Password Required

Please enter the password to view this page

The Best Augmented Reality SDKs

This week, we’re taking a deep dive into Augmented Reality Software Development Kits — also known as ARSDKs by those obsessed with brevity. Apple’s ARKit and Google’s ARCore are the most well-known ARSDKs, with a combined user base that stretches into the hundreds of millions. But there are a significant number of other SDK options on the market, and figuring out which one to build an AR project around can take a little digging.

SHOW NOTES

TRANSCRIPT

00:00 Joe Johnson: Welcome To the In Reality Podcast now starting in three, two, one. Welcome to the In reality podcast where we cover all things augmented and virtual reality. The In Reality Podcast is hosted by Joe Bardi and Joe Johnson and features news, commentary and perspective from industry veterans and experts. First up, introductions, I’m Joe Johnson Creative Director of Marxent Labs.

00:22 Joe Bardi: And I’m Joe Bardi Communications Director here at Marxent.

00:25 JJ: This week we’re taking a deep dive into augmented reality software development kits also known as AR SDKs by those obsessed with brevity. Apple’s ARKit and Google’s ARCore, the most well-known AR SDKs with a combined user base that stretches into the hundreds of millions but there are a significant number of other SDK options on the market and figuring out which one to build an AR project around can take a little bit of digging.

00:45 JB: Thankfully, Joe, you strapped on your miner’s helmet and did a mountain of research so that we could do this overview of what AR options are available right now and what changes to the SDKs might be coming just over the horizon.

00:57 JJ: Did you write this pun?

00:58 JB: I did.

01:00 JJ: Shall we dig in?

01:01 JB: I thought you already did that.

[music]

01:08 JJ: Augmented reality is ready for prime time after overcoming some early technical limitations like the need for an image tracker and fumbling through some, let’s call them, interesting use cases. [chuckle] And you’ll always have a weird spot in our heart MUSHA. AR is now finding its way in applications that span e-commerce, education, training, gaming and more. Most of these AR experiences are currently delivered via smart phone but the tech industry is currently working overtime to produce workable AR-enabled smart glasses. So no matter the gizmo you’re using for AR, there’s some kind of software development kit that’s creating an underlying experience. So we’re going to dig into the most popular AR SDKs in the market with an eye towards matching them to their best use cases. This list is in no way comprehensive Joe, but it’s also fair to say that the two primary options for AR SDKs have such market dominance that you’re probably better off focusing on them. And what are those guys?

01:54 JB: They would be Apple’s ARKit and Google’s ARCore.

01:57 JJ: Shocker right?

01:58 JB: Yes.

02:00 JJ: Alright, so, let’s just dig in I guess.

02:00 JB: Yeah, yeah. So let’s start with… ARKit is obviously the, it’s probably the most well-known at this point because it has the most market penetration. It’s good on every Apple phone for the last two years plus…

02:14 JJ: Yeah apple does a really good job of keeping their library of devices…

02:19 JB: As current as possible and they make sure that they obsolete stuff that will not fit their vision.

02:22 JJ: They do do that…

02:23 JJ: They do do that, yeah.

02:25 JB: And so because Apple enjoys control over the hardware that Google does not, they have been able to deliver a more consistent experience across a wider range of devices…

02:35 JJ: Are we focusing on ARKit right now or…

02:36 JB: No.

02:37 JJ: Or hang on, are we going to do the ARKit versus ARCore thing the whole time?

02:40 JB: You stopped me before I could finish that sentence, which was other than that they’re pretty much exactly the same.

02:46 JJ: [chuckle] That’s fine, I’m just… Well, I mean, you know, you always sort of evaluate these guys in relation to other things. And really the biggest obstacle I think to ARCore becoming a thing is the fact that there’s no unity in that user base. I have a really nice OnePlus 6 that’s like a Chinese phone that’s basically identical to an iPhone X, it just happens to run Android. I don’t know if ARCore runs on it or not, I’m pretty sure it does but my brother has an old Samsung. Like a seven or whatever…

03:13 JB: Okay.

03:13 JJ: And I’m sure that it won’t run on that.

03:15 JB: Yeah.

03:15 JJ: So you know, I just, I know that ARCore is positioning itself, I just watched a video a couple of days ago where they were talking about enabling all of their software on everybody’s platform, including iOS, right? But I can’t see, I honestly can’t see developers who are working for IOS using ARCore to do their AR SDK. Even though that’s definitely what Google wants.

03:36 JB: Of course of course.

03:37 JJ: They were saying something along the lines of like if you’re gonna play a multi-player AR game, you shouldn’t have to worry about who has what phone. And I was like, good luck with that buddy.

03:45 JB: Yeah.

03:45 JJ: Welcome to all the walled guards and…

03:46 JB: Yeah, yeah, that’s actually a really sort of interesting thing too. The idea of how niches may stand in the way of sort of widespread adoption. Although someone will figure that out. Like sort of the new Harry Potter game is gonna come out…

04:02 JJ: I think Google claim that they did.

04:03 JB: Well so ARKit will…

04:05 JJ: ARCore.

04:06 JB: ARCore will work on an IPhone.

04:09 JJ: That’s what they showed in the last… So, when doing the research, there are mountains and mountains and mountains of videos that show off features.

04:18 JB: Right.

