what are differences between t point and star pont in form modelling

what are differences between t point and star pont in form modelling

dawnny68
Participant Participant
700 Views
3 Replies
Message 1 of 4

what are differences between t point and star pont in form modelling

dawnny68
Participant
Participant

Hi there,

 

I'd like to know what does t point and star pont mean exactly.

 

Also want to know differences between both two fucntions.

 

thanks in advance.

0 Likes
Accepted solutions (1)
701 Views
3 Replies
Replies (3)
Message 2 of 4

laughingcreek
Mentor
Mentor
Accepted solution

good question. possibly the most important thing to understand to use t-splines successfully.

it's all about edge/loop control.  you want your edges to form complete "loops", and you want to control the flow of these loops around your part.

 

a "complete loop" either flows continuously around a part if closed, (or from edge to edge if just a plane)-

laughingcreek_4-1698337239245.png

laughingcreek_5-1698337354347.png

 

OR

it should flow from star point to star point (or from a star point to the edge if a plane)

 

 

laughingcreek_0-1698337141306.png

laughingcreek_6-1698337436776.png

the star point is used to change the direction of a loop so yo can get smooth transitions (it's hard to see, but the edge along the right and bottom is highlighted yellow and forms a continuous loop)-

laughingcreek_7-1698337540968.png

example on a solid-

laughingcreek_8-1698337617063.png

 

 

a T point allows you to have only part of a loop visible and available for control.  under the hood t-splines will complete the loop and control the edges positioning so the result is the same as it would be if there weren't edges there at all.  this allows you to add detail to an area, without having to control the edge all around the part manually.  bellow the t-point is circled.  The dotted yellow line represents the edge that t-splines is controlling for you.  you can use the little bit of edge to control the shape at that area without having to make adjustments to the rest of the loop to make things look ok-

laughingcreek_9-1698337722954.png

as a rule of thumb, you only want to use a t point where it can flow continuously around the part (or edge to edge on a plane.) the first example above.  You don't want to use a t point where the loop flows to a star point.

 

you also generally don't want to use the "crease" function on loops that flow to star points.

 

hope that helps.

Message 3 of 4

adam.helps
Autodesk
Autodesk

You can think of spline surfaces (NURBS, subds, or T-splines) as being made of a network of curves (splines) woven together in a grid. It's similar to the way fine fabric is created by weaving thousands of threads together in a grid pattern.

 

The spline curves themselves are a series of "perfectly smooth" regions with "breaks" between the regions. For many surfaces, the smoothness becomes "only" C2 at those breaks. The edges on the surface tell you where the breaks are. Explaining this in detail is difficult, but if you have a little patience, this video explains it beautifully:

https://www.youtube.com/watch?v=jvPPXbo87ds

 

The difference between star points and T-points is the way that their continuity breaks behave. You can think of areas where the fabric has been cut at the corners and glued together; the woven pattern "stops" at the star point. This can create little lumps in the surface. You can see a video explanation of this behavior here:

https://youtu.be/ckOTl2GcS-E?si=IrESxouTihmMxc_T&t=217

 

On NURBS and subds, the only way to have the splines continue smoothly through the surface is for you to make all of your vertices have four edges. T-junctions are a little different; they behave like there are four edges, even though there are only three. The spline curve ("thread") keeps going straight through the T-junction, as if there were an invisible edge that continued out just a little further. The continuity break caused by that invisible edge fades out to nothing within two faces, and beyond that the surface is perfectly smooth. This makes T-junctions suitable for isolating detail.

This also effects the definition of a "quad" on a T-spline surface. T-junctions don't count as corners, so it's possible for a face to behave like a quad even when it has more than four vertices; only the corner vertices matter. You can have as many T-points as you want and as long as there are only four corners, it's still going to behave like a quad.

Message 4 of 4

glenn-chun
Autodesk
Autodesk

Here's a clarification of @adam.helps's comment "The continuity break caused by that invisible edge fades out to nothing within two faces", using the last T-spline surface posted by @laughingcreek.

t-point.png



Glenn Chun
Sr. Principal Engineer