Hacker News Reader

Comment by idopmstuff | Hacker News Reader

I bought one and was really disappointed - the visual didn't feel as 3D as I expected. Their marketing videos make it look like things really feel like they're coming up out of the board, and to me it just felt like the board was sort of a TV laid flat on the table. Things were 3D only in the same sense that they are on my monitor.

Also some really bad UX stuff - you have to have the glasses plugged into an Android phone or Windows PC. I have a Windows laptop, but the cord to plug in was so short I almost pulled it off the table a couple of times. I Googled, and that was a pretty common complaint that came up, which is just such a silly thing to get wrong.

The controls also weren't that good - when pointing sometimes it'd just lose track of my controller. I tried with a few games and never had a particularly good experience.

Overall, this was one of those things where the marketing made it feel like I should expect a magical experience, and the reality was very disappointing. Thankfully, they have a return policy - it's buried in their site and confusingly worded, but I was able to ship the thing back to them and get a refund.

View on HN

> Their marketing videos make it look like things really feel like they're coming up out of the board, and to me it just felt like the board was sort of a TV laid flat on the table. Things were 3D only in the same sense that they are on my monitor.

You don't by any chance have an undiagnosed stereoblindness do you? The promo videos make it look like extremely classic fiducial marker positioning of 3D models. The rest is up to your eyes and brain to be able to see depth from stereo.

Have you been to a 3D movie? If so did it feel like 3D to you?

The only scenarios that make sense from your description of it looking the same as a computer monitor are that either you got a broken one or you personally cannot see depth from stereo. I want to add that brains are very good at hiding this. If you've never used VR or been to a 3D film, its entirely possible that you could be stereoblind and just don't know.

I've done 3D movies, and they look as expected.

I guess I'm not quite being accurate when I say it was exactly like stuff in 3D on my monitor - it did sort of look like things were 3D, but meaningfully less than what I expected. Honestly, I'm having a hard time describing it accurately, but suffice to say it was extremely underwhelming.

Good to know! Thanks for the response.

This was also my takeaway from trying it at AWE2022. 3D works great on me, and I’ve always loved stereo 3D stuff even just for the gimmick value. But Tilt5 didn’t feel close enough to the experience of viewing through a hololens or looking at a real 3D display. I think that was especially obvious at the show, where vendors like Looking Glass and Sony had 3D displays with much more impressive depth and rendering capabilities.

I get that Tilt5 is taking a different approach and has a reason for it - I’m just not sure that approach is actually good enough in practice. I’ve only used it for the simple demos at the show, but I remember thinking the resolution looked like it might be a limitation for doing more complex games too. Most of the board games I enjoy the most but which are complex enough I’d want a digital alternative involve lots of cards with small text on them.

I mentioned this in another comment below, but it likely applies here as well. We tune the interpupillary distance (IPD) software setting down a bit to accommodate a wider variety of face shapes. If the IPD is set to a distance greater than the physical distance between their eyes, then the effect is very uncomfortable to look at as their brain has trouble "unifying" the two images each eye is seeing. If the IPD is set to a distance less than the physical distance between their eyes, then the effect is perfectly comfortable, but there is reduced depth perceived from the images.

We err on the side of comfortable but "flatter" for our default setting and for trade shows and conventions, but adjusting the setting to be appropriate for your face is just a slider in our control panel. We also have a calibration app in our latest public beta drivers for people to find their IPD if they don't know it.

I hadn't heard people complaining about this at trade shows before, but it makes a lot of sense. We should probably incorporate a quick calibration at the start of our demos. Thanks for the feedback!

Aha, this makes sense - especially because I have a relatively narrow IPD! We also weren’t using the mat with the tilted sides, as I recall, which probably makes an impact. Thank you for the explanation, and good luck with the platform! It was really exciting to finally use the tech after watching it develop for so long and I do expect it’ll be part of our tabletop group one day. I mentioned it elsewhere, but if you folks could license some of those complicated, mini-heavy board games like big trouble in little china, it would be a dream. Finally play the complex games without the complex setup.

Our company started building AR software around the time of the Tilt5 kickstarter and seeing your team’s efforts was really inspirational to us. Congrats on all your success so far, I look forward to giving it another try…and calibrating my eyeballs first.

I have also tried one and I had a similar experience but only in some scenarios. It might be a different effect to what you had but I’ll try to explain.

In my experience I got a good 3D effect for objects that were right in the center but the moment things got close to the edge the illusion broke. At times it did look a bit like 3D content on a 2D screen.

