Відео

@@diegox7767 That’s a good question! That video is a bit old, and my approach to modeling in Houdini has evolved since then. Despite that, these days I still use multiple objects in my workflow, and here’s why: 1. Ease of Management: Houdini’s Node Viewer has a Tree View mode that lets me organize my model as a simple hierarchy list, making it easy to manage and hide unnecessary parts and reference images. Modeler for Houdini also makes switching between objects almost instantaneous. 2. Enhanced Viewing Modes: Working with objects gives me extra viewing modes. I can lock shading on objects and have separate shading for current and template geometry, which helps with seeing complex details. 3. World and Local Coordinates Access: I like working with both world and local coordinates. This lets me place a cylindrical geometry in world space, then dive inside and work on it as if it’s centered, which simplifies many modeling operations and works with symmetry. Also, I can have multiple OBJ subnets to build a layered hierarchy of transformations. 4. Performance Considerations: Houdini's performance will slow down if you place too many models inside one object and apply a template flag to each. While you can pack unused geometry, it may lead you to attempt to replicate an object workflow within the SOP context, which is counterproductive. However, on a micro level, whether you model a simple model in the OBJ or SOP context is entirely up to you, because Modeler for Houdini makes it easy to do both. Also, the OBJ workflow does have some downsides, like duplicating objects, which is smoother at the SOP level. And since modern rigging is based on the SOP context, you’ll eventually need to consolidate everything back into one object. Regarding SOP Subnets, I’m avoiding them since you dived inside SOP subnet you don’t see a template flagged geometry of the upper levels like moved to a new object. In my production stages, I model complex assets using the OBJ workflow first because it’s convenient, and then simply with only one “Object merge” I merge everything into one separate object for rigging or export in USD for texturing and rendering. With object merge I can do it procedurally without breaking the obj hierarchy I’ve built, so I can return back to the previous stage to expand or modify my model at any time. A bit about Obj and Lop context similarities: While the LOP context can potentially replace OBJ in the future, it’s still rough for this type of work in almost all areas I described earlier, not especially logically but more technically and I don’t want to describe all of the issues at the text, but the situation is slowly improving. For now, I believe, the OBJ context remains rock solid for modeling.

@@FaitelTech Thank you for the thoughtful answer and for taking the time to explain this. And must say I agree with you. I come from a decade of modeling in Rhino, and file and objects organization is a high priority in my workflow, and essencial to work in a smart way. In Rhino you can model "semi-procedurally" using the objects itself and the history of commands, and you can nest layers and sublayers. In Houdini, we have the nodes and contexts, but in fact the geo nodes are nothing more than another layer really, if you can merge everything after and just "deal" with the output in a higher context. Thank you for expanding my view on this matter.

Yeah, I was surprised too when I found it by pure accident while reading the documentation for Subdivide SOP

Sure thing!

Yeah, and you can actually do that with OD tools. In the network view, select the node setup you'd like to save as a text file, and press RMB - OD Tools Import/Export/Bridge - Export Node Network. In the opened file, browse and choose a path to save the network preset as filename.txt . The file will contain well-documented code with instructions on how to recreate the selected network. Then, in the network view, go to the top menu - OD Tools - Import Node Network and select the file. You'll get the exact setup with all nodes, flags, keyframes, parameters, etc, even if you load them into another subnetwork.

Здорово

You're welcome 🙂

Hi, that panel was part of the Modeler 2022 interface, and it was removed with the release of the 2023 version.

Hi, thank you! I think I had a first generation of neuron 32, and it was pretty hard to calibrate, especially for feet and hand positions. So, I had to do a lot of manual adjustments in the post-capturing. I didn't try the last generation, but from the experience I had, I can tell that there is a lot of marketing in that field where everything seems decisively quick and smooth to set and use, but you shouldn’t rely on it too much and spend some time to find actual users of the specific model you are interested in and ask for their first-hand experience. In general, I’ve enjoyed motion caption as a tool for animation, but I’ll be more assertive with buying a suite next time.

@@FaitelTech dude thanks so much for your thorough response! I appreciate it so much. I'm on the fence about getting it, so this definitely helps out things into perspective.

I may not be the best judge of speed or make comparisons with other softwares since I’m a generalist primarily using Houdini. What's important to me is the fact that I can do modeling in Houdini with the help of Modeler plugin. This plugin allows me to model using shortcuts or by choosing tools from the launcher menu, and what more important to use a lot of modeling tools which are common for other dcc’s. When it comes to sculpting, since Houdini doesn't have that feature, I use ZBrush whenever I need to work on organic models. This combination of tools satisfies all my creative needs, and I don't feel the need for any additional software for modeling. Regarding of speed I believe it depends more on the type of model itself, as character modeling, props, automotive and environments all present different challenges that require different techniques and approaches to modeling and optimization.

@@FaitelTech Thank you for your answer! I'm asking you this because I'm looking to use as little softwares as possible and use only Houdini so I improve much faster. I didn't think about using it for modeling, but since it looks like it works really well with the plugin, I'm actually going too.

Добро, я запишу видео.

Да, извините, большинство из этих видео записано в контексте для ребят из дискорда. Какие проблемы у вас возникли?

@@FaitelTech А, тогда понятно для кого. Просто ютьюб вывел в рекомендации, а для тутора это очень странно оказалось.

Thank you! You can already download SD Image Roll and SD Shapes to Points from the repo: github.com/Faitel/MLOPsLC

Hello Howard, I am eager to create a tutorial but unfortunately, I am fully occupied with a project for the next two weeks. Would it be possible for you to provide a more comprehensive overview of your interests and the challenges you have encountered? I may be able to assist you sooner and more specifically.

Everything is possible if you're brave enough😄

Thank you! It's menu from plugin Modeler 2022 but the interface was completely redone in Modeler 2023: alexeyvanzhula.gumroad.com/l/modeler2023

Sorry for the late reply. Glad the video was helpful. The shelf layout is still under development. I’ll post link here when it will be available.

It's true what Houdini isn't widely used for direct modeling and most of the people who do have experience with modeling in other software packages but that doesn't mean you can't. First of all, you can get pretty far with just default Houdini's set of tools like Igor Severic did here: www.sidefx.com/tutorials/project-cabin-direct-modelling/ or Houdini Element here: ua-cam.com/video/cRnpXWUzBvU/v-deo.html . Also, whatever you do in Houdini and especially when you want to do direct or procedural modeling, a strong understanding of SOP context and its components is required, so there are www.hipflask.how/ courses which widely covers the field. Second, you can purchase Modeler 2023 plugin and get set of tools and workspace made with direct modeling in mind: ua-cam.com/video/Y_SI5MMWkYs/v-deo.html and join its discord community of Houdini modeling enthusiasts: discord.gg/gwE8TryC . And if you have interest in direct modeling in general, learn more about its theory and common subdivisions modeling practices, there is John Dickinson's course MILG11 johndickinson.gumroad.com/l/eLGkh made for Cinema 4D but the knowledge is mostly software agnostic.

@@FaitelTech do you have another link for the DIscord? That one doesn't work. Thank you!

@@thehomedepothouseplant9986 discord.gg/UHgRk99pzt

Hello, thanks! My discord and email are written in the bottom left of the video.