#147 – Transcript – Absorptive Augmented Actuality

Episode Name: Absorptive Augmented Actuality
Episode: #: 147
Date Aired: 5/27/2013
Guest: Jeri Ellsworth
Episode Link: https://theamphour.com/the-amp-hour-147-absorptive-augmented-actuality/

DJ
I’m Dave Jones from the EEVblog.
CG
I’m Chris Gammell of Chris Gammell’s Analog Life.
Psst … and your line is ..
JE
And I’m Jeri Ellsworth and I missed my cue!
CG
Hey Jeri! She’s back!
DJ
It’s been a long time – what is it … 2 years?
JE
Something like that. Someone on Twitter was saying it has been 100 episodes so it’s not a surprise that I missed my cue to go!
CG
You’re fired Jeri, you’re fired.
DJ
Well thank you very much for joining us because we know that you are super-duper busy.
JE
Oh yes, things are just going nuts over here right now.
DJ
How many hours sleep a day are you getting?
JE
Well, er, we just did a public reveal at Maker Fair and leading up to that we were getting, like 5 or 6 hours a night. It was like 16 and 18 hour days, it was ridiculous. But now that I’ve come back I’ve slept, like, 8 or 9 hours a day for the last 3 days – and took ALL WEEKEND off! It felt weird.

CG
That’s awesome. Luxury.

JE
Yeah. We’ve been going like this for 3-3.5 months.
CG
That really starts to wear on you. That can make you go crazy at certain points.
JE
Yeah, there were times that my partner Rick and I were a little snippy with each other.
DJ
Yeah, but you’re still young and carefree, you can handle it. Come on, suck it up!
JE
Yeah, that’s what I keep telling myself.
CG
So for the 3 or 4 people in the world who didn’t see the article or hear us talking about it last week, could you tell us just about the project you have been working on so that we could have a baseline of what you have been doing?
JE
Absolutely. So I started working with Valve Software (***) about a year and a half ago. And when I got there I was given the mandate to research everything video game-related in the hardware space. So we started looking at input devices, output devices, virtual reality, augmented reality, and a few months into working there I got really excited about augmented reality. And started looking at every type of AR glasses that were out there.
DJ
Why is that – why did AR excite you?
JE
At first it didn’t. I thought it was kind of a limited medium to work with, but as we started researching it a bit more I got excited and thought hey there is a lot of things that you can do in this space and it looks like there might be some possibility for lightweight glasses that someday will be on the market.
DJ
Right.
JE
Yeah, so I really started researching that a lot and then part-way through last year, I kind of stumbled onto this technique that to do AR glasses in a very low-cost way, which instead of trying to put the light directly into your eyes, which is the most common approach, we project from the glasses using little micro projectors to a special surface that is highly directional. It’s called Retro Reflective.
DJ
It’s on the moon. The Apollo 11 put a retro reflector on the moon and regardless of what angle you fire a laser at, it comes back directly to you.
CG
[laughs] Right into your eyeball actually!
DJ
Yeah! It’s great!
JE
Exactly. So we are using this retro reflective material that they use on road signs – regular freeway signs. They look really bright at a distance. It is because they have this retro reflective property to them. So by using that in conjunction with these little tiny projectors on glasses, we can project out to the surface and 90% of the light that leaves the glasses comes back.
So you don’t have to project very many photons out to the surface to get a very bright, vivid image coming back.
DJ
Right.
JE
Upside is the glasses can be extremely small and lightweight. Downside is you just have to roll out the surface before you start playing your video games.
And so that’s what I worked on at Valve, that was my primary project. There’s a lot of challenges around that, so we several of us started working on the head tracker. You have to track the user’s head position, so that was a critical piece of it.
We developed a head tracker – most of this actually happened outside of Valve after getting laid off. We developed a head tracker that uses cell phone image sensors that are super low-cost, like $1-2 image sensors, and then we did some tricks on the optical paths so that we could detect LEDs with a super high precision. So at 3 metres distance we can do sub millimetre accuracy.
DJ
Nice!
JE
Yeah! A funny story around developing that we put this whole camera system together and it was, like, wow this seems pretty accurate. I gave it to Rick my business partner, and he started doing the software side of it and he was like, Jeri, no, I keep seeing noise, I think there’s something wrong with your circuit. So I get up and I start walking over there and he’s like wait a minute. Jump up and down and so I jump up and down … it was just me tapping my feet and rolling my chair!
DJ
Oh no!
CG
It was so accurate that the ground, the table did it. Is that the idea?
JE
Yeah, yeah it was just like vibrations through the floor.
DJ
Wow!
JE
So to have to actually maybe I should take a step back before getting into all the tech. maybe I should explain what you see when you look through the glasses.
CG
Yeah.
JE
And why it’s important to have this super high accuracy head position tracker. So you roll out the surface, it’s just kind of a matt grey surface, and we can project 3D objects onto the surface, below the surface, at the surface or above the surface.
CG
Below the surface, that’s interesting.
JE
Yeah. Say you roll out a matt – imagine like a flat surface – you can project as much as you want onto the surface so you could make it look like a table. There’s this huge void that opens up and extends for infinity. You can have stuff drop down in there. One of our experiences that we show is a block-type game, like a Jenga-type game where blocks are stacked above the surface.
CG
But not Jenga of course, right? That’s what you guys said in the video.
JE
Yeah. Rick called it Jurga. I guess I suggested Jenga and he knew he would get sued into the ground if we used Jenga.
CG
You started a good idea.
JE
So the blocks are rendered above the surface, it looks like they are sitting above the surface.
