

So then do you generally go to F-stop four shutter speed, one over 80, and then adjust your ISO for the scene?

Jeff (01:18:24):

And so like I said, at least the calibrations I typically use are between five, 10 and 20, and that's the calibration I'm using for Nitere. And so I can always change it and recalibrate in Nitere. But you also, again, it depends on how you display it. And so you want to make sure that that display is fair and accurate. And so if you took measurements or video at 10,000 and that's a little too bright, but when you display it, you can bring down that total contrast by just changing the brightness of the monitor. And so it's not that we're altering the video or it's just we're displaying it more appropriately. And you can measure that and you can quantify it, which is typically how I do it. Like I said, one trial, they actually let me calibrate the monitor and spend an hour with it before the jury came in. But other than that, it's just me and my office doing it for me.

Lou (01:19:21):

Right. Yeah, I mean it does allow you to at least come to your conclusion. And my preference is, and I'm sure yours always is too, is I don't want the jury to have to trust me. I want them to be able to see what I did, understand why I did it, and then hopefully take the same thing from my analysis that I got from it. But in some cases, on those ones where you can't show it to them, at least you can explain your process and hopefully you do have a very trustworthy face. So I have to imagine that everybody believes you when you speak. So people can send you calibration, a camera, you'll calibrate it, you'll give them this best fit. Well, this is my question. I guess you get the coefficients for this best fit and then they can do their own stuff, or do they get a file that works with your software as well? How does that work? Are they coefficients that you enter into Nitere?

Jeff (01:20:15):

We really send them a fit file for Nitere, which is the coefficients that you're talking about for the model equation to fit that data. It's just a text file. It's named .fit for Nitere, but you can open it in Notepad and it gives you all the values. And I normally put in a screenshot too of the software when we've calibrated it so you have all the values. This initially was done MATLAB, the initial work my brother did. He's the programmer. I just told him what I wanted and he did it, which is really nice way of operating.

Lou (01:20:55):

Yeah, my brothers are useless in that respect.

Jeff (01:20:58):

Really?

Lou (01:20:59):

You got better brothers.

Jeff (01:21:01):

Yeah. Well, yeah. I was telling you, I was joking with you that I'm the dumb Sue way out of the two of us. Yeah.

Lou (01:21:09):

That's scary.

Jeff (01:21:10):

And like I said, he's really good at this stuff, and so he did it in MATLAB. I know that there were other people that have kind of taken the concepts that we've talked about in our papers and done it themselves. And you can do it between a bunch of different software. You can even do it in Excel. It just MATLAB was easy. And then ultimately to kind of package it up in this prettier version, Nitere, he redid it in Python, but it was really a project because I'm telling him the struggles I'm having and how I can't stand in the middle of the five freeway at night, and I'd love to be able to do the luminance camera, but in video. And he's like, "Oh, I can analyze that in MATLAB." I bet I can fit it, fit the data. And that's kind of how it started, which was really nice. And so I use it for almost all of my... Anytime I'm doing anything with contrast or a lighting simulation, I use it. In all of those cases.

Lou (01:22:06):

The deliverables are really beautiful. It seems like it's easy to use. We'll put a link to it in the show notes. It's what, 500 bucks a year? And then you also offer training and then the camera calibrations. So it doesn't sound like... If people are directed to the appropriate literature where you are taking the contrast and comparing that to detectability or discernibility, whatever you human factors experts call it, then it seems like it's kind of plug and play for a lot of people if you have... It's really a video skill. You need the video skill, which we just went over the settings. So if you get that camera and throw those settings up and the video looks good and you calibrate the camera, then you're in a position to have luminance measurements for the whole thing. And then by comparing to a few papers, they're in a good position to talk about contrast and detectability.

(01:23:02):

I got to see that firsthand in a case that you and I worked together, a tragic case where a kid on a dirt bike tried to enter a trail that was supposedly off limits at the time, there's a chain across it. He runs into the chain. I think the rider lived right, but the passenger died or vice versa. And there was a lot of debate whether the chain was painted, whether it was silver, whether it was rusty. And we went out there and did some analysis. Can you just talk about what you did in that case and what your ultimate deliverables would've been had it not settled out?

