[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/3/ - 3DCG

[Advertise on 4chan]

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • There are 36 posters in this thread.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: Screenshot.png (367 KB, 792x438)
367 KB
367 KB PNG
The other thread is too far down the page now.

Get started for free. Houdini is fun and easy:
https://www.sidefx.com/products/houdini-apprentice/
>>
> Can't even export you mesh
In the trash it goes
>>
So I've been getting into RBD lately and I have a question for more experienced users:
the material fracture node is relatively new and the fractures look pretty decent to me, but I'm not sure how much I should be leaning on it.

Do people mostly stick with the material fracture node now or are they still relying on their own fracturing set ups (for speed or quality reasons?). How much time should I invest learning about manual fracturing workflows?
>>
>>891220
Learn the manual way (using voronoi / boolean / etc.). It will help you understand the principles and techniques of doing fracturing / RBD, rather than knowing which button to press.

Once you understand how to build the basics from the ground up, it's perfectly fine to use material fracture as a shortcut.

Same goes for any of those huge complicated new nodes that SESI has added over the years (Vellum solver, SOP level pyro solver, etc.)
>>
>>891111
See if this works, it did the last time I used apprentice: Right click the sop you want to export from, choose save as.
>>
File: vellum.jpg (183 KB, 2204x744)
183 KB
183 KB JPG
>>891370
yeah bro haha who doesn't roll their own xpbd solver? vellum? pssshh amateurs..
>>
>>891414
i think he was talking about preset dopnets that are in sops these days, not the solver itself.
>>
>>891414
I don't mean you need to rebuild the actual solver from scratch, but try to do the basics manually first. In case of vellum, use the vellum solver in DOPs instead of the SOP wrapper to understand what kind of geo & attributes are needed.

Problem with "only" knowing those giant SOP level wrappers is that once you run into an issue, you're basically fucked.

I guess it also depends on what your goal is. FX TD? Then knowing all the low-level stuff helps a lot or else you're gonna be a monkey pulling levers for a long time (at least in bigger studios).
I like to think of FX as creative problem solving. The better you know the simple building blocks, the more flexible you are in coming up with a solution.
>>
>>891370
>>891428
thenks men. don't know why i've avoided rbd for as long as i have, but just sucking it up and learning it now. i usually set up dopnets for vellum or pyro myself, so not worried about dopland. got any tips for setting up attributes early that make managing constraints etc easier down the line? is part_name and material_type usually enough?
>>
>>891102
Why doesn't Houdini have edge attributes? I've always found this omission very strange. You could use them for edge weights for bevels and subdiv.
>>
>>891446
Because edges aren't a building block of geometry. Every element in houdini is a building block. Points define the 'skeleton' for lack of a better word, vertices allow for UV representation of an object (why a single point can be at any number of places on the UV space), faces define how the points are woven together, and prims define the individual objects.
You don't need edges to represent the geometry at all.

Even in subdiv, for example OpenSubdiv doesn't use edges as far as I know. All the creasing values are stored on the vertices. The workflow where you select an edge loop and give it a creasing value for example in max with the crease set modifier, it's just a user friendly way of working, but you're setting the attributes to the vertices internally.
>>
>>891479
BTW I'm not against the idea of them adding edges. The problem is that they seem to be pretty much convinced they aren't needed. I get where they're coming from, but I'm not sure if they won't end up adding them somewhere down the road, especially if they want to make Houdini accessible to traditional modelers. It's a necessity vs user friendliness situation.
>>
>>891444
Not sure I understand what you mean. But in terms of how I manage my stuff, I like to keep a separate piece attrib for each fracture. So "piece_01" would be unique for each piece of my initial fracture. I then fracture each of those pieces, assigning a unique "piece_02" attrib and so on. Then depending on what you are doing, you can easily construct an attrib to iterate over. For example, wanna create constrains between the big pieces? Use
name = piece_01

Wanna add clustering to the second fracture? use
name = piece_01 + "_" + piece_02
while using piece_01 to iterate.