DJ
You’re not thinking big enough Jeri! [0:55:25]
CG
I was going to say that! I can’t help thinking that eventually someone’s going to look back and say what the hell were they thinking not charging more for this kind of thing? I know that’s my inner business brain talking and it’s like really terrible but honestly you guys are making some really cool stuff and it seems really reasonably priced and I guess that’s a really good way to get the market though. [0:55:49]
JE
Well it’s really early, you know. I don’t want anyone to commit us to that price point. We may discover that we botch our first 3A6 and now we have to charge more or something. [0:55:02]
Well we envisioned folks will have their glasses, they’ll go over to their buddy’s house and they’ll all sit around the table and they’ll hook it up to their cell phone and play together. But if it becomes too expensive for a system then … [0:56:21]
CG
Yeah, it’s the doom of a lot of systems actually. I remember hearing about some of those where they were $600 a system of something and no-one bought it. You can’t get any foothold then. Games are usually where they make the money anyway. [0:56:35]
DJ
Well they reckon that $300 is the consumer price point where if it is under that it is an impulse buy, if it is over that they um and ah. [0:56:51]
CG
I have a question about the cell phones because I was intrigued by you saying you are hooking up to those. Cell phones are notorious for having very slow interfaces to them. How the heck are you going to actually talk to them? Are you planning on doing wireless eventually? [0:57:06]
JE
Oh, so a lot more cell phones and tablets are coming out with H IOUT so that’s what we need to do to explore the video interface. And since all of our tracking is done in hardware it is actually a very, very slow data stream going back. The only thing that we need to be very careful about is that there is very low latency going back. [0:57:29]
One of the challenges in augmented reality is that if there is latency, say from the head tracker, you move your head and it takes 4 or 5 frames to update, it makes it look like all the objects that you are rendering in the world just jumped to the side. They don’t track smoothly. So we have done a lot of work to keep the frame rate up to 120 hertz, the tracker is running at 120 hertz. [0:57:56]
CG
Which doesn’t sound like much but for video that’s a ton. Most videos are what 30 hertz? Is that right? [0:58:03]
DJ
In Australia it’s 25 hertz. [0:58:05]
CG
Does that translate to frames per second? That’s what I’m thinking. [0:58:11]
JE
Yes. [0:58:12]
CG
Because movies are 24, home videos are 29 or 30. [0:58:16]
JE
The real magic started for us over 100 hertz. That’s where it really started feeling like these Jenga blocks or Jerga blocks are on the surface. Along with some predictions. So you can predict where you head is moving. So with prediction and having very low latency it really looks like as you walk around the table that those blocks are really there. When you smash them with the wand it looks like they are really flying up towards your face. [0:58:45]
CG
Why 100 hertz then? Is that just based on human reaction times? [0:58:50]
JE
I’m not really sure. We tried 60 hertz, that was pretty good, but getting over 100 hertz really looked good. And so all the demos that we presented at Maker Fair were at least over 100 hertz. [0:59:05]
And modern cell phones are getting really powerful. There is some hot stuff in the pipeline coming out for chips. [0:59:15]
CG
So you’re saying that data going back down the pipe to the cell phone – you can’t talk through HDMI can you? How are you actually inputting data? That’s what I’m always not sure about. There’s no way to hook to data. [0:59:31]
DJ
You’ve got to have a standardized interface. For these glasses to work with any gaming system are you looking to provide an interface for each system or are you going to leave that up to the manufacturers? [0:59:43]
JE
Many cell phones these days have USB on the go and they have digital video out, so our plan is to use USB and HDMI out. [0:59:56]
CG
OK. I didn’t know that cell phones were using USB on the go. Because that’s the one where it goes, it can be host or device, right? [1:00:05]
JE
Yeah, we’re going to see more and more devices like that. I can describe what the glasses look like – the current ones are kind of big and weigh several hundred grams. Lots of hot glue on them! So we have mocked up, using real projection engines, what we think we can do for the glasses. There are about 60 grams of weight. Then with the serializers we have reduced it down to just a few pairs of micro coax down the cable, so the cable that we are currently using is about what you would think of for, like, ear buds. Like iPod earbuds. It’s got that thickness. And that’s what we demonstrated at Maker Fair. [1:01:00]
And then it goes to a box and this box doesn’t have to be very big but this is where if there is a battery that’s where it would live and there’s some receivers for the HDMI videos. There’s not much going on there. And then from there that’s where the connections go to your PC or your tablet or your phone. [1:01:20]
DJ
Right, so all the processing is still done in the glasses? [1:01:23]
JE
Yeah, and we need it close to the image sensor because we pull the data off the image sensor and we process it immediately and we get the latency really low so it is right up next to the image sensor, just screaming along as fast as it can go. [1:01:36]
DJ
Yeah, you can’t send it down a couple of metres of cable, right? [1:01:38]
JE
No! Unfortunately not. [1:01:39]
CG
Yeah. Cable gets real expensive then. Even then, even with optical, you wouldn’t be able to do it, there’s still latency. [1:01:49]
JE
Yeah, so there’s going to be one ASIC that’s sitting up on the glasses that’s doing all this image tracking. And that just sends point data down to the cell phone. That’s also a huge thing, doing most of the processing on the ASIC. The phones are still a little bit anaemic as far as processing is concerned. We can’t have them processing entire frames of video, there would just be no way to do it. It has to be done up on the glasses. [1:02:19]
DJ
Ultimately what’s the data rate coming back from the glasses after it gets out of the ASIC? You said it’s low but is it kilobits, is it megabits? [1:02:27]
JE
On our current prototype we are just using 115 2 RS – we are not even close to saturating that. [1:02:38]
DJ
So serial? Right. [1:02:39]
JE
Yeah, and we’re not even close to saturating that. [02:40]
DJ
Right, old school. ASCI position. Great. I love it. [1:02:46]
JE
Yeah, we may just leave that exposed, who knows? We were talking to some folks at Maker Fair who were looking at our tracker and they want to do robotics with that. We said oh sure, maybe the tracker we will spit out and let people do other stuff with it. [1:03:03]
CG
Number one feature request in Maker Fair is hackable, right? That seems to fit. [1:03:08]
JE
Well if anyone knows my projects I always try to make them hackable. This one will certainly have Easter eggs. [1:03:18]
CG
Yes, we need to ask – are there any fart jokes in this one that we should expect? [1:03:22]
JE
Ah … well [1:03:27]
CG
Like you hit a Jenga block and it melts into a puke or something?! [1:03:28]
JE
You’ll just have to come to the Kickstarter so you can see! [1:03:35]
DJ
And join one of the road shows?! [1:03:38]
JE
This will be different though. I was always getting myself in trouble with other companies by adding Easter eggs, but it is Rick and I’s company so it is just us. [1:03:48]
CG
I’m the boss! [1:03:50]
JE
Yeah! [1:03:52]
DJ
So do you think you will have to bring on other software people? Because there’s an awful lot of software – if you want to develop real apps and real demo games and stuff like that, that is a lot of work. I would suspect, being a non game programmer … [1:04:08]
JE
You’re absolutely right. As soon as we get funding we are going to have to start staffing up. We’re staffing up right now but it is coming out of our own pocket so it is kind of limited. But this will turn into more of a software project than a hardware project in the end. Just like what you said earlier, the revenue streams will probably come from the software more than hardware. [1:04:35]
CG
Yeah, that’s just the game industry though. [1:04:36]
JE
Our hope is to offer a place to curate other people’s stuff. We want people to make money on this. We want people to make things and give it away. [1:04:48]
CG
Like a marketplace kind of thing? [1:04:49]
JE
Yeah. At least direct people where they can go and get it. We don’t want to be a walled-off garden like some people – Apple and some of these places where I can’t even … if I write a program I can’t even give it to you guys I have to go through Apple to deliver it. We hope to be more open than that. [1:05:09]
DJ
Yeah. That’s good. I think that’ll pay dividends. [1:05:13]
CG
Once you get funding is this like Jeri hops on a plane? Are you going to go to China for a while do you think? Is that looking too far forward? [1:05:27]
JE
Well, I’m accustomed to doing manufacturing in China and I’d like to do it domestically but I just don’t know if that’s possible yet. That’s something we’re going to look at. [1:05:41]
CG
I wasn’t even asking about that. I expect people to go to China, honestly. I mean, like, that’s kind of the Kickstarter thing, obviously you have done it before but that would be the different point from a lot of Kickstarter projects. People are shocked by having to go to China. You’re like, oh I have to go back to China. [1:06:00]
JE
Actually I have been kind of following a bunch of Kickstarters over the last few months and it’s kind of a pitfall for some of those guys. They’re like, we’re going to make this thing, but they’ve never done it before and they find out that China’s hard! [1:06:16]
DJ
Even if you’ve done it before! I mean, things can go wrong and you can be back to square one and oh damn. Not quite square one but just little things can mess things up. [1:06:30]
JE
Yeah, that’s why we suspect that we are going to take secondary funding besides Kickstarter. We suspect, especially since Maker Fair, and how people were really positive about it, that we will probably do fairly well on Kickstarter. [1:06:42]
CG
Oh hell, yeah! [1:06:44]
DJ
I guarantee you will! [1:06:47]
JE
But the software side is what’s really going to make it magical and we’ll probably have to take on some investment from VCs. It will be interesting, not many people have gone down this route before. So it will be interesting to see how VCs react to us saying we have got so many hundred thousand in the bank or whatever that money we get from Kickstarter is. It would be sure nice to have another 5 million. [1:07:19]
CG
Yeah! I could get a lot of chips for that! [1:07:22]
DJ
I’ll trade you all my C64 chips, you can tell them that, right? Use that as collateral. [1:07:28]
JE
Oh there you go! My retirement fund is on my C64 chips. [1:07:32]
CG
Maybe while you’re over there you can get them all packaged up too. You can be since I’ve already done the sweetheart deal we’ve got all these chips you can just package those old C64 chips. [1:07:46]
JE
I’d love to do something with those someday. But it looks like it’s going to be off in the future somewhere. [1:07:51]
CG
Oh yeah. You’re going to be busy. I’m guessing this is going to be the last time we talk to you for a couple of years. [1:07:56]
JE
Maybe. It’s going to be pretty crazy. It’s going to be fun though. [1:08:01]
CG
Good crazy. [1:08:02]
JE
It’s interesting being on this side of a start up. I’ve done contract work for a good part of a decade and walked into a lot of start ups. Been number 5 or 6 person in the door and we’re starting from scratch and they’re just getting their funding. Seeing in from that perspective, now reversing the roles and being the one out trying to scrounge up the funding is quite different. [1:08:28]
Not only do you have to do all the design work, you have to do the funding part of it too. [1:08:32]
CG
I think you’re doing it right though, I mean having the prototypes to show is ahead of a lot of start ups. A lot of them are vapour ware, not a lot of them, but some of them are. You’re just trying to go on an idea first because you don’t have the money to even bootstrap the stuff and the fact that you guys are bootstrapping the early protos, that’s a really good sign. [1:08:53]
JE
Yeah, Rick and I thought that was very important, that we get it out there really early and show that it’s real. Because there is so much vapour ware out there and especially around gaming hardware. We were talking to some folks that wanted to look at it. They were really very dubious because we were trying to get some meetings going, we were like hey guys this is what it is. Stop by the Maker Fair booth and kick the tires. They were wanting us to do some videos and supply them with a bunch of stuff. We said just come and take a look at it, it’s real! [1:09:35]
And it was pretty nice when they came out and tried it for the first time. They were like, holy, woo. [1:09:44]
CG
They were probably pissed that they didn’t get to see it privately first but, you know, that’s a good thing. That’s really different, that’s cool.