04:18 JJ: It’s like tech demos basically, and almost all of them are like, you know it’s a small feature and it doesn’t look like much but one of the ones that ARCore had was they were showing their software working on an iOS device, right? Whether or not that’s gonna become ubiquitous or not is another story entirely.

04:35 JB: Right Yeah, no, I mean, it makes sense.

04:39 JJ: Let’s do some feature overviews.

04:41 JB: That’s what I was gonna say yeah.

04:42 JJ: I just cut you off.

04:43 JB: No, no, no, that’s fine, that’s fine. I would start with, both have Markerless Tracking.

04:47 JJ: Yeah.

04:47 JB: Which is pretty much…

04:49 JJ: It’s like [04:49] ____.

04:50 JB: Yeah, I mean, you can’t not have Markerless tracking I would say.

04:53 JJ: You can. There are a couple of SDKs on here that clearly do a bunch of marker-based tracking. I think Vuforia still does a lot of it.

05:00 JB: I will refrain from being overly negative about the fact that you need a piece of paper to make your phone work.

05:03 JJ: Well, there are some cases to be made for using trackers that are specifically designed for it as anchors for experiences.

05:10 JB: Yeah, I could see in a real world setting, that has been set up for AR like a museum or something like that or in certain types of advertising where the print already exists and you have simply coded it on top.

05:26 JJ: Yeah.

05:26 JB: That I could see.

05:27 JJ: Okay, so they’ve got markerless tracking, they’ve got extended tracking. Do you know what extended tracking is?

05:30 JB: Is that, it keeps them in place after they…

05:33 JJ: Yeah, basically it’s some sort of persistence where if you move past the original mandate of the tracking zone, it’s using an accelerometer and a bunch of other stuff, to keep track where you are.

05:40 JB: Right, to keep track of stuff.

05:42 JJ: Yeah.

05:42 JB: And other users will know?

05:45 JJ: So that actually gets to the next feature. Multiuser AR Environment.

05:48 JB: Oh okay.

05:49 JJ: Yeah the holy grail has always been tracking stuff and then communicating that to other things so that other people can see what you’re tracking or working on.

05:56 JB: Gotcha.

05:57 JJ: And iOS, I think with ARKit2 on iOS 12, their apps can now do that.

06:01 JB: Okay.

06:02 JJ: They have multiuser AR environments and, additional feature, persistent AR environments and objects. So basically, you can save scenes and stuff, it’ll keep track of that state in between. So if you come back to a place that you tracked previously or whatever, it can theoretically remember where you were at there.

06:18 JB: Are we headed for this weird Brave New World future where I’m gonna go back to my AR or can find out someone has someone else has defaced it.

06:26 JJ: Defaced it. So there’s an app called Just a Line.

06:29 JB: Okay.

06:29 JJ: For ARCore where basically that’s how it works. It keeps track of geographical locations and you can draw line on top of basically a video feed.

06:40 JB: Sure.

06:41 JJ: Right? And then, since it’s persistent AR environments and objects, it actually communicates that to other users and they can see basically tags, graffiti and stuff like that. That’s an ARCore app as far as I know, there are no ARKit versions of Just a Line.

06:54 JB: Advantage ARCore.

06:55 JJ: Advantage ARCore. But ARKit can do the exact same thing. I just think that Apple’s focus is different.

07:01 JB: Yeah, that’s an interesting way to look at it too. They do have… Both companies, even though they’re building essentially the same thing. Sort of have different designs on what it’s for.

07:09 JJ: Yeah. And so Apple seems like it’s focusing on enterprise clients and stuff like that, providing tools and services. And what I’ve seen from Google, I think because of the way that the Google development environment is, I think you have a lot more people that are not enterprise level users coming up with tools and toys and stuff. It’s probably… It feels a little more like an open source scenario.

07:29 JB: Gotcha.

07:30 JJ: Which makes sense, Android is.

07:32 JB: Yes. [chuckle]

07:33 JJ: You’ll probably… I don’t always wanna make statements like this but I feel like the killer app is not gonna come out of iOS stuff. I feel like someone’s gonna find something that AR’s really good at. And it’s gonna happen because people are experimenting and a lot of people are experimenting with the ARCore.

07:50 JB: Yes. We’ll see, but that’s entirely possible and even almost seems likely.

07:55 JJ: You gotta talk in the mic chief.

07:58 JB: Yeah, I know, I’m trying to think what else. What were some of the other features here? Device motion…

08:02 JJ: Are you talking about other features for ARKit?

08:04 JB: Well, so yeah. ARCore has device motion tracking.

08:06 JJ: Yeah, so the lists are actually formed from features that the companies themselves self-reported. There’s a crap load of overlap actually.

08:16 JB: Okay.

08:16 JJ: Between these are AR SDKs. Let me get this out of the way. Almost all of these SDKs claim to do all of the same things.

08:23 JB: Okay. [chuckle]

08:24 JJ: I’m not kidding. So the feature list for ARCore says device motion tracking. ARKit does that. It did it in version 1.0. Horizontal and vertical plane tracking, ARKit has horizontal plane tracking. Now, they don’t list vertical plane tracking on ARKit and I don’t know if they’ve cracked that or not yet. We had a conversation about that the last time we talked about the features of ARKit.

08:46 JB: Yes.

