edward
I should point out that the TAB menu does already keep track of your most recently used items (during that session). The number of most recently used items is also configurable in the preferences. To shorten the mouse movement down, one can completely forego the default menus as well. I still think the fastest way to model is by becoming an expert user and assigning hotkeys to SOPs though. Marc H. has his own set using the numeric keypad.

Does anyone keep a permanent tree view open in their desktop for quick jumping around? There's the autodive option too.

I am agree with edward, for me if you want to be productive you should use hotkeys and forget TAB menu/hotbox/popupmenu/etc …
And using the Toolbar Manager is quite easy to create a custom menu entry with your most used tools.
I think that the majors improvements that are needed in the modeling field are:
- Improve the existing tools: better poly tools like an improvement in the PolySplit tool.
- Improve the OpenGL viewport performance.
- Study new workflow options when working from the viewport: this is the part, in my experience as an instructor, that most of the new users hate when beggining to model in Houdini. Definitely is more intuitive the workflow: Select geo -> Select tool, rather the select tool->select geo. I think that it would be very useful to have a quick access to the network, for example a floating network editor, like the parameter view you open pressing p in the network editor, but in the vewer pane.
So you can work with the viewer pane maximized and take a quick look to your nodes without needed of restoring the viewer pane. It could be activated using a hotkey or when the cursor is very close to the vewport edge for example.
An of course, somo tool to create poligons, more intuitive than the Add SOP and better that the Poly Knit.

lisux
- Study new workflow options when working from the viewport: this is the part, in my experience as an instructor, that most of the new users hate when beggining to model in Houdini. Definitely is more intuitive the workflow: Select geo -> Select tool, rather the select tool->select geo.

You know you can do this already by hitting “s” first to select and then choosing your tool. Is that what you meant?

Wolfwood
If you are in one SOP network, say object A, and you have display everything turned on, you can right click on another object and pick “Edit Geometry” which will jump you into the other Object. Or you can turn on handles from the other object by picking “Enable Handles”.

Strange, that has never worked for me It jumps to the same network)…neither Alt+CLick, as you would do from OBJ->SOP … That's why I suggested the key+click jump method in the first place

Cheers

sorry, double post

From my point of view it looks like this:

-Houdini is not a fast modeler, and it'd probably never be because of the main idea of this software. As raging Reggie says among many mistakes… it's often not neccesary to have history in modelling and it's faster to delete/rebuild than to search for a node/change parameter/ohmygodnowthatswhatididn'texpect/repair errors.
I often find myself trying create something too clever and complex for the current needs.

-What is very powerful you can model in any soft you like and plug the geo into whatever doing network (say scattering spots), if you like to change the model, the network is reused and you keep free.
Any modeler can work as a plugin for houdini, treat file SOP as model SOP

-If there is a need to experiment with different modelling ideas, no software gives as much freedom as houdini. You are able to build extremely complex geometries using a few sops, apply custom networks to different models, test, change, create and this is SO VERY UNIQUE.

Good discussion. I beleive, and it has been mentioned before, that most of the people that want to change Houdini way of modeling are mainly organic modelers and other modelers coming from a background in another package where they have seen better/faster ways of working. All of those who have followed the increased popular discussion the last couple of years about subd techniques know that there is alot of development going on for Modo,Silo,Hexagon etc. and the main thing that they all have in common is that they increase the speed and makes the user work interactively at almost the speed of thought (no kidding it can go extremelly fast).

So from my own experience in organic modeling I think that the best way to improve Houdini modeling is to add Interactive workflow combined with the current nodebased. More operators would of course be beneficial but making so that the current toolset is faster to work with would make a whole lot of difference. Here is a few.

-Pre Highlight Selection, including brush picking.
-Multi Component Selection Filter
-Color feedback and Edge distance for Soft Radius in EditSOP.
-Interactive Path Extrusion.
-Pre visibility in tools such a Bevel,edgesplit etc. combined with cancel option.
-Sticky Keys (editable)
-Interactive Slide (probably a function in EditSOP).
-Tweak/Drag as and alternate in EditSOP
-Isoline Subd Display.

etc.

These are functionalities that of course would be optional to give everyone their choice of workflow and it would still allow for procedural topology if so desired.

Perhaps that we that have experience from other packages should put our genious brains together and do a proposal backed up with problem analysis and proposed implementations with images/videos on a single location? It could perhaps be worth the effort.

SoulVector
All of those who have followed the increased popular discussion the last couple of years about subd techniques know that there is alot of development going on for Modo,Silo,Hexagon etc. and the main thing that they all have in common is that they increase the speed and makes the user work interactively at almost the speed of thought (no kidding it can go extremelly fast).

Off but on topic question. As more and more tools for SubDs and Hierarchical SubDs become more robust and widespread…does anyone still do extensive NURBS modeling? In Houdini I sometimes use the NURBS tools to provide a rough base which I then almost always convert to Polygons, but aside from that I never touch the toolset. What about everyone else?

(I only bring this up because NURBS modeling has been ignored through out this thread.)

SoulVector
-Tweak/Drag as and alternate in EditSOP

Have you noticed that you can just turn off the handle while in the Edit tool and drag the highlighted geometry? Follows the same rules as dragging the centre of the translate handle (ie. obeys construction plane, etc)

Simon

lisux
- Study new workflow options when working from the viewport: this is the part, in my experience as an instructor, that most of the new users hate when beggining to model in Houdini. Definitely is more intuitive the workflow: Select geo -> Select tool, rather the select tool->select geo.

You know you can do this already by hitting “s” first to select and then choosing your tool. Is that what you meant?

Yes, Simon I now it, but the problem is that you have to pres the s prior to select .
You can always disable secure selection and follow the Houdini workflow, select tool-> select geo, that is good for me because I am used to this workflow but understand the benefits of other workflows, if you want to use other ways of doing things you have to press s, select geo and then select tool, even using only hotkeys I thing is not the best way.
For example, an option in the main preferences to select the default behaviour for disabled secure selection in SOP will be great, with a couple of options:
1 - Use Houdini workflow, the same as current.
2 - When you select new geo Houdini automatically outs from the current operator, ans suppose that you want to add a new one, not to work with the current, and ask you to select the tool, then you can press q to use the last or select another. A quick solution I think to use the same workflow as other tools.

Another suggest is for subd modelling, the workflow using a subdivision SOP at the end of the chain, and playing with templates, is not bad but poor in interactivity, many new users have problems with this, I think that it would be great to have the option to have an “underground” subdivision SOP, for example we are modeling and we can select to enable background subdivision, and enable the preview of this subdivided geo with a hotkey, so anytime we need it, and we are everywhere in our chain when can press for example o and enable the view of a templated subdivided version of our model.

Wolfwood: I do use Nurbs for some models. I usually plan my nurbs model in another way since there are more strict rules but some of the enhancement for polygon workflow can also be applied for Nurbs.

edward
Have you noticed that you can just turn off the handle while in the Edit tool and drag the highlighted geometry?

Yes I am aware of it but I also have an idea on how to improve it. For each time that you select a new point the operator dialogue comes up and when you start to drag it creates a pop in the viewport. Not a big one but noticable which delayes the interactivity. Make an option to turn it off and combine it with multi component selection filter and pre highlight selection and you have a winner.

A short explaination for those who have not used multi component sel.filter it means that it can allow for selection of faces/edges/points without having to switch between them so you can have face/edge in the left stowbar or points/face or whatever you like. Combine it with pre highlight selection that gives a visual feedback by highlighting the component that you hover over, or from a optional distance in the view, then you will instantly know what component you are going to modify with minimal selection errors so you dont have to be careful.

Another,perhaps not obvious, thing that could minimize the steps to modify geometry is that since you can access the point data in each node´s spreadsheet I think it would be great to be able to manipulate point transformation not only by numerically typing it into the spreadsheet but also allowing for viewport manipulation in current node. The node gets softlocked in spreadsheet and the same could be done for the viewport.

lisux
it would be great to have the option to have an “underground” subdivision SOP

Yes agreed, it would be great.

lisux
have an “underground” subdivision SOP, for example we are modeling and we can select to enable background subdivision, and enable the preview of this subdivided geo with a hotkey, so anytime we need it, and we are everywhere in our chain when can press for example o and enable the view of a templated subdivided version of our model.

What would you see in the viewport? Can you see both your base polygon object and its subd version? Or just the subd? If you can see both are they both wireframe? Or is the base object semi transparent? How do they handle it in other packages?

I think what lisux means,if i understand right, is that we have Subdivision built into polygons and assign hotkeys so we can move up and down in Subd level in current SOP. In addition there should preferably be an option for cage or Isoline display. In other packages this is how it is done;

-XSI have +/- on to change level. Isoline display.
-3DSmax name is Nurms meshsmooth. It works the same way. Polygons with Isoline display
-Maya does not have Subd level swithing on current poly but have something called smooth proxy (works a bit like houdini)but they have hiearchial Subd´s as a unique geometry type. KBD keys 1,2,3 to change level.
-Wings3D has Subd on current polygons but you only have 1 level. Isoline display but points on cage. Shift+Tab for level switching.
-Silo has Subd on current polygon with Isoline display and point/edge/face manipulation on Isoline. C+V keys as default.

The key is to be able move the components on the lowres geometry but to have the “cage” wrapped around the divided polygons so it looks like the edges are curves that makes the boundaries similar to how it is displayed for Nurbs. Having so that you move the components on the Isoline surface and not on the cage gives you the freedom to work on the surface and not around the surface.

SoulVector
A short explaination for those who have not used multi component sel.filter it means that it can allow for selection of faces/edges/points without having to switch between them so you can have face/edge in the left stowbar or points/face or whatever you like. Combine it with pre highlight selection that gives a visual feedback by highlighting the component that you hover over, or from a optional distance in the view, then you will instantly know what component you are going to modify with minimal selection errors so you dont have to be careful.

Yes, yes, and yes - these would be really good enhancements for interactivity and speed. Please RFE them, or I will if you like.

Simon
Yes, yes, and yes - these would be really good enhancements for interactivity and speed. Please RFE them, or I will if you like.

Do feel free to RFE it Simon

Wolfwood

lisux
have an “underground” subdivision SOP, for example we are modeling and we can select to enable background subdivision, and enable the preview of this subdivided geo with a hotkey, so anytime we need it, and we are everywhere in our chain when can press for example o and enable the view of a templated subdivided version of our model.

What would you see in the viewport? Can you see both your base polygon object and its subd version? Or just the subd? If you can see both are they both wireframe? Or is the base object semi transparent? How do they handle it in other packages?

Yes is more like soulvector have explained. Very similar to the ZBrush approach.
Where you hit d (is I remember correctly) and then you subdivide your model.
Is like to have an implicit subdivision SOP at the end of your chain, you can have options for example like:
-Preview sub in wireframe, flat or smooth shaded, with the option of use a semi transoarent shading for example.
-An option to set that the subd is evaluated every any time, for example, if you are moving points around and in every action the subd version is calculated then you lose interactivity, the app maybe goues to slow, but if houdini calculates the new subd when it detects that you have stopped for, for example a couple of seconds, is like to cook in Changes instead of always.
Then you can enable the preview of your sudb along with the depth of sudb wit a hotkey, for example one hotkey loops along all the depth levels (maybe only 1,2,3 are enough), and another key switch between normal poly and subd poly.
It will put Houdini subd modellin one ster forward I think.

i watched the hexagon quicktimes and thought.. that really looks like Houdini, that could be Houdini… I still have to get the download working on my Mac (ducking) and try it out.

we use the polyAsSubD at render time for a lot of things and it does take the modelers a while to get the hang of modeling, rendering, modeling, rendering to see where things are lining up and resolving. it would be GREAT if we didn't have to rely on that method and had a in viewer display of the surface.

-k

thekenny
we use the polyAsSubD at render time for a lot of things and it does take the modelers a while to get the hang of modeling, rendering, modeling, rendering to see where things are lining up and resolving. it would be GREAT if we didn't have to rely on that method and had a in viewer display of the surface.

-k

Um, you mean like adding a subD SOP?

jesta
Um, you mean like adding a subD SOP?

For me, performance on my older machine is pretty poor for subdividing semi-complex geometry (especially if there's more than a couple of attributes on the mesh). Being able to toggle the limit surface with a hotkey, with better performance than Subdivide SOP and possibly skipping subdivision/interpolation of attributes, would be nice IMO.

you're funny jerry

no.

i mean actually implementing subdivision surface as a legitimate geometry type, as previous post(s) describe.. interacting and modeling with a real surface and seeing it in the viewport… sliding points on from the hull on that surface etc…

I also liked the brush work ala z-brush to paint things in hexagon.
if they went that far with the brush work in Houdini, you could even paint texture there.

-k

picture is worth a thousand words … when viewed as subd the base cage assumes the shape of the subd version and you can edit points on that version. For render you'd switch to your base cage but when rendered as subd surface it will pretty much look like what you saw in viewport.

At the moment you can sorta duplicate this workflow by putting down a subdivide sop, looking at your base cage in wireframe and templating the subd version and showing it as smooth shaded. Though you are still picking points on the rough base cage and things get pretty slow pretty fast even with fairly low res geo.