IKEA VR, AR, MR and Meatballs

MARTIN ENTHED: Thank you for joining for this session. I'm Martin Enthed. I'm from IKEA Communications in Sweden. And my time zone is 5 PM. So this is an afternoon session for me, and I've been up for four hours already. So my third cup of coffee's already been swallowed down. I'm happy to be starting out this day.

What I'm hoping to do during this 50 minutes that we have is to show you what we have been doing in the last 10 years when it comes to computer graphics for the first 15 minutes. Then I will show you what we have been doing when it comes to real time stuff. And at the end, I will show you what we are looking for in the future, and what we would like the industry to look at in more detail. And I hope it would be a little bit of fun.

As a startup question, is there anyone who hasn't been in an IKEA store in here? Please raise your hand.

[LAUGHTER]

I know you have. So then I can jump 30 slides about what IKEA is.

[LAUGHTER]

That's quite nice. I have some company slides. That's three of them. So I will start with something that you might not know. Let's see if this clicker works.

One thing that you might not know is our vision is nothing about furniture. It's to create a better everyday life for the many people. And for me, the really interesting part of here is the many part. We are trying to reach many. That means that we can't really go with a solution where we can have one person at a time. Not really. We can try stuff. But what we really would like to be able to do is give it in the hands of all our customers. And to give a number there, we have 100 million people in our loyalty program. So trying to reach 100 million is like some kind of rule of thumb for me. Then it's becoming really interesting.

Some other numbers to give you a size of the company. We have over 400 stores in the world. Almost a billion people visit our stores every year. Over 2 billion people are visiting our web page. And we have a turnover that is quite big. Last year, been growing. And we are almost 200,000 employees. So it's a huge company. And we are also present in a big part of the world. It's actually easier to say where we are not. And we are not in South America, and we are almost not in Africa. We are in some small places, but almost not in Africa.

Last year we opened 14 stores. So little bit more than one store a month. And this pace is growing because we are going into other areas. And just for you guys, IKEA is here and not in the middle of Europe. We are not Switzerland. We are Sweden.

[LAUGHTER]

Just so you know. It's in the south of Sweden. It's the middle of the woods. It's nothing else there than IKEA. And that's where I'm working.

IKEA is a company who is divided into many companies. And I'm working for a company who is called IKEA Communications AB. And we do the content for the web page. You have it up here. We do the IKEA catalog that you might have seen today. We also do the printing and all the paper work and everything. That's our work. We do the assembly instructions in our building. Please don't come with questions about that.

[LAUGHTER]

I know how they are done. I don't know them, just so you know. That's the normal question. And I don't have an Allen key in my back pockets.

[LAUGHTER]

We do about 300 films last year. And we are about 300 coworkers. So it's a small company. You could call it an in-house agency for IKEA for everything that is communication on a global scale. So those were company slides.

What I'm going to talk about now is 3D. And everything I show you now are renderings. This is computer graphics, all of it. It's nothing else. So the reason why we started working with 3D from the beginning is that we had a problem. We had these kind of images. It's combination of images. Images of a lot of products in a combination to a bigger thing. So here you have drawer fronts, you have drawers, you have a cabinet. And that makes an image. And we needed to do a lot of those because IKEA had a web page back in 2005. We needed to fill it with all the combinations.

So this then multiplies into these when you're starting to take all the drawer fronts. And then just for this combination, you have to do these. And then you multiply that by 300, and then you are up to 18,000 just to be able to cover the kitchen range in US. And then we have a kitchen range in Europe. And then we have other ranges that are combination like this. And of course that runs into problems.

So the basic idea was back in 2005 to start using computer graphics. And then we worked for a year and we made one image. So one whole image. They did a lot more images. But one image that took its way all the way to the customers in the IKEA catalog actually. That was back in 2006. Then we struggled and we did a lot of work during 2005, '06, '07, '08, and up to '09. We did one room. The first room that was in the IKEA catalog that was fully CG was in 2009.