08:47 JJ: But ARCore is claiming vertical plane tracking and I can’t imagine that Apple hasn’t figured that part out yet.

08:53 JB: Yeah, wait for June.

08:54 JJ: Yeah. So ARCore 1.0 had light estimation. It was featured in all their videos. I don’t know if ARKit does that. I don’t think it does yet.

09:01 JB: I don’t think it did.

09:02 JJ: Yeah. So that’s actually a big differentiator for ARCore. Markerless tracking listed for ARCore. Shocker.

09:08 JB: Yeah. [chuckle]

09:08 JJ: Basically they at least have it. And they list multiuser AR environments. And I think that’s like a catch-all for all of the persistent AR environment and object stuff that’s listed in ARKit. I honestly… I don’t know what distinguishes these two things other than what we’ve already talked about in terms of features.

09:24 JB: I’ll throw one additional thing out here just… Because we always talk about this when we talk about it especially here at Marxent and we talk about this a lot. The one bugaboo for both solutions is slow initialization.

09:37 JJ: That’s not really a thing so much anymore.

09:39 JB: Really?

09:40 JJ: Yeah, so IKEA Place is built directly on the bones of ARKit 2.0.

09:44 JB: Okay.

09:44 JJ: They don’t have any initialization timing projects.

09:47 JB: Oh, they’ve killed it.

09:48 JJ: I don’t think they’ve killed it. I think that it’s up to the individual developer to come up with solutions for their initialization time stuff.

09:54 JB: Okay.

09:54 JJ: So I would say that there’s probably a non-standard set of solutions for ARKit and ARCore’s initialization times.

10:01 JB: I see.

10:01 JJ: But I think it’s probably handled at this point by external tools.

10:03 JB: Right. So in the way that MXT tracking which is Marxent’s SDK.

10:08 JJ: Yeah, that’s our system… That’s our solution to their initialization problem.

10:10 JB: We were doing a lot of pairing it with ARKit and ARCore to just because it’s tracking started almost instantaneously whereas ARKit took a couple of seconds to finally…

10:18 JJ: And when you say it, you mean our tracking logarithm tracks and…

10:21 JB: Yeah. Correct. Yeah, MXT tracking algorithm.

10:23 JJ: Yeah, so I’ve looked over these SDKs, almost all of them. It just went like EasyAR, EasyAR.

10:29 JB: EasyAR, yes.

10:29 JJ: Which we’ll get to later. A truly unfortunate name. [chuckle] They have listed on their feature set instant tracking, stuff like that. I think that it’s important to note that while ARKit and ARCore are really solid platforms, the fact of the matter is they’re very DIY. They’re not… Whatever comes out of the box is what you get with them. They’re not gonna do anything custom for you. So there is something to be said for going with an AR SDK that utilizes ARKit and ARCore but does not… But is basically built on top of those frameworks.

11:00 JB: And so if we’re talking about businesses that are looking to adopt some kind of… Or use AR in their apps or whatever, ARKit and ARCore are really for companies that already have an internal team that is working on technology like that.

11:12 JJ: Honestly they’re for companies like us or companies like EasyAR or DeepAR or something like that.

11:15 JB: It is not sort of an off the rack solution where you’re like, “I would like AR in my app,” and just…

11:20 JJ: Well, you can do that. The bigger company… Obviously I mentioned IKEA and there are a bunch of other like Wayfair, basically, I think built their straight on ARKit.

11:27 JB: But these are all companies that have teams.

11:28 JJ: Yeah. Big teams. So DIY solutions are always sort of like that. You need to have some sort of technical knowledge in-house. And I don’t wanna get ahead of myself but there’s always gonna be a place for houses that understand the tech and are willing to sort of translate for companies that don’t wanna invest.

11:47 JB: Yeah. Yeah, yeah.

11:48 JJ: Hundreds and hundreds of thousands of dollars [chuckle] on building an internal team.

11:51 JB: Yeah, yeah, yeah. So…

11:52 JJ: Speaking of DIY stuff, there’s Wikitude which offers the same sort of DIY experience as ARKit, ARCore and stuff like that. Their features include object and scene recognition. You’ll hear a lot of these over and over again. Markerless tracking. They have image tracking, which is the Vuforia style image features tracking. Location services, which is a thing that’s sort of built on top of everything else. That could mean a lot of things but probably it’s just an extension of geographic AR.

12:21 JB: Right.

12:21 JJ: Multiple-image tracking which is another Vuforia style thing. Extended tracking, which we’ve talked about. And here’s a new one, cloud recognition.

12:27 JB: Yes. Explain cloud recognition, chef.

12:29 JJ: So cloud recognition is actually not as cool as it sounds. It’s basically just a… Vuforia did this, I don’t know, three years ago, four years ago, five years ago. Where you would upload the image that you want to track to their online database. And it would establish whether or not it was trackable, do all that stuff. And the next step of that was just storing that information in a database and when you… Instead of having it built into the app, the image that your app knows it wants to look for, it just fetches that from a cloud. It’s nothing crazy.

13:02 JB: Right. From a resource standpoint, if you have a fast connection, it means that the… Whatever app you’re running can be much leaner and whatever.

13:09 JJ: Little slimmer. Yeah. It’s such a minor feature.

13:11 JB: But, it is interesting.