Jeff (01:23:37):

And so you are obviously doing the recon and had the nice motorcycle set up with the right aftermarket headlight, and you kind of had that all set up for me.

Lou (01:23:47):

I forgot about that. That's right. Yeah. I got to buy the little motorcycle and start wrenching on it, putting on the right headlight and all that. That was a good time.

Jeff (01:23:54):

And we got to hang out at night in the parking lot with a bunch of interesting people.

Lou (01:24:01):

Yes, there was some drugs, there were some alcohol, there were fireworks, if I remember correctly.

Jeff (01:24:06):

A couple officers came by.

Lou (01:24:07):

Yep. Yeah, that was cool of them. That's right. They shooed everybody away when they found out what we were doing.

Jeff (01:24:13):

Right, exactly. And so my role in that case was really taking calibrated photographs that we could show to the jury or trier fact, whoever, to show what this chain would've looked like in the different configurations that it possibly could have been. And part of that was, again, the chain is super narrow. It was maybe what? An inch and width. And so in the distances you care about riding a motorcycle, a spot meter is going to be very difficult to measure just this one inch width. You're going to get everything in that. And also, again, we're in a parking lot with drugs and alcohol and fireworks and everything going on where-

Lou (01:24:58):

You want to get out of there.

Jeff (01:24:59):

Well, one, we want to get out of there, but two, I guess we could have brought a generator and set up a luminance camera and done 32 second exposures, but it just becomes more of a hassle. And so the process there was let's take these photographs with specific camera settings at specific distances, and then we can go back. Not only do we have a photograph that we can calibrate or display properly to show the trier fact what it would look like at different distances, but we also, what I was doing is putting it into Nitere and then like we said, I can draw a polygon and say, okay, the luminance of the chain is this and the luminance of the background behind it or in front of it or both is that, and now let's see what that matches for or looks like from a contrast detection model.

(01:25:52):

And then again, it's hard to describe that, I think. If you're using a spot meter, I'm a miserable artist trying to draw the chain and everything, and where did I take the measurement exactly? Then my chicken scratch of the actual measurements. This is now a printout, like you said, it's a false color image, right? It's in some version of blues and greens and reds, right?

Lou (01:26:17):

Yeah. Like heat map looking.

Jeff (01:26:19):

Right, exactly. A heat map or a false color image that allows you to show, okay, I took the measurement here and I took the measurement there, and those were the values, and this is why. And it also is, like you said, a visual representation of relative brightness. It's not an image that's just a color image, it's now that heat map that's giving you some sort of scale for relative brightness or luminance. And so that was kind of the idea of that case. I don't know if I told you this, but we were told the case settled and I assume it settled. But maybe the next month I was on a conference call with the opposing attorney and they brought up a declaration from their expert in that case that I had never seen as an example of something. And I'm like, I'm trying to back... You can't show this to me. And they're like, oh, no, no, it settled. It settled. They're like, wait, you were on the other side. They were kind of upset.

Lou (01:27:24):

Oh my gosh. I guess we never got the designation on that case.

Jeff (01:27:28):

Someone on the other side wrote a declaration or a report that they were showing me. I didn't read it because again I was like, I don't know that I should see this.

Lou (01:27:38):

Oh man, I'd be really curious to see what they did. We obviously went to great lengths to very accurately determine what the view of the rider would've been and how they could have detected that. So I would've loved to see what the other team did and see yeah, if they came up with a similar answer. That's funny. It ended up that case got kicked out of court completely. The judge just tossed it.

Jeff (01:28:02):

Oh. I didn't know that.

Lou (01:28:02):

It didn't actually settle. They got nothing. It's kind of Willy Wonka style. You get nothing, sir.

Jeff (01:28:10):

Okay. All right.

