Lost Marble’s Moho 14.4 update, charmingly titled “Side Quest,” arrives as a free upgrade for all Moho 14 owners. It’s more than a maintenance patch, it’s an expansion that connects 2D animation workflows to 3D pipelines and game engines via native glTF/GLB export.

The new exporter allows animators to bring Moho rigs, skeletons, and animation directly into Unreal Engine, Unity, or Godot, as well as 3D applications such as Blender. For production artists juggling both dimensions, this means Moho’s bone-based 2D animation can now live inside fully interactive environments without tedious conversion work.

The glTF Portal

At the centre of the Side Quest update lies the new glTF/GLB exporter. It preserves key rigging data such as bones, mainline animation, Smart Bones, Dynamic Bones, and Actions and transforms them into reusable animation clips ready for playback inside real-time engines. Each Action created in Moho, such as walk, jump or idle, is saved into the glTF file as independent animation data, making it possible to drive game character behaviour through player input.

Smart Bones, Moho’s custom rigging controls that manipulate complex deformations, now translate into Morph Targets (also known as Shape Keys) on export. This mapping ensures expressive deformations, like mouth shapes or limb warps, survive the jump into 3D software. Lost Marble notes that while not all Moho features are supported by glTF, the exporter retains the most essential rigging and animation data for cross-environment use.

Smart Bones Go 360°

Beyond the export layer, Moho 14.4 introduces a new 360° Smart Bone type. If a bone rotation reaches exactly 360 degrees, Moho automatically identifies it as a full-rotation Smart Bone, enabling artists to control the appearance of artwork at any viewing angle. That means animators can now smoothly rotate limbs, faces, or even entire character poses around a full turn, which is ideal for complex movements such as spins or walk cycles that require full rotational control.

Particles Behaving Better

Particle simulation sees a notable upgrade. Particles can now follow curve-based source layers, creating organic arrangements such as hair strands, tentacles, or tails. A new playback offset lets artists stagger particle timing to produce natural wave-like motion. Distribution has been improved to eliminate uneven gaps, and Moho now allows particle interpolation at variable frame intervals (1s, 2s, 3s, or custom) supporting both traditional and stylised timing.

Workflow Polishing and Quality-of-Life Fixes

The Side Quest update doesn’t stop at headline features. Lost Marble has tweaked dozens of small but significant details that make day-to-day work smoother. Freehand strokes are now more stable and visually consistent, with an improved smoothing algorithm and consolidated undo steps for pen input. Layer caching has been refined, reducing redraw lag when editing complex scenes. Vector layers now cache even when influenced by bones, and timeline zoom levels persist between sessions. Layer selection, particle rendering on macOS, and PSD import behaviour have all been cleaned up. Undo behaviour, motion blur logic, and bone visibility in Vitruvian rigs have been refined, while Moho’s scripting API gains new Lua functions for developers. New device presets have also been added for TourBox and Xencelabs Quick Keys, expanding the software’s hardware compatibility out of the box.

Styling, Simplified

A small but welcome improvement: users can now create a new Style directly from an existing shape. Instead of defining styles before drawing, artists can select any shape and automatically group others sharing its visual properties like fill, line, width, and brush. This makes it trivial to recolour or globally adjust line attributes across entire characters, without manual layer-hunting.

A Long List of Fixes, and a Few Quiet Triumphs

The update also addresses long-standing quirks such as accidental timeline drags, over-sensitive pen inputs, and incorrect particle rendering on macOS. Multi-threaded rendering now supports higher resolutions, and Liquid Shapes gain an obscure but potentially life-saving correction script.

These aren’t glamorous additions, but for working animators, stability improvements often matter more than new toys and Lost Marble’s changelog reads like a response to real production feedback.

Pricing and Availability

Moho Pro 14 retails for USD $399.99, and Moho Debut 14 for USD $59.99. The Side Quest 14.4 update is free for all Moho 14 owners, available now through the Lost Marble website or via Moho’s in-app update check. As always, users should verify compatibility within their specific production pipelines before deploying it in live projects.

Lost Marble Official Announcement: Moho 14.4 Side Quest

The post Moho 14.4 Goes on a Side Quest: 2D Meets Game Engines first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

Faceform has shipped Wrap 2025.9 and Wrap4D 2025.9. Both applications update monthly, and the September builds are now available. Pricing remains unchanged: Wrap subscriptions start at $50/month or $360/year, while Wrap4D is available at $90/month or $600/year.

Both tools are widely used for 3D scan processing. Wrap targets single-frame retopology and scan alignment, while Wrap4D extends the process to temporal data such as performance capture. The new releases focus more on bug fixes, interface tweaks, and incremental improvements than on large new features.

Wrap 2025.9: quicker, lighter, more predictable

The Wrap 2025.9 changelog shows that the main addition is a Viewport panel. This panel displays the UV map of any geometry node directly in the application, allowing artists to verify unwraps without exporting to an external DCC. Another significant change is the option for Instant Project Open. By enabling this in the preferences, heavy compute nodes are disabled when loading a project. This reduces startup time for large projects. Nodes can then be selectively re-enabled when needed, giving artists more control over system resources.