13:12 JJ: Yeah, So the advantage that Wikitude I think offers over these other… The core AR SDKs, Wikitude has enabled their software over a wide variety of platforms. Apparently, they work obviously in native languages. They have a JavaScript API, they work with unity which we also do, they work with Cordova and Xamarin extensions, which are honestly two things, never heard of. [chuckle] But I think their goal is to take ARKit and ARCore’s like device-specific, ecosystem-specific stuff and then extend it out over basically any device.

13:47 JB: Gotcha. Yeah, yeah, yeah. And what is the… It feels like there’s a wider range of companies or businesses or whatever who could actually utilize Wikitude than maybe even ARKit and ARCore.

14:01 JJ: Yeah, I would agree with that. And the advantage of Wikitude is obviously… So they have a free SDK, which means you can get access to their library of software for personal use or for whatever. And then they also have enterprise level support for things. I don’t know. I obviously haven’t contacted them to talk about what they can do for me. But I imagine that if…

14:21 JB: Right. You already know people.

14:22 JJ: If you’ve got enough money, they’ll help you out with a custom solution but they’re definitely positioned as a DIY offering.

14:27 JB: Okay. It’s good to break this up to DIY and more for feature.

14:32 JJ: To a certain extent, everything’s a little DIY. Unless you get to a company like us where we’re digging into your business problems and then trying to solve… Find a way to make AR specifically solve your business problems instead of just being like, “Hey, I wanna do AR on my app and it’s cool and new”. You know what I mean?

14:49 JB: Right. Yes.

14:49 JJ: I don’t usually talk so much about what Marxent does for clients but…

14:54 JB: I like to… I like to… I know that we… As we do these podcasts, we talk about us. But then we don’t talk about us, we talk about us, whatever.

15:01 JJ: I love not talking about us.

15:02 JB: Right. But for this particular topic…

15:05 JJ: It does seem appropriate, yeah.

15:05 JB: Because we do make an AR SDK and we do build all kinds of AR applications and have been doing a lot of this stuff, I’m gonna keep working us back.

15:13 JJ: Solid plug.

15:14 JB: I don’t even mean it as a plug. [chuckle] We’re just another company that fits in with all of these other companies that are doing some variation on this.

15:20 JJ: Speaking of fitting in with other companies, let’s talk about Vuforia.

15:23 JB: Yeah, since you’ve mentioned them several times already.

15:25 JJ: So, Vuforia’s great. I love Vuforia. When we got started way back in, I don’t know, the dark ages of the early 2010s.

15:31 JB: My God. [chuckle]

15:34 JJ: Our AR solution was basically built on Vuforia, on Qualcomm’s Vuforia offering. Which was a DIY solution that let you use marker-based tracking, right?

15:43 JB: Right, yes.

15:43 JJ: And it was a pretty… It was already solid and it was definitely built for enterprise clients, they had a lot of support for it. So it’s definitely on of the longest-lived AR SDKs. They say it was developed in 2015 but it was definitely developed before that.

15:56 JB: Yeah.

15:57 JJ: So they’ve evolved from a DIY marker-based AR system into an industrial enterprised-focus SDK focusing on mixed reality headsets. Their features include number one, mixed reality headset support to your HoloLenses.

16:10 JB: Right.

16:10 JJ: And…

16:10 JB: Magic leap.

16:12 JJ: Yeah, I’m blanking on the other ones right now because there isn’t anything for mixed reality headsets that’s as important as the HoloLens.

16:17 JB: Are you ripping on the Vuzix Blade right now? [chuckle]

16:20 JJ: No, I’m not ripping on the Vuzix Blade. I’m just being realistic about…

16:25 JB: It does feel like…

16:25 JJ: Which mixed reality headset hardware is doing well and it’s Microsoft.

16:28 JB: Yes. I think it’s fair to say that all of them are still at the developer kit level. HoloLens is nominally a commercial product.

16:36 JJ: Yeah. And if you watch some of Vuforia’s videos for mixed reality headset support, they’re using Hololenses and it looks a little nascent.

16:43 JB: Right, yeah.

16:43 JJ: But there are some good things there. Other features include, shocker, markerless tracking.

16:46 JB: No way.

16:47 JJ: Here’s an interesting one that other places don’t list a ton but I bet they can all do, integrated audio and video capture from the mixed reality headset.

16:55 JB: How interesting.

16:55 JJ: So one of the videos for Vuforia shows a guy, an industrial worker looking at a piece of machinery and then identifying something that has to be fixed. He circles it in real time and then later somebody comes back. There’s an AR anchor that some other worker comes along later and sees that somebody left it.

17:08 JB: Gotcha.

17:09 JJ: These are obviously variations on the Just A Line idea from ARCore but with a more industrial bend.

17:16 JB: Still, I think it’s interesting that you said it that way because when we were talking about the idea of persistent objects in AR before, I don’t think I said it out loud but in my mind, I was kinda like, “Who cares?”.

17:25 JJ: You know on a podcast you have to say things out loud.

17:27 JB: Well, I didn’t because I didn’t feel confident in what I was going to say in that moment, so I decided…

17:30 JJ: You look so confident right now. Let’s keep going.

17:31 JB: ‘Cause I do. I feel much more confident in this statement which is, I see how having persistence of work in a enterprise environment where people are working collaboratively, it would be important that whatever I did stay there for the next course of the file.