Then we started to build our 3D model library. And it took us about three years to come up to about 50%, 60% of the full range. All the things that are hard-surface stuff. We didn't do towels, we didn't do flowers, we didn't do the fluffy animals that we sell as toys for children. Those things we left out. We did all the hard body stuff. And we also, back in 2009, set a library standard. So we had a texture library where we capture all our textures in very high resolutions. If you don't know what a texture is, you can call it a pattern. So here you see they are textile patterns. You see wood, you see everything. And we do it in very high resolution in very, very high detail. And the corresponding normal maps, speculum maps, all the other things for it.

We also built a material library on top of that. And a material library for us, then, is set in a specific lighting conditions. And then we use those textures when needed to be able to build this library. And these libraries are not connected to specific product. These are the ones that we use for every product. So we do one oak of a certain time. And then we use that for everything. That's the basic idea.

And materials is made in a very high detail. It's meant to look photo real. That means that you go into very high detail when you build them. This is a about 40 centimeters wide sample. That means little bit more than a foot. Here you see the corresponding normal map and height map for it. And here we go closer and closer into that textile. And this is 2 by 2 millimeters, that thing. And this is then 10 centimeter across. Half an inch across. To be able to do this and the data we need, we need to capture it like this and create it.

Then we create 3D models. And all the way back since 2009 we have created four LODs. So the leftmost one have a 10 centimeter accuracy, and the rightmost one has an accuracy down to 0.2 millimeters. And that's to be able to cover the renderings we need to do. We render up to 5 by 5 K. And to be able to do that and then get photo real quality, you need very high textures, and you also need all the polygons there to be able to render.

IKEA has a style when we are finished filling rooms and building scenes that it should look lived in. It should look like you are just leaving the room. That means that we need to build stuff that we don't sell. So here you see a selection of fruit from our propping library. We have never shown this before, so it could be nice to see that. We were, in the beginning, buying this from outside. But we couldn't get the quality level we needed. So we had to build it ourselves. And I happen to like raspberries a lot, so I'll just show you the raspberries.

[LAUGHTER]

And you see the detail level we have of these to be able to do what we need to do. The base for everything is a 3D model library where we are up to about 33,000 models at the moment. The texture library and material library and the propping library. And when we started it all, it was mainly to create high end images and also fill 3D models to our home planner, our online planner for kitchens. But today we are using these high end models for many different things. And we're just adding things to this. So we have everything from SketchUp over here to room renderings, Revit, home planners, AR catalog app, VR apps, and also things going on in the different parts of our marketing things that happens in every country.

So we have a lot of different use cases for the same model. And that's the reason why we do with a very high end one, and then we use it for different use cases. So use cases. Of course we do room renderings like this. And room renderings like this, we make about 10,000 to 17,000 a year. And we would guess that we are up to about 20% of them are made in 3D. And the rest is taken in photos. So somewhere 20%, 25%. That's what we are at.

Then we do these, what we call product pictures. About 25,000 last year, and different types of product pictures. I would guess that we were up to about half of them are made in computer graphics at the moment. And then the huge portion of these combinational images where we do the same kind of thing, but with different door fronts and so on. And I would say that 95% or almost 100% of those are created in computer graphics. And this is, if you remember back, why we started this from the beginning. This is why we did that.

But, of course, we are looking into new ways of communicating and having computer generated images. We were able to do these kind of things. And that's hard to do in real life because you get a lot of people and hands in the way if you want to do it. These are tested at the moment at the Swedish IKEA website.

So to be able to do this, we create about 3,000 3D models a year. This is actually also 3D. And we are also now looking into going and doing the models we have skipped before. And this is the kind of model we have skipped before. So up to 2020, we are thinking about doing about 6,000 3D models a year to be able to cover for the rest. And the rest is products like these. And if you have done some 3D modeling, you know that these are hard. That are stupid hard. And of course these are hard because you don't really know how to build them, because when you build them, you build in some kind of setting for it. So you would like to be able to simulate it.