I have to say I want you to be careful with VCs, they seem, I don’t know. It seems like some of them are really good but some of them just, meh. [1:10:00]
DJ
ALL of them! Come on! [1:10:01]
CG
Not all of them. Some of them do good work. There’s some good ones out there, but some are just … [1:10:06]
DJ
But ultimately it doesn’t matter who they are they are in it for the money. They want to make a massive profit for their risk. End of story. [1:10:16]
CG
Yes, you’re not wrong. [1:10:17]
JE
We have all heard the horror stories. [1:10:20]
DJ
There are no nice VCs, they’re all the same. Some are nicer people than others, but ultimately no they invest in risky stuff and they expect a large return for that large amount of risk. And they will pressure you into doing all sorts of things to achieve that end. [1:10:38]
JE
And they are very good a lot of times to give you just enough to get you on the edge of success and then when you are desperate for that last little bit of money then it’s like, well for another 50% of your company we will … [1:10:54]
DJ
Yes. And you’re left holding the bag. [1:10:57]
JE
I’ve certainly seen that. [1:10:58]
CG
That’s no fun. So be careful! [1:11:01]
JE
Well that’s why I’m excited about the crowd sourcing. We want to show this is for real. We have this credibility. We’ve done this before. It works. [1:11:15]
DJ
But you’ve got the pressure to produce then if you get 10,000 people backing it then you’ve got 10,000 people screaming at you, where’s my glasses?! [1:11:23]
JE
Which is kind of a scary thing. The production runs that I’m used to are in the hundreds of thousands in toys. [1:11:35]
DJ
This will be under 10K. [1:11:36]
JE
Yeah, I mean I don’t know in the history of Kickstarter if there’s any projects that have sold more than like 10K but that’s probably pretty typical for a Kickstarter that’s very successful. [1:11:49]
DJ
It’s usually a couple of thousand I have been seeing for hardware-type projects. But because this is a gamer and gamers go crazy about this sort of stuff you could reach 5 figures. [1:11:58]
CG
One was 60,000 watches I think. That was a lower cost as well. [1:12:05]
JE
So we are crunching a bunch of numbers. It’s scenarios, what if. What if only 10 people order, what if 100,000 order. We want to try to be prepared. [1:12:32]
DJ
Yep, it’s important because that changes the entire game, doesn’t it? In terms of manufacturing and design for manufacturing. Really. [1:12:40]
JE
Exactly. [1:12:41]
DJ
I mean you may have to make some large design changes based on how many orders you get. [1:12:50]
JE
It’s unfortunate that the difference between NRE for manufacturing 1,000 units versus 100,000 units is not much. It’s all the same amount of work. [1:13:05]
CG
Yep. And then they want to charge you again when you go back after that first 1,000. They’re like, oh you want another 1,000? There’s a fee. [1:13:15]
JE
Yeah. [1:13:16]
CG
You could just charge $1,500 per prototype effectively. Suckers like me might pay it. [1:13:25]
DJ
You’ve been suckered, yes Chris. You suckered into winning. [1:13:31]
CG
Yeah, you should have people winning Jeri. [1:13:33]
JE
I missed that, was it the glass you had to win? [1:13:37]
CG
Yeah, you had to do stuff on Twitter and Google Plus and if you said what you were going to use it for they would pick something. I just put some random stuff on there, building hardware around it, and I got yeah you won. But it didn’t matter, a lot of people said I would use it to video record stuff. Anyways. [1:13:55]
DJ
But by ‘winning’ they meant you get the opportunity to buy it first for $1,500. You don’t actually win anything, you just win the opportunity to buy it. [1:14:03]
JE
Gotcha.
Now you’ve brought up Google Glass, the number one thing that we get asked is what’s the difference between this and Google Glass. [1:15:12]
CG
Everything, right? The similarity starts at the Pico projector and ends at the Pico projector. [1:15:19]
JE
Yeah, pretty much. [1:14:20]
DJ
Oh goodness. Do you actually have any competitors in this space at the moment? Are there any other Kickstarter projects or start ups or people doing similar? [1:14:34]
JE
I think Occulis is probably the adjacent, they are a different experience but there’s a lot of overlap so that’s pretty cool. There are some consumer products out there but they are all really expensive. They are usually typically the collimated displays, set to infinity and they project directly into your eyes. They are usually a very limited field of view. Quite different. [1:15:04]
The best thing that we saw there was like a Korean company, I forget the name of it, but they had kind of heavy but pretty wide field of view glasses that worked fairly well. But they had no tracking, no audio on them. [1:15:21]
Oh the audio, oh boy, I can’t wait. It’s the thing I’m the most excited about since we have head position we can do all kinds of fun things with audio. You can get little zombies standing on the table, you could put your head right down next to the zombie and hear him. [1:15:38]
CG
You think he’s whispering or something? [1:15:40]
DJ
That’s awesome! [1:15:44]
JE
I suspect that there’s going to be entire games that just used head position and sound and almost no graphics at all. [1:15:24]
DJ
Yeah I can imagine that. [1:15:54]
CG
Oh man, like where in the world is Carmen Sandiago? You could totally replicate that. All the board games of our youth you could totally replicate those, which is so much fun. Like Risk, oh my God, Risk could be amazing. [1:16:06]
What about the tracking though. You said you have the vision tracking, you are actually looking at these modulated LEDs and stuff. I do have a question about that. Are you using like a 9 axis controller, is that the idea for actually moving your head and everything where it streams out data? [1:16:30]
JE
Yeah. Sure, I can talk about it. It sends out points which are just X Y and some other magic source that we have in there and that goes out to the host, which then solves for world position so there’s some math behind that that figures out where you are in the world from all these points that are coming down. That’s pretty much it. [1:16:54]
So if you rotate your head we can detect the rotation. If you move laterally we can pick that up. Anywhere within a volume we can figure out where you are at. [1:17:05]
DJ
And this is just done with the IR? [1:17:09}
JE
Yes, IR LEDs. [1:17:12]
DJ
Right, OK. So it is just IR transmitting from the surface and then the … [1:17:17]
JE
The camera is on the glasses. [1:17:19]
DJ
What type of sensor? Right, you’ve just got a camera. How many pixels? [1:17:22]
JE
It really doesn’t matter on the pixels all that much. It does affect the precision. We’re just using a standard HD image sensor right now from a cell phone.