17:42 JJ: Yeah. Or let’s say, just for funsies, let’s imagine a mixed reality world where I’ve got a nice pair of glasses or whatever and I walk into the conference room and there’s stuff up from a meeting where I get to view it at my leisure.

17:55 JB: Sure.

17:55 JJ: It’s an extension of the whiteboard or whatever that doesn’t require… That you can change files out for. It’s just a fancier whiteboard. And sometimes we talk about killer apps but other times I think that there are good iterative steps that just make mixed reality a better option than other things.

18:11 JB: You had me at fancier whiteboard.

18:13 JJ: Fancier whiteboard. Other features on Vuforia include object recognition, which we haven’t talked about yet. Object recognition is…

18:22 JB: The future.

18:23 JJ: Well, some people say that it’s the future of visual search or whatever. I know Ben’s talked about.

18:27 JB: That’s my feeling. I feel I’m with him.

18:28 JJ: Basically, it just means, instead of recognizing a flat image, it recognizes a specific 3D configuration of an object. And in addition to that, it has model targeting. Model targeting is basically the same. It’s basically the exact same feature. It’s just a matter of whether or not it can understand that what it’s looking at is a computer-generated 3D object or a real 3D object.

18:51 JB: Gotcha.

18:52 JJ: And then, here’s the big one for Vuforia. They have Occlusion modelling. Do you know what that is?

18:55 JB: No.

18:57 JJ: Here, I’ll explain it for all the listeners…

18:57 JB: Please explain that one ’cause I don’t wanna do it badly.

19:00 JJ: So when somebody’s… You’re looking at somebody in real life and then somebody walks in front of you, then they block them. That is called Occlusion.

19:07 JB: Yes.

19:07 JJ: So Occlusion modeling is a way for software developers to say, “Hey, I want something to be in front of something else in this digital world.” That’s what Occlusion modeling is.

19:17 JB: Okay.

19:17 JJ: Now a bunch of these companies don’t talk about Occlusion modelling. It’s super important, obviously, for AR.

19:21 JB: If you’re placing multiple things into a scene it’s probably necessary.

19:23 JJ: Or if real life things need to be in front of something that you need to be able… Here’s something super important about Vuforia’s offering that I have to get through. In our last episode, Ken may have talked at length about…

19:38 JB: May.

19:39 JJ: About how important it would be to change our method of consuming augmented reality from a light projection model, like a light projected on a lens in front of you, to a full headset experience that uses the camera as your means of generating the world, so Vuforia already does that. Vuforia has a solution for that. They actually highlight it in a few of their videos where someone puts on a headset and the camera feeds them everything from outside and then, obviously, at that point, objects… Occlusion becomes much more important there because your only source of contact with the outside world is what the computer is generating for your eyes and he talks about being blacked out or something and ends up being in danger.

20:24 JJ: That’s why you need occlusion modeling for Digital AR. So, Vuforia’s focus is clearly on enterprise headset users. They focus on medical, they focus on industrial and I think that probably, that’s a good play for them.

20:36 JB: Yeah, military, they’ll wind up building stuff for the military for the HoloLens.

20:40 JJ: Qualcomm’s never done anything like that.

20:42 JB: Yeah. No, it never happens.

20:43 JJ: Yeah, so obviously, if you’re looking to integrate AR to workflows using mixed-reality headsets or if you just wanna work with a big, big company, V4 is a good option ’cause Qualcomm’s their backer.

20:54 JB: Okay.

20:55 JJ: Yeah. Let’s move on to some lesser known options. There’s a company called MAXST, which is a Korean company. They have feature parity with Wikitude. And if you go through their features, they’ve added mixed reality headset support where Wikitude doesn’t claim that.

21:10 JB: Okay.

21:11 JJ: That may or may not matter. They’re a relative unknown in the US market. There’s a lot of Korean, Chinese companies, Korean and Chinese companies, probably Japanese companies too, that do a lot of this stuff but we don’t really hear about all that.

21:23 JB: Right, yeah.

21:24 JJ: But I imagine they’re cheaper.

21:25 JB: Yes. [chuckle]

21:27 JJ: I know that they offer themselves that way.

21:31 JB: Were they the one that had won the two AWS Awards?

21:33 JJ: No, no, actually, let’s move forward a little bit to another… Let’s move forward a little bit to another SDK, that’s EasyAR. EasyAR?

21:41 JB: So we’re now moving into the adjective AR companies.

21:44 JJ: Yeah, adjective AR companies. So EasyAR’s an interesting case. They won AWE Asia’s best software two years in a row for their SDK.

21:54 JB: Gotcha.

21:54 JJ: I also cannot find a single Western company using EasyAR’s SDK. I don’t know if that’s due to it being difficult to research overseas news. If I don’t speak Korean, I’m not searching in Korean terms and stuff like that. But the fact of the matter is, I haven’t found anything. Now, I’m not an amazing researcher and I’m sure somebody out there can do a much better job of finding out people who’s using EasyAR.

22:15 JB: Unless it doesn’t exist.

22:16 JJ: Unless it doesn’t exist. So looking at those things, MAXST and EasyAR, there’s definitely foreign support out there that might be cheaper and may be just as good. Me, personally, I can’t recommend them ’cause I don’t know what level of search… Or anything like that.

