Public Beta: Trestle & Scrim

[Vizfizz]

How about birail? We plan it as one of next 2 modeling plug-ins. NOTE: up to now it’s just a plan/idea and nothing more. The birail is a “specialized lofting”, it uses 2 rails paths and 1 or more curves between them. See more details in big maya’s handbook. How do you estimate birail’s perspectives/usability? What should be a set of options? There is enough time to think and discuss all this.

I see birailing as a hybrid of lofting and a modified path extrusion. I’m all for creating a birail module for this new modeling family. Maya offers a Birail 1, 2, & 3 option. It uses two rail curves and either 1, 2, or 3 profile curves to define the surface. However I don’t believe it works with a closed profile and Maya’s implementation requires the profile curve must directly touch the rail curve in order to work. However, birailing can achieve forms that mere lofting can not.

As modeling advances within this family of plugins… one of the critical issues I can see coming to the forefront is getting specific trestle points of one instance to properly align with other trestle points of another instance. Perhaps some additional Trestle alignment tools are in order.

[Igors]

Hi, Brian

As modeling advances within this family of plugins… one of the critical issues I can see coming to the forefront is getting specific trestle points of one instance to properly with other trestle points of another instance. Perhaps some additional Trestle alignment tools are in order.

Fully understood (in context of patches Uwe said). But IMO it has no valuable effect if it’s not together with combined splines. We think it deserves further investigation/thinking/analyze - IMO there is too few material to make conclusions.

I see birailing as a hybrid of lofting and a modified path extrusion. I’m all for creating a birail module for this new modeling family. Maya offers a Birail 1, 2, & 3 option. It uses two rail curves and either 1, 2, or 3 profile curves to define the surface. However I don’t believe it works with a closed profile and Maya’s implementation requires the profile curve must directly touch the rail curve in order to work. However, birailing can achieve forms that mere lofting can not.

We showed all maya’s Birail 1-2-3 options to professional maths. They said like: “that’s very natural and there is nothing redundant and strange as you (Igors) see. The theory of curves operates with very specific terms, so a some simplification is necessary”. Hmm… we are still confused cause we guess such numerous sub-options would be not intuitive for normal user.

[bronco]

As modeling advances within this family of plugins… one of the critical issues I can see coming to the forefront is getting specific trestle points of one instance to properly align with other trestle points of another instance. Perhaps some additional Trestle alignment tools are in order.

i totally agree. power is nothing without control.
one of the main reasons i bought up the non-modal window discussion was, that it is difficult to line up different Trestle/Scrim instances without visual feedback. i don’t know if there is a real easy way to do that. “line up control point 3 of Trestle_A with control point 28 of Trestle_B” seems pretty difficult to me.

[Igors]

Hi, Uwe

i simply don’t have the time for another photo-real project right now. these tests are fast and i have some fun doing them.

Hehe, don’t worry, this kind of surrealism is fully understandable for us But let us ask you: ok, you’ve found an usable aspect, really usable. But how to make this simple and intuitive for user? Or he should repeat a whole way you’ve done? (note: don’t hope too much on documentation/instructions, an artist relies on his intuition first). And the second question: how to control such shapes? Maybe a some example with animation? Of course, all these problems are ours, not yours (we are just moaning how hard our work is ). Maybe some utility plug-in?

[bronco]

Hi, UweHehe, don’t worry, this kind of surrealism is fully understandable for us But let us ask you: ok, you’ve found an usable aspect, really usable. But how to make this simple and intuitive for user? Or he should repeat a whole way you’ve done? (note: don’t hope too much on documentation/instructions, an artist relies on his intuition first). And the second question: how to control such shapes? Maybe a some example with animation? Of course, all these problems are ours, not yours (we are just moaning how hard our work is ). Maybe some utility plug-in?

it is simple and intuitive right now. at least this example is. scrim eats everything that is a wire. nice and simple. if the user doesn’t read documentation (just like me ), than let him explore the possibilities on his on. that’s the easy way.

if you are willing to help the user and build some form of utility plugin, then i would love to see the z-projection thing we discussed.
for every other idea i have right now we would need some form of vertice/edge/polygon selection of existing geometrie.

[Vizfizz]

i totally agree. power is nothing without control.
one of the main reasons i bought up the non-modal window discussion was, that it is difficult to line up different Trestle/Scrim instances without visual feedback. i don’t know if there is a real easy way to do that. “line up control point 3 of Trestle_A with control point 28 of Trestle_B” seems pretty difficult to me.

Perhaps some kind of secondary utility/align plugin would be a good idea. Designed to be point sensitive for Trestle users… but could be used for any other objects in the scene. We really don’t have good object alignment tools in EI.

Edit: For example… just having the ability to select 10 different Trestle instances and parent them to an alignment plugin and select “Align Midpoint to Horizontal or Vertical Plane” would be great. The plugin would read the local center of all the Trestle instances and match accordingly…but point to point alignment… now that would be handy.

[Igors]

Hi, Uwe, Brian

it is simple and intuitive right now. at least this example is. scrim eats everything that is a wire. nice and simple. if the user doesn’t read documentation (just like me ), than let him explore the possibilities on his on. that’s the easy way.

Ok, clear :sad:

for every other idea i have right now we would need some form of vertice/edge/polygon selection of existing geometrie.

That’s not available in our host.

For example… just having the ability to select 10 different Trestle instances and parent them to an alignment plugin and select “Align Midpoint to Horizontal or Vertical Plane” would be great. The plugin would read the local center of all the Trestle instances and match accordingly…but point to point alignment… now that would be handy.

Hmmm… enough “expensive pleasure” to build a hierarchy (and write a plug-in) only to perfrom alignment, isn’t it?

We really don’t have good object alignment tools in EI.

