Noname's displayed method actually isn't that time consuming. Its pretty much the same method I use for braided cable/wicker basket's, takes like 15-20 minutes to crank out an entire weave.
For that bed its even simpler, I mean you just need the first one then smart move (or equiv in max) the number of copies you need to make 1 row, then its just a matter of taking the entire row, dup and snap the number of times you need to make the entire weave, combine and weld.
Theres no need to do bends for a cable or a corner weave so really a fast method.
[Maya] Braided Cable Tut 600 mb download for a 20minute video...but if anyone cares enough or has a fast enough connect meh. as long as you have xnormal you should have the codec. (edit: I made several mistakes while making the video so it takes longer than normal, but you get the picture. also, done a while ago)
What advantages are there of making a braid or repeating texture in max over using xnormal and heightmap to normals? Than overlaying the normal over the other normal with a soft light blend?
I don't usually find myself needing to ask for help with this stuff, but after about 10 tries I've had no success modeling the front of this train.
Pretty sure this is the right thread for this kind of thing.
Anyway, this is what I have so far.
I haven't got any idea how to start it off, and every time I do the shape just seems to fall apart, never really coming close to what it should look like.
I really dont want to be the one who tells what to do and so but does not it better starting from the most complex parts (front of the loc. here) and then do the other parts?
Ooh.. Lattice Modifier is a fine idea. What about this array tool? whats that do?
There's a better way of doing this if you can use max;
Create a plane with same amount of divisions as the strings, apply a tesselate modifier to get diagonals. Covert to poly and select the diagonal edges, create linear shape from selection and mirror it horizontally.
*You will still have the ability to adjust the editable splines so it's flexible as apposed to the already polygonal geometry in the tutorial.
i tried a bit & finally managed to model it..but i used a different image in the front view which actually was the F430's steering wheel, so my model looks a bit off..anyway i'm glad itried....thanks pedro amorim for motivating me..here are two screenshots-
Hello. I'm making a high poly desert eagle but I'm pretty new to modeling so I need some help. I'm having troubles with the area where trigger guard connects to the grip. How would u model that transition?
I have that n-gon there which I also need a little help with. I don't want to continue these edges to the other side of the mesh, how should I terminate them?
It's difficult to say without seeing the whole reference and your model so we know exactly what were looking at. Just continue blocking out the basic shape, later create your extrusions/depressions and finally tighten up edges by adding loops or insetting groups of polys as you switch between smoothed and unsmoothed modes to see the changes.
Pedro is right, most of the time ngons don't pose a threat to a mesh smoothed model as long as they are planar and it looks that way in your example.
ps: one thing to watch out for is when multiple edges converge, even two can cause smoothing issues.
If is smooths right.. leave the ngons there. No point in trying to making shit clean that smooths ok.
+1
Tri's & N-Gons don't mean shit as long as they aren't causing smoothing errors. In fact if left alone they generally lead to much easier oportunities of adding detail into a mesh as the LoD progresses.
ps: one thing to watch out for is when multiple edges converge, even two can cause smoothing issues.
+1
Adding to that a similar smoothing issue that occurs when multiple edges converge near a bend can occur with n-gons ending near a bend.
The easiest way to fix it is to either drag your initial control edge in slightly deeper and place a second one between it and the actual edge thats being sharpened and that will fix both problems. Though that isn't Always the solution.
You'll get a feel for it as time goes on, when you need to do stuff like that and when not to. atm it looks like its not really causing problems.
got a question. when working with hardsurface, when it comes to creating the lowres mesh and smoothing groups, is it better to give an object control edges for the object to have just 1 smoothing group, or use multiple smoothing groups for hard edges with no controlling edges?
Good question.
Chamfering an area on the low poly to allow a single smoothing group is a good idea as apposed to spliting uvs which produces obvious baking seams.
It's a matter of balancing the creation of seams against a triangle budget, where the model may have to be reduced in less visible areas to allow for seam correction in focal points.
Good question.
Chamfering an area on the low poly to allow a single smoothing group is a good idea as apposed to spliting uvs which produces obvious baking seams.
It's a matter of balancing the creation of seams against a triangle budget, where the model may have to be reduced in less visible areas to allow for seam correction in focal points.
Im not sure if you mean diffuse or normal seams?
Splitting the UVs where hard edges are, actually reduces the normal map seams and are generally in the same places that you would also want UV seams too be anyway.
lets not forget that adding a bevel or using smoth groups, doubles the vert count in the same way.
+1
The cost is identical whether you bevel or add hard edges to a mesh and which to add would depend entirely on the model you are working on.
Dont some engines render long thin tris less efficiently than others too?
Its always better to use 1 smoothing group if you can, but that generally only works well if the tangent basis of the baker matches the engine you are working with.
I tend to use smoothing groups and bevels depending on the area im working on but honestly i prefer smoothing groups as the resulting normal maps are easier for me to edit in PS later.
What program do most of you Polycounters(?) use for modeling? I have some ideas for weapons for Team Fortress 2, but I don't know how to model them. Thanks.
