News Meet the Team #6- Ben Parry (Graphics Programmer)
- Thread starter Ashley Barley
- Start date
- Status
Schmergonomics
I'd like to know if you are happy with the height of your monitor.
I could never quite get mine high enough in the last office I worked in.
Have you got your eye on another fatter book?
I had to settle for the fattest book that nobody in the office actually wanted to read - so I ended up with a very practical (elevation-wise) MS Access tome.
(This was after an office move before which I had used a ream of laser paper for many months, with a taller stand which got "lost" in the move.)
I'd like to know if you are happy with the height of your monitor.
I could never quite get mine high enough in the last office I worked in.
Have you got your eye on another fatter book?
I had to settle for the fattest book that nobody in the office actually wanted to read - so I ended up with a very practical (elevation-wise) MS Access tome.
(This was after an office move before which I had used a ream of laser paper for many months, with a taller stand which got "lost" in the move.)
Okay, for a bit I was confusing John Kelly with John Laws, so sorry John (you can decide between yourselves who's taking more offence, and apportion the apology apropropriately). So 3D-artist questions... <scratches head until blood appears>
How aware are you of keeping everything to scale, especially regarding future updates? To what extent are you modelling interiors currently (e.g. blocking out/lovingly placing every nut & bolt/leaving it to chance)? What about tall humans from low-gravity planets, and Bernie Ecclestone sized humans from high gravity planets?
Are you working on anything other than space-ships?
Are you working on space-ships?
What's the most obscure/random thing you've modelled a) for Elite: Dangerous and b) for any game.
Can you draw/paint/sculpt/tap-dance?
Are you looking forward to potentially seeing your handiwork through the Oculus Rift?
Do curvy Imperial ships have more polygons than boxy Federation/Alliance ships? Does that make them better?
How aware are you of keeping everything to scale, especially regarding future updates? To what extent are you modelling interiors currently (e.g. blocking out/lovingly placing every nut & bolt/leaving it to chance)? What about tall humans from low-gravity planets, and Bernie Ecclestone sized humans from high gravity planets?
Are you working on anything other than space-ships?
Are you working on space-ships?
What's the most obscure/random thing you've modelled a) for Elite: Dangerous and b) for any game.
Can you draw/paint/sculpt/tap-dance?
Are you looking forward to potentially seeing your handiwork through the Oculus Rift?
Do curvy Imperial ships have more polygons than boxy Federation/Alliance ships? Does that make them better?
The answer to that starts getting complicated so I'll wait till someone does a feature on them
Slawkenbergius/Maxeren:
We're currently storing the planets in spherical coordinates, so the poles are actually get an unfair share of the detail. I'm worried that the trig operations needed to decode them are going to be excessively expensive to run per-pixel though, especially if something needs to sample a few different locations, so we've been discussing reprojecting them to something that only looks good on one side, but doing it on the fly so that once you move a couple of degrees round it redraws it to keep the ugly side away from you.
Dave Leestir:
I like my monitor at this height. What you can't see is the biro poked in at the back, if it falls out I suddenly get loads of glare from the lights above me (or possibly the sun). At home I have a stack of four programming books under the monitor, and another three on top of my PC case holding up one end of my desk. One of these days I'm going to need to look something up and it's going to get messy.
Last edited:
Interesting approach.. if i understand correctly its kind of like culling, where you with trig find which areas of the sphere needs to get culled, right?
flavien brebion, the guy behind the I-novea engine had alot of issues regarding rendering of planets for "infinity: the quest for earth" and i think he took a somewhat similar approach if i remember correctly.
If you haven't sen their tech demo then here it is.. kind of cool
http://www.youtube.com/watch?v=fO7XhaTGDYg
Last edited:
The answer to that starts getting complicated so I'll wait till someone does a feature on them
Good! A simple yes would have made me very disappointed.
I have an idea of what I think/hope you are doing in regards to this, but obviously I couldn't fool you into saying what that might be.
It's not so much about culling as it's about reducing the number of operations needed to get from 3D space to 2D space. Conventional texturing can just store a 2D location at each vertex and stretch the image in between, but spherical coordinates would end up with a lot of error if you tried to do it per-vertex and instead have to do the operations per pixel despite sin/cos etc not being operations to use lightly.Interesting approach.. if i understand correctly its kind of like culling, where you with trig find which areas of the sphere needs to get culled, right?
flavien brebion, the guy behind the I-novea engine had alot of issues regarding rendering of planets for "infinity: the quest for earth" and i think he took a somewhat similar approach if i remember correctly.
If you haven't sen their tech demo then here it is.. kind of cool
http://www.youtube.com/watch?v=fO7XhaTGDYg
Luckily, "formula for mapping a nearly-spherical object to 2D" is something humanity has put a lot of work into over the years.
Obviously you're right though, there's probably about 40% of a planet that we don't need to generate until you're looking at it.
Thanks for the replies Ben. They're very interesting, even if the technicalities are a bit over my head in places.
The mention of 'nearly spherical' brings up another question I'd been wondering about. Images of Saturn show that it deviates quite a bit from a perfect sphere. Presumably there are even more extreme examples out there in the (real) galaxy, with superfast-spinning gas giants displaying very noticeable equatorial bulges.
Is this something that the team will be modelling?
That can be done mathematically by changing the spherical positions of the vertices in the sphere..
That can be done mathematically by changing the spherical positions of the vertices in the sphere..
Cool. I'm pretty sure I know how that would work/look with naked geometry in Blender, but I've never got as far as texturing to understand the issues there.
While I'm on a roll with asking questions, I might as well ask a really awkward one. Will there be a 3D element to clouds so that they 'float' above the planet? As seen here, or here. I realise that might be expecting too much, which is why I'm not expecting it...
It think it quite likely that clouds will be 3D to some degree. Gas giants have been mentioned before and once we get planetary landings it will need to be handled in some realistic way.
I think I've seen code to do that hanging around. Certainly the hardest part of doing that is knowing how much a planet ought to be squashed, so that's more a question for procedural-galaxy type people.
I think the biggest problem we have with realism is that even if we get things perfect people will think we didn't - Saturn clearly has some texturing bugs on it:
(from Cassini)
Well DB did show clouds in one of the DEV videos, but yes volumetric clouds would be a given if the planet has to be somewhat realistic
That thing is so cool.. wonder how and why nature is doing this?
My desktop background
So beautiful..
I think the biggest problem we have with realism is that even if we get things perfect people will think we didn't - Saturn clearly has some texturing bugs on it:
(from Cassini)
Wait a minute! What if Elite has already been released and we are all playing it with a Oculus Rift Mk15!?
*starts checking for a plug at the back of the head*
This sort of thing came up with random play on MP3 players. Truly random does not always seem like it is. The random functions had to be enhanced so that they seemed more so when in fact they were less so. (Fantastic picture, by the way)I think I've seen code to do that hanging around. Certainly the hardest part of doing that is knowing how much a planet ought to be squashed, so that's more a question for procedural-galaxy type people.
I think the biggest problem we have with realism is that even if we get things perfect people will think we didn't - Saturn clearly has some texturing bugs on it:
(from Cassini)
We've currently got 2D clouds on a floating shell around the planet, but from most distances you might not even be able to tell they're not just on the surface (on an Earth-like planet). Obviously when you get very close, landing-close, you're going to need a lot more detail but that's the kind of thing we're going to need to intelligently swap in as you approach.
- Status