Humor Random Comments Thread


Oh, no. With comments last week in the Random Comments Thread, could unleash a third world war.

GIF_101977_bombardeo_masivo.gif

Image of a educated discussion about quantum mechanics in the Japanese anime produced in Random Comments Thread.

All Right. Anyway...

We sell weapons of war in MS.
Are you a world leader of [insert country name here] and want to make a third world war, just because you (and your other friends world leaders are bored)? Do not think twice (good overall if you want to declare a war because you are bored should not use the brain at all). Fool population with false nationalism and spend all public coffers [insert country here] buying weapons of war in MS.
 
"Even old New York was once New Amsterdam", hmm seems like I've heard that somewhere before.

"Why they changed it I can't say"

he he he
 
On a random note.
Want a cheap and exciting roller-coaster ride with no lines to wait?
Get on a bus in Reñaca, Chile.
vRovCOO.jpg
 
And the skill of the artist.

I can recommend reading OrbiterConfig.pdf page 33 "Performance optimisation".

Yeah, I have found that even moderate sized meshes made in Anim8or can become quite large in size even with moderate polycounts, depending on the workflow you use to create it. My biggest one so far was at least 5ish mb in size last I recall.
 
And the skill of the artist.

I can recommend reading OrbiterConfig.pdf page 33 "Performance optimisation".

Actually mesh filesize is directly proportional to polycount (and split edges count). Polycount depends on the skill of artist, however there is no simple rule here. There are few different aproaches for models to look good in final product.

You can make 100k plus poly model with lots of meshgroups devoted solely to make wires, bolts and other stuff on the inside of gear flap or you can simply make 2 poly rectangle and put nice texture and normal map on it.

In the Orbiter we have luxury of having only few objects in the scene at the single time so having highly detailed mesh (50k - 100k triangles) in not an issue (unless you put 100 of them side by side).

My last orbiter mesh was somewhere in between. It was fairly complex geometry but if I'd go all the way with meshing cooling tubes, my RL-10 would be probably 25k polys instead of ~6k.

rl10-0001B.jpg

Rl-10-renderC.jpg
 
Last edited:
Actually mesh filesize is directly proportional to polycount (and split edges count). Polycount depends on the skill of artist, however there is no simple rule here. There are few different aproaches for models to look good in final product.

You can make 100k plus poly model with lots of meshgroups devotet solely to make wires, bolts and other stuff on the inside of gear flap or you can simply make 2 poly rectangle and put nice texture and normal map on it.

In the Orbiter we have luxury of having only few objects in the scene at the single time so having highly detailed mesh (50k - 100k triangles) in not an issue (unless you put 100 of them side by side).

I should have added "care of the artist" to that. Polycount is a always a compromise. I don't like to see meshes with 10's of groups using the same material (unless they are animated) or a lot of simple textures that could be made into one (or a few) larger without loosing the visual effect.
 
I should have added "care of the artist" to that. Polycount is a always a compromise. I don't like to see meshes with 10's of groups using the same material (unless they are animated) or a lot of simple textures that could be made into one (or a few) larger without loosing the visual effect.

I agree. Mesh/texture optimization is vital part of 3d work. Excellent example was XR5 Mesh optimization . It halved polycount without quality loss (even added minor improvements).

I can understand multiple textures issue. UVW unwrapping isn't fast and sometimes it's easier to use single unwrap (planar or cylindrical) on the element and call it done.

As for objects sharing same material - I think there is no excuse to do it unless they're animated.

Took me few hours to unwrap this engine and I'm still thinking I could do it better:
rl10-UVW.jpg


---------- Post added at 12:38 AM ---------- Previous post was at 12:31 AM ----------

And now someone please tell me how do I make decent looking, detail rich, 1km in size sci-fi space station with less than 10k polys and no more than 3 2048x2048 textures.
 
Last edited:

And now someone please tell me how do I make decent looking, detail rich, 1km in size sci-fi space station with less than 10k polys and no more than 3 2048x2048 textures.

Is that limit in total or in one scene? If the latter, you could sure employ a lot of LOD calculations to reduce the polycount to the really necessary, though maybe a bit ugly because of the size.
 
And now someone please tell me how do I make decent looking, detail rich, 1km in size sci-fi space station with less than 10k polys and no more than 3 2048x2048 textures.

First, use mip mapping. Not only for performance, it also looks better (hi-res textures without mip-maps tend to look really uggly from a distance.

Second, build your 10k poly mesh, store it, and continue building up to a few million polies. Then bake that down into a texture and normal map (and lighting map if you have support for that). I utterly hate that process, which is why I never really got further into modelling although I noticed that I can produce decent results. It's just waaaay too tedious work for me :shifty:
 
Last edited:
Is that limit in total or in one scene? If the latter, you could sure employ a lot of LOD calculations to reduce the polycount to the really necessary, though maybe a bit ugly because of the size.

First, use mip mapping. Not only for performance, it also looks better (hi-res textures without mip-maps tend to look really uggly from a distance.

Second, build your 10k poly mesh, store it, and continue building up to a few million polies. Then bake that down into a texture and normal map (and lighting map if you have support for that). I utterly hate that process, which is why I never really got further into modelling although I noticed that I can produce decent results. It's just waaaay too tedious work for me :shifty:

Lol. I was asking from design/concept POV rather than execution. I'm already familiar with those. :P

@Urwumpe: that's no external limit but I like to keep my meshes low poly to allow more of them in single scene (more massive space batlles \o/)
 

Eh. If it was a real bot answering on Tinder instead of some intern in PR department, that would be a real achievement...

It could revolutionize online dating. Why waste your time messaging members of the opposite sex when you can set up a bot to do this, and the bot will e-mail you after the date is set? Of course, it gets a bit creepy once you realize that the opposite party could set up a bot also, and you have two bots talking to each other. Then the next logical step would be to send a bot on a date instead of yourself...
 
Last edited:
@Urwumpe: that's no external limit but I like to keep my meshes low poly to allow more of them in single scene (more massive space batlles \o/)

Well, for space battles, you can still use LOD calculations for selecting the level of detail that is visible in the distance to reduce the total poly count of the scene.

If you allow a total of 250k poly for example, you can still use 100k for the station, if the camera is near it, and just the other 150k for the remaining vessels, which are kilometres in the distance.

---------- Post added at 10:43 AM ---------- Previous post was at 10:09 AM ----------

Doppler Radar unchained...

The new Swiss air surveillance radar (build by Airbus Defence and Space, not by somebody competent about the physics in Switzerland) is mistaking cows for attacking jets. Simply because nobody ever considered that happy Swiss cows could actually be at elevations above the radar...

http://www.faz.net/aktuell/gesellsc...r-armee-haelt-kuehe-fuer-feinde-13483032.html
 
I think I just might have right idea for that:

Frigate visible in the picture is 150 meters long

slayton_concept01.jpg


I'll need to use 2 UVW channels for fine detailing but that's acceptable. Asteroid is only 1100 polys.
 
I think I just might have right idea for that:

Frigate visible in the picture is 150 meters long



I'll need to use 2 UVW channels for fine detailing but that's acceptable. Asteroid is only 1100 polys.

You could also improve this by using a geometry shader for generating finer details on the surface of the asteroid by a height map. Would have the advantage that all such calculations happen on the GPU, can be adapted to distance to the camera origin and can be simply turned off.
 
The new Swiss air surveillance radar is mistaking cows for attacking jets.
:rofl:

I doubt we'd have enough ammo in stock to eliminate all the threats... :shifty:
 
If they are really Switzerland, the cows would be part of some sinister defence system. :hmm:
 
Back
Top