polygon model subdivision setting

[Flash FROG]

All I can really say is that when I am modeling in maya with subdivisions I usually use polygon proxy mode where the subs are smooth and there is still a rough polygon cage around the subdivision surface. Then to convert back to the original polygon, in the sub the poly menu just choose verticies so that it doesn't smooth it more.
I hope that this helps.

[magilla]

I'm also at the experimental stage with subd modelling in Maya but....

I don't think there are any settings for rendering subd resolution. You can choose to view the model at the various quality levels whilst working on it - but it will always render at the smoothest possible resolution. I can't see why you'd render at anything less.

I also use the poly mode (you can switch between poly and standard) which brings up the poly version as a transparent cage if I need to do splits etc. You are right though - it is slow. However I've found that working on a poly model and using the cps script (to approximate what the subd will look like), then converting to subd at the latest possible stage is a much more efficient way to work.

Once converted you can add your full/partial creases and any higher level detail using refine etc rather than any "drastic" modifications such as poly splitting.

As a side-topic to this thread:

does anyone know how to preserve creases if you have to convert your subd back to a poly model for any reason?

--magilla

[RichSuchy]

1. on the channel bar or in the atrivute editor you can set the render resolution to 0.

I find it easier/faster. (I worked that way in Lightwave too) to just work with my cage and transform it to see how its smoothing periodically.

2. though you can model in proxy mode, I avoid it in general, because (answer is above under number 1)

3. See answer one above. BUT you can make a button and put it in your tool bar (or two buttons if its two hard to write mel you can just record the command to a button) to do the transforming.

4. Render tesselation and view tesselation are controllable from the attribute editor.

Note: though you can make a button to switch modes you may want to use smooth mesh. Going back and forth between sub-d and not sub-d tends to add history nodes that you need to delete in the hypergraph which accounts for many slowdowns for those that dont realize it.

Again, you would only use this to check smoothing.

Note 2.

people often keep maya tesselation (at render time too high) If you do not modify at level two, turn that off for render tesselation (set to base level ) and set tesselation as low as required.

[magilla]

i think you are confusing the term "render".

The render resolution is set at render time - it doesn't change - it renders the subd surface as smooth as it can - regardless of the "display" level you set. The render is your final output, ie "batch render"

For realtime display smoothing thare are two different concepts -

The "display level" (the one in your channel editor) merely shows you in the panel view what level you are editing at (from 0 to 13). You can leave your model at 0 and still render it completely smooth.

To confuse this further you also have 3 levels of " hardware smoothing" to play with - this does not affect your editing display level. So you can set your editing level to 13 and view at a "display smoothing" of 1.

again, this does not affect your final render.

confusing enough!?


--magilla

[RichSuchy]

Originally posted by magilla
i think you are confusing the term "render".

The render resolution is set at render time - it doesn't change - it renders the subd surface as smooth as it can - regardless of the "display" level you set. The render is your final output, ie "batch render"



--magilla

In renderman yes, In maya no. there are explicit tesselation settings for Sub-Ds. In maya you have depth and sample count. Depth is how many levels of tweeks to considder in the render, and sample count defines how many levels of subdivision to preform. Load up a cube turn it into a sub-d model and look for yourself by tweeking those settings.

In Renderman you have true Sub-D rendering and yes it makes it as smooth as needed for the camera.

[magilla]

I just did this with a sphere before I sent the last post.

I can render a perfectly smooth sphere with it set to 0, but I do see what you are saying.

If I switch the tesselation format to uniform I then have to select a depth at which to tesselate the subd......

does this affect the realtime model though?

aha
:surprised:


--magilla

[rock]

Looking back at my questions posed above, I think it's too long and unfocused. Therefore, I will rephrase and shorten it:

In Maya, I would like to model only at subdivision = 0 in two modes unsmoothed and smooth. The smooth version will look like the final version. Then when I am ready to render, I render the model at subdivision = 3 for greater smoothness.

So the question is can I do this without ever seeing a smoothed model at subdivision = 3 in my perspective view?

Thanks,
Rock

[roger]

2 ways to do this.
1) Model your poly model. It will be at subdivision 0. When your ready to render, turn it into a Maya sub-d. You can always undo the Maya sub-d....there are more than one undo in Maya. :)
2) Use Smooth Proxy. In the html help files it tells you how to turn off the visibility of the smooth mesh layer. You should keep subdivision levels low while modeling and increasing it just before rendering and the help files give you a little mel script to do this automatically for a given object.

[playmesumch00ns]

There's also my "shut the lightwavers up" script in the MEL scripts sticky for toggling between smooth/unsmooth which you can bind to a hotkey.

Subdivision surfaces ARE NOT POLYGONS. They model a lot like polygons, and you can change polygons into subdivs very easily, but they ARE NOT POLYGONS. Changing a polygon model into a subdivision surface IS NOT THE SAME as performing a poly smooth

Maya's render tesselation settings for subdivs do the same as for NURBS: they will try and give you as close to sub-pixel accuracy as possible without exploding the renderer. PRMan succeeds a lot better at this because it dices everything to sub-pixels anyway, but that's just cos RenderMan's lovely.