Not sure if that answers you question, I might have misunderstood it?
>>
>>891479
Does that mean the edge weight in Houdini is an average of its point weights? I haven't tried it in Houdini, but in Blender vertex and edge weights clearly have different effects, and you can't emulate the later using vertex weights.
>>
>>891535
In Houdini speak points and vertices are two different things. Crease weights set on edges are stored in vertices but not points. If a crease is created on a point group I bet it behaves exactly like Blender "vertex" creasing.
>>
>>891529
i was thinking more about setting material properties and constraint properties in big easy batches down the line. so tag material type early on as a point attrib and it flows through into constraints and proxy geo for easy adjustment pre-sim. t bh i'm still very early into rbd so it'll take me a few projects to just organise my thoughts and come up with a decent workflow. tyvm.
>>
File: hi.jpg (68 KB, 777x788)
68 KB
68 KB JPG
Gentlemen, post more research and development.
>>
File: aEM7sea.png (189 KB, 620x501)
189 KB
189 KB PNG
i want these big box things to be not colliding with anything and be constrained to the car. i added collisionignore attribute and that kinda worked without constraints, but as soon as i glue those boxes to the car, they start collide with everything (car, ground) again
any help?
>>
>>893120
Is it necessary for the boxes to be part of the sim? If not, I'd use the extract transform SOP to get the car's position and rotation and merge the boxes with the car's transform applied after the sim.
>>
>>893121
i'm using them as proxies to be able to open/close doors, so yes, they need to be the part of the simulation
>>
Does someone know if there is a "copy flipped keyframe" function? Doing it in the graph editor is taking a lot of time and I was wondering if there's something to help me animate faster
>>
has anyone downloaded that matrix project and looked at the hip files yet?
>>
File: Houdini Pasta.jpg (227 KB, 1659x1644)
227 KB
227 KB JPG
>>891102
How long does it take to learn Houdini in order to make six figures?
>>
>>893778
Depends, how good are you at math?
>>
>>891102

Why would I pay for Houdini when Blender is free and better?
>>
>>893790
You'd need to have more than 70IQ to understand why
>>
>>893792

Blender is free and it has geometry nodes just like houdini. Don't act like you're the smartest asshole in here. If you were that smart you'd stop sending money to big corporations and you'd stick to FOSS.
>>
File: 1648001530675.gif (838 KB, 440x330)
838 KB
838 KB GIF
>>893799
>>
>>893790
Efficiency and better simulations.

Until Geometry Nodes has something close to VEX, I feel like Blender procedural modelling is far behind.
>>
>>893802
>Efficiency and better simulations.

Bullshit. Blender is better at this too because it's free and it does a better job.
>>
don't reply to him.
>>
File: cry.jpg (5 KB, 350x245)
5 KB
5 KB JPG
>>893802
>>893800
>>893792
b-but it's freeeeee!!!
>>
>>893817

Yes, rent free in your head. Keep paying for inferior software faggot. Blender is quickly becoming an industry standard in most studios. It was used in Man in the High castle. Ian Hubert uses it, Tangent Animation made Next Gen and it looked way better than Big Hero 6 made on Maya. Cope seethe and dilate.
>>
>>893844
Piss off to the Blender containment thread and leave the chads alone you dumb piece of trash.
You're to retarded to even attempt a conversation in this thread.
>>
File: ocl.jpg (9 KB, 396x187)
9 KB
9 KB JPG
Is it better than VEX? Can I brick my GPU if I play with it too carelessly?
>>
>>893850

Houdini users are just like Quora users: they're a bunch of fart sniffers who think they're smart. And I'm not going to the blender thread. I'm staying here as long as you won't admit that Blender is superior to your paid crap. Deal with it.
>>
Is Modeler 2022 worth it?
>>
>>893867
>that Blender is superior to your paid crap
lmao
>>
>>894153
buy it and release it to us
>>
>>893867
what the fuck are you even trying to achieve here you retarded freegger? fuck off back to figuring out keyboard shortcuts
>>
Do I need to learn VEX to do procedural modelling?
>>
>>894210
hot take: you need to learn vex to make houdini worth using, period. good thing it's very easy to pick up and a little goes a long way.
>>
>>894219
help me learn vex, senpai. I need tuts
>>
>>894225
pick your poison:

https://www.youtube.com/watch?v=k4q1UQp4x6U

https://www.bilibili.com/video/BV13K4y1J7Fh

https://www.tokeru.com/cgwiki/index.php?title=JoyOfVex

https://www.youtube.com/playlist?list=PLzRzqTjuGIDhiXsP0hN3qBxAZ6lkVfGDI
>>
File: Sleep-Concept.jpg (135 KB, 2000x1333)
135 KB
135 KB JPG
>mfw COPS
fix it you filthy canucks
>>
File: bluedabadee.jpg (47 KB, 575x413)
47 KB
47 KB JPG
How do I turn off this blue highlighting of unshared edges? Never noticed it before, must be a new feature.
>>
File: 1639423624110.png (6 KB, 142x221)
6 KB
6 KB PNG
>>894518
it's not new, but now it can be toggled with the button in the middle of this pic. previously you had to use the viewport options
>>
>>894608
Thanks, I may have accidentally clicked that and forgot.
>>
got tired of lookin at houdini so been doin sum p5js. it's good fun; would recommend if you're already used to vex.
>>
It might be a stupid question but is there a valid usecase for making animations in Houdini for games instead of doing the traditional way in Maya or whatever?
>>
>>894898
Valid usecase = cleaning up mocap en masse
Would only work with a really good pipeline.
Wouldn't recommend for keyframed animation.. trust me.
>>
>>891111
>>891400
Use a File SOP. Set mode to Write Files. Choose destination, file name and extension, check display flag on node to save the file. This is the simplest way for OBJs, but you can save FBX and Alembic with it too.

For FBX or Alembic with more control, use a ROP FBX/Alembic node.
>>
Does anyone here know how to assign textures with mtlx? Currently attempting to use several mtlx image nodes into a standard surface as my output (in a blender sort of fashion) but I can't seem to get it to stop tiling after trying several methods.
If anyone knows a standard or good sort of way to make wrapping work I would appreciate it.
I've saved a UV from the geo/mesh but I'm not sure how to properly apply this using mtlx
>>
Houdini is for nasty chudcels who can't into 3d. Blender all the way.
>>
>>895215
Post work chud.
>>
anyone know how vellum constraints are "attached" to their corresponding geo?
i can't find anything like a name attribute.

weirdly enough if i do a copy and transform on geo and constraints after creating them, it just werks. don't get me wrong, that's a nice surprise, but why does it work?
>>
>>895850
vellum constraints use the same points as your input geo, check point numbers. there is no attachment, it just werks because it's the same geo, all the constraint specific attributes are stored in primitives which are ofc different.
>>
>>895855
oh fug, that's very clever.
>>
>>895217
This board is dead with the influx of braindead conservatives. you guys dont deserve to see anything made by anyone ever.
>>
keepin my bois fed:

r'way - advanced magic
https://www.bilibili.com/video/BV1uT4y167eW
>>
>>896166
cool thanks
can we follow along without assets?
>>
>>896192
I've only had a quick skim and the legit course itself doesn't seem to provide any assets at all - there's a video telling you to just buy props you need. The ship he's using is from turbosquid.
Plus I think its missing half the course videos, but they'll show up in a week or two
>>
>>896232
>Plus I think its missing half the course videos, but they'll show up in a week or two
Actually, I think I'm wrong about that
>>
File: 1649783703790.gif (699 KB, 160x120)
699 KB
699 KB GIF
>>896166
Many thanks, friend.
Have a great one
>>
>>896340
> that gif
Thanks for the nostalgia
>>
File: file.png (128 KB, 540x360)
128 KB
128 KB PNG
>mfw vellum sack full of other vellum objects
aaaaahhhh go faster
>>
Any recommended webinars or shows where the tutor actually tries to teach something instead of flexing his skills like a faggot?
>>
>>896427
Applied Houdini.
>>
>>896427
webinars are for flexing, tutorials are for teaching.
>>
>oh boy a houdini thread, theres gonna be some cool WIPs to look at it in here.
>>
>>896434
nobody is going to post their hard work and node setups on 4chan of all places. Get a grip.
>>
>>896435
sir please post mesh developability implementation in opencl
>>
File: headshot_hover.png (214 KB, 400x400)
214 KB
214 KB PNG
>>896439
no. :^)
>>
>>896427
houdini simon and official houdini youtube channel are pretty good. I don't think I've had to look elsewhere for anything in 2+ years of using houdini
>>
>>891102
Why are women such bosses in Houdini?
https://youtu.be/9MEs4cezPGk?t=595