But just to show you what we are thinking about doing when it comes to these kind of models, this is a test we did. And you will see it here. Unfortunately, there's no really standard way of setting here. So we will have to go with what we can do. But it's a quite nice rendering. And just to show you some images on what we are doing, these are 3D then. Again. And the main reason why we do this is that we do these kind of changes and images. And if you look at a web page somewhere in the world, you will probably see one of these. But for another country, you see something else. Again, I'm stressing that this is all 3D, all computer graphics.

I think you will see an example here that it's painful to do in the studio. You see the worktop is a little bit thinner. And kitchen is a specifically hard area to do in the studio. So here we are doing a lot. Here, you also see a changes. You see the fridge and the hood is going, and other things there. Of course, we can do these kind of beauty shots. We have all the 3D models. So we can do it.

And just to show you that it doesn't have to go on the web page, this is a brochure for US. I think I have it in my bag and can show it. You can pass it around. I would love to get it back because I'm not in US that often, so I can't get my hands on this one. And if you look at that one, you would say that about 95% of all the images in this one is CG. And the prices are super low here, if you see that.

[LAUGHTER]

It's before we send it to the countries and they set the price. So it's not that low, normally.

AUDIENCE: I'd pay a lot more than that for my CG.

MARTIN ENTHED: I hope.

[LAUGHTER]

And just to show you some little, little thing in there. If you see this to the fellow here. He's actually a 3D model, our first little cat 3D modeling in our propping mag. You will find him on a page somewhere in there.

This is something we did from 2005 up to 2016. It really started to take off 2012. So it took us about six, seven years before we could start to volume. So that's where we are at the moment. And the reason why I'm talking about that now is that this is what we be building the next steps on. And that's real time.

So we really started looking at real time back in 2010. This is a test we did internally. It was really early versions of WebGL. So what you see is a sketch-like thing in the web browser. And then we sent information about where everything is standing. And a web server on the other end is loading everything in high-res, rendering, and sending it back in 10 seconds. And this worked. The only problem is that we have a little bit too many people on our web page. So we couldn't really cater for this. We would have needed at least 1,000 servers standing by, and we couldn't really scale it in a good way. So it worked, but we had to put it on hold.

So we started again in 2012. And this is WebGL here on the client side. And here you see it in a game. And this is Unity. And it worked for about a third of our range. And the rest didn't really work. And the reason is that it didn't look good. So some of this we have materials on it that could look fairly OK in an online rendering.

Just as a side note, this was the same time we did the first augmented reality thing when the IKEA catalog. I don't know if you remember, but we had a catalog out. You could use the phone and look at the catalog and get a 3D model on top of that. So we did augmented reality at the same time as we did these internal tests.

So then in 2014, a lot of things happened at the same time. I don't know if you remember this, but this was one of the Unreal Engine YouTube videos that came out showing off interiors that looked fairly OK in real time. We had things going on with Oculus, and Sony had their Morpheus thing coming. Google were talking about their capture device Tango, and Realsense was doing the same thing. Both of these are maybe not there yet, and maybe never will get there. But this was happening at that time. So we said, OK, we think that real time will be very important for IKEA somewhere 2019, 2020. So we started a project internally. We call real time 3D.

Then during that fall 2014, I had a very personal turning point. I started with computer graphics in 1990 and I've been very passionate about it ever since. But I got a refuel on my passion when I went to a place the 4th of December, 2014 at 1:36 PM, Pacific time.

[LAUGHTER]

I was led into this room, and I was told to put this on. I felt a little bit scared. This was up in Seattle at the Valve's office. And this is the pre-version of the HTC Vive. And for me, this was a total turning point. Because for the first time, I could spend two hours on the other side of the screen without getting motion sick. And I get motion sick very easily. The graphics wasn't great, but it was stable and it worked. And when I came out of those two hours, they asked me if IKEA would like to be one of the first companies getting a headset and trying it out. And I said yes.

