Anybody good at creating cartoon heads?
-
spogadog said "Thanks for clarifying the 'remote-move' thing. I may be wrong, but this does not seem to work with single points selected. The Move tool stays put, then 'bollocks' you if you try to pull away from it. If this is the case, perhaps the folk at SU could iron out this anomally?"
This is not an anomaly. The move tool must be on the endpoint (vertex) to be moved, as you cannot select an endpoint. You can only select an entity.
However, there is a trick I use when I need to move an endpoint with the "Remote Move" technique.
Draw a short line from the vertex endpoint and away from it ending in space, not connected to anything else in the model. Then select that line and perform the remote move. It will drag the vertex with it. Simply delete that line when done.
-
Thanks, another 'beneficial nugget'!
Ok, how's about this: you can use the smoove tool on a polygon surface, but it does not appear possible to select and area and then 'add detail'. Try this, and the whole shape gets taken on to the next iteration of 'multifacetedness'?
Do I take it the only way to selectively to increase the number of polygons in a given area is to draw each one in 'manually'? Or is there, perhaps, some means by which you can DE-select an area back by one iteration?
@jgb said:
spogadog said "Thanks for clarifying the 'remote-move' thing. I may be wrong, but this does not seem to work with single points selected. The Move tool stays put, then 'bollocks' you if you try to pull away from it. If this is the case, perhaps the folk at SU could iron out this anomally?"
This is not an anomaly. The move tool must be on the endpoint (vertex) to be moved, as you cannot select an endpoint. You can only select an entity.
However, there is a trick I use when I need to move an endpoint with the "Remote Move" technique.
Draw a short line from the vertex endpoint and away from it ending in space, not connected to anything else in the model. Then select that line and perform the remote move. It will drag the vertex with it. Simply delete that line when done.
-
Select the area that you want to be selectively subdivided, group it, then go into the group and use the subdivide script on it. Always remember the maxim that Grouping separates geometry. You can't use Subdivide and Smooth, however...because the smoothing part will distort the edges of the selection and the group then won't connect with the surrounding geometry.
EDIT: After actually playing around with the settings after I wrote the above, it seems that if you follow the above procedure and only then apply smoothing to the geometry (using the Smooth all Connected Geometry tool on the SDS toolbar...not sure what it is in Artisan, as I haven't upgraded) this results in a much less abbreviated perimeter than if you just dived straight in with SDS. This means that you can effectively subdivide and smooth a small part of the model, then reconnect that section to the rest of the geometry with a minimal amount of tidying up. These illustations both used 2 iterations.
-
I found all I needed to do was select a facet then click the (brown/mesh)'add detail' icon . I'm assuming this is what was meant by 'subdivide', as the neasrest command I could get to that was 'subdivide selection', which did b*gger-all!
What still perplexes me however, is SU's ready-made sphere which seems to live by its own rules compared to all the other 3D shapes in the programme. Why, for example, does the above 'subdivide-move' not work the sphere? More confusingly: where are the triangular facets with this shape? You can stretch and squash the thing all you like, but its not until you go the next iteration (and beyond)that battalions of our three-sided chums finally show up.
Finally, how do you guys get your models/pictures onto this forum? Please don't say it's via the warehouse, as I've well & trully screwed the whole password/username deal as far as that is concerned.
Thanks for your patience with my ineptness, both of you!
@alan fraser said:
Select the area that you want to be selectively subdivided, group it, then go into the group and use the subdivide script on it. Always remember the maxim that Grouping separates geometry. You can't use Subdivide and Smooth, however...because the smoothing part will distort the edges of the selection and the group then won't connect with the surrounding geometry.[attachment=1:1ni507wb]<!-- ia1 -->heado_thumb.jpg<!-- ia1 -->[/attachment:1ni507wb]
EDIT: After actually playing around with the settings after I wrote the above, it seems that if you follow the above procedure and only then apply smoothing to the geometry (using the Smooth all Connected Geometry tool on the SDS toolbar...not sure what it is in Artisan, as I haven't upgraded) this results in a much less abbreviated perimeter than if you just dived straight in with SDS. This means that you can effectively subdivide and smooth a small part of the model, then reconnect that section to the rest of the geometry with a minimal amount of tidying up. These illustations both used 2 iterations.
-
No, I did actually mean subdivide, not add detail...but add detail works just as well, if not easier. I forgot you probably don't have subdivide, as it's part of the Subdivide & Smooth plugin, which is commercial.
True, Add detail won't work on the default SU sphere (where is that hiding nowadays?...it's not on the draw menu...I usually just make my own). It won't work, because Add Detail doesn't work on quad faces. It's part of the Sandbox set and is designed to work on terrain meshes which are already triangulated. If you use it on the sphere you'll see it only subdivides the triangles at the poles, not all the latitude and longitude stuff in between. There is however a couple of triangulate scripts by TIG that'll do the job just fine.
-
"Sub-Divide" is something I never used, in fact (2 actually) never knew about it till I read this thread;
and, I am usually cursed with too much detail and I spend a lot of time reducing complexity.
I get all this extra detail from Intersecting and using Fredo's Curviloft, which, after you get a bit more experience with native SU, you should seriously look at for your curved mesh generation.As for getting files and pictures loaded into a thread reply; if you look at the bottom left of the reply layout where you see the OPTIONS list, right beside it is the UPLOAD ATTACHMENTS tab. There you can point to the file/pic you want to upload, and add a comment.