22:31 JB: Sure, sure, sure. And we should say, we’re making almost no recommendations here. We’re not about recommendations.

22:36 JJ: Yeah, this is not about recommendations. This is just talking about stuff.

22:38 JB: It’s really just… I think that what you did over the last few days and sort of I picked up a little on the backend of, but this is stuff that people were doing at businesses and all the time now as they know that there’s this thing out there, AR, that maybe they need to use or maybe they don’t. They don’t know the use cases and they start researching it. And you get companies from Korea that have won awards from AWS Asia.

23:00 JJ: Yeah, and then we get back to the big advantage of ARKit and ARCore, which is these are huge, huge… ARKit and ARCore, which is that the companies that are backing them already have a bunch of penetration into the devices, everyday devices, that people use. So it’s a really hard sell to be like, “Hey, you should definitely go work with X company to develop your app,” when if you wanna have an in-house team, you do it yourself. You really can.

23:28 JB: Yeah.

23:29 JJ: Probably, the advantage of places like EasyAR and MAXST and stuff is that you’ll get much better support than you will from Google or Apple because you mean more to them.

23:40 JB: A little more personal attention, you feel? Maybe?

23:42 JJ: Yeah, yeah, yeah. Smaller companies, smaller vendors are definitely better for that sort of thing.

[overlapping conversation]

23:46 JB: Yeah, it does. Yes, that’s right. [chuckle]

23:47 JJ: So speaking of crappy names, DeepAR, DeepAR… So, this is an interesting SDK. This is specifically for facial tracking. Their features include… And these are just the only ones they’ve listed. I’m sure they can do better than this: Facial recognition, facial tracking and multi-face recognition. Now, they can do a bunch of other things like render stuff on top of faces and deform faces. But really, what they’re for, it looks like, is mobile-optimized Snapchat and Facebook style 3D realistic face lenses, masks, etcetera, like special effects.

24:16 JB: So if for some reason, you have a mask that you’d like to…

24:21 JJ: Yeah, that you wanna use, yeah.

24:21 JB: I could see sports teams doing this, or whatever, to…

24:25 JJ: This seems like a really weird play for an SDK.

24:27 JB: It does. Or it’s very specific, or it’s very limited.

24:31 JJ: It is, it’s super specific. If you think about it like this. ARKit also enables those things, Animojis exist in ARKit. The software can do that sort of thing, I think. I don’t know exactly what they’re offering to developers.

24:41 JB: And it’s not like there’s a whole bunch of Snapchat competitors out there.

24:45 JJ: Yeah, I mean…

24:46 JB: That don’t have billions of dollars or the need…

24:49 JJ: Yeah…

24:49 JB: The wherewithal to build this themselves.

24:51 JJ: DeepAR still exists, and I’m assuming that they do business with somebody. I’m just not entirely sure who it is. But obviously, if you’re looking to horn in on Snapchat’s lens game, DeepAR is probably your best bet. They’ve got a free… DeepAR has a free SDK that you can use. I don’t know what their limitations are on that. But probably, if you’re looking to leverage a head-mounted mixed-reality display for guiding on-site employees for machine op and maintenance, you probably wanna look elsewhere.

25:14 JB: Yeah. [chuckle]

25:15 JJ: Again, you can see videos of all these SDKs attached to a post that Joe Bardi’s gonna handle for me.

25:20 JB: Yes, yes, that’s right. Yeah, I’ll have that in the show notes.

25:24 JJ: So where are we at in this list?

25:26 JB: We are at my favorite name.

25:29 JJ: This is also my favorite name. When I saw it, I have literally no idea how to pronounce it.

25:34 JB: Yes, I can’t help you.

25:35 JJ: And I’m more than willing to throw shade on this SDK. We’ll get into it.

25:38 JB: Yeah.

25:38 JJ: It’s X-Z-I-M-G. XZIMG?

25:42 JB: XZIMG.

[vocalization]

25:44 JJ: I don’t know. This is another facial recognition SDK. Interesting to note, it appears that there is only one employee of XZIMG. And he’s in all of their videos. And I suspect it might be just some guy who’s figured out, and he has pretty good knowledge of the facial tracking software and stuff, he might have curbed it from somewhere else or some other company.

26:04 JB: He has a way of naming things.

26:05 JJ: He does have a way of naming things.

26:06 JB: Because in addition to XZIMG, they also have a product called…

26:09 JJ: Magic Face.

26:10 JB: Magic Face.

26:11 JJ: Yeah, that offers developers more augmented reality lenses and whatnot. So, I imagine that if you wanna work with XZIMG, I bet they’re a fraction of the cost of any other place. [chuckle] But the fact of the matter is there’s a free SDK for DeepAR. And I imagine you could probably use ARKit for free to get sort of the same features. So, it’s another hard sell. Sometimes, these companies are too small to take seriously.

26:34 JB: Sure, sure, but…

26:36 JJ: And I’ll be honest, I’m gonna go out on a limb, hot take, I do not take this company seriously. [chuckle]

26:41 JB: Which means, in two years they’ll have developed something that no one else has.

26:44 JJ: Oh yeah, they’re gonna take over the business [chuckle] and I’m going to look like an idiot. But that’s fine. I always look like an idiot.