And in January 2015, we got a headset, in late January. In February, we took this. This is from a kitchen brochure. And we put it into Unreal Engine. And then I managed to get a meeting with one of our top managers. And this is him. You see that it's early equipment because you see that this 3D printed hand controller sends cables everywhere. And I'm putting him into the room here, and he's standing here at the stove here. And I'm asking him to go up to the micro and read on it. But he's holding both hand controllers in one hand, and he has one hand free. But he's so immersed. So he is grabbing the micro. And we happened to be there with a camera.

[LAUGHTER]

And when he stepped out of this, he said, this will be big. We need to do something. So I was asked to start our Virtual Reality 2 project, and that's what then later became the IKEA VR Experience. And the basic idea here was to try to see if we could get image quality that was OK, if we could see how you interact and move. This is early 2015, so we didn't have that much to look at. And then I wanted to do something that you can't do in real life.

Becoming a child was an idea we had. And the reason why this is interesting for IKEA is that we have a long-term priorities, living with children. So could you shrink yourself to child size and feel that? And then can we get a correct feeling of size? Of course, that's super important because then you can take a buying decision in this. And of course, in some way, that is interesting for us.

So we looked at what we could find, and we found this kitchen here in the kitchen catalog 2016. We could get it to render at about 10 minutes per image in the normal flow, the high end flow that we talked about. And we needed to go down to 11 milliseconds, 90 frames per second. So we needed to run 60,000 times faster. And that was the problem, really. And this we did during June and July that summer. And what you see here is directly in the game engine. Of course, we used every trick we could to be able to do this.

But the experience by itself is something that IKEA has done for many, many years. I would say that IKEA has been running virtual reality for at least four years. If you go into the second floor in every store, you have rooms up there. Have you seen them? You've seen them. Great. Those are not real. Nobody lives there. So those are virtual rooms that you can visit. So what we really did was taking that concept and just moving into virtual reality, nothing else. Just a short little thing. That means if you spend more than five minutes in here, either you're very interested in this kitchen, or there's something else that you are doing. Normally it's a five minute thing.

What we also did was doing a huge internal test. So we were having four rooms in our basement of IKEA 1. That's the original IKEA store. And we have over 1,000 coworkers going through tests. And I was sitting back there in the sofa having interviews four by four with these people. And we were expecting 50% thumbs up, 30% mmm, and 20% we don't like it. Because you should know that I think we had two people in here that said that they were gamers. So there were non-gaming people. Most of them are about 30 years of age, and half of them are women, half of them are male. Mainly managers within IKEA. But we got 99% thumbs up all over.

And I would say that if you have a device that works like the HTC Vive and you have non-gamers going in there, the room scale tracking makes it very easy to do this for the first time. I don't have to tell anyone to go two steps forward, and lean over to see the other side of an object. They get that directly. The only thing that is really important is that you set the floor height correctly, so you're not 1 centimeter off up or down. Because the size gets into trouble directly.

Then fully tracked hand controllers, of course, very interesting, in the same space as the other thing. We had to teach them which button did what. But then after that, reaching out for a drawer to open it, then that was natural. So then that worked. And then, of course, the virtual size and the real size the same. In this case, it actually works very well. We are working in 3 to 5 meter range. In this range the parallax of your two eyes is helping you a lot with the size thing. So we got a really good experience out of this.

We actually also put a virtual share in the same position as the real share. So you could reach out for that and do that. That's actually also still available in the Steam version.

So as this went very well, we were allowed to publish this on Steam. And we did that the 5th of April, the same day as all the other Steam games for the HTC Vive was coming live. And we did it with this little thing on Steam, and a press release text thing. And we did it at 8:00 AM in the morning. And I was expecting a few hundred downloads, but it boomed. It just took off. I have no clue, really, why. Maybe it wasn't that much content there. I don't know. But it was really nice seeing how many people were downloading. And I will show you a short film here that went up on YouTube at about 11:30. So 3 and 1/2 hours after I pushed publish, somebody downloaded, played, recorded, and uploaded this thing.