It’s a really crappy rolling shutter image sensor which there was a lot of debate within Valve whether it would work and we never got that far within the company but as soon as we got out, we were like alright, they’ve got their tech, we can’t use that, so we’ve gotta do our own. So we went down this other path and holy smokes, it worked! [1:17:54]
CG
So you don’t have external like an IMU type of chip externally? [1:18:03]
JE
No giros, nothing. [1:18:06]
DJ
It’s just the angles and math? That’s it. [1:18:11]
CG
Lots of math! [1:18:12]
DJ
Yeah, lots of math. Well basic trigonometry right? [1:18:18]
CG
The other question I have is so the camera is only looking at the IR LEDs is that the idea? Is that what it does? [1:18:26]
JE
The reason we didn’t use IMUs, they are pretty good at very short timespans like they know that you moved but they are not good at anything, they drift extremely badly. [1:18:43]
You can’t do any kind of translation with them, they can’t figure out if you moved laterally. You can figure out rotation and you can use gravity to figure out that. There’s magnetometers which you can kind of figure out rotation with but … [1:19:01]
DJ
Now it’s getting ugly. [1:19:02]
JE
Yeah, it’s kind of a dead en d path using that. It’s not accurate. [1:19:06]
CG
Yeah, you need to filter the crap out of them. My buddy did a project on that and basically the math on that is accelerometers you need to basically integrate them twice and then if you think about any jitter in the acceleration then that translates in to huge errors in your integration. [1:19:23]
DJ
As Jeri said it’s a complete dead end. You just cannot make it work. Simple as that. [1:19:29]
JE
Yeah, it works great for cell phones and things like that. [1:19:32]
DJ
Oh yeah, for those limited things. [1:19:33]
JE
Where you don’t want things to look like they are really in the world but if you want to have something around your figurine and you want things locked to the figurine you have gotta have absolute measurements so that’s why were’ doing pure optical.
So the majority of our time after leaving Valve was just refining this tracker and that’s what’s really special about our system. At Valve we had a system that was like $20-30,000, it was a really military grade tracking system. We are achieving that with $8 worth of parts! [1:20:06]
CG
That is magical! Really impressive Jeri. [1:20:10]
JE
It surprised me too that we got that kind of accuracy. We just kept pushing it and we kept fussing with the optics and it just got better. Oh, hey it just got better.
Our very first experiments, the jitter was so bad. We were doing full frame captures, sending it down to the PC. The PC was looking at these paper markers. The result looked like you were in a bouncy car or something. These little characters were jumping all over and bouncing and now it is just rock solid. You can get down and really close to them. There’s just a tiny bit of jitter but it is acceptable. [1:20:55]
CG
OK. So another question. With the IR LEDs that’s how you track it. You said you can move down towards the characters and you can they can whisper to you or whatever. As you move towards the surface do you ever lose track of the LEDs? Do things get in the way depending how you move your head? Do you lose track of absolute position because you aren’t pointed right at the other user? What happens then? [1:21:29]
JE
Right now we have one cluster of LEDs, just mainly because we didn’t have time to work on putting them all around the edge. So we have over 100 degree field of view on our camera and so it is really wide so you can be way off axis and still track. But if you put your hand in front of the camera, which is dead centre between your eyes, you can lose track that way. You can get within inches of the surface and currently with only one cluster of LEDs on there you lose track, like 4 inches away or something. It’s pretty good. Definitely needs some improvement there as far as having robustness by having more LEDs around the surface. [1:22:11]
CG
OK. So as the playing field evolves you might be able to change that? [1:22:16]
JE
You need a certain number of them, you need about 5 points to keep track of the surface. So you get below that and things go bad really fast. [1:22:31]
DJ
So your key technology here is the imaging, ASIC and processing? [1:22:39]
JE
That’s it! Actually, with all the work, it’s been like a year and a half worth of work. We looked at everything, laser projectors right in the eye, all the display tech. In the end the display tech ended up being the simplest thing. It was just dual Pico projectors or a stereo image. It all boiled down to $8 worth of electronics for the tracker. That’s the magic. It’s like get the total building materials down to a point where we can sell it for this low cost. [1:23:16]
So yeah, it’s disappointingly simple in the end! [1:23:21]
CG
No, there’s a lot of magic in there. That’s a lot of math. It sounds like a lot of math, math. [1:23:29]
DJ
No, but as a start up you’ve got to have a sellable core technology. That’s quite important, not just oh we got our costs down to $8 or whatever. You said it before, a lot of people could use this in other areas, not just glasses and gaming. They could use it for other types of tracking. That’s where you could potentially sell the imaging tracking ASIC for example. [1:23:39]
JE
Exactly. Who knows where it’s going to extend itself beyond. And we hope to be on the gaming side of the revenue stream too. We think we’re creative and have good game idea. [1:24:12]
CG
Definitely. [1:24:14]
DJ
So does everyone though! But gaming, you know, how many bloody games are there? It’s incredible. [1:24:21]
CG
I’m amazed that this market continues to be so big. I shouldn’t be, but I grew up with a lot of games around me. It is big though, people just keep spending money on it. [1:24:33]
JE
It’s a dream that people have painted since we were kids on Star Wars and finally all the tech is here so we can have a comfortable non headache-inducing experience. People have tried it in the past and haven’t quite been able to pull it off with the tech. With cell phone cameras pushing everything forwards and micro projectors for cell phones. [1:24:57]
DJ
The question remains, though, will people just go wow for a little while and then cast it aside – see what I did there?! Or will it be a long-term thing? That’s the question. Will it be long-term or will they just play this is fun, this is fantastic, wow, but I don’t want to play it all day kind of thing. [1:25:22]
JE
Well personally I’m biased of course but I think people are looking for the next gaming experience that’s more sociable. You see board games are really hot, Warhammer’s really hot, all these figurine games. [1:25:40]
DJ
They’re very credibly low tech you see? They are old school, right? Rolling dice and stuff. [1:25:45]
JE
Exactly, but for me personally I would love to play like a Warhammer figurine-type game but my friends have been doing it since they were in High School and they understand it and I don’t. The hurdle of trying to figure out all the stuff about it keeps me from doing it. [1:26:07]
DJ
There’s a lot of culture involved in that. It’s not just the game itself. [1:26:13]
JE
If I don’t have to set up for 2 hours, if I can within 30 seconds roll a mat out and start playing some kind of figurine game with my glasses, that’s very appealing to me. [1:26:24]
CG
Are we going to see this in a pinball game sometime soon too? [1:26:30]
JE
Yeah, we should hook it up to visual pinball. Why don’t we throw in the home chip lab too? We could use it to monitor. [1:26:38]
CG
Oh Jeri, you cut me so deep! You’re supposed to be on my side here. [1:26:45]
DJ
All of our new audience today who are listening, all the gamers who have no idea what we are talking about, and we weren’t even trying to explain. It’s a bit of an in-joke here on the show. [1:26:53]
CG
Yes, come back and you’ll hear Dave rant on it for ever. How wrong I am.