Lou (01:28:11):

So what are the bare essentials for a recon trying to do nighttime stuff?

Jeff (01:28:17):

So the couple of things that I really like, and I use obviously a good night camera as we talked about, like Sony, Alpha 7S II, super cheap on used, the 7S III is the new version of it. That's really nice ice. Any manual, fully manual camera that you can take proper images. Again, I'm a little partial to the Sony camera. I can handhold it most of the time at night and not worry about the longer shutter exposure and get really nice images without, I'm sure everyone's been at these inspections where no one move, no one move, and you're setting up a tripod and taking these long exposures and then some light in the background changes and there's a blur through it. And so I also have to imagine that, especially in the beginning, I was frustrating other experts at joint inspections where they've set up their camera and they're about to do it, and I've taken a picture or five of what they're doing and the scene, and then I've taken light measurements and I've done all this while they haven't even taken one photo yet. I really like this camera, especially at the price point.

Lou (01:29:30):

It's funny the same way some of the things that I do at inspections make me uncomfortable just because I know I'm taking up a lot of time from the other experts, which is a big reason why I love running the Leica RTC360 now, because I'm like, I could just get that thing in and out. Even two Faro's takes longer than one of those things. But for now, this is being recorded in what, September, 2024. So we'll see what FARO does. But anyway, back to your tools.

Jeff (01:29:58):

Yeah. And so an illuminance meter I think is important and relatively inexpensive.

(01:30:03):

An illuminance meter I think is important and relatively inexpensive. Like I said before, there's lots of good research on the amount of illumination you need on a pedestrian that's going to allow you to detect it. You can go through that response software that Muttart puts out, and there's really good stuff in there for illuminance. It's relatively cheap, and then you get an idea of the amount of light and you can talk about, instead of just, "It was dark."

(01:30:29):

The other relatively cheap thing I like is like a color checker or a color chart, Macbeth chart. It really allows you to say why the color in your photo is right. Some of them start white and have grayscale to black and some of them have different colors, but essentially I use it to white balance the photos or the video. Each headlight's going to be different and each overhead light's going to be different. Again, the idea is to get the white balance to look proper. It saved me a couple times where attorneys have tried to ask silly questions in depo or on trial and it's like, "No, no, no. Look at this photo. I've color corrected this and I made sure."

Lou (01:31:11):

Yeah, I keep meaning to buy one of those. They sell them. I imagine you're using a bigger one. The ones that I've seen are say, six by six. Then if you find a material transfer on a car and you want to prove that it's yellow, then you put that up next to the shot, take the photo, and you've got this base.

Jeff (01:31:28):

Exactly.

Lou (01:31:29):

How big are the ones you're using?

Jeff (01:31:30):

So I have little ones like that, but they also do a... I guess it's more of an eight and a half by 11 size. Again, I'm using it more of a close up, and also color of car paint. Especially if we have to go back and we're doing a lighting simulation and we need to get the color of the car right, a stopped car on the roadway that you want to see what that looks like, you want the color right. If it's not right, you're going to get beaten up. So just putting a color checker on the car and taking a photo will allow you to white balance it and ensure that what you're using in your lighting simulation animation is accurate. It's like 50 or 25 bucks depending on the size. It's a really cheap, easy thing to throw in your kit.

(01:32:18):

So we talked about in Nitere, part of that was design the price point, whatever, and everything was designed where there's a certain amount of people that aren't going to spend the five grand on a spot meter, and there's certain people that don't care about the price. If it was 10 grand, they're going to pay 10 grand for it. We're trying to play both worlds where someone's not going to spend five grand, they still can do the analysis and they can send us the camera to calibrate it. Then there's also outfits that they have the spot meters, they have the luminance cameras and they have this as another tool. You have Leica and FARO and all of the scanners.

Lou (01:32:59):

I'm actually pure Leica now, but for a long time I had both, and I mainly switched over just so that we could have the same processing system. Now we have two Leicas, so that one is out for calibration, we still have one in the works. But yeah, I have a lot of redundant tools here. They have different strengths and weaknesses and they work better in some situations. So I can imagine that to your point, somebody who's in a really volatile environment would rather use a video camera and Nitere compared to something that's going to need 30-second exposures.

