Feedback about T-Splines 3 for Rhino

[Matt Sederberg]

Hi all,

T-Splines 3 for Rhino has been released for a month or so.

We've gotten a lot of helpful feedback on the release from a number of channels, but I just wanted to put out this general invitation for feedback - what have you really liked in T-Splines 3? Are there things you can do now that were painful or impossible before? What tweaks would you like to see? Which bugs do you want to be absolutely sure that we fix in 3.1 (which is due out in a few weeks)?

We're listening. And thanks for using our product.

Thanks,
Matt

[evert4130]

I watched the intro to tsplines3 webinar and did not get the same results from tsMatch as in the video. Maybe it's just me, but after messing around with it for a while I found that it only works if I rebuild the surface to degree 3. The u and v directions of the surface also seem to sometimes cause unexpected behaviour. Hope the pictures explains it better.

tsMatch2.JPG (33.05 KiB) Viewed 2418 times

[Matt Sederberg]

I watched the intro to tsplines3 webinar and did not get the same results from tsMatch as in the video. Maybe it's just me, but after messing around with it for a while I found that it only works if I rebuild the surface to degree 3. The u and v directions of the surface also seem to sometimes cause unexpected behaviour. Hope the pictures explains it better.

tsMatch.JPG

tsMatch2.JPG

Hi Evert,

Thanks for the response. Could you post the file that's giving you problems?

Thx,
Matt

[evert4130]

Hi Matt,
It is not really a problem, it is just that I found tsMatch does not always work as expected. It seems like it only works if the surface is degree 3 in the direction of the match tangency and the U and V directions must be right.

I simply tried what I saw in the webinar by using tsMatch to create a blend between a tspline and a extruded nurbs surface, but did not get the expected results. I was trying G2 tangency with refinement on, 0.001 units tolerance, 20 maxrefinements.

In all other cases tsMatch works exactly as expected and better.

Ps: In the tsMatch command, when choosing the "midsurface" option, the command line still asks for "border edge loop"

[laborda]

Hi all,

For me V3 is great ! I've been using the pipe command and pull a lot. The addition of UV maps support was something I've been waiting for a long time too. The improvements in match are more than I expected and make T-Splines more useful when presicion is needed.
Discover symmetry is also incredible.

I found 1 simple bug that you might know, and it's about the toggle smooth mode. Sometimes it refuces to work. I know that changing from the view mode, eg. compatible to fast or other, will make this work again.

Other bug is the draft angle anaysis with other modes than compatible. If you use fast mode and try to analyze, first the colours don't change even though you move some vertex, and if you close the anaysis function, the model will keep painted untill you change to compatible mode. Perhaps if it's dificult to integrate in fast modes, the analysis command can be left only for compatible mode. I found a workaround for this, you have to change in the TS object properties from use globlal settings, to the other option an apply. This forces a remesh of this specific object and the analysis updates.

I also found what Evert mentioned about normal directions in match command.

This is not a bug, is a suggestion. Can we have the option to copy with alt using the manipulator, as we can do dragging?

Other suggestion is that we need a way to smooth the mesh, the moveUVN command from rhino has a good interface and that could be great, but I don't know how easy could be to integrate such a feature.

2 final suggestions, in this case something I'm not sure if many people think this could be useful. First an option to make planar pipes and the final to add an option to project a tspline object to a surface. Something like the pull command but in only one direction, the pull command does a great job but it warps the points to the surface and sometimes you may need a planar pull. I think this could be added as an option to pull.

Again, I think T-Splines is getting every version better and this is related with one of the best communication with users I ever seen.

Thanks,

Joaquín

[knead]

I'm using V4 SR9 (Hotfix 1) and V5 in Windows 7 (64bit)

(1) Control Points size are NOT same size.
I WISH same size.

(2) Tangency handles are "ALWAYS" same color.
I WISH different color (Moved and not Moved).

(3) Command: _tsOptions Speed is very slow.
I WISH always quick speed.

(4) I don't Docking window Dragging.
I WISH Docking window Dragging.

(5) Speed Options icon color (with shadow) is all same color.
I WISH different color.