Jeri, once this is done you have made your gazillions can you go off and fix that and make that happen next? [1:27:07]
JE
Absolutely! [1:27:09]
CG
Thank you Jeri! [1:27:10]
JE
I’m so happy to be back. You know, working at Valve I had to be so secretive. [1:27:18]
DJ
I see, outside of Valve? Right. [1:27:19]
JE
It was kind of like radio silence. All I could do was go on Twitter every once in a while. I was like I’m working on something cool, and I couldn’t talk about it. Really I’m a nerd at heart, I love coming on shows like yours and talking about what board rate we are using to communicate point data down to the host. Oh yeah, that’s the communication I should have been having all the time. [1:27:45]
DJ
That’s the problem with all large corporations like this, they really don’t like you doing stuff outside or talking about the company ‘s stuff. [1:27:54]
JE
It’s risky. [1:27:55]
DJ
Is it more secretive at Valve? [1:28:00]
CG
She can’t say, it’s secretive. [1:28:01]
DJ
The first day do you have to sign a form saying I will not say anything on Twitter? [1:28:12]
JE
Well Valve is far more open than other companies, it encourages the employees to talk to their fanbase. Unfortunately it’s no management in there to guide us so it becomes a debate, it’s like I just said something on a podcast and your colleagues are like oh my God you just revealed something. It’s like, you’re more accountable to your colleagues and then it becomes awkward trying to talk about stuff. [1:28:48]
CG
Oh there’s the talker over there! Let’s not sit with her at lunch today. [1:28:52]
DJ
So did you feel bad about that, not being able to a) release any of and b) have any time outside of Valve to work on your own stuff that interests you? [1:29:06]
JE
Going to Valve I took it very seriously. Building that team. So I put in insane amount of hours, it was incredible the amount of hours I was putting in trying to build the team up and work on these projects also. So like I said earlier, often we would recruit during the day, have our various meetings during the day, and then we would work way into the night because we were all passionate about building things. [1:29:36]
DJ
How many people did you end up with in the hardware team? Or is that proprietary company information? [1:29:43]
JE
I probably shouldn’t say that. [1:29:44]
DJ
Yeah, probably. [1:29:48]
JE
They’re all my friends. [1:29:49]
DJ
Well we know that. There were like 4 who were on this show. [1:29:55]
JE
I wish them all luck. They’re all my friends in there I want them to succeed. There is some cool stuff they’re working on and I hope it makes it out there because I want to own some of the stuff they are working on.
A lot of it has my DNA. [1:30:12]
DJ
Even if it is the competing VR. [1:30:15]
CG
So you are in charge now, you and Rick? You like have a company now. Do you have to be, like, a finance person too or what? [1:30:25]
JE
Rick is officially the Treasurer and Vice President. I am Madam President and … we had to divide this stuff up. [1:30:43]
It is really small. It’s actually really fun there’s like 5 of us working on it pretty much full-time. It feels very much family-like. It’s all in Rick’s living room. He has this big family living room. The day after we were let go I grabbed up all of my scopes and my laser cutter and we were like, dammit we are going to do this thing, I don’t care! [1:31:14]
CG
Sleepover! [1:31:15]
JE
Yeah! [1:31:16]
DJ
So you stole all the company labs did you? [1:31:21]
JE
My own stuff! [1:31:23]
DJ
Rick do you have room for 80 pinball machines? [1:31:26]
JE
I had a bunch of pinball machines at Valve, they are all sitting in his entryway right now because that’s as far as they made it. They’re all folded up. [1:31:38]
CG
They’re not active? [1:31:40]
JE
They’re not, sadly, there’s no room. So there’s 5 of us working out of his house. The laser cutter’s here, I have my little mini lathe I have to turn a few things to do some of the optics and we are all just piled in his living room. So it feels really family-like because it will get to be about 7 or 8 o’clock at night and we will cook some food and we’ll sit around and watch – we’ve got a DVD set of Scrubs. [1:32:09]
CG
Nice. I love that show. That got me through Korea! [1:32:14]
JE
So we end up watching a couple of those a night. It’s great. The creativity is awesome because we are all on top of each other. And that was one of the great things about Valve, they encourage people to be very close to each other, like artists, we’re close to programmers, and electrical engineers near mechanical engineers and stuff. And that’s exactly how we’re structuring our group. It’s like everyone’s just right there in the mix and we hope to continue you that after funding. We’ll get a real office and staff up. [1:32:47]
CG
It is interesting how that stuff changes over time. [1:32:52]
DJ
You pretty much have to devote your life to this, don’t you? You can’t really just do it, punch in 9-5. [1:32:58]
JE
We could but it would take forever. I used to always say at Valve. The team was so small there that each of us were doing the work of 5 engineers. It is still the same outside in our group. It’s like Rick is doing amazing stuff with the software, I can’t believe how much software he churns out but it’s taking him 16-18 hours a day to do it. [1:33:24]
DJ
Jeez. To someone who actually has a life outside of my blog, yeah. [1:33:31]
JE
It’s all temporary though. We know that there’s an angle here of Kickstarter and after that then we can get real staff and we can ease the load up and maybe we can do some 9-5s every once in a while. [1:33:44]
CG
Yeah, or take a day off. You just had a day off, right? [1:33:48]
JE
I did! I took all weekend off. It felt weird. So the first day I’m like, oh this is so awesome, I just slept 8 hours. And then by Saturday I was like, I had to grab my laptop and start working. By Sunday I’m like, well, we agreed that we wouldn’t work on it so I can’t go over to Rick’s and I have kind of done everything I can on my laptop. I was getting pretty antsy, so I’m glad to be back. [1:34:15]
DJ
Because time off is more important than just doing a continuous 9-5 because anyone who is developing, any engineer who is developing something, knows that if you just go oh I’ve gotta stop now when you’re on a roll you can miss some good stuff. You’ve got to reset yourself for the next day or the next week and it takes you ages to get back into it. [1:34:40]
JE
Exactly. But this is such a fun space, it is so fun to come in and try these experiments as far as user interface. It is like sometimes we just put LEDs on a box and we use that for a prop and move it around the surface. It is like oh my God look at these lemmings are getting crushed by a box. [1:35:00]
CG
That sounds really fun. [1:35:03]
jE
It’s really fun and super hard work. [1:35:05]
CG
What about timelines? Are we saying Kickstarter in a month, 3 months, what are you thinking? [1:35:11]
JE
We really don’t know. We want to make sure we have our act together. Again the game experiences and we want to be damn sure we can hit the price point. So we have to actually start doing the dance with vendors right now. The dance of here’s our projected volume. It is so tedious and annoying. [1:35:31]
CG
Oh God, I hate that stuff. They want all the details up front. You are like, I don’t know, this might not sell. [1:35:39]
JE
Exactly. So that’s part of my task, to do the vendor dance! [1:35:47]
DJ
What about a better prototype? Because if you show that sort of prototype on the Kickstarter video people might go oh I don’t know. Because you have to show your real prototype, that’s part of the Kickstarter rules now. You can’t just do that nice 3D rendered product video anymore. [1:36:09]
JE
Yeah, I think we can ride on our old prototype for a bit. We have some parts being assembled for the next revision, which is going to be closer to the 60 gram weight prototype. We’re not going to hit the 60 grams in the next one. [1:36:27]
DJ
No, but it will look more self-contained and professional, right? [1:36:31]
JE
Right, I’ve gotta do a new board layout. Right now I didn’t really try to miniaturize the boards too much so they were just 2-layer boards with a TQFP144FPGA on them. So they are huge. It is pretty amazing, we jammed them down into this size. So now I’ve got to do a multi-layer board and put a micro FPGA down there and hopefully go to see if I can talk to one of these companies that have an FPGA that will run cooler than the Altera one that we’ve got. [1:37:05]
That was an issue. Right between your eyes is where the FPGA was sitting. [1:37:15]
CG
That’s not a feature Jeri, as you kill zombies you heat up! I can feel the juices splashing on me. [1:37:20]
JE
In the lab it wasn’t too bad because we would turn it on for an hour and turn it off but when we were at Maker Fair it was running all day and it was heat soaking for hours. It wasn’t uncomfortably warm but people commented on it. So we need to show that it is not going to run warm at all. [1:37:45]
Really when we do the shrink down to the metalizer gate array it is going to run way cooler. [1:37:52]
CG
I know. It is amazing how much energy you waste on FPGAs but you just get the flexibility so it is that much better. [1:37:59]
JE
And it is kind of Band Aided together right now. All my RTL is just … I was grabbing an IP that I generated back in the Commodore 64 joystick days. It has a little 6502 hanging out, it’s got my custom pixel pipeline going in there. Nothing is gated so even in the blanking it’s like I’m happily figuring out where points are even though there’s no data there. [1:38:25]
CG
That’s awesome. [1:38:28]
DJ
We have a whole bunch of Reddit questions. [1:38:33]
CG
I think we did a lot of these. [1:38:34]
DJ
Yeah, we did. Are there any we haven’t answered? [1:38:36]
CG
Some people were asking about the source code. I was going to ask about this. Are you going to patent anything? [1:38:44]
DJ
Oh please don’t! [1:38:45]
JE
I’m sorry Dave, we do have some patents! [1:38:48]
DJ
Oh no! [1:38:49]
JE
Yeah, we’re working on patents. [1:38:54]
DJ
No, you evil person. [1:38:56]
JE
I know, but you know we have to also look at the value of our company.