[VIDEO PLAYING]

CHRISTIAN DONLAN: Whoa. I have that whisk at home. Very exciting. This is actually made by IKEA. And it seems very much in line with the way-- In the background you can hear children. Delighted children living in a Scandinavian paradise. That's not going to get old. In Scandinavia, this is what the view looks like. It's just a socialist paradise stretching out in every direction. OK, let's have a go at this. I'm tangling myself up terribly, so I might die while doing something with this. Whoa, this is amazing. Are you witnessing this? Oh, that's not so. No one's happy there.

[LAUGHTER]

You might not want to eat this when it's done. What's that say? That says shoveling snow. You. See, a little bit of family tension there. I like it. It's crazy. We have this. This thing, I have this. I put it together. And let me tell you, the wheels are a bit of a pain to get on. They've got books in it. That's a nice idea. We've just got old packets of Lady Grey Tea in ours. [INAUDIBLE] Where are you, Chris? Are you over here? Would you like some of this? I made it.

CHRIS: [INAUDIBLE]

CHRISTIAN DONLAN: It's fresh. It's lovely. That looks like-- is that cilantro on top?

[LAUGHTER]

MARTIN ENTHED: So this was the 5th of April. And in the press release, we said we're doing this in the name of co-creation. We've listened to the customers who download and what they would like to have. And so we were listening in on Twitter, Facebook, Reddit, everywhere we could find. And I did that for one day, and two days, and three days. Then I knew what everybody wanted. But I didn't really know how we would do this, because the overwhelming response was, where is the meatballs.

[LAUGHTER]

You can't have an IKEA kitchen without meatballs. This doesn't feel real. Where's the meatballs? All those kind of things. And I was really not expecting that. I was expecting, I can't open that many drawers, or I would like some other colors, or something like that. But you couldn't really neglect that. So I called the manager you saw earlier, Jesper, and I told him, we know what they want. Yep, give it to them. No, wait. They want meatballs. What? Yeah, but I have an idea. We have just started to sell veggie bowls. So if we put a bowl of vegetables and a bowl of meatballs in there, then we get some thing across here and we can-- Yeah, do it. But don't tell anyone.

So we did this.

[LAUGHTER]

This is the only communication we did. Nothing else. No press releases, nothing. And social media went boom.

[LAUGHTER]

And more YouTube videos, of course.

[LAUGHTER]

Launching something on Steam is a little bit scary for anyone, and especially for big--

[LAUGHTER]

Especially for big brands. You get a community site.

[LAUGHTER]

You get a community site for free, and you can't really turn it off. And I just want to highlight something that somebody said on that community site.

[LAUGHTER]

If you know about Steam, then you know that the audience here is people who should be 13. I don't know if they are.

[LAUGHTER]

And they are up to 30. And mainly males. The people you normally drag into an IKEA store. So of course, this is in the audience that we would love to get more in contact with. The only problem here also is that there is not that many devices out there. There's not millions of devices out there. So for us, it's even more important the videos people make of this. Because those will be shown by more people than are able to actually experience it.

So we played along. So we did the stealing pencil update little bit later. We added the IKEA catalog in there. We also made it possible for you to be a cat size. And we added a lot of other things in there. And in August, three months after this, another guy made a video. And first he's playing Star Wars for five, six minutes. Then he's in the IKEA kitchen for 30 minutes. And then he goes underwater for a few minutes. And in the last three minutes, he's going really berserk. And I'm just going to excuse his language. I didn't put beeps in here. So this is him, not us. I will show it now. You will see. It's about three minutes long. And I love it.

[VIDEO PLAYING]

GAMER: But my favorite thing of all-- actually, this is how they [INAUDIBLE]-- That's fine-- is you can become cat-size. And everything's massive. It's huge. And you can jump onto surfaces, et cetera, and have a look around as a cat would. We can be a cat on a catalog.