26:48 JB: Yeah no, go out on a limb, [laughter]

26:49 JJ: That’s so, I really like this next one ’cause, in the video for it, I got to talk about GitHub being dirty.

26:55 JB: Yes, that’s right. Dirty it up Joe.

26:58 JJ: I’m gonna dirty it up… So if you wanna get your hands like GitHub dirty, there are always open source tools, like ARToolKit or its, I guess younger brother artoolkitX, which was from the same guys. ARToolkit is an SDK that’s maintained by a open source project on GitHub. Technically, it’s owned by Daqri, but I don’t think they pay a ton of attention to it.

27:19 JB: Right.

27:19 JJ: I think they bought it for something that it did and they’re like, “Whatever, I’ll just let this stuff stay out here as a tool that people can use.”

27:23 JB: It’s very kind of them.

27:24 JJ: Yeah, so it’s basically just an AR tracking library. It’s got over of 160,000 downloads since it was released in 2004, so it’s pretty old.

27:31 JB: I was going to say.

27:32 JJ: Features include camera position tracking, square marker image tracking, planner image tracking, trackable image generation utilities and it obviously, it’s a free open-source SDK. It’s a little bit like old Vuforia and it never really got a new set of features but there’s something to be said for toolkits that don’t have a ton of stuff in them. If you just wanna do one thing and you just wanna play around with it, ARToolKit’s is a good idea.

27:53 JB: Yeah. And you talked about, when you were doing this research or whatever, about how it’s primarily for research and experimentation. That makes a lot of sense to me in the sense that it started in 2004, which is right along the edge of the creation of AR.

28:05 JJ: Yeah.

28:05 JB: And so it’s been a playground, it’s been a sandbox for developing a lot of the… As you read the sort of planar image tracking and trackable image generation, blah blah blah, it’s… All of the features are almost like a triptych through the history of AR.

28:17 JJ: Yup.

28:18 JB: Leading up to now.

28:19 JJ: Yeah.

28:19 JB: And it’s cool that it’s still out there and it’s still available.

28:23 JJ: Yeah, so I always like free resources that don’t require you to engage with a corporate entity, ’cause it basically means hey, do you just wanna screw around with it you don’t want pressure or time pressure or anything like that. You don’t want anybody trying to sell you something.

28:34 JB: Yes.

28:34 JJ: Cause you just wanted to try something out.

28:36 JB: Just try to learn.

28:37 JJ: Yeah, ARToolKit’s probably a good choice for that. And now, I guess, we get to the part where we talk about us.

28:42 JB: We could do that.

28:44 JJ: Talk all about us.

28:44 JB: I’ve been weaving us in already, but let’s talk about us.

28:46 JJ: Weaving us in already.

28:47 JB: In one shot.

28:48 JJ: So, how do we really fit into this whole thing? I mean, do we really have an AR SDK that we offer?

28:54 JB: We do.

28:55 JJ: We do? Okay.

28:56 JB: Yeah, so I would say… I would say so the interesting thing about what we offer is…

29:00 JJ: Gotta look up.

29:00 JB: Oh yes, sorry about that.

29:01 JJ: Gotta talk in the mic.

29:02 JB: I would say the interesting thing about what we offer is that we offer feature parity with a lot of these other companies but we’re also… I mean, we’re providing just a wider service as far as in addition to the SDK, we will build the entire app. We will build other uses for the content that you were putting in your AR app.

29:23 JJ: You mean like VR in-house sales tool?

29:25 JB: I mean, like a 3D room planner, I mean like a VR in-store show room display.

29:30 JJ: So this is a good time to talk about shops that are specializing at this point. Cause that’s what we’re starting to see in the industry. We’re starting to see shops that specialize in a specific vertical. We have two SDKs on here for… Technically, three on here for facial tracking, right? And you think, is there enough room in the business for those guys? And I’m like, I guess there is.

29:47 JB: I guess.

29:47 JJ: Yeah, there’s enough room in the business right now. I think that companies that wanna… So, we’re focused on, what, furniture these days?

29:53 JB: Yeah.

29:54 JJ: Furniture, cabinets, stuff like that.

29:55 JB: Yes.

29:55 JJ: High consideration products. It’s interesting to see what happens as a technology matures, you see all of the other companies that are starting to focus on, “Well, how do I use this technology to get the results that I’m looking for, for my business vertical?”

30:06 JB: Right, right.

30:07 JJ: And that’s really the advantage of using somebody like Marxent and I’m trying to think of other ones that are doing that.

30:12 JB: There are other companies that will sell you like a suite of solutions that may also include AR.

30:17 JJ: Yeah.

30:18 JB: But oftentimes, they’re using off-the-shelf components of other peoples other AR solutions ot whatever.

30:24 JJ: Sure.

30:25 JB: And we use ARKit and ARCore too, because there’s certain advantages to doing that, especially if you, again, with the widest number of devices…

30:32 JJ: This might actually be a good topic for a later podcast like companies that have specialized in a given vertical.

30:37 JB: Sure.

30:37 JJ: Within AR or VR.

30:38 JB: Actually yeah, that would be a great topic.

30:40 JJ: That would be a good one, yeah.

30:40 JB: We should write that down somewhere.

30:42 JJ: I should definitely write that down.

30:43 JB: Record that somewhere.

30:43 JJ: I’m gonna record that somewhere.