Same as we really don’t have many and many other things. But please agree: it’s a bit side away from model plug-ins family.

[bronco]

it is simple and intuitive right now. at least this example is. scrim eats everything that is a wire. nice and simple. if the user doesn’t read documentation (just like me ), than let him explore the possibilities on his on. that’s the easy way.

at least that’s the way i see it. but, we have some more beta-testers here. what are they thinking? maybe you have some idea what else is needed for trestle and scrim.

That’s not available in our host.

i know. it’s too bad.

But please agree: it’s a bit side away from model plug-ins family.

here i have to disagree. alignment and snapping, or all these control functions for that matter, are crucial for modeling. way more than for animation.
if you want to build a real world object you have to deal with measures and proportions.

[halfworld]

a little something for the graphic designer among us. i am not, but i hope you get the picture:

It has practical uses… Your images reminded me of the Model 172…

I got pretty close with my first attempt, will post an image later.
Ian

[bronco]

It has practical uses… Your images reminded me of the Model 172…

I got pretty close with my first attempt, will post an image later.
Ian

hehe, i don’t know model 172, but Ikea has something pretty similar.
now, i rendered some simple moving forms out of this overnight and what do you think? right, there’s a problem.
shadows seem to like to dance with this setup:

uberScrim MOVIE

uberScrim PROJECT FILE

i tried a little sequence without GI, just to eliminate the possibility of GI flicker, but the shadows keep dancing.

[Igors]

Hi, Uwe

here i have to disagree. alignment and snapping, or all these control functions for that matter, are crucial for modeling. way more than for animation.
if you want to build a real world object you have to deal with measures and proportions.

True, but it’s not a task of concrete modeling tool/module like Scrim

uberScrim PROJECT FILE

i tried a little sequence without GI, just to eliminate the possibility of GI flicker, but the shadows keep dancing.

Looking into this problem… Again: it’s a very good style to post prj. Keep up it.

[Igors]

Hi, Uwe

now, i rendered some simple moving forms out of this overnight and what do you think? right, there’s a problem.
shadows seem to like to dance with this setup:

You’ve not enough “U” vertices, thus deforms produce jittering. Simply set U Adaptive Angle = -1 (create extra U vertices always) and U steps like 5

[halfworld]

I’ve given up, this looks more like an Ikea knock off then the real thing but something like this… or that…

Okay, back to the Moai.
Ian

[bronco]

little question regarding UVs in Trestle.
i can see why only U makes sense when dealing with a wire, but i don’t see why it shouldn’t be possible to generate a real UV (or UVW) space for polygon output.

on this subject: i can’t get to show up a texture when rendered on a trestle polygon. in animator it shows up just fine.

SIGN PROJECT FILE

[Igors]

Hi, Uwe

little question regarding UVs in Trestle.
i can see why only U makes sense when dealing with a wire, but i don’t see why it shouldn’t be possible to generate a real UV (or UVW) space for polygon output.

Possible, but such UVs would be exactly what default flat mapping (UV = OFF) does - not a big feature

on this subject: i can’t get to show up a texture when rendered on a trestle polygon. in animator it shows up just fine.

SIGN PROJECT FILE

Common problem. For plug-in group set “use UV space” = ON and then check/uncheck “Disable UV” in Texture Window to switch UV’s on/off (last section in Scrim draft doc).

A plug-in can enable or disable the usage of UV coordinates, but… not just follow by “use UV space”, the host doesn’t pass this flag to Camera correctly :sad: An additional checkbox “generate UVs” (in plug-in’s interface) would be redundant/confused cause it does exactly what “use UV space” should.

BTW: new utility plug-in comes tomorrow…

[bronco]

Possible, but such UVs would be exactly what default flat mapping (UV = OFF) does - not a big feature

i guess there will be differences when the control points are animated - maybe not a big feature, but a little

thanks for the further info. and i am looking forward to tomorrow.

[bronco]

my little observation about performance:

one word: great!

my car project is now well into the 500 + plugins realm with 650k polys (not optimiced for performance so far).
first slowdowns were noticable at about 400 plugin instances.
adding additional plugins now takes about 1-2 seconds, which isn’t to bad.
when moving the cam in the viewports i can see the plugins do their work for a fraction of seconds, again nothing you wouldn’t expect from many many plugins working together.
overall i have to say trestle and scrim are really a joy to work with. great job, igors!

…now i just need an idea how the front and back lights should look like…

[Vizfizz]

Impressive Uwe…

Obviously as geometry was finalized, exports of various portions of the car would be exported out and reimported in to ease the plugin overhead on Animator. Although the number of plugs your pushing is pretty impressive. For true modeling purposes, I’d doubt you would keep things in plugin form unless you were planning on morphing the car’s shape by animating control points…

As more modules come available for this modeling family… the easier this will get.

[bronco]

sure, if this would be a real project i would have exported at least the tires and rims (each tire & rim combo = 66 plugins). most of the time i turn them simply off.
but for testing reasons i thought it would be a good idea to try to push the limits a little.
and i wanted to see how well this performs.

if i compare it to 3ds max (worked with it for some time) i would say the modeling performance wouldn’t be better, really.
(with no modifier stack collapsed, taking the same modeling approach)

[Igors]

Hi, Uwe

my car project is now well into the 500 + plugins realm with 650k polys

Well, what to say? We are “interested party” and our opinion cannot be objective. But we want to say: we like your car. Thank you. BTW: 600-800K of polys is quite “standard” size for car model.

if i compare it to 3ds max (worked with it for some time) i would say the modeling performance wouldn’t be better, really. (with no modifier stack collapsed, taking the same modeling approach)

3ds has a modifier called “Lofter” but in our terms it’s “extrusion along path” right? 3ds never was a champion of speed/safety, but it has a lot of usable modeling tools that EI has not. Hmm… yet