Look at her vexing the fuck outta this battle. How can I learn this power?
>>
File: 368991826478.jpg (95 KB, 682x790)
95 KB
95 KB JPG
Does this look right for a minimum viable manual convex decomposition set up with copies or am I complicating things?
>>
>>891102
anon, requesting
ODTOOLS
MOPS+
please
>>
>>897015
You going to make Henry sad :(
He swings by sometimes
>>
Neato, Steve knipping did the new project titan tutorial. Basically a free applied Houdini

https://youtube.com/playlist?list=PLXNFA1EysfYkQeOKNeETuaiR3zw7sFFLE
>>
Is it a good idea to start making HDAs for Unreal Engine 5 yet or should I stick with UE4? I learned making them with UE4 but most of my HDAs are very basic procedural assets. Now that I'm tackling serious stuff and want to start making a prototype for my game, I was wondering if it was safe to start using UE5. Any experience someone can share?
>>
How does it feel knowing SideFX and Houdini will inevitably be acquired by Autodesk?
>>
>>897185
SideFX is a private company. Fuck off.
What is it with you faggot, shitting up this board? Kill yourself.
>>
>>897384
>SideFX is a private company
So what you stupid idiot? You don't think private companies can be bought by larger corporations? How stupid are you exactly?
>>
File: houdini34.jpg (516 KB, 1544x1008)
516 KB
516 KB JPG
Beginner here.

Can someone please explain to me why this subdivide node isn't working?
>>
>>898333
Check revolve is outputting polygons by checking surface and primitive type
https://www.sidefx.com/docs/houdini/nodes/sop/revolve.html
>>
>>898333
this >>898337
either your revolve is set to output "Bilinear Mesh" (or for some reason it does automatically) or there's a crease attribute (even though there shouldn't be, since the crease sop is set to bypass)
can't re-create the issue otherwise.
also reverse cross sections in the revolve so you get prims flipped the right way.
>>
What is the difference between a point and a vertice?
>>
>>898352
https://www.sidefx.com/docs/houdini/model/points.html
>>
>>891102
how long does it take before I can understand and apply this to my work
https://www.youtube.com/watch?v=ysujwGvr2_8
https://www.youtube.com/watch?v=Xw6i9gFtsgo
>>
>>898369
3+ years
-1 year if experienced 3d before houdini
-1 year if some programming experience
>>
File: 195725858470.jpg (199 KB, 569x567)
199 KB
199 KB JPG
https://mega.nz/file/svgggAoS#GMqn9oPszeGIqAwdyj9sgX3wMMzr8URxXGsNiAz2MPE

can someone have a look at this rbd emission set up and tell me where this weird offset is coming from?

i need precise collisions from the emission objects so there's a convex decomp step that might be causing the centroids to shift?
>>
>>898752
If the "center of mass" method gets the centroid by nonweighted averaging of point positions (docs say "like the bullet solver" which isn't helpful) then it is sensitive to topology.

Tried using the Bound SOP? If you calculate the centroid from the bounding box then topology doesn't affect anything. Extract transform and Match size probably have a slightly different idea where the centroid is so it's better to only use one method throughout.
>>
>>898756
lol, just tried a bunch of different things that didn't work.

turns out the problem was happening when i rotated the letters.
the centroids shifted off the origin.
fixed that and the standard dopimport points method works fine now.

ta for looking into it.
>>
File: flat_n.jpg (28 KB, 549x264)
28 KB
28 KB JPG
How do you compute proper normals on meshes that have perfectly flat folds? This is what the standard normal sop gives on a flattened sphere.

I'm thinking an intermediate step is required, maybe spherically interpolating the normals of two adjacent faces before the final sum, but it's a coin toss if the interpolated normals point outward or inward so I dunno..