-
Right, I've just tried attatching something.....
@jgb said:
"Sub-Divide" is something I never used, in fact (2 actually) never knew about it till I read this thread;
and, I am usually cursed with too much detail and I spend a lot of time reducing complexity.
I get all this extra detail from Intersecting and using Fredo's Curviloft, which, after you get a bit more experience with native SU, you should seriously look at for your curved mesh generation.As for getting files and pictures loaded into a thread reply; if you look at the bottom left of the reply layout where you see the OPTIONS list, right beside it is the UPLOAD ATTACHMENTS tab. There you can point to the file/pic you want to upload, and add a comment.
-
I'm certainly starting to appreciate the significance of overdoing the poly-count. Try as I might, I still find mtself knocking up stuff that is way too complicated. For instance, my latest creation has more faces than Tony Blair, and I've still got the jaw and ears to add!
Today's 'Eejit-special' is on the subject of components. Does Sketchup 'regard' these as less complex entities? I'm just trying to think of ways to avoid over-simplifying my models.
@alan fraser said:
The same head, a little further advanced, then mirrored. 390 faces so far...just about on course for 500 by the time I've added ears and some suggestion of hair.
-
No they're clones that reduce file size because SU only has to know where in 3d space they reside not how much geometry they contain. That's a rudimentary answer without getting too in-depth.
For example make a sphere and group it. Then make 10 copies of that group and save it. Check the file size.
Now make a sphere a component and make 10 copies and save then check the file size.
So if your making a head model half and mirror a component for the other half.
-
Though Alan has a file somewhere that's only 100kb in size yet makes SU hang. So components can also cause issues if you go OTT with them
-
That's interesting. I'm actually using SU to create virtual mannequins and save time on 'old school' construction techniques (a lot of folk appear to be doing that?). To that end, I'd like to keep the surface detail and contours 'to the max' as much as possible. I know that such an approach can seriously slow down the PC in the construction phase, but are you saying that once these complex shapes have been 'componented' you can effectively have the penny and the bun?
@unknownuser said:
No they're clones that reduce file size because SU only has to know where in 3d space they reside not how much geometry they contain. That's a rudimentary answer without getting too in-depth.
For example make a sphere and group it. Then make 10 copies of that group and save it. Check the file size.
Now make a sphere a component and make 10 copies and save then check the file size.
So if your making a head model half and mirror a component for the other half.
-
@spogadog said:
....but are you saying that once these complex shapes have been 'componented' you can effectively have the penny and the bun?
Only as far as file size in concerned...not as far as navigability is concerned. The only thing SU is concerned with in terms of being dragged to its knees is how many faces and edges the model contains...it doesn't matter how you dress them up. They can be groups, cloned components or even just completely exploded raw geometry, there'll be little if any difference in performance. In fact the exploded one might even be easier to work with, as SU doesn't have the extra hassle of burrowing down into nested groups and components.
-
Alan's spot on there. Going OTT on components does more harm than good. It's a balancing act at times.
Is it something like this that you're modelling?
-
I'm actually modelling the cartoon form itself, i.e: a 3D Mickey Mouse, Donald Duck, etc. As previously mentioned, this saves so much time I'm sure I can't be the only person doing this?
Apologies for not being able to show this stuff, but I'm having major problems accessing the Sketchup Warehouse. Having screwed up+++ with the password/username malarkey I now don't know what to do. I was using this facility on an almost daily basis about 8 months ago and could really do with it now. Can anyone help me to start over with it?
@unknownuser said:
Alan's spot on there. Going OTT on components does more harm than good. It's a balancing act at times.
Is it something like this that you're modelling?
[attachment=0:3slleag8]<!-- ia0 -->Mannequin.png<!-- ia0 -->[/attachment:3slleag8] -
Something like this?
-
-
I'm staring to make pretty good progress now, but am finding sketchup slows alarmingly if I import a scanned profile to follow. Should that be happening?
@alan fraser said:
@jgb said:
Using SU to do what you are attempting is akin to using a screwdriver to hammer nails. Doable, but not well. SU will frustrate you in short order trying to model/manipulate life shapes.
Trust me, I tried....
On the contrary, I've just spent most of this thread so far demonstrating exactly the opposite.
The figure 5th post down was constructed entirely in SketchUp. Sure it takes time if you're constructing it from scratch...but it does in Max or Maya too. Daz or Poser models...even if exported at low-res are MBs in size and almost unusable unless you go in for serious poly-reduction and cleaning up(which can take as long as actually building them yourself.)I've already demonstrated how to make a head. Justin Chin demonstrates here how to make the rest of the body. This is exactly the way I do it...but with fewer polys. It's basically just a series of layered cylinders which are then manipulated to the correct proportion and angle.
[flash=640,390:29gjob29]http://www.youtube.com/v/ZzDM_JrHCGM?version=3[/flash:29gjob29]
[flash=640,390:29gjob29]http://www.youtube.com/v/8pz7IXpTMMk?version=3[/flash:29gjob29]
-
Some of my characters are also created from images I got from a client for a theme park in Japan. Do you first draw the characters before modeling them? If yes, got an image to show?
-
At the moment I can't get sketchup to work above crawling speed as soon as import an image to trace. This is despite cropping the shite out of the scanned image and pulling the resolution right down to the lowest setting. Is this common?
@solo said:
Some of my characters are also created from images I got from a client for a theme park in Japan. Do you first draw the characters before modeling them? If yes, got an image to show?
-
Advertisement