Objects near the edge of the retroreflective area disappear when you would normally expect them to appear ‘above’ the mat. They somewhat get around this by presenting most information below the surface. Apps like board games were the worst for it because they typically use the whole mat and they typically appear on or above the surface. Anything above the surface just gets cut off around the edge. The larger board that can be tilted up at the back helped but only from one angle.

I don't have stereoblindness and I agree with idopmstuff. There are several problems that I think stack up to make a relatively bland experience.

1) The cable is indeed too short. And when I got mine, they didn't have Android support at all. You were expected to have a full PC for every user. That was not clear at all at any point through the Kickstarter or the subsequent marketing they did, showing people playing games together. I could see having people plug devices into phones in that scenario, but dragging out their own gaming laptop for a game of "D&D with animations" is absurd.

2) The display resolution is actually pretty low. While the solidity of the graphics and FOV are both a bit better than, say, the HoloLens 1/2 or Magic Leap 1 with their janky waveguide displays, the degrees per pixel is relatively high, making the images kind of blurry and pixelated. This also means that there is a limit to the distance at which it can resolve stereo differences significant enough to be visible.

3) Their concept of placing the retroreflector pad on a table is just wrong. It's intended that way to allow multiple people to play around the pad on top of a table. But that biases your view of the pad, giving you a narrow frustum to look through. There's also the fact that the graphics don't "pop" above the pad like they show in all their promotional material (though I know enough to have not expected it, it's still disappointing to see them make those misrepresentations). The device is actually much better if you mount the pad on a wall. Then you have full view of it.

4) Points 2 and 3 above combine to make a graphics miserably hard to read. The display resolution is independent of the size of, orientation of, and distance to the pad, which is good. But your displayable area is limited by both the pad and the display. You can only see graphics in the Venn AND diagram of those two. Which can often be a very narrow area. If there's any text or small details on the graphics, they get very, very hard to read.

And that's just the display. I could go on at length about how awful the controllers are. They may be the worst controllers I've ever used. I'll just leave it at that.

Jeri Ellsworth was kind of a hero of mine, post-college, discovering my own hacker-ness. I still follow her on Twitter and appreciate her take on a lot of things, but it's definitely disappointing to see how much she pushes this device that is really substandard in a lot of ways.

> There's also the fact that the graphics don't "pop" above the pad like they show in all their promotional material (though I know enough to have not expected it, it's still disappointing to see them make those misrepresentations)

This was a tough call. Originally we shot 100% of the CastAR video through the glasses and it worked great, but didn't really explain the 3D "pop". So in our second pitch video we sometimes allowed a small amount of synthetic image beyond the edge of the display which gave people a better clue as to the 3D nature of the image, but also for some people did overpromise. Because of that I always look for this in T5 videos and in general I think they don't drift any or only a little, but it's not like I have seen them all. Also note that the T5 board can tilt up at the end (don't know if this is released) so you actually can look above the tabletop and see things, though again, I don't know if this has been demoed in any videos (I'm sure it's been demoed at some conferences).

Deciding to compose the image rather than shoot through the glasses can also be misleading, but for various technical reasons shooting through the glasses is also negatively misleading because the camera and the eye don't work the same (among other things the image is brighter in your eye). There's no perfect answer, but I think Jer and the rest of the team do want to get this right.

At the end of the day it's an arcade and tabletop game system, not a AAA gaming rig, and doesn't try to be the latter. Perhaps they should make that more clear. I actually plan to use it in my current work for looking at CFD* video as a group (each of us with a headset) because we can walk around the table, lean in, and point things out to each other.

I also like to put the game board on the wall BTW!

* computational fluid dynamics -- analyzing plume performance in the atmosphere

Also, no support at all for Mac OS or iOS seems like a deal killer at the price point.

Maybe it’s just the circles I run with, but I think it would actually be pretty unlikely to select a group of (say) three friends chosen at random and have each of them have use daily a Windows laptop or an Android device.

For instance I have my gaming desktop and work laptop (both Windows), but I doubt I can install the software on the laptop and I’m not running a cable from the desktop. And for me, 100% of family members and 90% of friends, no joy on the android device front.

In theory, you can plug in 4 to a single PC.

So at least only one of you would need the computing end of it.

Co-founder and CTO of T5 here.

> 1) The cable is indeed too short.

Yup, fair enough. We designed the cable more around using the glasses with a mobile device, and then ended up focusing more on PC support on the software side. We have a longer cable in the works.

> And when I got mine, they didn't have Android support at all. You were expected to have a full PC for every user.

We now support plugging up to four glasses into one PC, and we have beta Android support available for devs to play with. Focusing on these was, to some degree, what delayed getting started on the longer cable, though now I wish I'd started pushing on the cable sooner. C'est la vie.

