Softimage to Houdini - Pros and Cons - What could be done?

   44884   63   3
User Avatar
Member
4189 posts
Joined: June 2012
Offline
pezetko
I cannot see reason to have such black boxed SuperEditSOP node.
It's against Houdini procedural nature. Instead of this blackboxed node you can use other application as well.

To consider is that Nuke is procedural too and has a similar tool. It adds to the workflow.
User Avatar
Member
4729 posts
Joined: Feb. 2012
Offline
pezetko
I cannot see reason to have such black boxed SuperEditSOP node.
It's against Houdini procedural nature. Instead of this blackboxed node you can use other application as well.

I am with you on this. The modeling problems are not because Houdini doesn't have a SuperEditSOP black box node but because the modeling workflow is not refined and the viewport interaction and control is limited. So it's not that Houdini doesn't have Slide Edges feature in a SuperEditSOP but that it doesn't have a Slide Edges SOP at all. Once you have these ops, then it's only the workflow and interaction that stands in the way, nothing else.

Also didn't Houdini move away from a black box modeling approach in the past?
Senior FX TD @ Industrial Light & Magic
Get to the NEXT level in Houdini & VEX with Pragmatic VEX! [www.pragmatic-vfx.com]

youtube.com/@pragmaticvfx | patreon.com/animatrix | pragmaticvfx.gumroad.com
User Avatar
Member
37 posts
Joined: July 2011
Offline
pezetko
I cannot see reason to have such black boxed SuperEditSOP node.
It's against Houdini procedural nature. Instead of this blackboxed node you can use other application as well.

You can you whatever modeling package you like as long as it has export feature. Then you can import obj/alembic/fbx or other file format directly to Houdini and you will get this one node modeling solution.

You can use for modeling what you like. Zbrush, 3DCoat, Silo, NVil, Modo, XSI…

It's even possible to have direct connection between modeling app and Houdini. Look at this:
https://vimeo.com/14448626 [vimeo.com]

Or some sort of clipboard (3D Applink, copy paste feature like in Nvil)


On the other side some better more interactive workflow for modeling would be plus. Something like this would be really useful:
http://www.youtube.com/watch?v=N2YprHoEWgE [youtube.com]
(Interactive modeling mode that create classic procedural nodes under it's asset, so you don't loose the possibility to modify history of operations).

I think a RMB menu based selection type (edge, primitive, points) will be a good starting point. This RMB will give all modeling tools(node) for edge, primitive or points. Some tools will be only a node configuration like polyextrude which can do extrude along normal, extrude along axis,.. But “Polyextrude node” should be compatible with points and edges.
Edited by - March 6, 2014 18:39:37

Attachments:
RMB_component_menu.jpg (370.1 KB)

User Avatar
Member
4189 posts
Joined: June 2012
Offline
pusat
I am with you on this. The modeling problems are not because Houdini doesn't have a SuperEditSOP black box node but because the modeling workflow is not refined and the viewport interaction and control is limited. So it's not that Houdini doesn't have Slide Edges feature in a SuperEditSOP but that it doesn't have a Slide Edges SOP at all. Once you have these ops, then it's only the workflow and interaction that stands in the way, nothing else.

This is a good point!

pusat
Also didn't Houdini move away from a black box modeling approach in the past?

Yeah but times and modelling competition has changed so the question is does Houdini have to change?
User Avatar
Member
207 posts
Joined:
Offline
tjeeds
Do you mean getting this data in VOPs? The PolyFrame SOP calculates these guys (in a variety of ways) and stores them as attributes that you can pull into VOPs.

Hi Jesse,

Well the PolyFrame SOP kind of does that, but (talking about curves here) what seems odd is the way that the resulting N or tangentv doesn't seem square with the tangentu (see attached… I *think* I tried a bunch of settings for it, 1st edge, etc), such that if you tell it to orthonormalize the result then the tangentu isn't even tangent to the curve anymore.

