John Petrus

Oh i forgot about most used tool by n00bs wich are help tips. When you hoover mouse over button, shelf preset ot weird paramater na parameter pane like "friction or turbulence" it shows black box with wery small fonts that quicly explains what does it do...

Who thought that such thing gona be easy readable? It should be way bigger with nice readable font and show right away without any lag. Same with documentation it takes ages before it loads... It is terrible!

Please fix it SideFX. For new people its most used stuff.

I think it should be dynamic and animatable, so first it shows short text as reminder for those who know it already but forgot, then it should expand into bigger window with more explanation perhaps with some illustrations and examples.

And final stage of this help popup should be generated by LLM wich will be context aware of connected nodes and can generate really helphull tips and techniques (fe. Using another node) AI become intelligent now we need it at least for help. Ah and this LLM should run externally and be used by entering API Key so everyne can use their favorite AI Agent for help. From open sorce fine tuned (wich SideFX could provide and run on cloud) to models like GPT-4o or Anthropic Opus.

That kind of future i want so see ASAP in mighty Houdini.

Houdini + AI is GOD MODE on steroids )))

I can live without compositing linea like rule of therd by just drowing them on monitor xD

but i cant live without camera sequencer!

And i cant live without ready to use camera rigs for differen scenarios.

Of course inncamera lines for compositing are way better that ones drawn on pc monitor xD

Moar camera tool.please! We need it for making layout an animatics and all

My list of wishes and priorytets for new release. Place matters. Houdini number do not. Its about priory of feautures:

1) ASAP Fully Finished Karma and Vulkan viewport. So they could work in tandem. Next step is to add ML to Karma renderer like it was done in Octane renerer. Where bidirectional ray traycing from light is accellereted thu ML model, alsa evenen ray distribition or ray scattering SSS is also done thru ML. Not mentioning denoicing. So this chapter have to be closed ASAP

2)ASAP finishing of MaterialX and making materials artist friendly with visial feedback on each step. It does serioussly help produce better looking materials and reduces time for making them dramaticly. Expirience artist want to see and check all ingradient of final meal not jus guess and play with numbers. feedback in middle stage of cooking materials is what Houdini missing most.

3) contine working on animation and rigging stuff including new ways on HUD redesigned controls and baziers. Making animation state of art! It is possible!

4) UI/UX with animation in mind. First of all it needs to catchup current standard in terms of aesthetics and functionality and animated feedback. Icons in maint menu should be flat, minimalistic and only use few colours for example when selected. Current icons are great but only for nodez where it it trys to explain what node do by making small ilustration in icon.

Next is parameters pane... Man its total chaos. It should be tided becouse ita all over place. I know whay it looks so shittty. Becouse makind UI entry you just expose simple line of vex code. And in this workflof there isnt artist (UI designer) involved. And thats whay it look so terrible sucks for artist. Those exposed GUI elements from VEX should be more inteligent and respect all rules that apply for good ergonomic design.

Overall interface refresh so it could be more touch friendly (there are rumours tham Apples M4 Ipad pro will have new iPAD OS that will merge macOP with iPAD OS so now it will be possible to rum Houdini on iPAD but with mature OS based on Macs. It should be flatter, less visible and less destructive. Mayby instead of tight shells there should be new idea to store such preset with better description, mayby animated illustration what doeas ot do (for example what type of flames it generate) new place for stuff from shelf is needed. It should be searchable and thanks to ai well understud its function so even it could be suggested by Houidini to use lates say texture map of explosion instead calculatong it becous it takes only 5 pixels of frame shelves should be just flat areas where you can place your own shortcuts and vex code etec. More like extra buttons not wholr presets of burning house

UI/UX for Animators. First of all speed how thousnds of bazier curves (channels) handle manipulation. In blender that was speed up imressive last update. Then there how each of this basier behaves (artist know and feel baziers from adobe illustrator, affinity designer) so theres not nee to reinvent well. Also custom proposition from Sidefx will be felt like counterituitive. So make those baziers behave like in bazier illustration software and artist will be happy. There should be skin colours, so bazier could look like in MAYA or other packages. Color is individual. Some like work with black hole background that tire eyes as hall when douing claßical animation other sont. It should be eassily changable from menue level

Keyframe UI manipulations. Just look at cascaader UO and how its solved. Animation is about constant moving keyframes of all parts of the body all around. It must be properly designes with lots of options to attrackt animators for hand animation tasks. But its doeabe

Some sort of track kind of menu, where you have block of data and can move it all aroound. Imagin it as those bars on after effect. It could be used fe. For camera seckencer and designing layot for whole movie. Or playing with motion capture datas in better way. Also other neat and unicke uscase could be applied. Just open your imagination. This kind of UI is still missing

Overall this UI have 15 years and need to catchup in every aspec. I hope it will be state of art like iPAD OS or macOS is. It will be coll if it was fully customizible so developers could i tegrate own icons styles and fu ction to it.

Then i wait for AI

First to have node similar to Nuck CAT. So it handles all the data types do the math behind the scenes, create tensaflow or whatever and split out usable model that can be reused easily. Also those models can be updated by checkpoints, and LoRA also divide model weights and other cool operation all thru nodes. So it will be general training model of houdini good for making skin and cloth also other effects on mesh, points, particles, material parameters, voxels. Basicly everything in houdini should be trainable.

Redo of compositing nodes so it could have tool from Nuke and Fusion and few others there also ML and all sort of generative ML models there too

Redo of modeling tools

Making it more collaborative with chat convestation on elements of UI and stuff. More lice collabotate tool for artist that have to do shit done on deadline and not chat with client and his stupid ideas.

Cache editore. After simulation it alwas is needed art direction, so hiding, moving objects, particles afer simulation is highly important. There are several software that do that but itll be nice to have in nativly in Houdini so it can be also used for hand editing animated characters, starting with facial expresion and ending on artistic overshoting poses wich makes it less 3ad and more hand drawn. Artist love it more than procerdualizm. Otoy have such cache modifier based on layers. In houdini those layers could be generated automaticly, lates say each layer for each muscle etc.

Ah and sorry for typos iam on medication (alcohol in my case xd) ill fix them later. Just wanted to be more relaxed and creative when writing this, sorry guys