Yes. If the vertices are not locked in place that will happen. However one could saturate the image with many vertices, and the effect would be deminished. What I do is move the grid in lockstep. I also expand the mesh in lockstep. But finally I juat moved dynamically creating a clipmap mesh. Not sure which is better, mine or this approach.
This looks cool, however I have to say after working on the same kind of thing, it's a whole different game, when you start adding things you need for a real playable world: Caves and overhangs, trees and other assets, rivers, collision, pathing etc.
i know this demonstrative, but one thing i see so many entrants not doing is using ridged noise 1 - abs(2 - n - n) like on the smaller octaves of water here.
hi sorry to disturb you guy's i'm a pretty noob in unreal engine I've already done the planet with the water and atmosphere but i don't know how to make the volumetric cloud around the planet visible from the space it show me a glitched grey thing if you could help me i would be really thankful
Hey, am I right in saying that the terrain vertices need to be recalculated every frame that the camera moves? I'd like to try something like this but I'm thinking that things could get "interesting" if I wanted to add some kind of physics collider. Also the volumetric clouds look amazing.
They definitely do not need to be recalculated every frame. However, when using a quad tree for the terrain, you would need to recalculate some meshes.
Hi Alexandr, please have a look at the paper (same name as this video) in the referenced library. In case of any questions please contact the authors listed in the header of the paper.
That is an interesting technique for the terrain generation. I like... but it means you have to constantly update all the terrain. I think it's possible to do this spread across multiple frames, though. Alternatively in a vertex shader, but that would not allow for physics interactions.
No! Collisions are very easy for such terrains. We only need to pass the physics body's vertex closest to the center(or probably do it for each vertex on physics body if it doesn't necessarily collide from single point) and pass its coordinates to the terrain generator function. If physics body lies closer to the center of planet than value returned by function, collision detected! Beware, avoid doing complex calculations else floating point values will show there magic.
Hi KJTR! You have to ask the Florian directly, he is the author of the software. You will find his email contact in the paper listed in the referenced library (search for the title of this video).
In VR? There is a FREE ONE, It is my favorite thing to relax, I go in the AIRCAR and fly around, go way above the city and look down, then zoom down to sreet level ad fly between building. If you buy an Oculus Rift S, there is a free game called: AIRCAR... There is an impanel screen you can place, that show your desktop, what I do, is put Netflix on my desktop, then in Aircar, I have a 10" tv tablet on my cockpit panel and can watch Star Trek or Friends Re-Runs while flying around making believe 'im a future cabby! www.oculus.com/experiences/rift/1204731822978255/ or search Aircar on TH-cam… me-> TH-cam.com/NextWorldVR I'm happy to help!
@@UrodCyka no fast as like in this video, i know. but theyre literally standing in nasa feeds. just pick a point and mait 2min. you cant tell me clouds are don move in 2 minutes.
@@UrodCyka you can name what ever cloud u know. clouds dont stand 5 minutes on the same spot. and where do i get that "basic" knowledge ? in school i guess, so it have to be true. pls stop eating shit from the goverment. free your mind.
this is a bad technique, you're going to get vertex swimming if the sample points move with the camera instead of being constant in world space, it's very telling that the camera is static
Hi doltBmB! This is described in the paper in section 3. We prevent morphing vertices by not updating the projection in each frame which results in popping vertices. With tessellation it is barely visible anymore. However, we later improved the projection technique by snapping the grid on the planets ground level to get rid of morphing and popping vertices. See our latest publication "Real-time Rendering of Procedural Planets at Arbitrary Altitudes".
broh the waves !! and the clouds !! well done, fantastic work !!
will the mesh move awkwardly as you explore the planet?
Yes. If the vertices are not locked in place that will happen. However one could saturate the image with many vertices, and the effect would be deminished.
What I do is move the grid in lockstep. I also expand the mesh in lockstep. But finally I juat moved dynamically creating a clipmap mesh. Not sure which is better, mine or this approach.
This looks cool, however I have to say after working on the same kind of thing, it's a whole different game, when you start adding things you need for a real playable world: Caves and overhangs, trees and other assets, rivers, collision, pathing etc.
Yeah I've seen this same demo from twenty failed projects over as many years.
i know this demonstrative, but one thing i see so many entrants not doing is using ridged noise 1 - abs(2 - n - n) like on the smaller octaves of water here.
hi sorry to disturb you guy's i'm a pretty noob in unreal engine
I've already done the planet with the water and atmosphere but i don't know how to make the volumetric cloud around the planet visible from the space it show me a glitched grey thing
if you could help me i would be really thankful
*_Please, source code github link_*
Hey, am I right in saying that the terrain vertices need to be recalculated every frame that the camera moves? I'd like to try something like this but I'm thinking that things could get "interesting" if I wanted to add some kind of physics collider.