Polymeshes seem to be pretty stable since the N (and first edge) is well-defined, but in the case of curves you only have the tangentu, maybe it would seem to make sense to force the curve's N (and tangentv) to conform to tangentu instead of the other way around.

Also I'm not sure why the tangentv seems identical to the normal (documentation says it should be perpendicular to both tangent u and N?)… So while I can take the cross product or do gram-schmidt myself, just having clean basis vectors to begin with would be really cool.

There was also a compound on Orbolt called “tangents” I saw that seems to produce the cleanest results (to me, anyway…), basically it path-animates an “L” shape along the curve and extracts the vectors off of that shape.
http://www.sidefx.com/exchange/info.php?fileid=90&versionid=90 [sidefx.com]

Still there is plenty to like… very cool that you can use the gradient of texture UVs or any attribute to produce a set of basis vectors (like an XSI “tangent map” which was useful, but came with a significant performance penalty).

cheers,

-T

Attachments:
tmp_PolyframeVectors.hip (73.2 KB)

User Avatar
Member
207 posts
Joined:
Offline
pezetko
You can, you just have to convert that matrix to quaternion and bind it to orient attribute. Then in the right menu of the viewport panel there is Display particle origin button. Look at attachment.

That's a cool trick, I'll have to remember that. Thanks!
(though I do wish there was a native matrix type for custom stuff…)

pezetko
takita
The other thing I really really miss from SI is getting stable solid orthonormal basis vectors from curves and polymesh faces for free.
Polyframe SOP

(see answer to previous post…) it seems to kind of work for curves but I think it could be cleaner and better (for meshes it seems pretty good except for the fact that tangentv and N are nearly identical and not perpendicular, the last vector I would have to do myself… I know it's lazy of me…).

I also wouldn't even mind if the result of the PolyFrame SOP were an m3 (even lazier, I know).

cheers,

-T
User Avatar
Member
53 posts
Joined: July 2013
Offline
hey katana is there anyway to shrink that image or make it a thumbnail. its killing this forum reading ability for me on this page
https://tekano.artstation.com/ [tekano.artstation.com]
User Avatar
Member
392 posts
Joined: Nov. 2008
Offline
On the other side of laziness, you have to do it only once, make otl (Houdini Digital Asset) and then you can reuse it forever.

For curves, there is edge_dir (edge force) trick. Look at scene how to get good orthogonal frame for curves.

Matrix custom display type for visualisation of custom frame is neat idea.

Attachments:
edge_dir_trick.hipnc (101.1 KB)

User Avatar
Member
207 posts
Joined:
Offline
pezetko
On the other side of laziness, you have to do it only once, make otl (Houdini Digital Asset) and then you can reuse it forever.

For curves, there is edge_dir (edge force) trick. Look at scene how to get good orthogonal frame for curves.

Oh neat, using “Edge Force” is a cool tip, I did not know that. Thanks!

And yes, I did build an OTL to construct them manually… although given a choice I would be very happy to be able to throw that away. Less junk, you know?

Thanks again,

-T
User Avatar
Member
392 posts
Joined: Nov. 2008
Offline
takita
Eventually I found the “~” key and figured out the “enter-key-to-commit” thing, but I certainly missed the simplicity of the old SI “add/remove to cluster” - I'm still not entirely sure what the tilde key really does or if there's a gui equivalent to it, although I did see what command that hotkey maps to…
Give a “~” key (or backtick `) a little time, it's nice tool but take a a little time to get used to it.
I'm also used to right click menu option to change cluster selection when you select new components and then deside what to do. Tilde works just in oposite order. First you deside to change selection (hit ~) than you select new components and confirm.

takita
And while I know that a group can be generated by any operator, and that it's maybe not as simple as putting in a or icon in the active toolbar, I'm pretty sure there's gotta be a good design solution for that somehow
It shouldn't be that hard to create a shelf script for that.

takita
In general, I think the component selection tools themselves could also benefit from some love… the tools I've missed are “select parallel edge loop”, “select border components” (it's there, but maybe broken for prims in 13.0.314?), “select border edges” (like open mesh edges, the kind you'd want to cap in order to close off a mesh), and while the existing edge loop tool “works” it's still a little funky sometimes.
For loop/ring selection there is tool on orbolt (not 100% sure about that)
For border components you need 2 groups, in first select components in viewport. In second one with same name and Entity: Edges you had to uncheck Enable on Number tab and check Enable and Unshared Edges on Edges tab and change Merge Operation to Intersect with existing.

