Hacker News new | more | comments | ask | show | jobs | submit login
Ford's designers are learning to create 3D cars in VR (cnet.com)
67 points by mstats 20 days ago | hide | past | web | favorite | 57 comments

They obviously showed a lot of meticulously mated actual 3D components but those were all prepared in a CAD program.

Sure, the VR sketch might be useful, but the visual part of the video is quite a lot of marketing spin when it’s not showing the abstract lines.

And I’m not sure that VR is the best way to “feel” a car in this way either. A car is an extension of your body, so it’s just as important where/what your shoulders, knees and legs are/feel as to what the eyes can see.

I can't imagine it being the most efficient development tool, but I think for overall "feel", VR could have a legitimate use case.

For example, you're designing the car on the computer. You put on the VR headset, and suddenly your computer chair becomes the driver's seat. You can look around at your surroundings. Do the dash controls and buttons feel like they're in a good location? Does reaching for the handbrake feel natural or is it too far forward or back? If you look in the rear view mirror do you have a good view, or is the back window causing it to be limited?

Then you can quickly jump out of VR, make a few adjustments, and jump back in. Or get in VR and quickly A/B test different multiple saved versions of the interior.

It would also be good for general feedback. Get some people off the street and toss them in VR to walk around your car. I could see it being more effective and accurate than looking at photos, or 3D renders.

>And I’m not sure that VR is the best way to “feel” a car in this way either. A car is an extension of your body, so it’s just as important where/what your shoulders, knees and legs are/feel as to what the eyes can see.

Compared to looking at a flat computer screen, I would think this would certainly be a move in the right direction.

It would be in comparison to a full scale clay model. Not a computer screen.

It's unclear from the article what Gravity Sketch offers that absolutely requires VR to get the several orders of magnitude improvement that is claimed:

> "If you wanted to make an interior and an exterior, you're still talking months," versus approximately 20 hours in the VR tool, he says.

If Gravity Sketch were on a traditional 2D screen instead of in VR, would the designers still accomplish in hours what used to take months? Conversely, if the VR visualization were to be added to the tools and processes that currently take months to prototype a model, would a similar boost in productivity be observed?

Basically any usage of VR outside of gaming right now is guaranteed to just be marketing nonsense. The hardware simply isn’t there for real world professional usage as a productivity enhancing device. Some day we’ll be there, but the current state of even the most advanced HMDs is little beyond toy prototypes at this point.

Lots of people are doing some really cool art using VR headsets. Those two Facebook groups are really nice to join if you’re interested.

Animations: https://www.facebook.com/groups/virtual.animation/

Paintings: https://www.facebook.com/groups/virtual.paintings/

I should have included art as well. There's definitely a lot of cool visual stuff being done. But the ergonomics just aren't there yet for the use cases constantly being touted in the media like virtual desktops and CAD modeling.

You could not be more wrong. A lot of new employee safety training is moving to VR, backup NFL quarterbacks are practicing in VR (since the starting quarterbacks get all the actual time in practice), this is already happening. Not vaporware.

> Basically any usage of VR outside of gaming right now is guaranteed to just be marketing nonsense.

I've seen some noise about VR used to locate pipes and wires when digging. That seems pretty legit.

I've also heard of VR used in conjunction with building construction, although it's less obvious to me just how useful it would be.

Both of those sound like applications for AR.

It seems likely we'll see widespread AR deployment first.

Seems like AR would be a bit better there.

The hardware is fine. (What's more arguably lacking is software)

Yeah resolution could be better but it's far from being a deal breaker for viz, modelling etc.

Curious to hear what you thought the blockers were?

>"Curious to hear what you thought the blockers were?"

1. Resolution is hopelessly too low for working with text

2. Ergonomics too awkward to wear for 8 hours at a time.

3. Total lack of multi-plane focus, no ability to sense depth beyond a few meters

These are strictly true for every HMD available, having owned and worked with them all extensively since DK2. I wish it wasn't true either, but it's just where we're at. One big thing that has been solved is tracking though. It's easy to forget what a quantum leap Lighthouse was in consumer grade tracking technology. Having sub-millimeter precise tracking and pose estimation was a huge part of getting VR to an MVP. That gives me hope the other problems can be solved, but there's a long way to go with the need for things like integrated eye tracking, foveated rendering, and full vision FOV being table stakes for a real consumer product.

All these things are solvable, and there are consumer hmds coming that will solve some of them. Resolution upgrades are happening, ergonomics will improve. One problem with getting these features fast tracked in an HMD is the industry focus on gaming and media and keeping costs low for consumers. A high price enterprise focused HMD with all of the features you mentioned might work with the right software and clients.

