Jump to content
Sign in to follow this  
dayone

Indents in Mouse

Recommended Posts

LOL, feeling yer pain , if it does come to cinema Im sure it won't be retrofitted to R18 so will be a while before I get too use it. I will investigate the other sud divs tho on an upcoming project.

 

Deck

Share this post


Link to post

OK, here's what we could have weighted in the final model...

 

The problem: In the reference, the end of the button split is totally straight and sharp. In order to achieve that with control loops I would have to added 2 additional loops (4 including the symmetry), both of which risk disrupting my curvature in both directions (!), even to the extent to which it has been already established by subdividing it to L1. So my initial solution ? Leave it rounded ! It's a minisule detail you just can't see from a distance - and nobody is going to complain about that if they look at my pictures, because it's too small to see.

 

But doing my best to aim for perfection, we should really nail that exactly as it is in real life, and this is where edge weighting can leap to the rescue...

 

Here is what we currently have with no weighting at all - a nicely pleasing and well supported, but actually incorrect rounded end to that split.

 

1022941096_Noweights.thumb.jpg.8e87ab3a948987e927739bbfe3de6c4f.jpg

 

Ignore my edge selection there, that is irrelevant ! :)

 

Now, let's look at the underlying topology of that and change it so the end is more straight, like so...

 

1108082040_Weightthese.thumb.jpg.9cdbbd037b987d9efcf92804ae3b50ad.jpg

 

If we activate SDS now, we will still get the rounding, even though it looks straight without it, because that sharpness is unsupported.

 

However, If we SDS weight these edges to 100% we can sharpen the hell out of that without having to add any more control loops...

(note we already have some, so adding weighting now is actually a hybrid method !) In fact the presence of existing proximal control edges is what is going to make all the difference as to quite how 'Pro' our final weighted surfaces look ! It is rare that you can subsist on weighting alone.

 

Here is our L4 subdivided result using Cinema's Catmull Clark SDS algorithm, so not even OSD this time ! Here it makes no difference but on lower res geometry it becomes apparent that the OSD catmull Clark is actually slightly different in its rounding characteristics, and about 30% superior if we're really being honest :)

 

Top Tip: whenever using edge or point weighting, use one more level of subdivision than normal. Look how sharp it is ! Almost like there was no SDS there ! But of course we can tell it is there from the curves in the surrounding lines.

 

2045487905_CatmullL4.thumb.jpg.b342c5d852389f939315cf6c9d3b949c.jpg

 

and there is zero perceivable effect on the mesh / curvature, as we can see in this more distant shot...

THAT is what the makes the choice to use it here a valid one. If it caused ANY pinching or other surface artefacts that would be the reason to kick it to the kerb and do it properly with traditional control loops.

 

76460858_Nodisturbanceintheforce.thumb.jpg.0cfe410bcf44418479e65c7398548c93.jpg

 

We could also do it to the button cutout, but those edges are rounded like that in the reference so no need. But the rounding wasn't quite tight enough, so in this instance it made more sense to simply bevel an edge loop running across the model to give me 2 additional inner control loops that also preserved the curvature of that surface.

 

But there you go - a legitimate use for edge weighting, and we've got away with it because we had already subdivided the base mesh once, and because we already had some control loops doing half the work. Had we tried that on the low poly original the disturbance to the force (and our curvature) would have been catastrophically noticeable as surface shading errors.

 

Hope that helps explain where weighting has a place in a traditional modelling workflow...

 

Actually in preparing this, I have hit on something of a personal revelation ! I didn't use OpenSubDiv for this ! I used bog standard Catmull Clark, and the great news about that is that not only do I have a client-worthy result using edge weighting, but I actually COULD UV  map this wihtout problems, because Cinemas own SDS doesn't have the UV distortion issue !! So maybe it's not quite as urgent as I thought to get it working with UVs in OSD, which is the bit that's broken at the moment.

 

CBR

 

Share this post


Link to post

Thank you so much cerbera I will have a go at the weekend.

Share this post


Link to post

You're welcome - good luck !

 

CBR

Share this post


Link to post

ok sooo got the first bit right but how the hek do I make the second bit?... I'm doing all this free style no books no tuts just me.

 

Any help much appreciated.. sorry I keep asking   

Share this post


Link to post
22 hours ago, dayone said:

how the hek do I make the second bit?

Select the border loop at the base of the top you already made, deselect the edges you don't need at the back / bottom of the model, then extrude that whole row of edges out from the main form once (to make the white rim bit) , then select that ring of new polys and choose split from the r-click menu to create the first poly row of the middle section. Don't forget to delete those polys from the original model - they don't automatically get deleted when you do the split...

 

Once you have that loop isolated in its own object, adjust its outside points to refine that rim, then extrude down another row to begin the side of the mouse, and carry on from there, using my topology pics earlier to help you place your additional polys and subtly change that edge flow so it goes round the back corner nicely...

 

If you struggle with that it might be a good idea to go back to that post where I showed you the base section, which contained a side view you could load as a viewport background for your model; then you could see pretty much exactly where each poly goes and line yours up over it...

 

When you start working with split objects like this the most important thing to watch out for is that you don't move any of the points at the seam lines, where the models join, although you can put them back using vertex snap if things go astray, it is better to take extra care not to move them in the first place ! That way you should retain perfect seams when you eventually add inner thickness. Btw, don't do that until you have all 3 sections perfect - far better to to be working on a single surface while are you are still refining forms...

 

CBR

Share this post


Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

Latest Topics

Latest Comments

JOBS / BUY & SELL

×
×
  • Create New...