[LAUGHTER]

Oh, it's much easier to read as a cat. Incredibly large amount of detail in that photograph. We could be a cat over here. We can be like, oh, meatballs, I'm a cat. Well, can we cook meatballs as a cat?

Let's reset everything. Change the thing. There's a reset button. I'm not paying attention, am I? OK, cat size again. The first thing I would need to do is I should get slightly a little bit over in that direction. I'll need to open this drawer and then pull out this absolutely massive frying pan. [INAUDIBLE] cat mode. I meant to grab the handle. Hang on. Other side. Shit, I'm in the wrong place. It's so hard to do on the teeny, tiniest little bit of a thing that I've got. Pick it up! Yes, I have a frying pan as tall as the devil [INAUDIBLE]. Hang on. Let's just close that up there. I managed to get it up there. I keep trying to fling it.

[LAUGHTER]

[INAUDIBLE] with this hand, but I'm holding it. I can. Oh, saves me time. Right. So we can teleport up to here. Marvelous, great. Boop.

[LAUGHTER]

Now I need to get to the meatballs which are slightly out of my little bit of range. But there we go. I'm actually playing a giant's cooking game.

[LAUGHTER]

Oh shit, don't break. Be careful with those things. This is the weirdest thing I've done in me home. I'm loving it.

[LAUGHTER]

It's huge. I feel like I'm some sort of Saturday morning TV show from the 90s. There you go. Now we just fry them up.

[LAUGHTER]

I could look around my land. Going to put the extractor fan on. So I go up here as a cat and lean over, and boop it from there. Oh my god, that is such a long drop. That's actually quite scary of how far a drop that is. I can't really estimate the distance other than very far, very far. It's nice up here, though. I think I can do being a cat. There's a window there. Can we have a nosy out the window? We can get quite close. It's all that fucking nothing we moved into. It's all that fucking nothing. Doing it as a cat. We should never move there.

Right, but my meatballs are burning. Shit! OK, right, I got my meatballs. And here, [? mistakes. ?] Here's your meatballs. Lovely, very lovely. Right, let's just get some help [INAUDIBLE] going on. [INAUDIBLE] Come on. Back up there. Ow, ow, ow. Still pretty warm from [INAUDIBLE] so we shouldn't go there.

That's the IKEA VR thing. You could dick around as a cat who's also a chef. I feel like I'm in a Pixar movie right now. Catatouille. We could call it Catatouille. To do a proper impersonation of a cat, I have to go, this looks like a thing that you like. We should put them on the floor.

[LAUGHTER]

Why are cats such pricks?

[LAUGHTER]

MARTIN ENTHED: It's quite fun. I hope you liked it. So why am I showing this? This has over 400,000 views. So just this video has over 400,000 views. And that's more people than have download the app we have made. So I guess these kind of videos, at least at the moment, gives more attention and get more people interested in what we're doing than the actual applications, at least when you are really early.

There are no official statistics on things, but I found two that I've tried. And the first one was after a month on all the VR thing that was available for HTC Vive, IKEA was actually number five. And this is a little bit embarrassing almost because I still think that Tilt Brush is one of the nicest ones in there, and we are before Tilt Brush. That doesn't really feel good, even if it feels a little bit good.

And then after five months, there was somebody who claimed that they had looked into [INAUDIBLE] all the branded apps from big companies. And that was five months in. And we were number one, and we were before Star Wars, and ILM and Disney. So I don't think that will happen ever again. So I will keep that forever. Also, just recently where we were nominated to the most creative use of marketing in VR among all these other guys as an event in Langban. And we won that. That felt really good. That also concludes that project. This was a three-year project, but really fun to do. And we were really early.

We do, of course, other things. We do 360 videos. This shows one that was part of the IKEA catalog. So here you see me running it in the view. Here you see the real renderings. And these are 3D created. These are not photos. We do them in 3D because it's easier to do. Of course, rendered time is long, but it's easier to control.