No, it's definitely the hardware. Hardware makers like to tout that VR is waiting for its "killer app." What's really missing still is the "killer hardware." They're all still devkits (speaking as someone who owns one of each of the top headsets), it's just that when FB bought Oculus, suddenly there was a mad rush to plop the "consumer-ready" label on them. But we're not there yet, which can be evidenced by the re-orgs at Oculus.

You can't make Angry Birds or w/e on the VR equivalent of a Blackberry. No one's made the iPhone yet.

Our large medical device customers for VR surgical training would beg to differ.

Next they can put the design on the blockchain so you can verify the car you buy is the car they designed.

Virtual reality and pain management: current trends and future directions - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3138477/

> They obviously showed a lot of meticulously mated actual 3D components but those were all prepared in a CAD program.

Do you have any reference for that?

If so, they wouldn't need gravity sketch at all. The most popular CAD programs already have VR support:

Rhino -> https://www.mindeskvr.com/

Maya -> https://www.marui-plugin.com

3dsMax -> 3dsMax Interactive (Autodesk)

Blender -> https://github.com/MARUI-PlugIn/BlenderXR

Those are 3D programs, but not CAD. For CAD you'd expect Solidworks, Catia and PRO-E.

Modeling programs make art, CAD is for engineering. They’re very different.

Yes, but I want to point out on odd edge-cases that's not always true.

For example, when designing something for 3dprinting, I sometimes prefer modeling programs over CAD programs because I have control over, for example, however a sphere is made up (the style of polygons, tesselation, that sort of thing). When you're slicing for a 3d printer, you can get odd bugs if your stl is made in a non-optimal way.

I've used most of those tools, they are all pretty terrible to use. You can't very easily take a rendering engine designed for massive geometric modelling and just plug VR into it without major performance sacrifices. I always got a headache after using them.

> I’m not sure that VR is the best way to “feel” a car in this way either.

I'm pretty sure it isn't. Just like about everything in human experience, almost nothing can be replicated just as well in VR. VR can give you a rough sense of space, but your body has no sensation, no touch, no feedback, it's so far from reality that you can't really just rely on VR to design things.

Plus, there are things that you don't notice in short time but need to "sink in" with everyday and long-term usage. This is also very hard to replicate in VR.

You're comparing VR to being in the car. The correct comparison is VR to a 2D screen, which is what this is replacing/augmenting.

It doesn't need to be the Holodeck to be useful.

I’m working on something similar (not announced yet).

Tools like Gravity Sketch, and the tool I’m making, are great for conceptualization and communication.

In seconds to minutes you can take an idea from your head and show it to someone else. You can have an idea for a product ready to show your peers and in moments they can offer feedback and make changes.

Tools like this aren’t replacing the final engineering yet, but they’re making the planning and communication stage dramatically more efficient.

we're using VR for some progressive design stuff, do you have a contact you can provide. It might be that we're a good opportunity to test out something as it nears announcement.

Sounds interesting. And I'd love to have more eyes on the product closer to release!

Reach out to me at my personal email: ian.kettlewell@gmail.com

That goes for anyone else reading this as well. If anyone is working in this space or wants to chat about VR design tools I am interested in talking to you.

I design mechanical stuff for a living and I can't imagine that designing cars in VR is more efficient than CAD. There isn't any barrier right now from idea to CAD. If an engineer (or a scientist) can imagine it, an engineer will be able to design and create it digitally with current technology. I would say the current limitation is the manufacturing tools rather than the designing tools.

Side note, I think the work process for designing the exterior of cars is to create a sculptures with clay, transfer to clay to CAD, and rinse and repeat. I wonder if VR will ever beat manual sculpting.

Also design mechanical stuff for a living. I’ve spent multiple days over the course of the past six months creating drawings to demonstrate how a design will ‘feel’ once made. A 1:1 3d model that other stakeholders could walk into would be so much more effective.

As a hobbyist - the CAD tools I've used (Fusion360 and Inventor) - are clunky with arbitrary 3D shapes (e.g. a Porsche door panel). How do people in the "real world" do it? Iterate in Maya / similar and then export? Think in parametric curves? Something else entirely?

An industrial designer would sketch the outline then the images would be brought into parametric CAD software as references. The 2D image can be built into 2D/3D lines, then these curves are used to create surfaces which finally are stitched together. The nature of complex surfaces interacting and joining seamlessly is where the difficulty lies. "A class" surfacing is an entire speciality and can take a career to master. https://en.wikipedia.org/wiki/Class_A_surface

Yeah but presentation is pretty different from designing. I know where you are coming from, I always tell people pretty pictures comes before practicality when it comes to meeting with management.