CG
You have to. [1:38:59]
JE
At some point VCs might … what kind of moat do we have around our attack? A patent on our portfolio is part of that. [1:39:12]
CG
Some of it you get to encapsulate in the ASIC and the FPGA right now, right? Some of that stuff is just a case of trade secret, you aren’t just going to give that stuff away anyway. I doubt you are patenting that side of it? [1:39:27]
JE
Yes, some of the algorithms and stuff that we are working on and some of the optics tricks and stuff we are trying to protect that stuff. [1:29:36]
DJ
Right, so it is more technical, not a pattern for AR glasses with 2 Pico projectors? [1:39:45]
JE
No. [1:39:46]
DJ
Well that’s a step in the right direction at least. [1:39:49]
JE
I try not to be an evil person most of the time! [1:39:52]
CG
Yeah. [1:39:56]
JE
We’re not going to open the source. [1:40:03]
DJ
Well you’re going open source in so far as the interface, right?
JE
Yeah, we want people to happily use our stuff. [1:40:09]
DJ
Yeah, but you won’t be able to download your VHDL and Verolog code for your gate array or something like that? Yeah, we want people to happily use our stuff. [1:40:09]40:17]
JE
No, that would probably put us out of business right away. That’s my theory. Yeah, we want people to happily use our stuff. [1:40:09]40:21]
DJ
And it would be pointless too because nobody’s going to want to – the only person who is going to want to develop, take that, is a competitor. Is somebody who wants to put you out of business. Yeah, we want people to happily use our stuff. [1:40:09]40:36]
CG
There’s no point. Yeah, we want people to happily use our stuff. [1:40:09]40:44]
JE
APIs are going to be open. We want people to develop for it. It is a chicken and egg. If there’s no software for it then why buy the hardware? And vice versa. Yeah, we want people to happily use our stuff. [1:40:09]40:56]
DJ
So do you see your revenue stream coming from the hardware? Or do you see like licence rights? There’s gotta be money coming in from somewhere. [1: 41:10]
JE
We see multiple revenue streams. I’m a toy designer, I like to work on some of the props and various things that go along with it. We see selling software and the actual hardware, the glasses themselves, there’s multiple angles that we can get revenue. [1:41:30]
DJ
Sometimes a lot of companies will not make a cent on the hardware, the hardware is basically at cost to enable you to make money on the software and the other spin-off things. [1:41:43]
CG
Or at a loss. [1:41:42]
DJ
Yeah, or at a loss because that may be required. If you find the hardware is more expensive than that magic $300 price point you may have to take loss on the hardware or break even in order to make your money elsewhere. [1:41:59]
JE
Yeah, that’s one of the things we were considering when we were under the umbrella of a bigger company that could do that, but it is a little too scary for us to think about that right now. It think everything’s got to generate a little bit of revenue or at least break us even. [1:42:17]
DJ
Yeah, you wouldn’t be selling it at a loss, you would be at minimum breaking even. [1:42:24]
JE
My bank account is going down quickly putting this together. Not for long! [1:42:28]
CG
Don’t worry, you will make it up on volume Jeri. It’s cool. [1:42:31]
JE
Like the underwear gnomes? [1:42:37]
DJ
It reminds me of the joke of how do you make money on free software? Volume! [1:42:46]
JE
Exactly. [1:42:50]
CG
I think the last question was how do people work for you? Should we set up a Google Docs, what people are willing to do to sell their soul to work for Technical Illusions on the Cast AR system. [1:43:14]
JE
Well, right now we’re being very conservative about letting people come work over here. It is kind of cramped! Folks that are fans and friends have been helping us out here and there on art and sound and programming but until we have funding and a real office it is going to be pretty limited. [1:43:39]
CG
Gotcha. You could pay people in pinball plays right now right? [1:43:43]
JE
Exactly. [1:43:44]
CG
You get 3 quarters but you keep cycling them through the machine. Spend that however you want to. [1:43:51]
JE
I’m really excited about getting our studio up and going because this is it, we get to choose all the coolest people that we want to work with and there’s not going to be the old guard that’s going to come through and mandate things. We get to set the culture right now. [1:44:09]
CG
You’re the old guard! That’s the best part. [1:44:13]
DJ
Seniority. [1:44:18]
JE
Back when I had my computer store it was one of my favourite times in my career. I had my computer stores I had some employees working with me, this feels a whole lot like that. Family. We sometimes squabble a little bit here and there, when things get tense, but we all really like each other. We have good personalities, if I don’t say so myself! [1:44:52]
DJ
I read that George Han has a question that we have pretty much covered but I think there’s an interesting angle to it. He is curious about the ASIC. How much functionality do you decide to put into the ASIC because the FPGAs have that refuelled reprogrammable reconfigurability, so I know it’s a trade off against power and cost and everything else but with that ASIC you do lose that flexibility. How are you treating that? Are you just going to go right we’ll put these features in and that’s it, too bad can’t do anything else in the future, we’ll have to spin a new ASIC. [1:45:36]
JE
Exactly. So for the ASIC we want to add as much flexibility as possible, that’s why we have that little 6502 hanging out in there that does a bunch of stuff in the blanking period. That way we don’t hard code anything. So we try to offload any kind of low speed stuff on it. [1:45:58]
DJ
So it is actually a processor? It’s all processor-driven in there essentially? [1:46:03]
JE
Status-driven by a processor. It is actually a pixel pipeline that is doing all the detection. So that’s actually a little scary. If we design this thing and a year down the road we decide if we had only changed this one parameter … [1:46:22]
DJ
Exactly. [1:46:24]
JE
Every ASIC that you design you have to try to figure out what’s going to be completely flexible and going completely flexible usually the trade off is lower performance and what’s going to be hard coded. [1:46:40]
DJ
So you absolutely cannot do this project without a gate array ASIC? Is there no avenue to do it with some power FPGA? [1:46:53]
JE
We could do it with an FPGA but it is cost. [1:46:56]
DJ
Yeah, it is cost but it depends on how big. You can get a dollar FPGA but you can’t put much in it. [1:47:01]
CG
No, you can’t do video. [1:47:02]
JE
Yeah, that’s the unfortunate part. Unless Altera or someone wants to [hint hint] get this design win and gives us a competitive price to persuade us not to go down the route of the metalized gate array then I don’t think there’s much option to hit our price point. [1:47:22]