Jeff (01:33:37):

Yeah. That was really the driver for me of really not wanting to play Frogger in the middle of the street, right?

Lou (01:33:45):

While you're talking about your kit and Frogger, we had a case, I was doing the recon, you were doing the nighttime stuff. It's like a tractor trailer backing in to make a delivery and this guy delivering bread just goes straight into the side, no breaking whatsoever. That tractor trailer had been there for a long time in motion, and I think you were hired to... I think we're going to do a lighting sim on that case. You showed up full head-to-toe high vis yellow suit. I had never seen that before. I had seen the vest. I rock the vest pretty consistently, but you had the full head-to-toe jumpsuit and I was like, "This guy, he really wants to get home to his kid tonight."

Jeff (01:34:25):

That's right. That's right. Yeah. So they sell those on Amazon and it's nice. It's got retro everywhere, your ankles, your legs, your waist, your arms, and it really falls into that nice biomotion pattern discussion where you can see the full pattern of the person. Like you said, yeah, I'd like to go home. I don't want to end up with a recon figuring out what went wrong.

Lou (01:34:52):

Yeah. That has happened several times to some of our colleagues, some of whom have died unfortunately, and then some of whom have just been injured to a certain degree. So I run the same filter through my head whenever I'm doing anything out there. I was like, "All right, no messing around," put on some high vis stuff and even if it's the daytime. So tripod presumably, you kind of mentioned that, good nighttime camera, hopefully one that takes video. Color stuff, what is that chart called? Macbeth chart, is that what it's called?

Jeff (01:35:26):

Macbeth chart. But yeah, I think if you just put on Amazon color checker, it shows up too. Yeah. Color checker I think is what shows up.

Lou (01:35:35):

Then illuminance measurement tool, which can be had for pretty cheap, like the Extech I think is the one that a lot of people seem to like, and that puts you in a good position right there. Monopod, and the monopod, by the way, you're driving and just hand-holding it in front of your head, peeking around the side of the camera?

Jeff (01:35:55):

Right. Again, that feels safer than playing Frogger. I get I'm still holding a camera in front of my face and driving with one hand, but it still feels safer to me.

Lou (01:36:06):

I've been there, man. Yeah. I've been there. It's something that we were trying to solve here for a long time. Because first of all, I love that look when you get the camera at the driver's head and it's the right field of view and you can see the gauges and the steering wheel and the radio and the windshield, it's beautiful. If you can stabilize that and make it exposed properly, it's such a beautiful exhibit. You wouldn't be taking that if the visibility didn't matter on some level, whether it was field of view, what they could see or contrast.

(01:36:41):

So we were working, doing the same thing that you did, like suction cups everywhere, all this rigging, and I think ultimately we just came to wedge a tripod in there. Then as the driver, you just got to put your head to the side. I'm not sure who has come up with an elegant solution to it, but it sounds like you and I arrived at a very similar solution.

Jeff (01:37:03):

So, I've done the wedge the tripod. The problem is that I have trouble getting my hands on the wheel and also moving my legs without disturbing it. I don't know, maybe you're just a more compact person than I am.

Lou (01:37:14):

No. That was not easy. That was not easy.

Jeff (01:37:17):

But that's why I went to the monopod. Again, I don't know, I buy everything Amazon, Black Friday sale or whatever. They have some nice foldable tripods that turn into a monopod. I can get it if you want, but basically the whole thing will fold out and then you can unscrew one leg, unscrew the top, and now it's a monopod. So I have both when I travel to the inspection, and it fits really nicely in my Pelican case that I can carry onto the plane and not worry about.

