hMonkey
I don’t remember which part, but it’s peppered through out the video, basically pivot around cursor is imprecise and wonky at times, and occasionally you have to wrestle with pan and zoom, all that falls under the viewport issues category.

I'm not going to sift through that video, but I'll take your word for it. Maybe I am fortunate, because I am approaching 5 years of using Houdini, and I have not experienced any issues whatsoever with viewport navigation.

It works very well and exactly as expected/advertised. I am able to navigate very tight spaces very cleanly - the tumbling/dollying around cursor is very precise - the double-click to close-up works great - framing components all good - even the temporary override to use the view center as the navigation pivot is nice.

Pavel Breiva
The only hope are labs team, which could come up with some additional tools..

Or they can reach out to people like @alexeyvanzhula1984 , @animatrix_, who already have a lot of experience (ui/ux/modeling), or to who ever made that game inside the node network, could probably come up with a way to have better auto layout and better organized / readable node tree. I don’t know why people keep thinking that it should be all-hands-on-deck for any given feature… And big studios still depend on single artists.

eikonoklastes
I'm not going to sift through that video, but I'll take your word for it. Maybe I am fortunate, because I am approaching 5 years of using Houdini, and I have not experienced any issues whatsoever with viewport navigation.

Again it depends on the platform, and it’s not always noticeable… maybe you are more focused on net view, and don’t consider speed bumps along the way as an issue… no offense.

hMonkey
Again it depends on the platform, and it’s not always noticeable… maybe you are more focused on net view, and don’t consider speed bumps along the way as an issue… no offense.

I don't know what you mean by "net" view. Do you mean I do my work in the network editor as opposed to the viewport? I wouldn't get much 3D work done, if that was the case - I interact frequently enough with the 3D view directly, including component manipulation.

Back on topic, you went from "terrible/imprecise viewport navigation and screwed up priority (navigation as a tool)." to "depends on platform and not always noticeable" which is quite a shift in stance, and you still haven't clarified what you mean by "screwed up priority (navigation as a tool)".

If you can provide a concrete example (that isn't buried in an 83 minute video that you yourself cannot remember the locations of) of this terrible, platform-dependent, not-always-noticeable viewport navigation, that might be a more constructive approach to the issue.

eikonoklastes

hMonkey
Again it depends on the platform, and it’s not always noticeable… maybe you are more focused on net view, and don’t consider speed bumps along the way as an issue… no offense.

I don't know what you mean by "net" view. Do you mean I do my work in the network editor as opposed to the viewport? I wouldn't get much 3D work done, if that was the case - I interact frequently enough with the 3D view directly, including component manipulation.

Back on topic, you went from "terrible/imprecise viewport navigation and screwed up priority (navigation as a tool)." to "depends on platform and not always noticeable" which is quite a shift in stance, and you still haven't clarified what you mean by "screwed up priority (navigation as a tool)".

If you can provide a concrete example (that isn't buried in an 83 minute video that you yourself cannot remember the locations of) of this terrible, platform-dependent, not-always-noticeable viewport navigation, that might be a more constructive approach to the issue.

He probably means that during navigation with alt+lmb, I accidentally click on an object with a blueprint that interacts with viewport pivot, and the camera doesn't rotate as needed. The issue is likely not with navigation, but with the fact that background images aren't suitable for my type of navigation. Houdini background images can't be displayed in the perspective viewport all at once. This is important to me because I'm used to aligning the camera with the shift key, like in ZBrush, instead of toggling between orthographic projections.

hMonkey
Because it’s pointless getting into details on these issues with you.

I'm not sure why you say that. All I'm asking for is the actual evidence to the claim you're making. It's an extremely reasonable thing to ask for.

hMonkey
Just curious what was you previous dcc of choice?

I started with 3ds max (back when it was Discreet 3D Studio Max), used a bit of XSI, moved to Maya for a few years, then to Cinema 4D for about 10 years as I shifted more into motion graphics, and then to Houdini for the last 5. I've modeled with Wings 3D, Max, Maya and Modo (which was easily the best modeling experience I've had). I stopped modeling heavily once I got into motion graphics, but I still do some basic modeling.

Why do you ask?

Interactivity with HDAs and viewport material previews are two aspects that I find myself increasingly missing.

We could benefit from a high-level tool that facilitates Python states to easily generate controllers for custom HDAs. Improved interactivity is necessary at both the SOP and OBJ levels, particularly within the OBJ context.

While Solaris might offer more power from this perspective, it still isn't well-suited to my needs, as I typically work on smaller environments.

SideFX has teased the introduction of Karma RT. It would be amazing to set up material previews and seamlessly integrate materialx materials all in real-time. This would allow us to move away from third-party rendering solutions. Personally, I've grown frustrated with Redshift over the last three years; it's been quite a hassle.
I hope this is the last year I have to throw away money for a Redshift maintenance license.

