The new speed race in 3D texture painting software
Modern 3D texture painting software is shifting from slow, CPU-heavy workflows to faster, more interactive tools that mix standardized PBR material creation, machine learning simulation, and web-based 3D tools accessed through the browser. This wave of updates is not about incremental polish; it is about cutting iteration time, reducing technical overhead, and letting artists stay in look‑dev longer and in pipeline plumbing less. Substance 3D Sampler, Mari, Nuke, Mixos, and Adonis now form a clear pattern: they treat materials, textures and soft tissue motion as problems that should feel responsive rather than laborious. The stakes are high, because whoever owns the fastest, clearest path from idea to rendered image effectively owns the next generation of production workflows.
Adobe’s Substance 3D Sampler 6.0 cements that direction by adopting OpenPBR as its default shading model, replacing older proprietary standards and aligning material authoring with the broader DCC ecosystem. Paired with new material-creation presets for common real-world surfaces, the tool now encourages artists to stop wrestling with shader compatibility and focus on look, relying on standardized inputs and exports for USD-based pipelines. That push toward OpenPBR is not neutral; it is a bet that shared material definitions will matter more than individual renderer quirks. With Substance 3D Texturing subscriptions priced at USD 24.99 (approx. RM115) per month or USD 249.99 (approx. RM1,145) per year, and the broader Substance 3D Collection at USD 59.99 (approx. RM275) per month or USD 599.99 (approx. RM2,755) per year, the message is clear: standardized, production-ready material creation is now a paid, but accessible, baseline.
Mari 8.0: smarter nodes, fewer excuses for repetitive textures
Mari 8.0’s open beta shows how traditional desktop tools are responding: not by reinventing painting, but by attacking the pain points that waste time. The expanded Node Graph receives over 50 new nodes, but the headline changes are practical, not flashy. The Hex Tile node attacks a classic texturing sin—visible repetition—by tiling and blending textures on a hexagonal grid instead of the usual rectangular one, aligning Mari with other VFX tools that now support hex tiling. Color Remap centralizes hue, saturation, exposure and gamma adjustments into a single color grading node, while Compare enables A/B wipe comparisons between node streams to evaluate material variants without duplicating graphs.
This is the kind of update that quietly pushes artists toward more disciplined, non-destructive workflows. Instead of reaching for another paint layer or texture hack, you are nudged into building clean, reusable networks that can be tested and iterated quickly. Mari’s camera navigation and project management refinements—trackball rotation and clearer indicators for archived projects—underscore the same point. The open beta status matters: Foundry has not set a firm release date, but based on previous cycles, users can expect a stable version in roughly three to six months. That window is an invitation to production teams to pressure‑test these nodes on real shows and decide whether the Node Graph finally earns its central role in their PBR material creation strategy.
Nuke 17.1: 3D Gaussian splats meet modern USD and Hydra 2.0
On the compositing side, Nuke 17.1’s beta makes a strong argument that 3D is no longer an optional bolt‑on to 2D workflows. By extending the 3D Gaussian Splatting system introduced in 17.0, Foundry is effectively saying that point‑cloud‑style scene captures are now first‑class citizens for environment building and set extensions. The new GeoSequencer node handles sequences of USD, PLY or SPLAT files, then outputs a time‑sampled USDC file ready for Nuke’s existing tools. Meanwhile, the updated SplatRender node supports 2D rendering of Direct, Point and Spot lights in the 3D system, enabling “basic relighting” of splats and turning previously static captures into malleable elements for look‑dev.
Hydra 2.0 support in the 3D viewer is the other quiet revolution. By updating to the current USD rendering framework, Nuke opens up richer display options such as previewing individual AOVs like depth and position. More importantly, Hydra 2.0 “will pave the way” for third‑party render delegates in the future, which is a clear signal that Nuke’s 3D view is meant to be a serious part of the pipeline, not just a rough preview. Non‑destructive USD export workflows, explicit sublayers, and overs for sparse overrides all serve the same agenda: stop flattening scenes, stop breaking interchange, and let departments round‑trip work reliably. Annual subscriptions underline the high‑end positioning: USD 3,839 (approx. RM17,670) for Nuke, USD 5,219 (approx. RM24,040) for NukeX, USD 6,379 (approx. RM29,410) for Nuke Studio, with Nuke Render licenses at USD 462 (approx. RM2,125) per year.
Mixos and web-based 3D tools: portable PBR texture painting
While heavyweight desktop tools chase deeper integration, Mixos argues for a different kind of speed: accessibility. Built on WebGPU, it lets artists create 4K PBR textures in a browser window, painting directly onto imported 3D models or mixing from a library of over 2,600 CC0 materials sourced from ambientCG and Poly Haven. It supports standard 3D formats such as FBX, OBJ and glTF/GLB, and can output albedo, normal, roughness, metallic, height and opacity maps, plus baked geometry-derived maps like ambient occlusion and curvature. The material‑mixing features are also exposed as a separate web app, Strata, further lowering the barrier for quick look experiments.
This is not a replacement for deep offline material-authoring tools like Substance or open‑source ArmorPaint, but it is a deliberate challenge to the assumption that meaningful texture work must happen on a workstation. Mixos runs in Chrome, Edge and Safari 18+ and is intended to work with GPUs from roughly the last six years. Using the editor is free and scenes can be saved locally; however, exported textures are watermarked, so professional use requires a paid Pro or Studio account. According to the developers, “If you often work remotely, tools like Mixos may provide a way to work on the move, or at least to make quick edits to materials when you don’t have access to your main machine”. In other words, web-based 3D tools are no longer toys—they are becoming a practical part of the texture pipeline.
Adonis 2.1 and ML-powered soft tissue simulation for animation
If texturing tools are racing toward faster iteration, character pipelines face the same pressure. Adonis 2.1 responds by folding machine learning into soft tissue simulation for Maya and Houdini, through an extended AdonisML edition. Originally built as a modular anatomy framework to fill the gap left when Ziva VFX stopped being commercially available, Adonis lets rigging artists stack physically simulated muscles, fat and skin in layers. The new AdonisML adds the ML Deformer and Smart Tissue systems, which train lightweight, interactive approximations of complex tissue setups so animators can pose characters more responsively. Smart Tissue then restores secondary motion—the “jiggle, settle and overshoot”—that can be lost in training.
This approach to machine learning simulation is honest about trade‑offs: you are not replacing high‑fidelity solves, you are staging them. Animators work with fast ML-driven deformations, then rely on Smart Tissue to bring back nuance before final renders. AdnMush, a new deformer for smoothing simulation artifacts while preserving anatomical detail, can run as a standalone post‑process or automatically inside other solvers, cutting the need for extra cleanup passes. Pricing again reveals the target market. Indie subscriptions for artists earning under £100,000 per year cost £25 per month or £250 per year per node‑locked license, while pro subscriptions run at £110 per month or £1,100 per year per node‑locked license and £150 per month or £1,500 per year for floating licenses. The message: high‑quality character motion is now expected to be both physically plausible and interactive, and studios unwilling to embrace ML-driven tools will feel that gap in animators’ productivity.