We could add $100. [1:47:25]
DJ
So it’s just price? [1:47:26]
JE
Yeah, we could add $100 retail. [1:47:28]
CG
I’ve never heard of a win like that. That would be a crazy win for an FPGA. I’ve never heard anything like that. Maybe volume. [1:47:36]
JE
I tried. [1:47:38]
DJ
Maybe they can sponsor it. It could become the Altera Cast AR glasses. [1:47:44]
CG
No way man. [1:47:48]
JE
I tried that with the toys. We were doing all these metalized gate arrays and I went to Altera and Zylinx and said hey you have hard copy for these really big FPGAs, your $1,000 FPGAs down to like $100 hard copy version of it. Would you ever consider doing like a cyclone hard copy? Go from $15 down to a buck? They pretty much just said get out of here kid. [1:48:17]
DJ
Yeah, right you’re wasting our time. Yeah. That’s chump change. [1:48:20]
JE
Unless something’s changed, they’re not interested. [1:48:22]
CG
They’re making good money. [1:48:24]
JE
That’s where Atmel and EASIC and some of these other companies step in and make it up on volume. It seems pretty lucrative to me, they made quite a revenue off of us on toys. They are making money on the masks and making money on the per part. [1:48:46]
CG
That’s interesting business, I think we could probably talk to you about that just for another 5 hours just cos … [1:48:54]
JE
I really wish there was something in between.. as a designer I would love to have something between metalized gate arrays and FPGA, something that would bridge the gap. I could foresee getting grid of the $20 FPGA and putting a $8 part in there. [1:49:11]
CG
Yeah. Well they are starting to do that too. [1:49:13]
DJ
So what is your price target for your ASIC? Can you get away with a $20 part or does it have to be sub $10? [1:49:43]
JE
We are looking for less than $5 on the ASIC. [1:49:45]
DJ
Wow, OK. Yeah, that’s pretty restrictive, isn’t it? Especially at that sort of volume. [1:49:50]
JE
So we are going to eliminate $30-40 worth of electronics on the headset just by doing this. Which will be huge as far as getting the cost down. [1:50:01]
CG
Yeah. [1:50:04]
DJ
So the electronics is a large cost component of the final product? It’s not the screen, it’s not the plastics to get the moulded glasses, it’s not the Pico projectors? It all adds up obviously. [1:50:21]
JE
Yeah nickels and dimes. By far the most expensive things in the whole design is the micro displays. We are still doing the dance on those. [1:50:32]
Cg
We’ll sell a million … maybe. We’d like to! (Over 10 years!) [1:50:40]
JE
If you give us the right price now we can sell a million, if you give us the wrong price, no. [1:50:54]
CG
Exactly. [1:50:46]
DJ
Engineers have a hard time lying when it comes to stuff like that to vendors, don’t we? We are just too honest for our own bloody good. [1:50:52]
JE
Yeah! I’m always in trouble because of that. [1:50:54]
CG
You need to put on a sales person. [1:50:55]
JE
I’m a very candid person and I’m sure Rick’s gonna listen to this podcast and be, like, you said what to them? You’re talking about our actual costs? [1:51:04]
CG
We’re sorry Rick! [1:51:05]
JE
Oh my God! [1:51:08]
CG
But she’s the President, she’s allowed to! You said you were President right? [1:51:11]
JE
Madam President! [1:51:12]
DJ
Any experienced hardware person can guess your costs anyway to within a reasonable margin. [1:51:18]
JE
Yeah, those that have been around the block a few times. [1:51:20]
DJ
It’s not proprietary. [1:51:22]
JE
If we’re going to do sub $200 you can guess exactly what the glasses … [1:51:26]
DJ
Yeah, I can guess that you need a bus $10 processor. It’s not hard. [1:51:31]
CG
Plus Dave will do a tear down anyways once he gets one. [1:51:35]
JE
There’s just one blob of epoxy under here! Damn you Jeri! [1:51:40]
DJ
It’s not as exciting!
What about tooling for glasses and stuff like that, for the plastic? How are you going to do the 10,000 for the Kickstarter?
CG
Things have gotten cheaper lately man. [1:51:56]
Je
Yeah, so we think we’ve got it figured out what we’re going to do as far as the sub assembly. It needs to be dimensionally stable, micro displays versus the camera. So there’ll be a bit of a floating assembly. The glasses themselves will be kind of decoupled from that. You can’t be bending the relationship. [1:52:22]
DJ
It’s that critical due to the nature of the tracking? [1:52:27]
JE
Yeah, so you can’t have the projector … [1:52:31]
DJ
You can’t have the plastic expand when you put it on your head and it heats up? [1:52:36]
JE
We can tolerate a bit but we can’t tolerate someone putting it over the top of the wire snow cap or something and bending the arms out and having the 2 projectors angle in or something. [1:52:49]
DJ
Ah, got it. Yeah. [1:52:50]
JE
That’s some of the challenges we have to work through. [1:52:54]
DJ
Is there any self-calibration? The first time you use it do you calibrate? Or is it factory-calibrated once and that’s it? Do they even need calibration? Or is it so dimensionally stable that you don’t need to? [1:53:08]
JE
There is a calibration process that we go through here but that’s because we hot glue all of them together so … [1:53:16]
DJ
Exactly. But in the final production do you envisage that there’s going to be a calibration step in either manufacturing or by the user? [1:53:24]
JE
No, we don’t expect that to be the case. We expect to manufacture in tolerance and that tolerance should be good enough. You don’t have to be accurate within 1 pixel accuracy with your magic wand that you are poking at things, but if you get off by 10 pixels that’s pretty bad, you notice that. [1:53:45]
As far as tooling, I imagine what we are going to do is soft tooling. We will do aluminium inserts for the plastics and since if we go the Kickstarter route it is going to be pretty low volume so we can just shoot out of an aluminium tool and it will wear out pretty quick but … [1:54:04]
CG
Then you cut steel later. And that’s when the fun begins. [1:54:09]
DJ
Are you going to have to bring on an industrial designer to do all the 3D caddy? [1:54:17]
JE
Absolutely. That’s one area that we don’ t have a mechanical engineer yet. Most of the work has been left over from Valve. Some of the frames and stuff came from our mechanical engineer who was also let go at the same time. [1:54:30]
DJ
Well just steal your mechanical engineer from Valve! [1:54:33]
JE
He was let go also unfortunately, kind of a bummer, he went off and got another job before we got our feet under us again. [1:43:44]
DJ
Right. Is there now a clause that you can’t – this often happens when you leave a company there’s a no poaching clause that they make you sign, or it is in your original agreement. [1:54:54]
JE
Yeah, I didn’t sign anything. [1:54:56]
CG
That doesn’t work in some states, though, California that doesn’t work. So they can’t be enforced, apparently. I’ve been told.