(6) T-Splines toolbar color is always Pink color.
I WISH toolbar of different color.

(7) "Detect concave corners when converting meshes"
Could you please attach a file for example?

(8) T-Splines Help file is only Image and text.
I WISH add "MOVIE" and "Toolbar icon location" etc... in Help file.

[Matt Sederberg]

Hi Matt,
It is not really a problem, it is just that I found tsMatch does not always work as expected. It seems like it only works if the surface is degree 3 in the direction of the match tangency and the U and V directions must be right.

I simply tried what I saw in the webinar by using tsMatch to create a blend between a tspline and a extruded nurbs surface, but did not get the expected results. I was trying G2 tangency with refinement on, 0.001 units tolerance, 20 maxrefinements.

In all other cases tsMatch works exactly as expected and better.

Ps: In the tsMatch command, when choosing the "midsurface" option, the command line still asks for "border edge loop"

Evert,

Thanks for this - it looks like we're not communicating correctly with Rhino on non-degree 3 surfaces. We'll look into this. Nice typo catch on the midsurface - we'll get that fixed.

[Matt Sederberg]

Thanks Knead! We'll review this internally.

I'm using V4 SR9 (Hotfix 1) and V5 in Windows 7 (64bit)

(1) Control Points size are NOT same size.
I WISH same size.

(2) Tangency handles are "ALWAYS" same color.
I WISH different color (Moved and not Moved).

(3) Command: _tsOptions Speed is very slow.
I WISH always quick speed.

(4) I don't Docking window Dragging.
I WISH Docking window Dragging.

(5) Speed Options icon color (with shadow) is all same color.
I WISH different color.

(6) T-Splines toolbar color is always Pink color.
I WISH toolbar of different color.

(7) "Detect concave corners when converting meshes"
Could you please attach a file for example?

(8) T-Splines Help file is only Image and text.
I WISH add "MOVIE" and "Toolbar icon location" etc... in Help file.

[knead]

(9) Found a problem with unselect object.
viewtopic.php?f=2&t=30237

[oliviertij]

Hello.

I found that v2 was already WONDERFUL, so words are now missing to qualify v3.
Congratulations, you should be proud of your work !
I'm not working with it, I'm playing

Laborda said : "The addition of UV maps support was something I've been waiting for a lont time too". I don't understand what it means, cans someone explain, please ?
I agree with him when he says : "First an option to make planar pipes and the final to add an option to project a tspline object to a surface. Something like the pull command but in only one direction, the pull command does a great job but it warps the points to the surface and sometimes you may need a planar pull. I think this could be added as an option to pull."

Matt said :"Which bugs do you want to be absolutely sure that we fix in 3.1 (which is due out in a few weeks)?" Well, maybe it's because of my graphic card, I don't know, but I often encounter problem with the manipulator : in translation, sometimes, the object is going the opposite direction I would like; and always, when I want to scale, it does anything : for example, I drag to make a little bit bigger, and the object goes so smaller it seems it disappeared . No problem with rotation.

And last thing, as I'm not speaking very good english, I'd like to know if there's a french forum better than another to post questions ? ( Or somebody as good as M Botha )

Have a nice week-end.

Olivier

[knead]

(10) Command: _Smooth does not operate.
I WISH Smooth function.

[laborda]

Hi Olivier, it seem I wrote extremely fast, is long time, not lont time... My post is edited now.

[oliviertij]

Thanks Laborda, and excuse-me, I didn't even pay attention to the spelling, I read "long", even the first, but I simply don't understand what it involves, that "addition of uv maps support".

[knead]

(11) For example: http://www.youtube.com/watch?v=Eti7ucZtrI0
I WISH Bevel

[Chris Botha]

(10) Command: _Smooth does not operate.
I WISH Smooth function.

I am not sure if we after the same thing but a "rebuild" command would be nice. grab pinched up edge/edgeloop and relax it.

the closest way I can explain this is by doing a surface rebuild in rhino to a similar number of points or less and with even spacing. With tsplines it seems rational that you can grab jsut an edge, or a series of edges and "rebuild" them to relax areas that become pinched up from tweaking.