What Houdini needs in viewport component interaction are (Custom) Selection Filters like in the XSI.

takita
In SI proportional modeling was a big, huge timesaver. and while softPeak and softTransform are really nice, it would be even slicker if I could then drop down a weightmap and paint the effect back down without leaving the 3d viewport (is there an existing way to do this without doing it manually?).
I don't know if there is build in way but you can always build your own simple HDA tool that combines softTransform, Paint and some blend (VopSop/Wrangler). It's very easy.

takita
Another minor annoyance is (transient) windows getting lost underneath the main viewport window. So if I'm browsing for a SOP path in a parameter field or doing that “edit parameter interface” thing and I make an errant click somewhere, then I have to go looking for the window I just opened which is now lost beneath the main UI. Anyway it would be nice if windows with “accept” or “cancel” just stayed on top until I was done (like the help window).
Actually that's nice thing that all windows aren't modal. You can work with that windows in behind and when you need it bring it into foreground. So you will not get your workspace cluttered. I really like that some windows aren't modal and I can make changes instead of being forced to click Ok/Cancle and reopen that windows later.

takita
Oh and passthrough (or “dot” like in Nuke) nodes would really help with node graph legibility and accidents when “drop on a wire” is active (which I like, but is really dangerous for big crazy graphs). Really miss those passthroughs…
There is Null node that work same, just passing data. It's bigger than Nuke dot node, but on other side, consistency is plus.


takita
And yes, I did build an OTL to construct them manually… although given a choice I would be very happy to be able to throw that away. Less junk, you know?
Actually it's more junk but just under the carpet. Heavier scene, slower processing. I think it's big plus to be able to put in only what you really need. Offten there is no need for orthogonal frame on the curves and so on for other tools. It's nice to have the power to be so close to the metal.

It's nice from start to have everything-can-do tools but in the end you want speed in the scene and it's nice to have option to remove something, or even not to add it in the first place.

For example fluid solvers. It's nice to have pyrosolver, but if you need just some basic smoke and you need it fast and iterate quickly on versions it's better to keep just necessary functionality (like smoke solver). You get better memory management, solving speed and you have more time to be creative.

If you got blackbox that does all the stuff like fire, fluid, smoke and create bunch of data just in the case it doesn't help and you eventually hit the hardware roof sooner.

I think it's better to have the possibility to build own tools exactly to workflow that you need then use all-can-do tools that you cannot strip down to be faster.
User Avatar
Member
17 posts
Joined: July 2006
Offline
jordibares
As an industry you can see lots of companies are finally embracing it and there is more demand than ever for all sorts of positions, if anything, it is the FX side that is the lesser problem as we try to move to areas Houdini is not very well known yet very capable.

For example, I took it as my challenge to get into character animation and rigging in Houdini given there are very very few people doing that. Then I discover how it works and was blown away by the possibilities.

Right now I am handling a complex shot with 100 characters, all mocap with different levels of filtering, constraints, multiple meshes per asset and works very well. In fact, I doubt I could have put together 100 characters in Softimage and move it. You know Softimage was never great with tons objects.

Good Lighters are also need, good animators too… I feel there will be a demand for houdini artists for quite a while so looks promising.

What advice can you give on facial rig in Houdini (muscle based or bone based)?