That’s fun though, doing that mechanical stuff. You’ve done mechanical stuff before Jeri, right? [1:55:12]
JE
Yeah, I have a good idea of it, I’ll be able to talk to our mechanical engineer and be able to relay what I’m imagining. [1:55:23]
DJ
But not do the grunt work?! [1:55:24]
JE
Yeah. It is better to have someone run the tool that actually knows it! So Terra, our graphics person and animator, also has a background in industrial design, which is quite different than mechanical design. So she is actually working on concepts of what the glasses might look like and then she’ll work closely with the person who comes on board to do the engineering and choice of materials. It’s going to take a little bit of research to figure out what the right plastics should be. Is it going to be a liquid crystal polymer? Is it going to be some kind of nylon? It has to be rigid enough that when some kid bends the arms out or something … [1:56:04]
CG
Yeah, polycarbonate or something? [1:56:05]
DJ
And herein lies the problem. You can’t get that person unless you get funding and you can’t get funding unless you do your Kickstarter campaign, unless you get funding from somewhere else, and then in your Kickstarter campaign you have to put a deadline to when you’re going to deliver this stuff, so you haven’t even made those decisions yet. That’s why 80% of Kickstarter campaigns are late. [1:56:31]
JE
Yeah, you get going and find out it’s difficult. It takes a long time. We hope that we can have a lot of our team lined up. As soon as we get Kickstarter funding we’re going to have to hit the ground running and start recruiting like mad, because people aren’t going to want to wait a long time for these glasses so we’re going to have to be aggressive and come up with some schedule that people will be happy with. So that means we’ve got to get people working on this immediately after we get funding. [1:57:06]
DJ
That’s right. But ultimately, unfortunately for a complex hardware project like this you essentially have to sell a promise, like you’ve got your prototype and everything else but there’s a lot of steps to actually deliver a final product. [1:57:22]
JE
Well luckily this isn’t my first time to the rodeo so we can at least be close to s ship date or if not hit it right on. Rick has a background in shipping games, I have a background in shipping consumer products and designing them, so within a few weeks we should be able to. within a few weeks of our deadline we believe we can hit it. [1:57:52]
CG
Plus if you don’t, what is the President going to fire you or something? [1:57:57]
JE
That’s right. I don’t want to be in the situation where I’m months and months overdue. I saw some Kickstarters and how upset people get when things are late. [1:58:09]
CG
Yes, very. It is better to be realistic. [1:58:12]
DJ
And if you don’t keep them constantly updated as well, then they get nervous and all the rumors start that you’ve done a runner! [1:58:21]
Je
I hope, I think everyone else is on board around here, that we are going to be very open about what’s going on. Because it is to our benefit. Here it is, this is the fact, this is what’s going on, these are our costs. [1:58:38]
CG
And it will be educational too. [1:58:39]
DJ
I don’t know why everyone doesn’t, why everyone who runs a Kickstarer campaign doesn’t do that, unless you’re actually running a con. Why wouldn’t you be 100% open all the time? I don’t understand that. If you know you’re not going to be able to hit your target, just say so. [1:58:56]
CG
Some of it could be because you don’t know yourself. If you don’t know what’s going to happen next. Jeri has the experience with hardware and Rick has experience with software but if you’re just kind of grasping at straws, oh I guess we’ll ship it next. No, you’ll design plastics next. That kind of thing. [1:59:14]
JE
Yeah, I don’t envy some of these peopled that are struggling with that for the first time. [1:59:21]
DJ
Because it is too easy to take people’s money with an idea. You sell an idea on Kickstarter and you take people’s money and you go oo OK I’ll just get somebody to do it all for me and it all falls in a heap when you don’t know what you’re doing.
Fortunately you guys do. [1:59:47]
JE
YES! And we’re excited. I can’t wait until you guys actually get to put the glasses on. [1:59:56]
DJ
Oh, you’re making a trip to Australia are you?! [1:59:58]
JE
Maybe! There’s a conference, I can’t remember the name of it but they were pestering me to come out. I don’t know if I will actually do it. [2:00:10]
DJ
Well if you can get a first class ticket out of them! It’s only a 15-hr flight. Lots of time to write code! [2:00:18]
JE
Yeah, I’ll have my FPGA sitting there on the little table. [2:00:21]
CG
Yeah, your programmer on your lap. [2:00:23]
DJ
And good luck getting that through airport security. Bringing all this rough and ready hardware on to the plane, all this hacked together improvised electronics. [2:00:32]
JE
You guys would like this story. It was back when I was doing a toy design, before they had all these power outlets on the plane, so you didn’t get power on the plane. I was doing this cross-country flight that was going to be 5 hours. I HAD to get some of this stuff done. So I went out and got a battery backup UPS and it has a little beeper speaker in there. I open it up and crunch the speak er out so it wouldn’t be annoying. I went through airport security with this UPS in my bag. And they stopped me, of course, and they’re like … [2:01:14]
CG
Because it looks like a bomb, right? It looks like a chemical chamber. [2:01:16]
JE
Right! They look at this thing, it’s weighs like 10 pounds, with a battery. They’re like we can’t X-ray through something in your bag, can we take a look at your bag. They pull this thing out and they’re like what’s this. I said that’s just a surge protector. They’re like, it’s warm. It was plugged in before I came here. They’re like OK. [2:01;42]
CG
Yeah, you know designers these days, they can’t do anything right. [2:01:44]
JE
So I get to my gate, I plug the UPS in and to get this thing going before I got on the plane you couldn’t just start the thing without it being plugged in so I had to plug it into an outlet put the power switch on, tape the power switch so it couldn’t get bumped and then pull the plug. And I ran off this UPS for like 2 hours so I got an additional 2 hours plus what my laptop had. What do you think, I’m as genius or what? [2:02:16]
CG
Genius because you got away with it! That’s how it works. That’s really good. [2:02:23]
DJ
Was this pre or post 9/11? [2:02:24]
JE
Post. [2:02:27]
DJ
Right, I’m surprised then. [2:02:25]
CG
Hey man when you need power you need power! That was very inventive. [2:02:36]
JE
Another post-9/11 thing that I did that was probably not very smart is I had an old cell phone, a flip phone. And somehow I managed to plug the wrong charger into it and blew up the charging circuit. So I put 2 alligator clips soldered to the battery and drilled a hole through the case and I would put it onto my bench supply and put a little charge on it. [2:02:58]
CG
Yeah, a trickle charge. [2:02:49]
JE
Yeah, trickle charge my lithium battery in the thing. So I went through security with this thing and I went through a couple of times and they never really looked at it but this one time they like saw the alligator clips and they freaked! They were like what is this? And I had to dismantle it for them. I said it’s really just a cell phone, it’s just hooked to the battery. I’m an electrical engineer, trust me! But they let me go. You just have to have conviction. [2:03:30]
Cg
Exactly, confidence. It’s all in the swagger. [2:03:34]
DJ
This is now our official longest – way longest – episode ever. We blew 2 hours. [2:03:46]
JE
Ouch, I’m sorry about that guys. [2:03:47]
DJ
That means 2 hours of work you didn’t get done! [2:03:48]
JE
Oh God, I gotta get out of here. We’ve got a Kickstarter to do. [2:03:54]
CG
Well you have the site now right? [2:02:03:58]
JE
Technicalillusions.com and we’re going to do blogs there and YouTube videos. [2:04:09]
DJ
Any live feeds? [2:04:11]
JE
That would be interesting haven’t done a live feed in a long time. [2:04:15]
CG
Old school, you get the chat to speech thing going on again. [2:04:19]
DJ
With my computer I had voice jumping. It’s awesome. [2:04:25]
JE
I miss those days. [2:04:27]
DJ
Sorry that’s an in joke for anyone who has seen Jeri’s old live shows. [2:04:34]
JE
We had sound bits from Dave and then the IRC chat room could choose words and phrases to say. Utter rubbish! [2:04:46]
CG
Yep. [2:04:50]
JE
Well before we go I feel like I might have been a little bit too hard on Valve. I should at least say that I’m disappointed that I’m not at Valve but a lot of my friends are there and I wish them the best of luck. If they are listening I love you guys, can’t wait to see what you come out with. [2:05:11]
CG
Yeah. It’s great that they gave you the technology, allowed you to keep the prototypes and everything. [2:05:20]
JE
Yes, we walked in a week later after getting a verbal agreement we walked out with the contents of our lab. They just let us take everything. [2:05:29]
CG
Wow, except the equipment! [2:05:30]
JE
Yeah, I couldn’t keep my $30,000 scope but now I’m back to , it’s really terrible I’m working with this old 100 megahertz scope that everything looks like DC. I’m working on these giga hertz serializers and I don’t even bother pulling the scope out, it would be just like a DC bias. [2:05:53]
CG
Yeah, I’ll just check that one bit to see if it fired. [2:05:56]
DJ
Terrific. [2:06:02
CG
Well thanks for being on the show Jeri, we’ll look forward to seeing what’s next.
JE
Well, when we get closer, once we get the … [2:06:09]
CG
We’ll throw a guest your way so you can hire them. [2:06:16]
JE
Yeah, there we go! [2:06:17]
CG
We will be the recruiting arm of Technical Illusions. [2:06:22]
DJ
Awesome, thanks Jeri. It’s great. I’m sure a lot of people are going to really appreciate this episode.
JE
Thanks guys. [2:06:34]
CG
Alright, we’ll talk to you soon. [2:06:36]
DJ
Now get back to work [whip sound] [2:06:37]