Also the volumetric clouds look amazing.
They definitely do not need to be recalculated every frame. However, when using a quad tree for the terrain, you would need to recalculate some meshes.
How to make this, using basic material shader or post processing
Hi Alexandr, please have a look at the paper (same name as this video) in the referenced library. In case of any questions please contact the authors listed in the header of the paper.
This is mind boggling. I’m working on Procedural Generation for this very thing. How on earth (pun intended) did you do all this?
That is an interesting technique for the terrain generation. I like... but it means you have to constantly update all the terrain. I think it's possible to do this spread across multiple frames, though. Alternatively in a vertex shader, but that would not allow for physics interactions.
Hey, you were on the space graphics forums (it's the greater game creator)
No! Collisions are very easy for such terrains. We only need to pass the physics body's vertex closest to the center(or probably do it for each vertex on physics body if it doesn't necessarily collide from single point) and pass its coordinates to the terrain generator function. If physics body lies closer to the center of planet than value returned by function, collision detected! Beware, avoid doing complex calculations else floating point values will show there magic.
Where would I go if I had questions about the contents of the paper?
Looks great! ;)
There is something about the clouds.. Shouldn't they kind of be a bit brighter with the sun behind them? Glow sort of? They seem so dense.
This is suuuuuper cool. Kerbal Space Program could learn a thing or two from you.
Can't find the paper on the link provided...
cescg.org/cescg_submission/real-time-rendering-of-procedurally-generated-planets/
Please read the paper in the description before commenting.
really i want this project file plzzz. great job dude.
Чувак, это мегакруто, хочу научиться так же, особенно, что планета получается огромной!
That water looks so yummy
fluffy clouds
I used to live in Arizona and the skies always had little fluffy clouds in em
can we use your plugins?
Hi KJTR! You have to ask the Florian directly, he is the author of the software. You will find his email contact in the paper listed in the referenced library (search for the title of this video).
amazing
cool !
We just need a cool sci fi city/planet with flying cars. I'd so pay for that
In VR? There is a FREE ONE, It is my favorite thing to relax, I go in the AIRCAR and fly around, go way above the city and look down, then zoom down to sreet level ad fly between building. If you buy an Oculus Rift S, there is a free game called: AIRCAR... There is an impanel screen you can place, that show your desktop, what I do, is put Netflix on my desktop, then in Aircar, I have a 10" tv tablet on my cockpit panel and can watch Star Trek or Friends Re-Runs while flying around making believe 'im a future cabby! www.oculus.com/experiences/rift/1204731822978255/ or search Aircar on TH-cam…
me-> TH-cam.com/NextWorldVR I'm happy to help!
This needs to be added to space engine
I’m pretty sure water reflection is planned and they are currently working on volumetric clouds which might be added next year
@@users4007 the water in space engine looks so bad, honestly... But clouds would add a whole new perspective...
awosome!
what the hell how he/she did this wow really love this project. i want to watch tutorial. can anyone plzzz give more info about this. thanks
in opposite to nasa videos, thhe clouds are moving here haha. greatk work.
because clouds dont move that fast
@@UrodCyka no fast as like in this video, i know. but theyre literally standing in nasa feeds. just pick a point and mait 2min. you cant tell me clouds are don move in 2 minutes.
@@svens3722 Were they cumulus clouds because clouds at high altitude work like that. not all clouds move at the same speed thats basic knowledge
@@UrodCyka you can name what ever cloud u know. clouds dont stand 5 minutes on the same spot. and where do i get that "basic" knowledge ? in school i guess, so it have to be true. pls stop eating shit from the goverment. free your mind.
this is a bad technique, you're going to get vertex swimming if the sample points move with the camera instead of being constant in world space, it's very telling that the camera is static
Hi doltBmB! This is described in the paper in section 3. We prevent morphing vertices by not updating the projection in each frame which results in popping vertices. With tessellation it is barely visible anymore. However, we later improved the projection technique by snapping the grid on the planets ground level to get rid of morphing and popping vertices. See our latest publication "Real-time Rendering of Procedural Planets at Arbitrary Altitudes".
I wonder how many computers were in the botnet he had run this
there's a lot more info in the paper in the description, this was reak time on a GTX 1060 in OpenGL