How do I make a vehicle with edgeflow this good? Just using subdiv/turbosmooth from a low poly mesh gives me a lot of trouble areas.Using Maya right now.
>>586486what do you mean. you can just build the goddamn car in nurbs if you find polygons to hard and convert it. also is tris mixed with quads acceptable now a days ? what the heck
>>586490From what I gather, rigid stuff doesn't need to be all quads, so whatever works works, as long as it looks good and keeps inside budget.
>>586490I'm looking for workflow tips and whatnot, most youtube tutorials, while very nice, don't seem to really care about polycount. OP looks great without wasting polys which is what I'd like to be able to do.
>>586486Well you're lucky, Maya's curve tools are some of the best for modeling car.https://www.youtube.com/watch?v=YveckLumNGU
>>586486>How do I make a vehicle with edgeflow this good?>edgeflow this good?... Huh?I'm not sure what you're basing our observations on here, but that's some of the worst vehicle topology I've ever seen.
>>586652>worst topologyIt's only nurbs surfaces that have been converted to mesh. If you want to animate it as an anthropomorphic car, retopo it.
>>586602took me a while to find ithttps://www.artstation.com/artwork/mYVE1
>>586490the all-quads thing is to facilitate good deformation. Although good practice it's a lot less relevant for rigid objects.
>>586490ALL QUADS is necessary for deformation AND if you intend to use subdivision modifiers.The car in the OP probably is a model for a game and won't be subdivided anymore and also doesn't have organic deformation so it's ok
>>586486>How do I make a vehicle with edgeflow this good?>How do I make a vehicleor>How do I master edge flowIf you want to make a vehicle, you don't mash a bunch of triangles together, you use CAD where the geometry looks correct because it is.
>>586486You could also try smoothing groups, don't know if they exist in maya though
>>586915>use CAD where the geometry looks correct because it isYou are aware that the model in OP's image is a game-ready mesh? Good luck trying to tesselate CAD with that kind of surface quality, you'll end up with that polycount times 5.
>>587190>You are aware that the model in OP's image is a game-ready mesh?I am. I am also aware that a "game-ready mesh" doesn't mean shit and does absolutely, positively, fuck all in the real world/production line>Good luck trying to tesselate CAD with that kind of surface qualityI just really hope you're trolling. Pick a car. Any car made in the last 20 years. Guess what type of program was used to design every single part and surface on that car. CAD. At the very least it has been a clay mold, scanned, only to be put in CAD and have the surfaces rebuilt to engineering specs. You can't and never will be able to use that fucked up batch of triangles you are defending for real world production beyond a 3d printed replica of a real object that's only use is to be printed and sit on a desk.>you'll end up with that polycount times 5.What is NURBS. So do you want to learn how to properly make a vehicle or do you want to learn to manipulate objects for a video game. Two different things.So yeah. Once again.If you want to make a vehicle, you don't mash a bunch of triangles together, you use CAD where the geometry looks correct because it is.
>>587199You're missing the point entirely. OP's car is hand-made to provide realtime shading as close to the real thing as possible.>NURBS surfaces are used in automotive productionThanks for explaining a concept to me which I use at work for 40h per week. Reply to this post when realtime engines start rendering NURBS surfaces for cars, my point was that you cannot tesselate CAD data to polygons at this level of surface quality and expect a similar polycount.Quit moving goalposts, the car in OP's post was obviously never meant to be used in any automotive production line.
>>586486>edgeflow this goodimplying>>586602>OP looks greatimplying>>586652>that's some of the worst vehicle topology I've ever seenwordThrow a reflective shader on it and prove that it actually looks good, because I doubt it does.
>>587423What constitutes good topo for a real time vehicle? Not trying to be a smartass, I'm genuinely curious. I'm trying to learn how to model and I'd like to avoid any bad habits if possible.The artist in the OP has an Artstation with a lot of vehicles like that and and a couple of renders btw
>>587608For realtime vehicles:1. No one gives a fuck about quads only2. Use control edgeloops instead of sharp edges, not for subdividing, but for improving shading quality.Also, according to my experience, in case you model a full interior - the exterior lacquer hull will make up for about 1/5 to 1/6 of your final polycount. I use this as an indicator of the final polycount since I usually start with the hull and do the interior later on when the exterior is complete.
>>587705So I guess it's not a good idea to use opensubdiv creases then?
>>587769I wouldn't use creases anywhere, they're a huge fucking mess in addition to you not being able to define curvature along the creased edge in subdivision.
>>587771Also, should I be gunning for a perfect subdivved mesh or is it normal to have areas where you have to do some cleaning/optimization afterwards?
>>587782Subdivision compatability doesn't matter for realtime cars for the most part, at least the lacquer and other "clean" hull parts don't need a normal map and threfore don't need a highpoly. Model a subdivision cage with decent density - most of the subdivision workflow rules also provide for good reflection quality - and then try saving polys where you can afterwards.