I don't know hat you mean by "level", but I assume you're talking about the "Display level". This has nothing to do with how smooth the subdiv is, if you render at level 0, it will still be perfectly smooth (yes, depending on your tesselation settings). The display level just lets you do heirachical edits: try and find steven stahlberg's subdiv rigging and modelling tutorials for a good explanation of why heirachical edits are so good.

[RichSuchy]

Originally posted by playmesumch00ns
There's also my "shut the lightwavers up" script in the MEL scripts sticky for toggling between smooth/unsmooth which you can bind to a hotkey.

Subdivision surfaces ARE NOT POLYGONS. They model a lot like polygons, and you can change polygons into subdivs very easily, but they ARE NOT POLYGONS. Changing a polygon model into a subdivision surface IS NOT THE SAME as performing a poly smooth

Maya's render tesselation settings for subdivs do the same as for NURBS: they will try and give you as close to sub-pixel accuracy as possible without exploding the renderer. PRMan succeeds a lot better at this because it dices everything to sub-pixels anyway, but that's just cos RenderMan's lovely.

I don't know hat you mean by "level", but I assume you're talking about the "Display level". This has nothing to do with how smooth the subdiv is, if you render at level 0, it will still be perfectly smooth (yes, depending on your tesselation settings). The display level just lets you do heirachical edits: try and find steven stahlberg's subdiv rigging and modelling tutorials for a good explanation of why heirachical edits are so good.

Sounds like a nifty script. :thumbsup:

Never said they were polygons... suggested that you could get a quick idea of how resultant sub-d would look while working in polygons and without inadvertantly ending up with extra invisable nodes to clean up. :)

I prefer to set it to uniform tesselation, for rendering.

Also when I refered to levels of tweeking, I meant increasing detail on the surface... you know, level 1 vertices, level 2 vertices, etc. You can set how deep the renderer will notice those higher level heirarchical tweaks.

[Nemoid]

I tried and i have to say that
modelling with polys and then
converting tham in sub-ds to see the results
is the best way.
undo conversion after
seeing results and keep onmodelling
proxy object make the perspect view too cluttered

and its not like in Lw that u can use current tools
in sub-d mode.

this is what i know at this moment.

[RichSuchy]

Originally posted by Nemoid
I tried and i have to say that
modelling with polys and then
converting tham in sub-ds to see the results
is the best way.
undo conversion after
seeing results and keep onmodelling
proxy object make the perspect view too cluttered

and its not like in Lw that u can use current tools
in sub-d mode.

this is what i know at this moment.

you can alwqays keep two perspective views open, one with sub-ds visable and one with polygons visable.

more work than I think its worth though

[playmesumch00ns]

Originally posted by Rich Suchy
Sounds like a nifty script. :thumbsup:

Never said they were polygons... suggested that you could get a quick idea of how resultant sub-d would look while working in polygons and without inadvertantly ending up with extra invisable nodes to clean up. :)

I prefer to set it to uniform tesselation, for rendering.

Also when I refered to levels of tweeking, I meant increasing detail on the surface... you know, level 1 vertices, level 2 vertices, etc. You can set how deep the renderer will notice those higher level heirarchical tweaks.

Wasn't having a dig at you mate:) I think I was replying to Rock - can't remember now, 'twas just general infos anyway, not meaning to correct anyone in particular:)

[RichSuchy]

Originally posted by playmesumch00ns
Wasn't having a dig at you mate:) I think I was replying to Rock - can't remember now, 'twas just general infos anyway, not meaning to correct anyone in particular:)

Oh, My mistake then ;)

I was more worried, my tired and hasty writing was misconstrued.

[Nemoid]

Originally posted by Rich Suchy
you can alwqays keep two perspective views open, one with sub-ds visable and one with polygons visable.

more work than I think its worth though

That's a nice tip!!!

hey , i have a question i made,about Maya's sub-D
but i rally don't understand till now.

when i refine some face,and i move them,
Maya subdividesalso regions near to those faces.
is there a way to control that easily?

i'm used to Lw sub-d wich doesn't make this kinda thing.
Thanx

[RichSuchy]

Originally posted by Nemoid
That's a nice tip!!!

hey , i have a question i made,about Maya's sub-D
but i rally don't understand till now.

when i refine some face,and i move them,
Maya subdividesalso regions near to those faces.
is there a way to control that easily?

i'm used to Lw sub-d wich doesn't make this kinda thing.
Thanx

I generally model for even subdivision... no level two tweeks... however I believe there is an option to kull all those >1 level non-adjusted verts.

[Nemoid]

Originally posted by Rich Suchy
I generally model for even subdivision... no level two tweeks... however I believe there is an option to kull all those >1 level non-adjusted verts.

yap, that's the way i model in Lw! and i like so much.

i hate allthose unwanted verts. i'll search that option!!

i was just trying these hierarchical Maya sub-ds
but i have to say that i have alot to
understand...
:rolleyes:

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.