> 2) The display resolution is actually pretty low... making the images kind of blurry and pixelated

There's a few things that were probably contributing to this. For one, we've had some manufacturing issues where our projectors could become defocused during shipment or as the glasses warm up during use. We have a much better handle on this now, and if your projectors are blurry I'd encourage you to contact our customer support so we can repair or replace them.

A second issue was that our SDK shipped with some bad defaults initially that caused apps to render a larger FOV in order to cover some of the pincushion-distortion corners of our projectors at the expense of resolution in the center. This was my fault, and we've since changed the defaults. There's still more to be done in this area, with things like adaptive view frusta focusing on the gameboard.

Yet another issue, I think, was apps (including some of our demos) not applying antialiasing in the best way. There's a temptation these days to just turn on TAA at some default settings and call it a day. That works alright for high resolution rendering, but for our glasses we've found that good ol' fashion MSAA can make a big difference in how the images look.

All that's to say that I think we've improved on image sharpness in a number of ways, and we get pretty good looking images on our glasses these days. I guess I'm kind of biased, though :)

> 3) Their concept of placing the retroreflector pad on a table is just wrong.

I would say this depends a bit on the use case. We've had a lot of fun getting people around a table for stereo couch-competitive games, but I was also talking with someone yesterday about using our glasses for CAD design reviews where a vertical or tilted gameboard would probably make more sense. We've added support for custom gameboard orientations in recent drivers, mainly to get feedback from developers and adventurous users on it (it's a little user-unfriendly at the moment). It's not clear right now exactly what user-facing functionality we'll do for this, but it's something we've talked a lot about, and I appreciate the feedback.

> I could go on at length about how awful the controllers are. They may be the worst controllers I've ever used. I'll just leave it at that.

Oof :)

I'll say this... The controller design was basically taken as-is from the CastAR design, as we just didn't feel we had the time or resources to do much to its design. For a long time I was pretty down on the design, as tracking it with LEDs in a straight line is tricky since it's missing a visible degree of freedom (roll). Our initial tracking implementation left a lot to be desired, but we've since gotten full 6dof tracking to work surprisingly well. I'm actually kind of a convert when it comes to using what's effectively a barbeque lighter to interact with virtual objects. It keeps the controller from occluding a lot of the view while providing a nice physical reference for pointing or interacting with things.

As for the concept of turning the controller sideways to use as a traditional controller... I think it could still use some work.

While it's good that you acknowledge the limitations and don't lie about things, at the same time, I'm getting anxious feelings from realizing you're also not telling me I'm completely off-base.

I really wanted to love Tilt5, both because of your background and because it's not an ad-riddled megacorp backing it. And I'm an XR maximalist--VR and AR have been the focus of my career for the last 7, 8 years. But at this point, for most of the things I could do with T5, I'd have a better experience mounting a projector on my ceiling.

I just... I don't know. I don't understand why you ever released this product. It wreaks of "yet another naive startup got ripped off by a some factory in China". I saw the posts about the difficulties you all had with getting the manufacturing right. And this is a common thread from everyone I've ever talked to who off-shores manufacturing. It's just frustrating that nobody in the startup consumer electronics space seems to be paying attention to any of the other consumer electronics startups. Was it a case of hubris? "Yes, but we are special and this won't happen to us?"

On the controllers, I'd almost rather have a nose pointer and a small, hand-held clicker button. I actually have several already: they are basically Bluetooth keyboards that have a single button that sends the Enter or Space key.

I wish I had many more options for pads. They don't seem like the expensive part of the BOM. IDK, maybe I'm underestimating the cost of bonding retro reflective paint to cardboard. But it seems silly to have to choose between pad sizes while purchasing. Just give both the small and large size pads.

I got the large one and it's really too large for single-person use. I also don't have a table that fits it. My coffee table is long enough, but not wide enough. My dinner table is wide enough, but way too long. Not a lot of square tables out there.

Better yet, be able to support multiple pads at the same time and create a system for mounting pads on the wall. I could see creating a desk setup where I have the desk top and the wall behind the desk as retro reflector pads. It'd make my whole desk into a stereo tank. Get rid of my monitors completely and not be strained trying to where a tyoical VR headset all day.

I'd probably get more use out of a ridiculously large pad that I could setup against the wall in my living room than the current, awkwardly-sized pad that I basically have to use on my floor.

And I'd immediately stop putting any effort into plugging multiple devices into a single PC. It's a hack that fixes things only on paper. Where am I going to put this PC so that all the players can sit around the board and not sweep their drinks off the table or yank the cable out? Who even has a PC powerful enough to run that many independent views (2 per person!) but XR enthusiasts like me who already have other, much more capable devices?