Lou (01:37:49):

That's a huge thing. I have a Pelican Air that is the exact dimensions that you can check, and if it doesn't fit in that, sometimes I'll ship stuff. If I have to weigh something, then I'm like, "All right, I'm shipping my scales. It is what it is." But generally I'm with you. I try to get everything into that Pelican. I inherited a Manfrotto monopod at my wedding because the wedding photographer left it behind and I pinged him like three times and he never responded. I was like, "All right, I got a high-end monopod now."

Jeff (01:38:18):

Yeah, nice.

Lou (01:38:18):

For whatever reason, it's one of my favorite tools. I'd love a monopod, because you get 90% to a tripod with it, but it's super compact and easy to handle.

Jeff (01:38:27):

Right. So talking to Muttart about the research and stuff you've done, I would've guessed it was your suction cups that was your favorite tool.

Lou (01:38:36):

I have a lot of suction cups and I'm a giant fan of suction cups. The amount of power that you can get from that vacuum, I'll tell you what, man, I can mount stuff anywhere.

Jeff (01:38:47):

Just like I have a retro fetish, I think suction cup boy was what comes to mind, was his name for you.

Lou (01:38:52):

Yeah, and now, it's suction cup man, told him the other day in an email.

Jeff (01:38:56):

Perfect.

Lou (01:38:57):

I've grown.

Jeff (01:38:57):

Perfect.

Lou (01:38:58):

But yeah, I have a lot of suction cups here and they're not great on motorcycles, but they're good on everything else. Motorcycles generally just don't have the spot for it. So 3M double-sided sticky tape is what I go to there and that was a rebellion. My suction cups were in direct rebellion to Wade Bartlett using duct tape to mount all of the things.

(01:39:19):

I was like, "This duct tape is not making it into the research paper, Wade. I'm whipping out the suction cups." I don't know if you heard this story, but the next morning, we had to start testing. We had humans subjects there, or a lot of people waiting for us. It's like 6:00 AM, I roll out of the hotel and I go up to my car, and it is completely covered in duct tape. Every seam is covered with duct tape, wrapped all the way around the trunk, and that was Wade's response to me dissing his duct tape.

Jeff (01:39:51):

I love it. That's fantastic.

Lou (01:39:52):

I think he still loves it. Yeah, he's got the duct tape fetish. I got the suction cup fetish, and you got the retro fetish. All right. So like I said, you're a man of many skills in this industry, and probably other stuff that I don't know about. I know you got a boat, so you're probably a good wakeboarder and all that too. So like I said earlier, you're always up to some tricky stuff. Are there any methods or tools that you've been experimenting with recently that we haven't discussed yet?

Jeff (01:40:24):

Yes and no.

Lou (01:40:25):

Always, always. I knew it. I can see it. Yeah.

Jeff (01:40:30):

The one that's still on the list that keeps... I now have some, or I'm about to have some data to work with is using Nitere to also estimate illuminance and not just luminance, with the idea that I could put a drone in the air and I could take photographs. A lot of these drones let you do manual settings, and so we have calibrated some drones for Nitere for luminance. The goal would then be, okay, what if we put the drone in the air, take some photographs of a roadway and we've mapped all the illuminance from the overhead lights on that roadway? That could be pretty cool.

(01:41:11):

Then it has some implications for headlights also, especially as we get swivel headlights or adaptive high beam and headlights that are changing based on the vehicle inputs. How do you map headlights if they're swiveling, but the swivel is dependent on the vehicle speed and the steering wheel angle? You can't do that in a parking lot. So trying to come up with ways of seeing the future of headlights and what's coming, especially adaptive high beams, right? Those high beam as it's blocking oncoming traffic or whatever it's going to block, we might want to know or need to know that for a certain case, and so that's the dynamic issue that you can't just do in a parking lot like you would traditionally do for mapping headlights.

Lou (01:42:06):

They keep making stuff difficult for us between all the data and all the modules we have to pay attention to and variable steering ratios, so when you're doing sims, that becomes really tough. It's becoming an ever increasingly sophisticated discipline that's tough to keep up with.

Jeff (01:42:23):

Eric just did an HV with a newer BMW and that has rear steer, and the best data out there is the rear steer could be up to this many degrees.