A "copy shape" command would be nice too. draw a shape (planer I would assume), lets use for aguments sake an ellipse, then grab an edgeloop on your model and "copy" the shape of the curve to the edge loop. I relalise that the implimentation would be hard but it would be an extremely usefull tool.

If a brigde command is going to break symmetry it should insert a middle section to maintain it. This trips me up all the time.

Luvvin the tsplines

regards
Chris

[Adam Helps]

A "copy shape" command would be nice too. draw a shape (planer I would assume), lets use for aguments sake an ellipse, then grab an edgeloop on your model and "copy" the shape of the curve to the edge loop. I relalise that the implimentation would be hard but it would be an extremely usefull tool.

Sounds kind of like tsMatch on an interior isoparm? Have you tried that?

[Chris Botha]

no i have not, i will now tho

Did not know you could do that..

never assume huh,

i thought tsmatch only worked from an exterior edge to another exterior edge or curve.

[edit] ok tested.. it works.. kind of.. (thats what the midsurface thing was hehehe)
but its a bit crazy this test resulted in the iso fitting, pretty well actually, but it destroyed the rest of the shape.. but i see how it works.. and yes that was what I was asking about, but with a more predictable result maybe? hmm how about an option to average/blend the neighbouring isoparms too so its more flowy?

is the result more predictable if you use XXX? number of points in the target curve?

Should I be reading the manual at last?

[paul]

Hi Chris --
I wrote the match command. Without going into too much detail, when matching an interior isoparm to a curve, control points from the neighboring two isoparms are also used to find a solution (to be more precise, all control points whose support crosses the iso-line being matched are involved in the computation). In your example, this would move pretty much all of the control points on the surface, as you observed. To make the effect more localized, you need to do some exact refinement to produce at least two additional isoparms which are "parallel" (one to either side) to the one you wish to match with. The shape of all three isoparms will be modified, but this should prevent the effects from propagating around the entire model.
Hope this helps

[Adam Helps]

It's a good idea to get the surface "close" to the target curve by hand; this will reduce the distance the control points have to move, and hopefully make the result a bit more predictable & controllable. It'll hit the target curve either way, but you get a lot more distortion from your intended shape if it has to go a long way. Paul's observation about inserting some extra curves to localize the motion is also a great idea.

Reading up on the match command is a good idea . It's pretty complex.

[Chris Botha]

Thanks Guys, I am going to RTFM

[Algus]

Hi Matt,

T-Spline V3 is a great software, a very important instrument for my work.
I appreciated especially the new editor mode and the new pipe command.
What I wish, if possible, is to have a command that lets me, (if and when I need at the end of a project), to transform an object containing t-points in another object without them.
The reason is that I could need to export a BOXED T-Splines object (in obj) to a software (as mudbox, for example) to add fast a number of details.
(The follwing second step is to import the file again in T-Splines as obj and getting a final nurbs surface).

Al

[laurentheid]

Hi,
Love V3
when moving or scaling a t.spline behind or into an other object could it be possible not to disappear but switch to Xray so to remain visible like rhino does?
Thanks.

[Joe]

What I wish, if possible, is to have a command that lets me, (if and when I need at the end of a project), to transform an object containing t-points in another object without them.

Algus, are you looking for a convert to subd command? It's on the wishlist . In the meantime, you will need to use the tsInsertPointExact to extend the T-points.

when moving or scaling a t.spline behind or into an other object could it be possible not to disappear but switch to Xray so to remain visible like rhino does?

Laurentheid, that's a good suggestion. I'll put it on the wishlist.

[Algus]

Hi Joe,

You're right: sometime I need to convert the T-spline file to subd.
I'm still using tsInsertPointExact, but it isn't always a fast procedure .

Al

[Adam Helps]

I simply tried what I saw in the webinar by using tsMatch to create a blend between a tspline and a extruded nurbs surface, but did not get the expected results. I was trying G2 tangency with refinement on, 0.001 units tolerance, 20 maxrefinements.

Hello Evert,

We just fixed this bug today. Thank you for reporting it! The tsMatch command in 3.1 should be more reliable.

[evert4130]

Thanks Adam, I'll give it a try.