Hey. I was just wondering how I would approach this:
A pistol grip with the diamond pattern. Not so much the big diamond in the centre but the lots of little diamonds. I've been thinking I could use bend modifiers but that wouldn't give me very much control over the shape.
Oh hah, thanks pern. I was just giving a quick how to start idea for Brandoom, there was no actual intention of it being useful other than a starting point for how to approach the problem. Its just lower geo is faster to poop out a result to get an idea across.
How would I go about rounding this corner without the use of chamfer? Whenever I use chamfer it gets really sloppy and I spend ten minutes touching up all the corners. I'm trying to make a floor panel with a rounded concave rectangle in the middle, but I'm still learning how to sub-d and I'm TURRIBLE at it so far.
@Snakedoctor:
You could use your approach, just make a small cylinder, go into polygon mode , select extrude along path(pick helix), set iterations to a reasonable amount and you are ready to go.
@Snakedoctor:
Consider you're in max, on the helix spline enable "renderable spline in viewport" and then trim unnecessary itterations via dot-ring-selection aviable in the graphite tools.
EDIT: oh the cable isn't round huh? I'd go with loft then.
I have another question, is there any way to apply turbo smooth to a specific part of a model? Everytime I use it, it applies itself to the entire thing and ruins most of the geometry. Am I not supposed to be using turbosmooth that often for high poly modeling?
No I'm not collapsing the stack, I switch back and forth when I need to make changes. I suppose I just need to get used to supporting edges with more loops if I want them to retain their shape. I'm not used to that yet. I'll mess with one area to make it look round, but then an area I haven't fiddled with that's supposed to be hard edged will lose all its shape due to turbo smooth.
I'd love something like Pixar's semi-sharp subdivision settings. If I recall correctly it's a 0-1 range and from what I could see it really takes away the time of doing supporting edges. Can't say what special features (or maybe it was in one of their online papers) I saw it in but man was it cool.
Yeah it's hard to make sure you're getting the right results from turbosmooth but once you start thinking in the right flow you get the hang of it (for the most part, heh).
im pretty new to modeling and am just wondering about the best way to make a rounded inner corner on a object.
i tried just using a subsurf modifier on the whole object and use loop cuts to better control the shape but this knocked the poly-count above what i would like since this model is for use in a the source game engine.
by the way im using blender 2.55 but im sure most methods of doing this in a other 3d package would still be valid in blender.
I'd love something like Pixar's semi-sharp subdivision settings. If I recall correctly it's a 0-1 range and from what I could see it really takes away the time of doing supporting edges. Can't say what special features (or maybe it was in one of their online papers) I saw it in but man was it cool.
Yeah it's hard to make sure you're getting the right results from turbosmooth but once you start thinking in the right flow you get the hang of it (for the most part, heh).
im pretty new to modeling and am just wondering about the best way to make a rounded inner corner on a object.
i tried just using a subsurf modifier on the whole object and use loop cuts to better control the shape but this knocked the poly-count above what i would like since this model is for use in a the source game engine.
by the way im using blender 2.55 but im sure most methods of doing this in a other 3d package would still be valid in blender.
The main point of hard-surface modelling is to get high detail, rounded edges 'n such, bake a normal map form those details, and then use the lowpoly in game. Don't worry about the tri count until you have made that lowpoly. And yes, source can use normal maps, so you are on the right track.
The main point of hard-surface modelling is to get high detail, rounded edges 'n such, bake a normal map form those details, and then use the lowpoly in game. Don't worry about the tri count until you have made that lowpoly. And yes, source can use normal maps, so you are on the right track.
ah so continue making it high poly with subsurf like i am with lot of loop cuts for control and bake my normal map, than remove the sursurf or decimate things to a low poly version.
Replies
For that bed its even simpler, I mean you just need the first one then smart move (or equiv in max) the number of copies you need to make 1 row, then its just a matter of taking the entire row, dup and snap the number of times you need to make the entire weave, combine and weld.
Theres no need to do bends for a cable or a corner weave so really a fast method.
[Maya] Braided Cable Tut 600 mb download for a 20minute video...but if anyone cares enough or has a fast enough connect meh. as long as you have xnormal you should have the codec. (edit: I made several mistakes while making the video so it takes longer than normal, but you get the picture. also, done a while ago)
for a lot of people it's just easier to do it in 3d.
I really dont want to be the one who tells what to do and so but does not it better starting from the most complex parts (front of the loc. here) and then do the other parts?
There's a better way of doing this if you can use max;
Create a plane with same amount of divisions as the strings, apply a tesselate modifier to get diagonals. Covert to poly and select the diagonal edges, create linear shape from selection and mirror it horizontally.
*You will still have the ability to adjust the editable splines so it's flexible as apposed to the already polygonal geometry in the tutorial.
eg:
But then, pls you can write what mod. methods you were using. (:
Good tuts on that would be a nice addition to this thread.
http://www.staticcurve.com/Artists/viewtopic.php?f=11&t=168
http://www.polycount.com/forum/showthread.php?t=62214
I have that n-gon there which I also need a little help with. I don't want to continue these edges to the other side of the mesh, how should I terminate them?