The Feather tech preview was awesome. It's great to see SideFX developing a new brush that directly interacts with meshes, bypassing vertex points. This innovation could potentially revolutionize texture painting in Houdini and synergize well with Karma, alongside an improved COP system.

And finally, keep it up with AI and innovate in this field as well.
I had a dream with one of my H fiends, think about training proper hda’s and create faster and better

alexeyvanzhula1984
He probably means that during navigation with alt+lmb, I accidentally click on an object with a blueprint that interacts with viewport pivot, and the camera doesn't rotate as needed. The issue is likely not with navigation, but with the fact that background images aren't suitable for my type of navigation. Houdini background images can't be displayed in the perspective viewport all at once. This is important to me because I'm used to aligning the camera with the shift key, like in ZBrush, instead of toggling between orthographic projections.

That's a good point. Having all reference images in the perspective view, and not have them interfere with view navigation is a reasonable thing to want. I'll RFE for it.

eikonoklastes

hMonkey
Because it’s pointless getting into details on these issues with you.

I'm not sure why you say that. All I'm asking for is the actual evidence to the claim you're making. It's an extremely reasonable thing to ask for.

Fine, I’ll find it, just not today. But what would that accomplish, I'd rather spend that time documenting the issue for a bug/rfe report.

eikonoklastes

hMonkey
Just curious what was you previous dcc of choice?

I started with 3ds max (back when it was Discreet 3D Studio Max), used a bit of XSI, moved to Maya for a few years, then to Cinema 4D for about 10 years as I shifted more into motion graphics, and then to Houdini for the last 5. I've modeled with Wings 3D, Max, Maya and Modo (which was easily the best modeling experience I've had). I stopped modeling heavily once I got into motion graphics, but I still do some basic modeling.

Why do you ask?

Well you have enough experience with both good and bad workflows/conventions to source from and not take things for granted when you know there’s a better way of accomplishing certain tasks, couple that with enough Houdini experience and you should know what would work better in context of Houdini, or rather what’s not working well. And your migration path tells me that you weren’t happy with most of them for whatever reason, then you hit a roadblock with c4d and moved to Houdini. Speaking of which c4d’s navigation is more reliable because it’s not a tool, it’s built-in. If I recall correctly Houdini added the alt nav. at some point, the viewport wasn’t built with that in mind from the get go, and due to the fragmentation of H the issues are intermittent...

If H20 would bring the ability to assign a sticky hotkey to snapping and an 'edge divide' tool that works decently, I'd do my modeling in H and even though that still wouldn't make it perfect (far from it), it's enough to at least not make me go outside for some trivial modeling tasks.

The viewport is definitely very bad for me. This is one of Houdini's biggest problems, I think. I'm a beginner, and every version was waiting for a miracle that the viewport would work fine. But no miracle happened. 18.5, 19, 19.5. At the beginning, I stopped studying Houdini several times in complete despair, because I didn't really understand what was happening from my actions. All my many years of experience in dozens of programs turned out to be useless. I read help, watch tutorials, look at the screen, do actions and categorically do not understand the consequences of 50% of my actions. I stopped, came back, stopped again. So are some of my friends. And only six months later I found out that I had to reset the viewport, because it lies very often. And even later I realized that the viewport lies very often, almost always, if you jump through the nodes. And only from that moment I was able to learn Houdini and understand the response to my actions. Now I have assigned a hotkey to reset and I always press it, almost after each transition through the tree. Reset, reset, reset, reset, and so on a thousand times a day. And now I understand a lot. But who among the newcomers knows about this life hack ?
But even now, I am constantly confused about how to light the dots in the group, they sometimes glow, then not, I do not understand the sequence required of me. I do the same thing, but the result on the screen is always different. Very annoying.
So the first problem of the viewport is the destruction of newcomers by the screen's lies. A unique situation for a graphic software product, don't you agree?
The second problem is the demotivation of the user with a very dull image. I feel it on myself all the time
So, the main thing that I'm waiting for from version 20, and then I'll probably wait from 20.5, 21, 21.5 - this is a new viewport.
This is a very serious drawback of Houdini, thawing newcomers, and it is very, very unfortunate that the developers ignore this. New users are a must for any product. And most newcomers are just disappointed and don't get to all of Houdini's unique things.

Houdini modeling tools are not that bad, enhancing the experience with Alexey tool sets make it way better.

I can live with the glitches, hell I have used Modo for years, and despite the crashes, slow down, bugs, it's still a more pleasant direct modelling experience than Houdini.

At the low level set of missing tools I miss partial edge loops, edge slides duplicate, and real metrics with loops in general (like real distance on an edge split, not in %), mirror loops options etc.

Temp pivot need to be badly streamlined (I might have miss something, but the whole detach handle, pick an element, attach handle, move the geometry is a tad slow :P))

Would that make Houdini on par with other direct modeling solutions ?

Probably not, as long as you have the node graph in your way, having to deal with display flags, managing merging of parts etc (vs a simple layer system), I guess it will always be a "slower process".