Lou (01:42:34):

Wow.

Jeff (01:42:34):

It's like, well, when does that happen and how does that happen at the speed? So it's variables, and Tesla's new drive-by-wire with their Cybertruck.

Lou (01:42:48):

Oh, right. Super variable.

Jeff (01:42:50):

Right. Incredibly variable. How do you do that?

Lou (01:42:55):

In HVE, can you steer the rear wheels?

Jeff (01:42:57):

Yeah. Absolutely.

Lou (01:42:58):

Oh, wow. What a program, dude. That's a very flexible program. On that note, so you've got obviously a family, you've got the business, and then I look at your CV and I speak with you and you're attending classes all the time. You're publishing all the time, you're presenting all the time. You're up-to-date on all the bleeding edge tech. What is your process for staying in touch with all of that and making time for all that? But more importantly, I guess, the question I'm really trying to get at is how the heck do you stay on top of all of the latest tech in this industry?

Jeff (01:43:40):

Talking to studs like you. I think we all have a lot of car time and I try and optimize that or maximize my effectiveness. If I can hang out on the phone with Tony Cornetto or Jason Arst, or you for an hour once a week, I'm going to learn a lot. So there's a bunch of people that I'll call and talk to. I'm just trying to understand what's going on, and then the other thing is if your hobby becomes your job, it's something you're always doing. I like it and I'm interested in it, and if I'm laying in bed, I'm going to be reading something helpful instead of like... I'm not scrolling TikTok.

(01:44:26):

Maybe I'm more of a nerd and more boring than others, but I don't know. That's the way that I can keep up with it. And having good people around you, besides being able to talk to you. Henry is keeping me up-to-date on audio analysis and reality capture stuff, and Justin's a stud with HVEDRs and everything black box data, and Eric Hunter's same thing with HVE and vehicle sim stuff. Just having good people that you can bounce ideas off of you is the only way to do this, right? Otherwise, I don't know. How many times do I text you with some stupid question at random times of the day like, "Hey, what do you think of this?" Just being able to do that.

Lou (01:45:09):

Same here, man. Yeah, I really appreciate having you there, like I said at the top of the show, to shoot you a text. Invariably, you have a legit answer. I'm like, "How the heck? I know he's out in the field right now. How in the world were you able to answer that off the top of your head?"

Jeff (01:45:27):

Oh, Dropbox. Dropbox saves me. I can pull the paper up on my phone, and look at it real quick on the side of the highway. Yeah, I'll shout out to Dropbox for that.

Lou (01:45:37):

Yeah, I'm with you. We're using Google Drive and just to be able to have your entire database of everything, casework papers, wherever you're going is awesome. I could get a whole new laptop while I'm on the road and be up and running with everything. Well, except for sim and stuff. What sim are you using, by the way? You mentioned PC-Crash, obviously HVE. Are you using Virtual CRASH as well?

Jeff (01:45:58):

We're not using Virtual CRASH, just PC-Crash and HVE. Virtual CRASH seems nice for obviously the visuals. It just seems like another thing to keep up with that I don't really have time for.

Lou (01:46:14):

You got to pick one and I mean, the physics of PC-Crash and Virtual CRASH are nearly identical. It becomes the UI and the features and you got to pick one, man. I'm an HVE guy and a Virtual CRASH guy, and the only reason I would ever have PC-Crash is just to check somebody else's file out. But I'll usually just call friends like you and just say, "Hey, man, can we screen share and go over this guy's file?" All right. A few rapid fire questions. I know you got a hard out in about 13 minutes, so we'll fire through some of these. What's your go-to camera right now? Is it the 7S II?

Jeff (01:46:53):

For night stuff, yeah. Otherwise I prefer a Nikon in the day for what it's worth.

Lou (01:46:58):

Yeah, I do too. I have a D750.

Jeff (01:47:00):

Yeah. Same camera.

