[OP] My Houdini 19 Wish List

Houdini 19 is almost here and before we find out anything about it I thought it'd be fun to share with you guys some of the improvements on my Houdini 19 wishlist.
Most of these are of course just personal wishes but I'd love to know if you guys agree, disagree, and what are your wishes for an upcoming Houdini update. Let me know in the comments below!

[@RMAFIRE]

Amazing, interesting, and joyful✨✔

[@MattPuchala]

I think they need to talk to Matt too 😁

[@RightBrainedTutorials]

Have you had any contact with SideFX? If not, they really need to get on the ball! (with you and with Henry from MOPs as well).

[@FaddyVFX]

Looks amazing, alot to learn in this software, i think your video is a good resource for us to learn from, keep up the good work

[@wijobu7385]

I find myself using Kinefx a lot lately, and I want to see the elimination of the match pose node (for retargeting animation data). If the re-orient joint node worked properly, we could theoretically just map points and let Houdini automatically (under the hood) copy the exact position and orientation of each joint to the one it targets. This removes the need for this work around I have to do - hiding a tpose on frame 0, and using the motionclip sequence to transition from tpose to the actual animation. If we can eliminate the need to match pose, we can create much more procedural animation pipelines that simply plug and play new animation data to the same targeted rig.

[@pattyfinnpadidiver]

Awesome to see you 😀👍👍👍

[@RightBrainedTutorials]

HA...thank you Patty, it's been a long time I hope you're doing well.

[@philippzakrzewski8180]

Amen! 😄

12:23 The good thing with Houdini is we can do it ourselves with HDAs

[@BrianHanke]

Nice list! I'm pinning all my hopes on Karma, would love to see it reach Maya/Arnold level power. TOPs makes my eyes glaze over. Everything in Houdini is complicated but... TOPS... man...

[@leestripp7387]

If they do simplify things, all good. BUT not if that dumbs down the flexibility and power Houdini has now. Copy to Points does need some extra love. I love LOPs so far, as its built around USD. They have already said that will get a lot of improvements. I wouldn't be suprised if we see them in 19. We can live in hope :D

[@RightBrainedTutorials]

I agree...more streamlining but no dumbing down.

[@lmahesh26]

Instant click

[@lewistaylorFX]

You didn't really clarify what is was about the animation editor you are fighting. The copy to points example you used, in H18.5 you have scatter and align SOP, that creates the orient, pscale, etc needs to do the randomizing. You also have attribute randomize in H18.5 to further vary the things you are copying, all of what you are asking for already exists in that respect. Native digesting of Qucktime, Mp4 is a bad idea. There is a memory overhead to doing it, and you know need to support every codec and configuration under the sun. Also, if the quicktime is 5000 frames long, but you are using only 100 frames, you still wear the overhead of digesting it. Regarding the pyro progress. Pretty much all of the SOP level, and DOP level solvers are working on the assumption you are caching to disk. Running things live uses a lot of memory and resources, the solvers are designed to fully utilize your CPU, so caching is a base workflow of any Artist working in houdini. Predicting the time is also near impossible, as the sim progresses the voxel count will potentially increase, collisions could get slow to calculate, a lot of variables in there. When you use the caching workflow you get a progress bar btw. File snapshot preview is a good idea, we have this at the Studio, very easy to implement. Regarding GPU openCL. H18.5 already gave you a fast pyro option like Matt's, but his stuff is really well tuned and getting better and better all the time. The thing to remember is almost all of a basic pyro is already openCL accelerated, it is only the microsolvers that don't have openCL equivalents that cause it to slow down. Axiom is great, but the base houdini pyro is still a bit more flexible in terms of colliders and forces, which I'm sure he's addressing. Take a look at the "minimal solve" to get a faster version of the standard pyro.

[@kasali2739]

some measurement tools would be good

[@JinsungB]

Super fast karma xpu...

[@carlosrivadulla8903]

when do u guess houdini19 ll realese? I feels the developement of houdini kinda slow, I think they should be bought and boosted by unreal

[@es_ina]

in the forum they are talking about late october. there was an upsy email from sidefx to some users.

[@kasali2739]

need vex function for intersection of 2 vectors. you could do a lot of useful things with that. i am Wondering why Vectors Expression Language have Not that?

[@lewistaylorFX]

you mean the intersect() vex function?

[@kasali2739]

@@lewistaylorFX no, intersect need input geo from the beginnig. You cant use that for new made vectors

[@lewistaylorFX]

@@kasali2739 but you have geometry. Even a single point. A vector can’t exist in isolation, the vector attributes need to exist on geometry. What are you trying to do?

[@kasali2739]

@@lewistaylorFX thank you for answering. I am looking at vector abstract, as direction, no need for Geometry actually, you can start with detail wrangle and create geometry out of nothing. Meanwhile I have written such a function myself, but its not perfect (some edge cases) Can be usefull for "freestyle" cutting, creeping on geometry (only VEX) and many other things.

[@lewistaylorFX]

@@kasali2739 sounds interesting. Do you have a link to the type of effect you are producing with this? There may be another approach.