Thanks.
User Avatar
Member
207 posts
Joined:
Offline
pezetko
takita
Another minor annoyance is (transient) windows getting lost underneath the main viewport window. So if I'm browsing for a SOP path in a parameter field or doing that “edit parameter interface” thing and I make an errant click somewhere, then I have to go looking for the window I just opened which is now lost beneath the main UI. Anyway it would be nice if windows with “accept” or “cancel” just stayed on top until I was done (like the help window).
Actually that's nice thing that all windows aren't modal. You can work with that windows in behind and when you need it bring it into foreground. So you will not get your workspace cluttered. I really like that some windows aren't modal and I can make changes instead of being forced to click Ok/Cancle and reopen that windows later.

Yeah… in my case half the time all these windows are popping up in the back and never even make it to the foreground at all. Not sure if it's because I'm on a dual monitor and I'm often invoking them from a floating window but it's really annoying. Tabbing through windows gets old real fast.

pezetko
takita
Oh and passthrough (or “dot” like in Nuke) nodes would really help with node graph legibility and accidents when “drop on a wire” is active (which I like, but is really dangerous for big crazy graphs). Really miss those passthroughs…
There is Null node that work same, just passing data. It's bigger than Nuke dot node, but on other side, consistency is plus.

Oh nice… that works.

pezetko
takita
And yes, I did build an OTL to construct them manually… although given a choice I would be very happy to be able to throw that away. Less junk, you know?
Actually it's more junk but just under the carpet. Heavier scene, slower processing. I think it's big plus to be able to put in only what you really need. Offten there is no need for orthogonal frame on the curves and so on for other tools. It's nice to have the power to be so close to the metal.

It's nice from start to have everything-can-do tools but in the end you want speed in the scene and it's nice to have option to remove something, or even not to add it in the first place.

While I like that I can work around things, I do think that as a general matter of principle the factory stuff should be tight whenever possible. So if it says orthogonal frame, tangent, or normal I'd like it to actually be that and not have to try to figure out whether it's actually doing the right thing or not.

Other than minor things here and there though, there's a lot to like.

Thanks for the tips!

cheers,

-T
User Avatar
Member
99 posts
Joined: July 2007
Offline
hi takita,
regarding windows popping in background i think its an issue with gpu drivers/ os, but it was supposed to be solved long ago, i don't recall exactly why it was happening, but i'll try to look for it on the forums when i'll get some free time.
User Avatar
Member
655 posts
Joined: Feb. 2006
Offline
I have not played with muscle based facial rigging yet, my inclination is still the one I apply for Softimage, great blend shapes layered with muscles to help avoiding linear blends.

I am of the opinion that the animator needs to see the outline and therefore muscle simulation is out of the equation for me, this is the reason I have not pushed there yet.

Can be done? of course, here we were using FEM to do skin simulation for the Ogre project and worked great but my point of view is the animator's point of view.

I will suggest start with FACS based shape animation and then start inspecting how to transition in a more clever way… that probably will be wisest and simplest approach, then start working with post-simulation or full muscle based FACS setups.

Regarding control panels, Houdini UI is quite old and limited so either you go for PyQT or simply put the control objects as normal 3D shapes (not renderable) in the viewport.. a bit old school I know but its very efficient.

hope it helps
jb


faizol
jordibares
As an industry you can see lots of companies are finally embracing it and there is more demand than ever for all sorts of positions, if anything, it is the FX side that is the lesser problem as we try to move to areas Houdini is not very well known yet very capable.

For example, I took it as my challenge to get into character animation and rigging in Houdini given there are very very few people doing that. Then I discover how it works and was blown away by the possibilities.

Right now I am handling a complex shot with 100 characters, all mocap with different levels of filtering, constraints, multiple meshes per asset and works very well. In fact, I doubt I could have put together 100 characters in Softimage and move it. You know Softimage was never great with tons objects.

Good Lighters are also need, good animators too… I feel there will be a demand for houdini artists for quite a while so looks promising.

What advice can you give on facial rig in Houdini (muscle based or bone based)?