Lou (01:47:02):

There we go. It's a good camera. Battery lasts forever. It's easy to handle.

Jeff (01:47:06):

Right. I don't want to carry four batteries. If the one battery is going to last me all day or multiple, that's-

Lou (01:47:12):

I was going to say, last year you were a Top 10 Lightpoint customer. I really appreciate that. What are you using the point clouds for, for the most part? It's a question I never ask my customers and I really should.

Jeff (01:47:25):

Most of the time, it's photogrammetry. We're bringing it into PhotoModeler and doing something for some photographs to determine crush. I know you now sell those mesh models and work with Tony for those. We'll use those for the animation side. I don't think we've really used the mesh model so much for... Oh, we've used them for some simulation stuff too. I don't know that we've really used it for photogrammetry so much. It's maybe a little overkill for just photogrammetry.

Lou (01:47:53):

Yeah, for sure. Yeah.

Jeff (01:47:55):

Yeah. Crush. Mostly crush. Yeah, I'd say.

Lou (01:48:01):

So when you're doing reverse projection photogrammetry for video analysis, how are you determining the position of the vehicle? Are you tracing tires or are you using point clouds for that too, to align with the pixels?

Jeff (01:48:12):

Normally the point cloud, trying to align the whole point cloud to those pixels. Sometimes if the video or whatever is really terrible, we'll draw a plane and trace what we can see for the tire on the roadway. But if we can use the whole point cloud, we'll do that.

Lou (01:48:29):

Same here. I just prefer to have that full fit. Then like we said, then you get the exhibit to go along with it to show to the jury, which is worth a lot. I don't think there's going to be a big difference between the two results, but there's a lot more constraints when you're trying to fit the whole point cloud as opposed to just picking two points. This one could be tough, but I pre-sent it to you, so let's see what we got. Best investment under $5k.

Jeff (01:48:55):

So, I already said it. It's that Sony camera. Absolutely. Like I said, the first one I got was like 4,800 bucks with a nice lens and the body. I use it for all my night inspections. I use it with Nitere. I use it for everything. Everything night, or when that matters.

Lou (01:49:15):

Man, I'm a Sony fan too. It's funny because like we both just said, we're using the Nikon's for general inspection stuff. But yesterday I wanted to take some beautiful photos while I was doing testing, and I didn't want to disassemble this whole thing. I'm shooting this right now on a Sony FX3, which is the video version of the a7S III. The camera's amazing, but I didn't want to disassemble it, especially when I had to come in first thing, and first thing to me is 10:00 AM, by the way, Jeff, and do the podcast. I was like, I'm not going to disassemble it. So I took out my a6000 with a Sigma 30 millimeter lens that goes to 1.4. Even that thing creates beautiful images.

(01:49:52):

I'm a big Sony fan and that thing's cheap, real cheap. I don't know if I sent you this one, for the listener by the way, I didn't send Jeff hardly anything, but these single questions that might be tough to answer on the spot, I gave him a day to prep, which is probably not enough time since he's running a business and a family, but most used paper. What paper do you find yourself referencing the most?

Jeff (01:50:16):

So you did send me that and I struggle very much because it's different for everything. If we wanted to talk about general human factor stuff, MuttarT has a 2005 paper from SAE about driver decisions based on video recorded crashes. Steer left, steer right, brake, use the horn, those kinds of things. I use that all the time because so many times someone says, "Oh, the guy could've just swerved to the right." It's like, well, they're in the right lane. No one's going to swerve into the shoulder or off the road, or into the parked cars. I use that paper a lot.

Lou (01:50:55):

The other thing I like about that paper is that it talks about the psychology of making an intuitive decision versus an algorithmic decision. I think it does, and you can't really have time to do that, to evaluate every potential situation and make a decision based on it there. So generally it's attorneys, my own client that asks me like, "Hey, could you say if he had just swerved to the right?" I'm like, "Well, that's not really fair to Monday morning quarterback it like that and here's why." Then the research from Daniel Kahneman I think is a big part of that too. If you Google that, it'll pop up. But Thinking Fast and Slow I think is his book, where he talks a lot about that, and won a Nobel Prize. But yeah, so if it's a different topic, sorry.

