VR essential to design of Nvidia’s new headquarters | Hacker News Reader

VR essential to design of Nvidia’s new headquarters

archpaper.com

53 pointsptrptr21 comments

A friend of mine's husband built a new location for his business and the architecture firm used VR to helped him see what his place would be like. It seems only logical for much larger initiatives like this one to ensure that a lot of plans get revised earlier on in the planning process, such as figuring out the optimal lighting (as mentioned in the article). I'm sure most of us live in cities where we know of a story of an expensive project that had a mishaps because of a lack of foresight.

Fascinating. What, exactly, did they use VR for? Were they actually doing design work with it, or just doing demos? What kind of hardware and software?

> I'm sure most of us live in cities where we know of a story of an expensive project that had a mishaps because of a lack of foresight.

Like most of Gerhy's designs?

I just talked to a friend of mine who is an architect. He said most buildings look only good as model and when they are build they are only ugly blocks.

Even when rendered most architects play around with perspective to get their "ugly blocks" to look nice.

I said they should force the architects to design with VR and let them see the finished building from above only after they got the important stuff done, so the focus of the design is the people walking around and not the top down view on the building.

GPUs are capable of highly-realistic rendering now, so this is finally a viable proposition. And the VR hype will definitely go a long way to making this a reality.

A lot of the same tricks with lighting and point of view work just as well in the real world as they do in the renderings.

It might be an ugly block 99% of the time, but it'll look great if you get a photo from the right angle at the right time of day with the right lights turned on in the right weather.

The first thing I thought of when I saw the photo of the interior of the building: panopticon.

https://en.wikipedia.org/wiki/Panopticon

Have some unrelated Google Maps trivia I just discovered.

1. Scroll down to the map and the end, and zoom all the way in. For me the tileset goes back in time to before the building was finished. I initially thought "wow, Google merged an overhead render from the architect!", but when I followed the "View larger map" link zooming in showed the completed building. TIL Maps uses different tilesets for different APIs!

2. See your avatar at the top-right of the embedded map view? Hover that, and you'll see "+YourUsername", just like in the old +Google days. Wow :P

I'm guessing this site is using a legacy embed API or something. These were fun glitches to stumble on though.

Looks like zooming in switches to 3/4 perspective view. The top-down and 3/4 perspective images were taken at different times. When you click through to Google Maps, it will use the WebGL true 3D version which is, understandably, a completely different tileset. Unless your computer doesn't support WebGL, in which case you'll see the same 3/4 perspective imagery as before. The embed never uses WebGL, perhaps because it would be rude for a small embedded map to use a ton of your GPU memory.

If historical imagery interests you, Google has several options available:

The desktop Google Earth client will show you historical top-down aerial imagery that you can scrub through with a time slider: https://support.google.com/earth/answer/148094?hl=en

You can access historical street view imagery through the main Google Maps web interface: https://googleblog.blogspot.com/2014/04/go-back-in-time-with...

For large-scale or global changes, there's https://earthengine.google.com/timelapse/ which is essentially a zoomable video of the entire Earth over the past 30 years at ~30m resolution.

For the true 3D imagery and 3/4 perspective view, I believe there's not currently a way to view historical versions (perhaps because stitching and reconstruction continues to improve and older versions would be low quality).

> Looks like zooming in switches to 3/4 perspective view. (...) When you click through to Google Maps, it will use the WebGL true 3D version which is, understandably, a completely different tileset. Unless your computer doesn't support WebGL, in which case you'll see the same 3/4 perspective imagery as before.

Problem: my old laptop doesn't support WebGL, and I'm seeing the new building at all zoom levels.

> The embed never uses WebGL, perhaps because it would be rude for a small embedded map to use a ton of your GPU memory.

That makes absolutely perfect sense - another thing is that the embed needs to "just work", and WebGL is in effect kicking in OpenGL; if the display stack isn't in sane condition (eg, misconfigured or buggy driver) it may cause lockups or other issues.

Thanks very much for the other links!

- Google Earth is on my todo list of things to play with when I'm back on my i3

- The historical street view thing is awesome and really useful, I agree

- TIL about Timelapse, that is awesome. Thanks very much!

Your theory about stitching/reconstruction of old tilesets sounds quite plausible.

Why are headquaters allways non-modular? Usually such structures get to small the day, the company moves in. So why not integrate growth-ability into a building?

Where is the parking for this new building?

If I recall, from the 2016 GTC keynote, the parking is in the basement. The CEO felt he was voting in favor of the environment by putting the cars underneath.

So where are all the computer generated images?

I'm genuinely not sure what you mean. The hero image and the 3rd placement image are both CGI renders of the final building, FWIW.

You are right. But those types of image are also typically made using conventional techniques (although more labour intensive). VR gives you the opportunity to show the building from various angles as people would normally see it, and also from the inside. Hence I was expecting much more images, and not the typical images you see on a billboard next to a construction site.

Very fair point.

I guess the article was trying to drum up hype about the possibility of being able to do what you describe.

IOW, this news article is saying "this should exist soon" in a "come on, get with the program" sense, because the industry is not quite there yet.

Having said that, I can perfectly understand your point of view.

I know this is pedantic, but Nvidia isn't really a "microchip maker" as the article says so much as designer since they're fab-less.

Message I've received: We thought you might like to know that we put https://news.ycombinator.com/item?id=14125108 in the second-chance pool, so it will get a random placement on the front page sometime in the next 24 hours.

This is part of an experiment in giving good HN submissions multiple chances at the front page. If you're curious, you can read about it at https://news.ycombinator.com/item?id=11662380 and other links there. And if you don't want these emails, sorry! Tell us and we won't do it again.

Thanks for posting good stories to Hacker News, Daniel (moderator)

I like this, but isn't it essentially the same as manual curation?

Not that I see anything wrong with that. I've lurked /new for a long time and noticed more and more spammy articles increasing the noise factor. I really think the whole thing is an arms race for views that can't really be stopped without a real person in the loop.

Perhaps you could start assigning reputation to a domain that's linked to how many good articles have come from there? This would encourage "friendly" SEO spam of good articles at least, lol.

Thanks so much for the first link, this is really interesting. A couple days ago I was even wondering what would happen if there was a site that aggregated data from here but had independent voting. That wouldn't be ideal though for many obvious reasons. I'm really happy to learn about this initiative!

Reply on news.ycombinator.com