First, let me set your mind at ease by telling you that you're completely, 100% off-base. Just in general. Very off-base. :)

> I wish I had many more options for pads.

Yup, we've heard this a lot, and it's definitely something we're looking to address. It is possible to lay down additional retroreflective material to get a larger viewable area. Folks in our discord can likely give you some tips if you're interested.

> I don't understand why you ever released this product. It wreaks of "yet another naive startup got ripped off by a some factory in China".

I didn't mean to imply at all that we got ripped off by a factory. Getting manufacturing right is on us. We had an issue that we didn't catch, and some people got shipped glasses that defocused. We've done our best to do right by them and repair or replace those glasses.

We released the product because we believe that despite some imperfections, doing AR like we do it is way more compelling than any other way possible today for certain use cases. We want to enable people to do cool things and enjoy this product, while we learn from how they use it so we can improve on it. I think we've improved it a lot already, and will continue to do so going forward (that was the point I was going for in my first reply).

Currently, doing XR hardware is all about trade-offs. Designing just the display system involves significant trade-offs between price, weight, brightness, field of view, vergence-accommodation conflict (eye strain / comfort), image resolution, eye box size (accommodating different face shapes & look-directions), depth of field (what range of distances are virtual objects in-focus across), latency, and power consumption just to name a few. Our trade-offs are pretty unique in the industry, and we've struggled a bit with how best to communicate about them to different audiences.

Basically, we give up the ability to draw virtual objects anywhere, relying on the retroreflector, and in exchange we get to make a pair of glasses that are light weight, bright, in focus for up-close objects, have a wide field of view up-close, don't cause eye-strain, have a large eye box, are translucent with minimal dimming of the real-world, all for a non-astronomical price. For clarity, the "up-close" qualifier for the "in focus" is a feature - most headsets are focused at infinite distance and close objects will be improperly focused. The "up-close" qualifier for the "field of view" is a limitation of the gameboard, assuming additional retroreflector isn't added.

On top of that, afaik, we're the only company looking to enable local multiplayer experiences for AR. There's something pretty special about interacting face-to-face with other people that are seeing the same things-that-aren't-there as you. It helps with the suspension of disbelieve in sort of a unique way. I know you said that we should give up on this so we don't catch everyone's computers on fire, but 4 pairs of our glasses have fewer pixels than a 4K display, and because we do the 180 Hz image stabilization on the headset, the PC only needs to hit 60 fps for smooth animations. We tend not to be entirely pixel-limited in the rendering, but as we optimize more that's the limit we'll approach.

Anyhoo, hopefully that wasn't too infomercial-ish, but that's some semblance of what excited me about this when I first met Jeri all those years ago, and it's what still excites me (and I think everyone at Tilt Five) about it today. And that, ultimately, is why we made the product.

Maybe that's the sort of answer you were looking for? Feel free to tell me I'm completely off-base ;)

I remember trying it at PAX and was disappointed the effect didn't feel that strong to me either. I almost felt like something was broken or not calibrated since I could sense some 3D-ness, but it didn't jump out as truly 3D.

I agree that the controls (a wand?) didn't feel so natural.

I saw videos and followed the project a bit before finally trying it in person, so I felt let down a little.

I believe we tuned the interpupillary distance (IPD) down a bit for the demos at PAX. The IPD is a software setting that changes how far apart the two stereo images are drawn from. If it's set to a distance larger than the actual distance between your eyes then you'll perceive it as very strange (each eye seeing a different thing that your brain can't unify), and it will be very uncomfortable. If the IPD setting is set to a distance less than the actual distance between your eyes then you won't perceive as much depth when looking at it, but it's still fairly comfortable. The extreme version of this is just looking at normal monoscopic images (e.g. watching a movie) - perfectly comfortable, but no depth perceived from it.

Rather than adjusting it for each person we gave the demo to, we set the IPD a bit low so that it'd be comfortable for everyone, but people with a larger physical IPD wouldn't perceive as much depth.

Sorry you felt let down by it.

Thank you for the explanation. If I get another chance, I'll give it another try with that in mind.

Really interesting post, thank you. I looked at the linked page, and had a really hard time figuring out what this, well, is, or how it’s supposed to work. It sounds to me like some of my naive initial reservations are impediments to actual use.

Also, I was shocked at the pricing. This is 2-3x what I would have expected. Not saying it isn’t worth it, but even as someone very much inclined to waste money on shiny electronics, the value proposition here seems extremely questionable.

I had the same disappointing experience unfortunately. The best experience I had was actually the Fantasy Forest tech demo, which was hard to find. https://www.tiltfive.com/lab/fantasy-forest-tech-demo

Reply on news.ycombinator.com