Jeff (01:51:42):

No, so I was thinking for recon, Cipriani's restitution paper I use a lot.

Lou (01:51:49):

Oh yeah.

Jeff (01:51:51):

And Sadal's stiffness coefficients we use a lot when there's nothing published. Then for night video and photographs, I use my paper. I don't have a better answer for that, but yeah, I use my paper on how to display night videos and photographs.

Lou (01:52:10):

Cool. I'll try to link all of those in the show notes there. The Cipriani one is Coefficient Restitution and Low Speed Impacts, and he correlates closing speed to coefficient restitution with an ugly equation, but not too hard to use. Yeah, so I'll reference those. That is a tough question, especially when you work in so many different disciplines of the field, like you do. One crash for you, like you said, while it might be a premise case where somebody trips on something, then you might be analyzing nighttime visibility. Then you might be doing a rollover, a motorcycle crash or a low speeder, you got a lot to handle.

Jeff (01:52:45):

Yeah, there's certain papers on premise cases I use a lot, like where people typically look when they're walking. Yeah, there are a lot of papers.

Lou (01:52:59):

Yeah. I was writing the Northwestern chapter on motorcycle recon, and I finished and it was like 110 references or something just to write a chapter. Not even a whole book on it. That's really, I think, one of the keys to being a good reconstructionist, is to know what literature is out there for you to reference. Because if you can find a paper that's really on point, then it's much more powerful than something that's a little off to the side. Final quick, rapid fire question is what are you most excited about? Or not most excited about. What are you excited about with respect to the future of recon and the evolution of the way things are heading?

Jeff (01:53:45):

I mean, I think it's just going to keep building on what we have, on ADAS systems or automatic emergency braking, getting video for every crash. If every vehicle's equipped with cameras, whether it's still images or video, that just becomes a lot of data to know very specifically. There's going to be less and less gray area, which I think is just good for everyone. I think the interesting part with AEB or ADAS systems is what that means for the driver too on the human factor side. We have data that the vehicle was braking or we have external video showing the vehicle started braking. Well, did the driver do that or did the car do that? It starts to become really interesting.

Lou (01:54:34):

Yeah, especially right now, we're at a period where that is not reported a lot of the times in the CDR report. We see that the vehicle is braking, but we don't know all the time if it was the operator or the vehicle itself that had that response. I normally would ask if there's anything else you'd like to discuss, but we got four minutes, so I'm not going to do that. But if people want to keep up with what you're putting out, where's the best place to find Jeff Suway?

Jeff (01:55:07):

I try and keep stuff on LinkedIn in general, just, "Hey, this is a paper we worked on," or something, but it doesn't get updated very much. I'm not very good at social media or putting stuff out. But now the Lou Peck Lightpoint podcast and LinkedIn.

Lou (01:55:26):

Anytime you publish something cool, I'll put it out on To The Point as well. I really appreciate. I'm so glad that we met a decade ago, and I appreciate being able to just text you and ask you questions. You're an invaluable resource and I really appreciate you taking the time to come on here and talk shop for a couple of hours. So thanks so much, man.

Jeff (01:55:48):

Yeah, same. Absolute same. Just to be able to like you said, send a text and get an answer that I know is legit is invaluable, and I do very much appreciate that. So yeah, same.

Lou (01:56:01):

Awesome. Thanks, Jeff.

(01:56:05):

Hey, everyone. One more thing before you get back to business, and that is my weekly bite-sized email, To The Point. Would you like to get an email from me every Friday discussing a single tool, paper, method, or update in the community? Past topics have covered Toyota's vehicle control history, including a coverage chart, ADAS, that's advanced driver assistance systems, Tesla vehicle data reports, free video analysis tools, and handheld scanners. If that sounds enjoyable and useful, head to lightpointdata.com/tothepoint to get the very next one.