W.I.P. Lime Rock

[Mr Whippy]

SpeedTree speeds it up, but I wasn't hugely impressed with the tree breed specifics when I last looked.

Also last time I looked it was only available to dev studios. Is Speed Tree now available for free, or for 'free' somewhere?


On grass and trees. Wasn't the issue up normals and round normals done via KS editor etc?

Ie, the grass shader calls up normals itself, rather than taking the ones from the input mesh.

Same with trees, the tree object has round normals applied to the object (single tree), and they're combined in the KS editor...

But a 3D/2D hybrid tree could use a different shader. I'm sure there must be a shader that suits the 3D/2D type trees?
Trunk/main branches are just diff/spec/normals. Tree boughs can be diff/spec/trans (alpha check trans), as long as the normal direction is passed.

I thought AC passed explicit normals in all shaders unless they're overriden (trees/grass defaults)?


I need to go read up or find a proper FAQ haha, I'm still not totally up to speed on them all.


Can't wait to see some more progress on this one

Dave

 

[Pixelchaser]

per pixel shader for the better than y trees.

have went and re did the whole area. 4 hours start to finish, nothing really new to see, but its smoother, nicer, and down to 119K polys. loads instantly in the editor too which is a big difference .

 

[Pixelchaser]

\0/ close, but no cigar. but wickedly close so far. visual and physics layers test. (no noise yet, cambers or widths)

 

[luchian]

We'll have to smoke those mx5's on this track and show iR how it's done

 

[Pixelchaser]

I think its really fantastic so far. the difference in working with this data as opposed to without it is compelling. I like all of it so far apart from the uphill rise. I know it does have a sort of jump to the end of it, but mine seems to have one before it . had to resort to centre splining and general lofting, tomorrow its cross sections. that should give me the cambers and widths nicely.

timing wise it seems really spot on too. casually doing 55-57`s without the chicane,

 

[Pixelchaser]

my graphics cards blown. must have been those 41 million polygons. and ive no money to get a new one. feel free to use the donate button. coz I'm rooked with out a graphics card.

 

[luchian]

What ? How is this possible ? No matter how much you stress it, it should not happen (unless you overclock).
Was it a 670 GTX ?

 

[Pixelchaser]

just tried to go for a drive onhighlands for first time. I cant say I have not been over stressing it. maybe even the big scenery I was making has taken it to the edge. I hear the new 1050 has same performance for 130 bucks. so maybe that soon.

 

[Pixelchaser]

new monitor required now too. as new g cards don't have analogue inputs anymore. my acer monitor is only analogue. (2009)

 

[LilSKi]

Digital to analog converters work fine for that. We use them at work all the time.

 

[Pixelchaser]

i believe this connection doesn`t accept any analogue inputs.... anyway its done now. well I have a 1050 ti for 130 bucks and a dell p2212h monitor for 50 bucks, 2 people have been rather kind coz I really have **** all. This 1050ti is actually an upgrade of sorts. but not much 2g upto 4 gig vram and maybe 5 - 10 more fps, power consumption will be much better too. no external input power required for it.

and the bizzare part, my dfgt wheel feels like a completely new wheel. I must have had power issues all along, now the 670 was removed from the power unit(550w).

edit. yeah I see some nice vga to hdmi inputs, the dvi input is i duno what but my vga to dvi connectors I have dont work. so yeah I guess I didn't need a new monitor but would have had to wait a bit to get some connectors suitable so no biggy. means I can actually get 2 monitors going for working eventually.

 

[Pixelchaser]

solution to my dilemma. spline mapping a surface is fantastic, so easy to setup in max. but it only really works with a centre spline. but I'm trying this spline method with one at each side of track to get the physics layer.

so i must conform 2 meshes, 1 formed from the outlines and 1 formed from the centre line spline. bottom layer is the outer line mesh so I can grab the crazy width changes and general cambers at the same time. but you cant actually uv map that surface very easily. so I combine both. the constraints with 3dsmax splines and their general tools mean I will slowly manually conform my visual layer down to what will be the base of the physics layer. complicated and time consuming but it will be worth it, instead of unravelling a uv map without a centre spline or any form for basis. using snap to edge function with axis constraints I must go around the verticals first then then lateral changes. but great thing is the aleady uv mapped layer stays uv mapped nicely , obviously there will be some stretching , but that would be there anyway. next video will be a nice track, with cambers and bumps. fingers crossed.

 

[Pixelchaser]

11 hours that took me to conform the surface down to ground level. so then I tested it in game. It has captured the cambers and widths amazingly. also the physics surface I made for this has bumps and feeling built in now due to how ive performed on the visual surface. the slight undulations present on the lidar transfer well in this method.

I used the new visual layer for physics, subdivided the hell out of it. and its quite nice(without applying noise)there are a few errors that I have made. and it needs a once or twice over to smooth a few sections out meaning the video isn't worth posting as at one point I'm flying in the air. but the main part is where its right its right and it performs really quite well.

 

[Pixelchaser]

<waves> AGAIN !i may wont be around for a while unless i have another miracle. 500 ml of cola dripped through my pc tonight. not even payed last weeks shenanigans off. hope this can dry up.

 

[Pixelchaser]

3,2,1 and hes back in the room. wow. coca cola aint so bad for now .

absolutely loving this project. never seen anything come together so well and nicely. this video is a test of cambers widths and heights, separate visual and physics layer. I havnt performed any noise yet, infact I'm not sure it even needs it to be honest about it, the feeling is already incredible and on par with anything else ive felt so far. (without going into the world of fictional noise)

my new graphics card doesn't record very well

 

[LilSKi]

Looks pretty good. Looks like a pretty big hump/jump where you would go under the bridge that should be smoothed out but other than that it looks great.

 

[Pixelchaser]

thanks man ! both uphill and downhill have issues so far. the uphill rise I quite like. but abump does come before the actual transition to flat. the down hill, that's so fucked at the moment . but I have not done the whole lets watch videos and analyse this thing yet. just relying and literally seeing what the outcome is laying stuff on processed lidar. the problem arose from the hill that supports the bridge, the optimising of polys in that area crept the height over into the track and I missed it, so I let it happen .

the techniques involved in driving lime rock do seem to be there. have a friend who drives iracing there and he thinks its really similar. and the mx5 cup car seems to drive like its counter part

 

[Pixelchaser]

object placement map and polygon area split indication so far.... rumble strips buildings and maarshall stands next .and tall fences with foliage breaking through, they will be a nightmare to simulate. tree types too as my imagery is from fall/autumn.

it is amazing how you familiarise yourself through this process. absolutely essential for a track build with an easy to streamline system. plenty of laps were watched

 

[Pixelchaser]

to build the autocross or to not build the autocross, that is the question.

regardless this is pretty epic.

 

[LilSKi]

Considering there is a drifting scene in AC it may be worth it. When I was deciding on the kart track for NJMP I decided against it as there are no good karts for AC anyway so I left it as visual only.