I'd do those trees as actual 3d boxes with 2d transparency fins and detailings. X or Y at that size and shape won't look so great. X looks ok for smaller less regular hedges, or at a push box hedge, but box trees won't work so well with Xs. Don't forget a bit of waviness in positioning either. PS, on re-positioning on vertical if ground is adjusted, if they're separate instances (and only joined for export pass), just lift them all 10m, then use drop to surface script (scriptspot). On my city tracks I was doing this re-dropping in Z many times for all track objects, road decals, barriers etc, as they were done without lidar etc, so quite often things were tweaked after sim testing or better data arrived or whatever. In fact most items were on separate build layers and placed layers. I'd just delete all placed objects and clone and re-drop everything from the build layer then 'combine' for export each time to assure everything was correct after a change.
Yup, my very first attempt was a box with wings for the top, separate from a Y mesh trunk, so I could mix and match them for more variation, I thought it might be too many polys so went back to Y planes with a squared off alpha which didn't look good at all. Still haven't really given it a good shot, I keep getting distracted by other parts of the map and I was using Forrest Pro to distribute them which doesn't work so well on single splines for distribution, it's a bit overkill for the job in hand. I use SplineTrackMaker MMTools for all positioning, there are some pretty powerful tools buried inside it, the random feature is awesome as is the move to surface function with Z offset, groups align to normals, FFD's. http://www.mj-multimedia.net/soft/mmtools/mmt_rnd.php http://www.mj-multimedia.net/soft/mmtools/mmt_move2surface.php
Yeah, not very well documented or advertised, it's a shame as there are some great tools in the package, I keep discovering new things every time I use it. I just pulled out for an overview render...I thought I'd done more
it was RTB v MMs 2 years ago. now ive had a lot more max experience I wish id opted for mm scripts. a while back I was criticised for suggesting rtb could have a set of companion scripts for 3dmax .
Those scripts look good. Probably some duplication of stuff now inside Max etc, or do-able in different ways, different random snippets of scripts here and there, but all in all they look like some must-have's for anyone doing a proper racing circuit etc. The move points/selections to spline, along with edge tools relax, would be a nice way to use sub-d and splines to do some jobs where you want the power of splines but the power of sub-d too. That said, I just want some sub-d system that works like point based nurbs. Why there isn't point based sub-d is beyond me. For technical work, the point nurbs seem so useful, so why no equivalent in sub-d? Hmmm Dave
I'm having to break it up into smaller sections and use the models.ini, that renders not even half the route, champs is over a mile long on it's own with tons of furniture, I'm already having to break up the meshes to keep within ksEditors limits.
I was bored, thought I'd start addressing the buildings on this track Because there are so many (literally 1000's) I really don't fancy messing with lods, I'm trying to come up with the most optimal way of making them, in theory, with this method I could fit most of the building maps onto one or two textures, and all the meshes an be one object. Variation is endless as I'm using simple uniform meshes and face mapping, a modular approach where I could mix and match windows, doorways and shop fronts, with no repetition...only a few hundred faces per building. Roofs, balconies, and top windows can be mix and match as well, chimneys can be a row on 2d planes with alpha. I'm just using rough google screen grabs for now, but it would be possible this way to have all the buildings near as damn look like they do in real life. Each shop front could be, 128 x 256, each window 256 x 128,I could fit tons on a single texture. ...thoughts?
I can't really tell without trying, but shop fronts sound a little small resolution, imo ? I mean it's what you would see most as it's same level as player's view. Are you able to make a harsh calculation of how many buildings you'll actually have ? (front row only). This way estimation of texture size can be made as well. Also, are you planning for that era's environment, or modern ? I'm thinking especially Champs changed quite some in the mean time. Other than this, I love all you modelling tricks - it's inspiring. Sent from my phone using Tapatalk
I'd run with that approach for this track too. Trying to do anything else, even if the end results might be better, is just going to take forever and add details you can't really appreciate when driving along any way. It'd be nice though if AC supported vert colouring of some nature. Or had more flexible shading for these kinds of things. In Free Racer for instance, in this case, I'd use vert channel for AO baking since your resultant mesh is quite uniform. A node based shader creation system like UDK would have been a great inclusion in the ks editor, but hey ho. I'd also just add 2 or 3 variants per window type, and intersperse the odd bit of open window, closed curtains, whatever. Also the ground floor part could be split up and make bits tileable so you can swizzle stuff even more just to break up repetition a bit more. But then also remember with tree lines, street furniture generally, you might hardly see this stuff, only a few locations at low speed might you get close to appreciate it. I still tend to block out tracks and get trees in before I go near texturing buildings etc because sometimes some buildings need lots more detail, others need hardly any. Hard to know till you drive around Great work though, can't wait to see more Dave
Cheers, 'll probably go for the 76 layout, the only downside is lack of reference. The champs tress are pruned high so the shop fronts and bottoms of the buildings are pretty visible. Number of buildings? hmmm...lots... on the champs around 200, beyond the Louvre...a few more. I'm just messing with ideas, something should come of it...eventually ...adapt workflow to suit, experiment, etc.
But thankfully it's a nice consistent style of buildings, with a fairly consistent heights etc. By all accounts it's quite well suited to being modelled/textured relatively easily compared to a similar length track in different cities with more variation and long sight lines etc.
I'm just going to take it street by street, the number of texture variations needed will become apparent over time, I'm trying to avoid any bland repetition. I don't think consistent belongs in a sentence describing Paris streets...you need to look closer
It's consistent looking. Everything looks to be within 6 storeys +/- 1 storey kinda thing. No gaps between buildings that you can see through, just down side streets. On textures and variants I agree. Just keep it simple on that front and add in what you need once it's clear it's not enough. Especially on a track this big. I doubt any one will know it well enough to know the little tricks and copies you use here and there I often drive in Gran Turismo cities and wonder how accurate they really are. They obviously get them close enough, but they must cut many corners too
Lol, I don't see consistent in this image: ...anyway...it is what it is I want the track to reflect the film as close as possible. This now also has my attention:
