davpe

hmm... channels in the buffer don't show up for me. odd. SOLVED: You have to enable Export>Use Render Elements on V-ray ROP

and I'm back once again! please, good people, how can you export standard aovs? like diffuse contribution, reflections... that kind of stuff? there is something called Render elements on the ROP node, and a button that generates another network in ROP context but I have no idea how to use it and how to display it in Render view or vray framebuffer? there is zero info on this in docs how that's supposed to work. thanks for any hints. D.

i'm not a freelancer, but here are some thoughts: from social platforms, Zerply seems to work pretty well for freelancers - i'm getting job offers from time to time even if I'm not actively seeking one. linked-in works too. if you want a job from a studio, you'll need a good looking showreel and a history of projects/jobs. you are unlikely to land a remote job for a studio without some sort of good looking "professional background" for this reason (not only) i'd recommend you to get a few jobs in studios first. that will make your life significantly easier bcs you'll make friends there, they'll seee you're doing good, you'll see what's the style of work in studio, and you'll have a better understanding how to deliver your work when working remotely. i know a few freelancers who can do gorgeous looking commercial or print on their own, but would never be able to work for a studio, bcs they developed a unique style work and software environment over the years, which no studio is compatible with. hope that helps. D.

as far as i know, producing assets and/or tools as a remote freelancer is not very common. this is something studio TD crew typically do, as they can test everything directly in the studio environment and pipeline, they are expected to maintain their tools over longer time period, and they don't have to care about NDA policies or remote data trasnfers. Having said that, I personally have gotten a few requests for building procedural systems in last 6 months, so there certainly is some market for that. But establishing yourself as a freelancer specialized in building Houdini tools/HDAs would be very hard endeavor I think. this is just an experience of a single person thou. you can always give it a shot and have a backup plan for case it doesn't work cheers, D.

in wrangle you just use node input numbers, instead of a network path: detail(0, 'radii', 1);

hi, a bit late to the party, but here is my take on this. i do majority of my texturing directly in Mantra shader netwroks and it's great. in terms of procedural stuff, Mantra by far outperforms any other renderer. Redshift is great for many reasons, but it's not for it's texturing capabilities, and I doubt it ever will be - nobody seems to be adding texturing tools to their renderers, beyond the most basic stuff like perlin noise or a curvature shader. Redshift, Vray, Octance have very limited texture handling capabilities, Arnold also very limited but offers a few very handy nodes, Renderman is the only 3rd party renderer that offers sort of an advanced procedural toolkit - but even thou I think it's a way behind Mantra in this respect. So Mantra for me works basically like a texturing application that does all the heavy lifting, collects and manipulates all the data from other applications, assembling the final result and then, optionaly, bakes the stuff out to bitmaps for use with 3rd party renderers. Love it. I very much hope we'll be able to use the same (or updated) workflows in Karma soon. cheers, D.

well, "switching" materials at render time is not possible I'm afraid. switching colors and wiring it into a shader parameter is something completely different. for you to be able to render different materials on objects, you either have to blend layers as in example i sent, or you have to assign materials using shop_materialpath attribute (preferable in most cases, i think) if you only want to have different shader parameters per object, but otherwise having the same general qualities (like 100 different flowers, or book covers or whatever), the best way is to use attribs to drive these things per primitive, using a single material/shader. then you can also use material stylesheets. options are plenty...

what you want to do is to create different materials first and then generate a shop_material path attribute inside your loop, that's pointing to them. alternatively you can choose a reverse approach where you have a single material and for each piece you generate different parameteres (like texture path) that are then piped into the material. depends what you need to do. the setup i see at the picture doesn't make much sense. what you are trying to do is to have multiple names for a single node, which can't be done. even if it somehow worked, you'd end up with having the same material for all pieces, only with a different name. so at the end of the day it'd be the same like having the same material on all pieces. cheers, D.

hey, there is a few issues with your shader setup. 1. you can't just use a switch to blend shaders. there is a layer mix VOP to do that. 2. if you do a custom stuff like this, you need to use Layer output nad plug it into Compute lighting VOP and Output surface variables VOP (this is otherwise embedded in the Principled shader if you use it's Surface output). 3. depends on what are u trying to achieve but if it's just a texture variation you can do this much more simply and efficiently with only one Principled VOP and string attribute that can overwrite a texture path per primitive. cheers. Cards_01_fix.hiplc

and what about, after removing non-quads, just testing (in loop per prim) whether the perimeter of a primitive (using measure SOP) is equal to a distance between any two points multiplied by 4? does it make sense?

it's in top menu under Render > Update textures. If you overwriting texture files you have to push that for textures to get updated.

just pick 20 pieces you like and instance them if you don't need to literally see how the original shape breaks. that'll be the biggest performance boost, even if you use your current geo.

ok i opened the file, and after trying to cache the stuff out for like 20 mins, i gave up with a conclusion that everything in you scene is unjustifiably heavy. technically its ok but just impossibly slow to work with. for that kind of shot shot with orbiting debris you can safely go with 10x lighter geometry and use displacement at render time. also, use instances for this, there's no need to have 5500 unique rocks. scene like this can run almost realtime in opengl if you spend some time on optimizing stuff. that would be my first advice. then try to get your finished sim into a fresh scene and try to render with default settings of everything and see how it goes. also check the polycount before rendering and do all the other things I suggested before cheers, D.

i don't see the file attached...

it's hard to advice anything without inspecting the hip file, but scene like what i see definitely shouldn't be overly heavy for even a mid-range pc. some basic debug tips: - do you experience lower memory footprint with other renderers? - i assume your sim is cached to disk and you only render that - is your 5500 pieces made of instaces or is it each unique packed geo? - if theyre unique, packed prims don't really offer any benefits. - if you unpack the geometry before rendering, is it eating the same amount of memory? - you can read a lot of info from console verbose if you enable it on Mantra node (like how much memory different things cost) - is memory consumption lower if you render without textures/shaders? - have you tried to load the cache into a fresh scene and render from there with default mantra settings? - unlikely, but don't you have some crazypants large scene bounding box? (i.e. any stray geo floating very far from where you want it to be) hope that helps. cheers. D.