Thanks.
User Avatar
Member
37 posts
Joined: July 2011
Offline
jordibares
Regarding control panels, Houdini UI is quite old and limited so either you go for PyQT or simply put the control objects as normal 3D shapes (not renderable) in the viewport.. a bit old school I know but its very efficient.

hope it helps
jb



Do you play a lot wit PyQT? If yes, do you know if it's possible to catch keyboard and mouse events of houdini windows with the PyQt interface?
User Avatar
Member
655 posts
Joined: Feb. 2006
Offline
I am afraid I haven't… I assume you will be able to.

Let me ask here

katana13
jordibares
Regarding control panels, Houdini UI is quite old and limited so either you go for PyQT or simply put the control objects as normal 3D shapes (not renderable) in the viewport.. a bit old school I know but its very efficient.

hope it helps
jb



Do you play a lot wit PyQT? If yes, do you know if it's possible to catch keyboard and mouse events of houdini windows with the PyQt interface?
User Avatar
Member
1708 posts
Joined: March 2020
Offline
jordibares
(…) I took it as my challenge to get into character animation and rigging in Houdini given there are very very few people doing that. Then I discover how it works and was blown away by the possibilities.

Right now I am handling a complex shot with 100 characters, all mocap with different levels of filtering, constraints, multiple meshes per asset and works very well. In fact, I doubt I could have put together 100 characters in Softimage and move it. You know Softimage was never great with tons objects.

I'm glad you're looking at it – I always wanted to do the same thing (I did a lot of rigging in Maya but I got kind of fed up with it, moved on to FX, found it much more interesting.)

There are strong preconceptions against Houdini in the character/rigging field, but all the tools (and more) are there, and my opinion is that much more capable rigs could be done than in Maya. (Just never had the incentive to spend my free time on rigging, of all things )

So I for one am pretty interested on how things go…
Imre Tuske
FX Supervisor | Senior FXTD @ Weta FX

qLib -- Houdini asset library
http://qlab.github.io/qLib/ [qlab.github.io]
https://www.facebook.com/qLibHoudini [www.facebook.com]
User Avatar
Member
207 posts
Joined:
Offline
owlYzarc
hi takita,
regarding windows popping in background i think its an issue with gpu drivers/ os, but it was supposed to be solved long ago, i don't recall exactly why it was happening, but i'll try to look for it on the forums when i'll get some free time.

owlYzarc -

I would love a solution to that…

cheers,

-T
User Avatar
Member
581 posts
Joined: July 2005
Offline
riviera
There are strong preconceptions against Houdini in the character/rigging field, but all the tools (and more) are there, and my opinion is that much more capable rigs could be done than in Maya. (Just never had the incentive to spend my free time on rigging, of all things )

So I for one am pretty interested on how things go…
Yep in terms of architecture houdini offers lots of possibilites for rigging but the IK Solver in CHOPs and the transform stack in OBJ context needs some love.
CHOPs needs to be updated for the 21st century, the concept is amazing the implementation is old and needs to be correctly integrated with character animation.
The transform stack in OBJ is slow specially when you have lots of nested subnets, this doesnt affects character animation too much but affects vehicle rigs.

My hopes for character animation is that SESI rewrite CHOPS for H14.
Un saludo
Best Regards

Pablo Giménez
User Avatar
Member
99 posts
Joined: July 2007
Offline
ok, i think i've finally managed to find the thread i've had in my mind, unfortunately there was no solution for this posted.

http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=27756&start=0&postdays=0&postorder=asc [sidefx.com]

I used to have the same issues some time ago, but then i did something
If I understand correctly it the problem with dual monitor set-up, could you describe what it consists of? what GPU(s) you have?
drivers?
what monitors?
are they identical?
how are they set up? A-B? which one is set as primary display in windows? you are using windows, right?

btw, Houdini really dislikes negative space (lower left corner of primary window is at (0, 0).

P.S. problem seems to be windows related some possible sollutions ->http://social.technet.microsoft.com/Forums/windows/en-US/7f4543cf-874d-4a39-bea5-34a824e4c0ce/pop-under-dialog-boxes-in-win-7-ultimate?forum=w7itproui [social.technet.microsoft.com]
  • Quick Links