Any thoughts, papers or blogs on the topic would be great.
>>
Do you guys use Houdini's material system? Or is it generally better to outsource shading to a more dedicated software like Toolbag, Substance or Unreal?
>>
What are the main differences between Core and FX?

From what I can see on the comparison page, they look almost exactly the same for procedural modeling, except FX includes the pipelines for vfx and physics simulations.

Also, when it comes to Houdini files and digital asset formats, is there any downside to using the indie version's .hiplc and .hdalc format compared to .hip and .hda?
>>
>>899162
>meshes that have perfectly flat folds
what does this even mean
>what the standard normal sop gives on a flattened sphere
yeah, well, a flattened sphere "is still a sphere" as far as normals go, it will just have "up" or "down" normals for the "northern" and "southern" hemispheres along the scaling axis (assuming you scaled it down on an axis to flatten it) instead of a smooth gradient, which is exactly what you have there.

>>899356
if you plan to render in Toolbag or Unreal, you don't really have a choice. You don't really render in Substance.
if you mean assembling your scene in Houdini instead of, like, Maya, to render with a commercial renderer (V-Ray, Arnold, Redshift) then do your assembling in Houdini, all the way. Fuck Maya. Don't use Mantra. Fuck Mantra.

>>899509
With Core you don't have access to DOPs, but you can open a scene with DOPs in them and even run the simulation, you just can't edit the DOPs. At studios people like FX leads/seniors set up custom dopnets with exposed values in FX sessions so that all the mids and juniors have a modicum of control in their Core sessions.
Some shit like the new Pyro SOPs are just that, wrapped dopnets with exposed values for easy usage, but "less" control. Vellum SOPs are all wrapped dopnets and you can do 99% of Vellum work in SOPs.
For procedural modelling, Core is enough, though why don't you just get an Indie license which is just an FX license at a 95% discount?

hiplc and hdalc can't be used in a full commercial license like Core or FX (-lc is limited-commercial). They actually can, but the full commercial session will automatically turn into a limited commercial session.
Indie can open full commercial scenes and assets just fine.
Full commercial also means you're making more than 100k a year through your Houdini work. You aren't. People who are have licenses provided for them, even freelancers.
>>
File: flatn2.jpg (48 KB, 228x1143)
48 KB
48 KB JPG
>>899523
>what does this even mean
I could have worded that better, I meant the surface is locally bent along an edge exactly 180 degrees so the triangles are coplanar.

The problem is evident in this picture. I can scale the sphere down on an axis to a tiny fraction and the normals along the rim behave well, once the scale hits 0 the normals flip out. There's got to be some way to avoid this degenerate case but I don't know what keywords to search, obviously no standard weighing scheme works.
>>
>>899536
>there's got to be some way to avoid this degenerate case
no, there isn't.
the closest you can have is have the Y scale be something like 0.00001 instead of 0 and the normal sop set to 180 degrees, depends on whether you absolutely need it to be 0 or you have some margin

if connected triangles are perfectly coplanar but face opposite directions then the angle between them is 180. it's either up or down.
it's in the name, coplanar. they're basically two flat planes pointing in opposite directions.

as long as the angle is something like 179.999, the normal vectors between faces define a single plane in space, however tight that is. you can easily compute an average of the two vectors that lies flat on that same plane between those two.
when they hit 180.00, they are exactly opposite one another in a straight line. a straight segment defines infinite planes. where do you place your average vector?

or, if you prefer points, imagine 1 point at the end of each vector and 1 at the origin of both.
only 3 non-aligned points define a single plane in space. 3 aligned points define infinite planes.

it's the geometric equivalent of dividing by zero. you can't avoid it, only come to a compromise.
>>
File: untitled.webm (174 KB, 500x500)
174 KB
174 KB WEBM
https://youtu.be/acVpyIVqmxM?t=1414

there's some neat VDB transitions / morphs happening here.

i'm currently looking mostly at the vdb morph sdf and a vdb advect based method for the shape morphing. are there any tools/approaches/talks/tuts i should be looking at?

is the growing stuff just vdb from particles with falloffs controlling pscale to create those streaks?

pic-related v.low rez advection test.
>>
>>899541
what kinda of money can you earn if you're able to do this?
>>
>>899545
i pay my mortgage with exposure tb h





Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.