30:44 JB: Yeah, that’s a good idea.

30:44 JJ: Alright, well, do we need to cover anything else? Did we get everything done?

30:48 JB: I don’t know, I mean I feel like…

30:49 JJ: Do we wanna cover our features? Here, let’s do it.

30:51 JB: Yeah, yes.

30:51 JJ: Instant On tracking via MXT initialization.

30:54 JB: That’s right.

30:55 JJ: Markerless tracking. Thanks, ARKit. I appreciate it. Extended tracking same.

30:58 JB: Yeah.

31:00 JJ: Persistent AR environments and objects, thanks…

31:01 JB: That’s right.

31:01 JJ: Thanks to the firm backbone of another AR toolkit that we put together…

31:04 JB: We also offer, we offer object interactivity as well, don’t we?

31:08 JJ: I don’t know what that means. Hang on hang on. I don’t know that means. What are we talking about here?

31:13 JB: For example, placing a lamp on a table.

31:14 JJ: Sure, so when I said earlier that a bunch of these features are sort of like hand-rolled… Or a bunch of these features are rolled into other descriptions of things…

31:23 JB: Sure.

31:23 JJ: It pretty much goes without saying at this point, that you’re able to interact with the objects that you place in AR. Almost all of these SDKs have interactivity.

31:30 JB: Right.

31:30 JJ: And honestly, they would be useless if they did not have them. So, it’s almost not even worth mentioning at this point, but we do do that.

31:36 JB: We do.

31:36 JJ: We do do that. You can put things into scenes and then interact with them.

31:39 JB: Yeah, I don’t, but have them interact with each other.

31:42 JJ: They can do that too.

31:43 JB: Yes, and I think that’s really what I was going for…

31:46 JJ: There’s an ARCore demo for I think it’s… It’s not Crate and Barrel, what’s the potter…

31:52 JB: Pottery Barn.

31:53 JJ: It’s Pottery Barn. So there’s a ARCore demo for Pottery Barn that has multiple objects going in the scene and then interacting with one another in some meaningful way, usually reflections or lighting or whatever. I think that stuff is starting to become standard, and I think what you need to see now from an AR SDK that you’re gonna go with, is that you have to start figuring out how the tool works for your business.

32:13 JB: Yes.

32:13 JJ: Rather than just like, how do I get AR?

32:15 JB: Yeah, one of the things and I know I work here, whatever, but this is true. It has never felt to me like in any way, we look at potential clients and say, “We’re just gonna ram AR… ”

32:27 JJ: We’ve literally never done that.

32:28 JB: We don’t do that. [chuckle] It’s like, we only wanna put it in places where we think it’s gonna work and where we think that there’s a legit use case.

32:34 JJ: We gotta look at it like this. So we have solutions for a variety of vendors. We have a solution for AZEK, which is, the decking is very different than furniture.

32:40 JB: Sure.

32:41 JJ: And their solution is definitely customized for that.

32:43 JB: Yes.

32:44 JJ: Macy’s is different than Bob’s. The way that we present stuff is different in both of those things.

32:47 JB: Yeah.

32:48 JJ: Yeah so, it’s the difference between a specialty shop or a full-service company and a DIY.

32:56 JB: So at the end of the day, what we’re saying, I think my takeaway from this is, there are some sort of over-the-counter all in one options.

33:05 JJ: Sure.

33:05 JB: But to really get something that’s gonna work for your business, for retail, for enterprise, for ecommerce, whatever…

33:10 JJ: Yeah. You’re gonna talk to a different man.

33:11 JB: You’re gonna have to… It’s gonna have to be personalized.

33:13 JJ: Unless you’re a huge company that’s got your own internal… Yeah.

33:16 JB: An internal. Yeah, an IKEA, which has an internal 3D development team…

33:20 JJ: Yeah.

33:20 JB: Which can, you know, take some bare bones coding.

33:23 JJ: Although, we’d be glad had to help, IKEA.

33:24 JB: Yes.

33:25 JJ: We will definitely help, if you want us to.

33:26 JB: We have a lovely CMS, IKEA, come call. [chuckle]

33:27 JJ: Alright, well, that was a pretty good one, man.

33:31 JB: Yeah, I thought so.

33:33 JJ: I had covered some stuff.

33:33 JB: You really carried the weight this week. I appreciate that.

33:34 JJ: I did. I really did. I sound authoritative.

33:35 JB: You do. And I sound like I have terrible allergies and have been dying all week, so I really appreciate you carrying the load.

33:41 JJ: More than happy to talk more than you sometime.

33:43 JB: Yes, thank you.

33:44 JJ: Alright, well, thanks for listening in and we’ll talk to everybody again next week.

33:48 JB: Yes, we’ll see you in a few weeks.

[music]

Most Recent Podcasts

The future of retail with Sterling Hawkins

This week, we’re wrapping up Season 2 of the podcast with another very special guest: Sterling Hawkins is a global… continue reading

App-less AR with Anne Flynn Wear of Furniture Today

This week, we’ve got a special guest: Anne Flynn Wear is Associate Editor at Furniture Today, the leading furniture industry… continue reading

Oculus Quest and Valve Index: The Best VR Ever?

Rumors of the death of the VR headset market have been greatly exaggerated. New Virtual Reality HMDs now available for… continue reading