Why do you assume CAD and VR are mutually exclusive? Interfaces of the future haven't been designed yet - but it's not hard to naively imagine taking any CAD program and turning the viewport into a 3D window. Something that allows you to use your head to look around it, use your fingers to spin the object around, pinch to zoom or other gestures. The rest of the interface could stay the same as it is now and interacted with via a mouse and keyboard, and it would be much better than the current CAD (if there was a VR HMD that has a very high resolution and allows some level of mixed reality for interaction with with desk objects as well as your colleagues)

I think 3D mouse does pretty much what you ask for though[1]. I'm sure the possibilities are beyond my imagination with VR but I personally just can't see the benefits. Most engineers are pretty good with their abstract of space so I can't imagine VR being a necessity for design.

[1] https://www.3dconnexion.com/

I've briefly used one of those. It's a cool little device once you get used to it, but there's definitely a learning curve. By comparison, it's hard to understate how intuitive VR headsets are for controlling your camera. That might not matter for professional drafters, but it would certainly make 3D modelling more accessible to the public in general. I can imagine a future in which 3D models are created and used much like images are used today.

I'm having trouble with the title.

Wouldn't the title be equally as correct written as "create VR cars in 3D", or even just "create cars in VR".

And that's before mentioning that 'create' is ambiguous. Build or design???

Could someone convince me on the use case? I get that designing a car in 20 hours is 'better' than in 20 weeks. But with all the 1000s of hours put into other considerations, is this going to make any appreciable difference to anything?

My takeaway was that the designs might be different not just the same but done more quickly in VR. They stressed the point that in VR the designer can place themselves in the car and design at a "human scale." "We can get ourselves into the mind and the body, the virtual body, of our customer."

As far as the title, I think just dropping the "3D" might make it better?

This is beautiful because I think this is just the beginning. Who knows if we'll still be designing cars in VR in the long run. But it is really nice that people are beginning to experiment and tools to help them do so are emerging.

As Paul Graham says, all good ideas have an "ugly duckling" phase.

Yes, I was thinking that many people thought that designing cars with a computer was silly in the early days of 3-D computer aided design. It seems the jump from creating tactile prototypes to designing on a computer is a much larger jump than moving from a flat computer screen to a VR HMD. I'm surprised there is such a negative response so far on HN.

Were they making 2D cars before?

My first thought exactly. Perhaps ‘3D models of cars in VR’ would’ve been a more solid title.

Maybe they're working up to 4D cars. That would get my attention.

So what? The industry is far too conservative to bring any sort of radical design to mass market. This may as well be "Ford designers are learning to use A1 paper instead of letter sized"

>> So what? The industry is far too conservative to bring any sort of radical design to mass market.

I don’t think we need a radically new design for cars? ‘The Homer’ comes to mind. >.>

Cool. I'm more of a mechanical designer so it is hard for me to understand why this is an improvement because I'm used to requiring a lot of precision in my modeling. That said when he mentioned the ability to get line-of-sight on different curves I could appreciate it a bit more.

I still don't think that the hand controllers they're using are the best input device. I can imagine a hybrid approach where design teams are using standard surface modeling like Alias/3Ds Max/Blender for the bulk of the design work and then VR systems like this for review.

I'm surprised there isn't actual CAD software for VR yet.

When I tried tiltbrush for the first time I was immediately struck by how easy it was to create some things, but how much of a pain it was to modify things, or make things like parallel lines or surfaces.

Maybe everyone is waiting for someone else to do it. Even the linked software (gravity sketch) seems to be missing constraints, dimensions, and snaps, though it has control points so modification of existing geometry is at least possible.

What gives?

Why bother just yet, when the most important part/killer feature will be subpar (for now) - reviewing the models in high res 3D.

When I was at Edmunds.com, I worked on a few "greenfield" projects, including VR. I'm generally bearish on VR, but I think if there's a space where VR can truly usher in a paradigm shift, it's auto -- both for design/development, as well as showcasing (think potential buyers, auto shows, etc.).

"for design/development"

Could you expand?

funny, I was thinking that CastAR (now Tilt Five) would be a good compromise for that. Having your CAD display in front of you of you, the reflecting mat on the side and no too intrusive glasses. So that you just have to turn your head to see the 3D. You don't really need virtual reality when you're just working on an object, you just need a bit of 3D to help you grasp the shapes.

Maybe they could use all that expertise to design a smaller truck.

What is a “3D” car? Seems needlessly redundant.

Too bad their engineers are not learning how to make cars that don't need several recalls.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact