Jump to content

Houdini 14 Wishlist


LaidlawFX

Recommended Posts

Ok so I need to take some poor man's Abmien, a.k.a. NyQuil PM, because I was thinking of Houdini of the future... and well... 13 comes out in 31days which is kind of the present so... Houdini 14, you ready?

Over the next few weeks the rest of the 13 tool set will be revealed, so only the big dreams now...

I think the flatning of Houdini's multiple contexts is something brought up in the last thread that I want to continue forward. It's the crazy ideas like this that take time to be managable, the equivalent program for conversation is nuke. There are about 9 majors contexts obj, sops, dop, chops, ?pops?, cops, shops, out, and all the vops contexts. You enter through Objs and then have to dive down, up, and sideways. Why not make the 1500+ nodes available in one context called global for conversation sake, and make the nodes unique designed so that they can be visually categorized and wired together. Say even coloring the wires too. In sops alone not all nodes can/should be wired together vdb don't all work with volumes, polygons don't all work with nurbs. The round nodes in Nuke per say let the wires go around, maybe a slipping interface where the top and bottom are just bunchs that rotate around the core, whether that be a circle, a octagon, a square, or a concave shape(pacman)...

This would introduce the concept of floating networks like a bunch of chilling light and rops nodes off to ones side and a sop tree to the other. Which you can imagine to explode to be very big in the first place. But you should still be able to organize with subnets. Possibly create/maintain the sop Dop Network and the sop Pop Network like managers, for the old school/organized to stay with the new way, or even maintain the current manager network style subnets that are in sops. One thing to think of an Object node is only the very last parent node of a sop network after all.

Also since networks flow down(sops), sideways(vops), or just float around(objs, out), possibly make the wire connections slide so you can work right to left, down to up, or diagnol, or alternate make all the networks go in one direction. I'm not against the flow now, but why not open the possibility, sometimes rigid is good, but sometimes down networks, left to right networks, or floating networks have soo many branches or side patterns they are not easily defined, even within these "rigid" contexts. Think of the width of some of your networks, when the width of your networks are actually longer than the height.

Structs... hopefully making its way from vex to vops in the 13... will open the door to alot of this wire branching of nodes. Where wires now have can pump anything through and not just one data set.

Also this would create the task of eliminating some of the redundant 1500+ nodes in houdini and making a more sophisticated tab menu. This alone may make the contextual issue of only wanting dop nodes per instance kill this idea in general, but if apple can come up with a stupid wheel that eilminates a few dozen buttons. I think the tab menu could use a next generation idea, too, but please spend the time on it, I don't want an autodesk widget I hate. It could probably go along with the idea of building tools that users visually like to use with the same friendly building mechanism(maybe after 13 we can talk about that too) we currently have.

What are your crazy dreams or just that extra feature you need to have for the next version?

  • Like 2
Link to comment
Share on other sites

Usually i would say i will wait for 13 to go gold first but bwah that's no fun :)

I'm guessing, COP's, easier/better UI tools, stuff like being able to insert an image as a button to the parameter interface would open a lot of things up and even if you use wx or PyQt for your UI's you can't integrate it into Houd's main UI and a better workflow for importing custom guide hairs for the fur toolset (even with workarounds, the interpolated guides don't yield the greatest results) oh and faster, faster, faster. That is my fantasy, but im sure H13 will have a lot of decent features.

Oh and if there is one thing that i would like to see an changed, its the material palette/gallery. It seems really clunky, the pull-downs for the node contexts are kinda small and don't seem to be heavily emphasized enough since that is one of its main abilities (to drop materials into a specific shop, not just the standard /shop). And the shader ball only updates manually :(. More OGL options for shader would also be very nice.

Link to comment
Share on other sites

Oh and if there is one thing that i would like to see an changed, its the material palette/gallery.

I would like to see the death of material palette... I use shaders all the time, and I replace it with a shelf in the tab menu, and avoid a whole interface. If be another thing if the shaders there were actually useful in a production environments. Like a good car paint shader, glass, competition to vray/arnold and modo. Hard to turn a company to using houdini in production rendering if they can't get just anybody to render something nice out of it. Especially with the turn around vray and arnold had on the lighting and shading community in the last few years, houdini is capable of the same thing, but you can get your average lighter and shader to flip over to it when it doesn't work out of the box as nicely as the competition.

  • Like 1
Link to comment
Share on other sites

In regards to becoming more like Nuke's DAG; Nuke node graphs get extremely unwieldy and they need to be organised when they get to any decent size. The current contexts, dops, sops etc in Houdini does that automatically.

Link to comment
Share on other sites

In regards to becoming more like Nuke's DAG; Nuke node graphs get extremely unwieldy and they need to be organised when they get to any decent size. The current contexts, dops, sops etc in Houdini does that automatically.

A Houdini scene will get unwieldy with the same amount of nodes too. It's always up to an artist to organize, which I will agree most fail to do. But the global level contexts are just giving a predisposed set of folders. You could do the same, currently, if you wanted to only work in the sop context with just managers networks.

I think especially when teaching new people that global network context actually hinder the organization structure. If you teach a new person to learn Houdini and try to explain the contexts to them with no 3-D experience you can see their eyes confused and they'll ask why it is that way. Especially when you ask them to drop a sphere, object level or sop and why you should work on it at each level. The reflective question to ask is why are there multiple contexts to someone who doesn't know why? How many lines of logic do you need to quote?

Edited by LaidlawFX
Link to comment
Share on other sites

The reflective question to ask is why are there multiple contexts to someone who doesn't know why? How many lines of logic do you need to quote?

Sure but it's all semiotics when you extrapolate it all out. ;) i.e. why are their points, vertex, polygons - I just want to make stuff!!! Why do I need to render to files - I just want to playback good looking stuff - why do I need OpenCL processing - I just bought the fastest computer and it should just work!!!! yadda yadda yaddda.

I'm not arguing against a universal node graph, I just think it's mistaken to assume to that it solves anything really. A forced set of contexts, i.e. Dops, Sops etc. is not all bad and has value that is often rejected in pursuit of the what really? as it currently stands, Houdini automatically organises so the artist does not have to, obviously to a point, once the software allows anything to happen hen the artist must be organised. Looking the fact that most artist do not organise then is it not better to have the machine do it?

Link to comment
Share on other sites

..

What are your crazy dreams or just that extra feature you need to have for the next version?

{ wake me uppp .. when October endsss .. } ♪ . ♫

=)

.cheers

Link to comment
Share on other sites

I'm not arguing against a universal node graph, I just think it's mistaken to assume to that it solves anything really. A forced set of contexts, i.e. Dops, Sops etc. is not all bad and has value that is often rejected in pursuit of the what really? as it currently stands, Houdini automatically organises so the artist does not have to, obviously to a point, once the software allows anything to happen hen the artist must be organised. Looking the fact that most artist do not organise then is it not better to have the machine do it?

semiotics, why didn't they use that word on sesame street. I'll agree it may not solve anything really(with the teaching point in particular), and further to the point the system does work quite fine. Never though about it until it was mentioned a little while back.

Maybe another thing would be to add it wouldn't hurt anything(Outside the programming hours), where there is a potential(I'm opinion very high) that it will add a small but significant update, not say like core speed optimization as was 12s main thing, but at the worst on par with the individual feature being added to 13 now.

A discussion we had a few days ago was similar with 3 digit padding vs 4 digit padding. 3 digit padding covers say the 90% of padding related issues with files and renders, but where as 4 digits padding will get that last 9%, and the drop increases significantly beyond that. So why not use 4 digit as opposed to 3, they changes are slightly a bit more, but end up making a difference down the road. The analogy may be a little general(like the teaching), but it's one of those gut feelings that makes sense.

Something on the reverse in an attempt to simplify stuff too much you may make it worse. Apple has shown that curve quite well. But there is a lot of Houdini that has so many niches and corners where it may do good to take an outside look at it and merge and carve the portions off of it to make it more useful. The lighting tools and methodology would be one, while being very versatile, and almost all features are technically covered you could def. compact them into one set like categories, light mask, groups, and bundles with pbr and non-pbr. into a more straightforward approach. Which I hope they get closer with 13, those are some more changes I'm looking forward to.

Link to comment
Share on other sites

bit soon for a H14 wishlist me thinks.... I mean this isn't a Autodesk forum :)

I was on the fence, too. It has been a long time since SideFX has announced a build so far in advanced, usually it's a "surprise" during an event like Siggraph. Plus it doesn't make much sense to make additional wishes for 13 when the only practical fixes would be bug fixes and those should be submitted via the bug data base. Also most of the post on the wishlist are major long standing RFE, but induced with group support. Or they are wishes that exist already that can be help be redirected. We may not know yet what all the new features will be, but its a safe bet the craziest dream will not come true in 13, within 28 days later... maybe a zombie or two though

Link to comment
Share on other sites

For the Global UI Tab menu that contains all the 1500+ nodes a starter idea would be to press tab then a circular menu of the different context could pop up, so that you could hover on the context area and the subcategory list can show up. It would be like spacebar Maya spread with an Apple like circular concept for the first step. Houdini is pretty close already. Traditionally you can still tab and type your node name, as the names shouldn't have over lap. Any original overlap would have to handled when the first configure the global context, but the normal name system can already handle it. But even with node name overlap there is the case already when a node name exist in multiple categories they append a bracket of which context NodeName(CategoryA) NodeName(CategoryB).

Also you can creatively bend the sub context menu from the initial context menu in something not always grid like... one aspect of the UI with a curve in it.

Link to comment
Share on other sites

The only thing I care about now is viewport interaction and user interface enhancements. I don't care if they pull the entire team just to do those for v14. I'm tired of having a flaky viewport and crappy late 90's nodal networks. Nuke, Katana et al show how a modern UI should behave.

SESI if you're listening, I know you're scared of the UI code.... but you can do it! We have confidence in you!

  • Like 6
Link to comment
Share on other sites

Hey Marty

We have a ton of RFE's and Bugs in with SESI about the viewport. It would take too long to go into here :), but there's a lot that could be done to improve the day to day interaction with the viewport.

Link to comment
Share on other sites

thanks Marc, I understand that. You probably saw the "stuck geo in Viewport 12.5" thread, if there is even just one bug you want smashed, post it up and I'll have a go at solving and submitting it. RFEs I can't help with.

That thread:

Link to comment
Share on other sites

Generalizing IMO the UI/Viewport needs another upgrade such as from v8 to v9. It's falling behind a lot of other new age packages. It lead the wave coming into this gen, but every other new age package(not autodesk products) has come with a way better ui handling and viewport design(and a bit of handling). Technically everything may work, but it has a real lack of an artist touch to the UI. A very nice 2000's technical design, but there are a shit load of prettier thing you can do to this. The simplest example is the Color Picker for the network pane nodes and the color dropper paletter for color parameters. For viewport it's just not really linked to the node network, it's the node network that drives the viewport, but you look at zbrush/mudbox and it's the complete opposite.

Link to comment
Share on other sites

Could do - though one still uses Houdini over Mudbox/zbrush for the productivity it provides, that's Houdini's core strength.

On the UI front I like what Croquet from Viewpoint Research Institutes provides. A user interface that can always be manipulated to suit your style.

http://www.vpri.org/

Link to comment
Share on other sites

I would like both Houdini (and Nuke) to change their nodes to (1) expand/collapse the node box interior so that it becomes the parameter pane at the scene level; and, (2) add a "tree" flag that aliases the entire underground node system up to scene level, same thing as hitting the w key but I could add connections. For example, I could connect geometry directly to a merge node in DOPs and it would automatically do the conversion in the same manner as the shelf tool.

Really, (1) would make my day just b/c I seem to jump between two nodes in different contexts quite often.

Cheers, Kevin

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...