Other things we do in VR. What you see here on the left is this VR 3 project we did. It's the IKEA VR pancake kitchen. And that was actually part of a store test where we tried to use VR to transfer knowledge over to the customers. We have also now released it on Steam. So if you want to do pancakes in VR, it's there so you can try it. There's also a mini game in there. But the reason here to do the pancakes was to teach people about corner storage, work triangle, waste sorting, those kind of things. So that that's really the reason for it. It's not really a game. So that's there.

What you see in the middle here is an internal tool we are starting to use with our designers. They can walk around in virtual reality with the new products and take images that will then be high-rendered on the server side. That's running actually in the early versions of Web VR. And what you see on the right side over there is IKEA Australia, who has been trying to do VR, or 360 images of a store. So if you go to IKEA Australia, I think it's still there, publicly available so you can walk around the store if you want to do that in VR.

More things during 2017. You might have heard that we are release something together with Apple, something we call IKEA Place. Really a follow up what we did back in 2012 when we did the [? ARF ?] with the catalog. Now we don't need the catalog. We have ARKit so we can track so we can set products in the space. And it's really, really good, actually. The size accuracy is much better than we ever hoped for. So that's out there, you can try it. About 2,000 products in there.

So this is where we are at the moment. But this doesn't tell anything about the future, and also doesn't tell why I'm at these kind of conventions talking. Because we don't sell services and we don't sell software. So why am I actually here? We are here to talk to you guys because we think that if we can put our problems into as many heads as possible, then people might want to solve the things we need in the future. And we don't have to hunt that down. So that's the reason why I'm here.

How do we attack this problem? And what needs to change? And why are we interested from the first case? The last one is easiest. We sell products and function in a home. We don't only sell products. We actually sell function. So there's a reason why this is at the end of this. We try to sell a functional home. And as we are doing that, we need to be in a three dimensional space. And of course virtual reality or mixed reality, augmented reality, is all about three dimensional space. So this is the reason why we are interested in this.

But what needs to change to be able to do this in the large scale? It's a lot of things, actually. And I've been talking about this before. If you've seen any of my talks before, you will recognize this one. This is our biggest problem at the moment. We can transfer 3D models over to almost any application, but we can't move the materials in a good way with the high fidelity we would like to. The base color, of course, but not really.

So we would love to have any generic material description created once, and then translators over to-- let's say that this is WebGL, and here you have Unity, Unreal, or Stingray, or Max Interactive. Or here you have V-Ray, or Arnold, or something. That's what we would love to have. But we haven't seen anyone in the industry trying to do this. And we really, really need this. And all companies who do what we do need to have something like this.

Then we would like to be able to do the LOD levels automatically. Just doing the rightmost one, the most high-dense, and then being able to get that down into different levels. And our biggest problem is if you have a shape like this, almost all LOD systems, or polygon reduction systems, don't create this. They don't go to the outer volume. They go to an inner volume. And that's our biggest problem. If you have a solution for that, then just come up to me, please. And then you have to keep the UVs, you have to handle all that. Of course it would be great if we could have everything in solids because then we could do defeaturing. But unfortunately, people don't like to do sofas and so on in solids. It doesn't really work. So we have to have a set that works for polygon models. Otherwise we can't really work with all the range that we have.

This is something we also are lacking. How do you describe combinations and connectivity? We don't find any standards out there for this. If I want to tell this where it fits on that back end, and where those fits in that so I can build this. Of course I can do it by myself, just three dimensional points. But then I have one way of describing it to one configurator, and then I have another way of describing it for another configurator. There's no real standard for this. I would love to have that. I think we internally already have 10 standards. This is within a company where we should be able to do this. But this is something we have trouble figuring out.

I have tried to make IKEA stop selling textiles for a while.