Modeling in Houdini is a very different way of working, it's actually really powerful to create 'freezed' mesh, like revolve/lofting curves, boolean stuff ... but for poly knitting perfect sds poly cage with accurate edge loop, I will just use something else ... (and use houdini as an amplifier).

It's a pity that Nurbs have been abandoned, because it fits the Houdini way of working so well.
Building surfaces from curves, blend them with continuity control, trim, booleans ... those are operations that don't require topology level knitting and could work very well with a nodal system (Grasshopper for rhino). But that's another story.

(Thrilled to see the new SoftBooleans from Alexey, that way of working really speaks to me and really fits Houdini mindset too)

A lot of great thoughts here. I really hope that SideFX is listening and reading these, but just to be sure, please do create FERs for these.

Don't think Houdini should invest too much resources in modelling, there are plenty of solutions on the market.
Look at the pre alpha of next Modo version, it's insane performing.
Houdini just needs some more functional sculpting tools, for generating general adjustments or doing easily morph maps
for face animations or stuff like that.

Houdini should invest in rendering and look dev, just adopt the best out of Clarisse.
In Lops, it would be nice not to move in a new world. I would like to have the same parent and target functions like on
GEO Level. Look dev should be easy, should be intuitiv, should be interactive.

Houdini won't be a Zbrush, 3D Coat or Modo and that's ok.

tomtm
Look at the pre alpha of next Modo version, it's insane performing.

can't find info about this

alexeyvanzhula1984

tomtm
Look at the pre alpha of next Modo version, it's insane performing.

can't find info about this

https://m.youtube.com/watch?v=XPTLyTy_QiM [m.youtube.com]

tomtm

alexeyvanzhula1984

tomtm
Look at the pre alpha of next Modo version, it's insane performing.

can't find info about this

https://m.youtube.com/watch?v=XPTLyTy_QiM [m.youtube.com]

thank you!

This eternal debate of Blender and Maya compared to Houdini is quite tiresome. Who wants to continue using destructive tools in 2023 that only have Ctrl+Z as a corrective element? Where nothing is procedural or with preconceived tools from a thousand third-party plugins that are not maintained in any pipeline.

Who wants to manually model a ship when you can design a system that generates infinite ships and their variations based on sequence characteristics? I only see Maya + ZBrush being used for characters and certain key assets.

Maya's way of functioning is outdated, painful, slow, and is only kept alive by a community of users who refuse to learn another program. As of today, Maya is only strong in animation. Blender, well... impossible to fit into any pipeline.

I believe that each software has its pros and cons. But comparing Houdini to any other software nowadays is laughable. The data management and the ability to do anything you can think of in a thousand different ways, non-destructively, is UNIQUE.

Things to improve? Of course, there are. A viewport that matches at least Clarisse's, improvements in LOPs to make it run faster since USD is the right direction to go, improvements in Karma, MaterialX, the ability to drive attributes in textures (supposedly coming in H20).
I don't know, the endless absurd debates are getting tiring.

Also, I don't know what kind of Mayas, Blenders, and Houdini setups you guys are using, but Houdini is much more stable than those software options. Review those configurations or scripts for those who complain about Houdini's stability. Because if you have things set up properly, it's as solid as a rock. (Except for viewport glitches). By the way, not everyone knows, but in the labs tab, there's a reset viewport option that cleans and reloads the viewport to eliminate glitches without having to close the tab.

Greetings, and let's hope some new features come out soon

Blender just started geometry and simulation nodes. Look at the interfaces. It's horrible.
I think it would take tens of years for Blender to catch up with the node-based system of Houdini.
It may seem like that you can quickly perform modelling in Maya or Blender.
But modelling is comprised of countless repetitions and corrections, and Houdini is the best in these procedures, not to mention recycling.
Of course, Houdini modelling can't be done in completely procedural ways, but should not be compared to Maya or Blender's way of modelling. Once accustomed to Houdini modelling, Maya modelling is just a pain, and I never used Maya for modelling in recent years.

Jackatack
This eternal debate of Blender and Maya compared to Houdini is quite tiresome. Who wants to continue using destructive tools in 2023 that only have Ctrl+Z as a corrective element? Where nothing is procedural or with preconceived tools from a thousand third-party plugins that are not maintained in any pipeline.

Who wants to manually model a ship when you can design a system that generates infinite ships and their variations based on sequence characteristics? I only see Maya + ZBrush being used for characters and certain key assets.

Greetings, and let's hope some new features come out soon

You probably don't know much about subdivision surface modelling ? (genuinely asking), not everyone is building giant ship or need to build parametric system for everything. Subdivision surface modelling require thousand of micro adjustment (edit), meaning your 12000 graph node is a giant undo stack, or more precisely a time line (which is already cool actually). But there is nothing you can adjust upstream that would get correctly propagated. So whats left is the quality of the tools, and how fast you can operate them in a non graph building context, and it that area there is lot on the table to improve imho. I'm fine with the idea to use something else for that part of the work, but if there is a will to continue to improve that direct edition experience, I would appreciate too.