It's difficult to say without seeing the whole reference and your model so we know exactly what were looking at. Just continue blocking out the basic shape, later create your extrusions/depressions and finally tighten up edges by adding loops or insetting groups of polys as you switch between smoothed and unsmoothed modes to see the changes.
Pedro is right, most of the time ngons don't pose a threat to a mesh smoothed model as long as they are planar and it looks that way in your example.
ps: one thing to watch out for is when multiple edges converge, even two can cause smoothing issues.
eg:
+1
Tri's & N-Gons don't mean shit as long as they aren't causing smoothing errors. In fact if left alone they generally lead to much easier oportunities of adding detail into a mesh as the LoD progresses.
+1
Adding to that a similar smoothing issue that occurs when multiple edges converge near a bend can occur with n-gons ending near a bend.
The easiest way to fix it is to either drag your initial control edge in slightly deeper and place a second one between it and the actual edge thats being sharpened and that will fix both problems. Though that isn't Always the solution.
You'll get a feel for it as time goes on, when you need to do stuff like that and when not to. atm it looks like its not really causing problems.
Just my experience anyways.
I have some triangles now but it seems to smooth ok.
Good question.
Chamfering an area on the low poly to allow a single smoothing group is a good idea as apposed to spliting uvs which produces obvious baking seams.
It's a matter of balancing the creation of seams against a triangle budget, where the model may have to be reduced in less visible areas to allow for seam correction in focal points.
Splitting the UVs where hard edges are, actually reduces the normal map seams and are generally in the same places that you would also want UV seams too be anyway.
+1
The cost is identical whether you bevel or add hard edges to a mesh and which to add would depend entirely on the model you are working on.
Dont some engines render long thin tris less efficiently than others too?
Its always better to use 1 smoothing group if you can, but that generally only works well if the tangent basis of the baker matches the engine you are working with.
I tend to use smoothing groups and bevels depending on the area im working on but honestly i prefer smoothing groups as the resulting normal maps are easier for me to edit in PS later.
So basically use what looks best
Sorry bit of a small image. This shape for whatever has my brain wracked. Not worried about the fancy details, just the main shape.
Obj File if you want/need to spin around a bit.
Any tips would greatly be appreciated.
http://www.polycount.com/forum/showpost.php?p=893340&postcount=233
Crappy resolution, but you can see how I did it on my revolver ;p
http://fc07.deviantart.net/fs70/f/2010/028/a/e/Brotherguard_Highpoly_Wire_by_The_5.jpg
I did it the way racer (was it racers?) did it in one of his "Hardsurface Elementals" videos. (currently cant find them, I downloaded them back then
A pistol grip with the diamond pattern. Not so much the big diamond in the centre but the lots of little diamonds. I've been thinking I could use bend modifiers but that wouldn't give me very much control over the shape.
Any ideas?
huh?
His OP
How would I go about rounding this corner without the use of chamfer? Whenever I use chamfer it gets really sloppy and I spend ten minutes touching up all the corners. I'm trying to make a floor panel with a rounded concave rectangle in the middle, but I'm still learning how to sub-d and I'm TURRIBLE at it so far.
there's a starting point for topology
I tried using a heliex with enough rotations and used a path deform but got stretching problems.
You could use your approach, just make a small cylinder, go into polygon mode , select extrude along path(pick helix), set iterations to a reasonable amount and you are ready to go.
Consider you're in max, on the helix spline enable "renderable spline in viewport" and then trim unnecessary itterations via dot-ring-selection aviable in the graphite tools.
EDIT: oh the cable isn't round huh? I'd go with loft then.
Thank you, that helped very much.
I have another question, is there any way to apply turbo smooth to a specific part of a model? Everytime I use it, it applies itself to the entire thing and ruins most of the geometry. Am I not supposed to be using turbosmooth that often for high poly modeling?
I always use turbosmooth and yes it's applied to the entire mesh.
Ah well.
Yeah it's hard to make sure you're getting the right results from turbosmooth but once you start thinking in the right flow you get the hang of it (for the most part, heh).
i tried just using a subsurf modifier on the whole object and use loop cuts to better control the shape but this knocked the poly-count above what i would like since this model is for use in a the source game engine.
by the way im using blender 2.55 but im sure most methods of doing this in a other 3d package would still be valid in blender.
Not shure what you mean, but reminds me of this.
Never tried the Crease Parameter myself, I just stumbled over this today (here at PC?).
[ame]
The main point of hard-surface modelling is to get high detail, rounded edges 'n such, bake a normal map form those details, and then use the lowpoly in game. Don't worry about the tri count until you have made that lowpoly. And yes, source can use normal maps, so you are on the right track.
ah so continue making it high poly with subsurf like i am with lot of loop cuts for control and bake my normal map, than remove the sursurf or decimate things to a low poly version.