[LAUGHTER]

I haven't given up yet, but I think I have to. We have to solve this. We have to solve to be able to do what she did there with the cushion and being able to arrange stuff. We are selling these things, so we can't just skip these products. And how do you do that in real time with 90 frames per second in a HoloLens that can't even have the data in there? Of course, you can't do it today. But we need to do it someday.

And then I'm coming back to this. We have the high end library here. We have to create the real time library today to be able to cover all these different things. So we need to do this in an automatic way, and we need to find formatting that could be handled by many. Unfortunately, there are no file format here that can be handled by many. And this is not really automatic yet, and we're trying to do that. And the real reason is that we would like to reconvert to real time maybe every three months or six months. Because real time quality is a moving target. It's getting better all the time. And these we do at very high def at the moment. So that is also moving.

So my biggest problem really is that we are not really fitting into other companies' verticals, or whatever they call it. So we are in the middle of game, film, CAD, and architecture. Film because we do the quality they do. Game because we use game engines and we store asset for a very long time. CAD because we have a lot of CAD data coming in. At least half of our products are coming from CAD. And then architecture. Everything we do is an architectural space, so we need doors, windows, power outlets, all those things.

So going real time, we are still aiming on the other side of 2020 for the large use cases. And if you remember the time line we had before for the computer graphics, for the JPEG creation thing, it has taken us 10 years. So we are on the first steps here. We are three years into something. So it will probably take a few years before we are going large scale.

And of course, IKEA is a company who is not on the stock market. We can have 30 year plans, and of course we have. And I will tell them-- no, I won't tell them.

[LAUGHTER]

No, I can't. But I think most of you in here think that it will be a huge change. And we think that, too. Just looking back 10 years or 13 years. Let's say if we look forward to 2030, that's 13 years. Going back 13 years, that's before the smartphones. It's when we had the Palm Pilots and thought they were the best things in the world. And then go up to somebody with a Palm Pilot and say, will this be for everyone? Probably they would say no. But look what happened.

So of course people won't buy hardware or devices if IKEA has an app. We don't think that. But there will be other things happening at the same time. And of course, we would like to be there when it happens. So what we have decided to do is start a little lab, and I'm running that at the moment. And that's what we call IKEA Digital Lab to keep track of this change. And to be able to keep track of this change, we have a lab. And the lab is working in nine areas. And this is the first time we tell about this.

So those nine areas are virtual reality, and mixed reality, and rooms. And for me, mixed reality is really glassware. I talk mostly about augmenter reality when I talk about phones. So of course it's interesting, that too, but I'm talking about glassware. We're not talking about mixed reality. So virtual reality, mixed reality, and rooms. We're talking architectural space. Doors, windows, those kinds of things. Then we have the product without the room. Of course that's interesting to look into as a special case.

Then we have the simulation of products. And simulation here can be simulation of a surface. One of the worst things we have is semi-transparent plastic with the subsurface scattering stuff going on in there. It needs to be doing that. Otherwise it can't really show off as a real thing. And then, of course, textile simulations, configurations, all those kind of things is in this one. That's the first three.

Then we would like to understand the room you're standing in. So we would like to somehow understand that you are standing in a kitchen, or in a living room, or in a bathroom, and using probably artificial intelligence to understand that information from a point cloud or from images, or whatever your device is giving us. And then we would like to recreate that space. Giving you new ideas and placing stuff in there, or replacing stuff in there. This is something we also are looking into. These five is mainly talking about the visual sense. But of course we need to add sound.

So sound simulations in 3D. How does sound behave? Sound rendering. Then we would like to look into experience as somebody else. We are talking about child mode. But we are talking also about if somebody is visually impaired, you don't see color. Those kind of things. Experience as somebody you care about that you would like to, OK, how does this space looks for this person?

Sharing experiences. Most of the virtual reality or mixed reality stuff that is happening at the moment are lonely experiences. And if you bought anything at IKEA that is a little bit bigger than a tea candle, then you probably have taken that decision with somebody else. And of course sharing that experience in a virtual space or in mixed reality space is important for us. So sharing is interesting.