The timeline has gained new playback shortcuts, with . and , keys for forward and backward playback and additional modifiers for fine control. Artists can also adjust the timeline range to match the selection of nodes with a single command, or modify playback frame rate with a dedicated shortcut. Python scripting stability has been improved. Recursive crashes in the nodeParamValue function have been fixed, scripts now properly update when nodes are renamed, and errors caused by re-running identical scripts with different results have been addressed.

In the Node Graph, lazy evaluation has been introduced. Nodes are not computed until they are actively needed, whether selected, displayed, or designated as save nodes. This prevents unnecessary computation and improves responsiveness in large graphs. Nodes with 2D image outputs can also be pinned for persistent viewing in the 2D viewport. A new colour copy function allows node colours to be quickly cloned across the graph, which can aid complex project organisation.

Drag-and-drop behaviour has been refined. Multiple files dropped simultaneously will now generate multiple nodes, making it easier to set up geometry and textures together. JSON and TXT files can be dropped to spawn loader nodes such as LoadCamera or LoadVertexMask. Shortcuts allow timeline or frame-rate adjustments during file import.

The panel system has been overhauled. Layouts can now be saved as presets, recalled later, and mirrored or re-oriented with new commands. A new Playback panel joins the UV viewport, and tab insertion has been made more predictable. Overall, Wrap 2025.9 is less about big new features than about making existing workflows lighter and more predictable.

Wrap4D 2025.9: the heavy-duty update

While Wrap gains convenience tools, Wrap4D 2025.9 receives the larger changelog. The only truly new node is ConvertNormalMap, which transfers normal maps between world and tangent space. This is a small but useful addition for pipelines that move assets between DCCs and game engines, where mismatched tangent bases often cause shading artefacts. The timeline system has gained new playback and adjustment commands, mirroring Wrap, but also allows adjusting FPS for video input nodes directly. This gives finer control when working with frame sequences from different capture devices.

The Node Graph benefits from lazy evaluation, reducing unnecessary computation in complex setups. A new pinning feature allows any node with image outputs to be displayed persistently in the 2D viewport. Error reporting has been refined: users can now select the exact error source by clicking icons in the graph. Colour coding has been made easier with a copy-and-apply system for node colours, and node grouping behaviour has been stabilised, fixing issues where moving nodes too quickly caused errors.

Python scripting is more reliable. A new interpretPath function translates path variables to absolute values, making project scripts more portable across machines. Fixes include proper handling of renamed nodes, consistent recomputation when rerunning scripts, and corrected variable cleanup. The user interface gains the same Viewport UV panel as Wrap, plus a multiple-input reorder widget in the node editor. Instant Project Open is also available here, letting artists delay heavy computations until needed. Texture previews in the 3D and 2D viewports no longer suffer from half-pixel offsets, and colour pasting now respects linear values when importing hex codes.

File handling is more flexible. Dropping multiple files spawns multiple nodes, while JSON and TXT files generate appropriate loader nodes. Shortcuts for adjusting FPS or timeline values are available during import. Gallery nodes now spawn within viewport bounds instead of off-screen, and group resizing has been corrected. Panel management has been expanded. Users can save and recall custom layouts, mirror panel arrangements, and insert tabs more precisely. Two new panels, UV and Playback, extend monitoring options.

The command line now supports a -t flag that prints the computation time for each node, sorted and expressed as percentages, which helps to identify bottlenecks. A -w flag prints absolute file paths written to disk. Return codes on computation failure have been corrected, timestamps in verbose mode are accurate, and crash conditions in WrapServer have been resolved.

Batch processing via the Rush GUI now tracks selected projects correctly and respects project priorities. The gallery assets have also been adjusted to anatomical scales, with eyeballs corrected to 24.3 mm and iris loops updated to 12 mm, eliminating previous inconsistencies in facial rigging setups.

Node refinements everywhere

Wrap4D 2025.9 makes dozens of small node-level adjustments. The Brush and VertexMask nodes now inherit wireframe and display settings from their inputs, reducing duplication. The ImageFacialWrapping node gains illumination control parameters and separate semantic weights for lips, eyes, and other regions, giving finer control of fit. The LoadImage node now supports RAW formats including Sony ARW and DNG. MergeImage gains Subtract and Divide blend modes, while its visual editor now shows image borders for debugging.

Camera behaviour has been corrected, preventing crashes when switching to invalid reference images, and fitting to view is now reliable even with hidden geometries. ExtrapolateImage introduces a new push-pull method. BlendRigToGeom adds a blending type parameter, letting users choose between fast or precise computation. FitBlendRig handles empty control strings more gracefully and now reports symmetry errors explicitly. FixSymmetry applies transforms correctly to input geometry, while the LoadBlendRig and corrective rig nodes fix duplicate naming and corrective loading errors.

SelectPolygons now supports direct keyboard shortcuts for grow and shrink operations, speeding up selection tasks. SelectPoints and SelectSmallComponents both receive reliability fixes, while ScreenPointsToImage now sorts points correctly when drawing. Additional refinements include support for two-channel images treated as colour plus alpha, corrected MSAA handling in Render, proper linearisation in TextImage, and crash fixes across many nodes. The PointDetector node now accepts optional initial target points to guide detection, and LoadFrameFromVideo no longer confuses file paths when aborted mid-operation.

Incremental but valuable

Neither Wrap 2025.9 nor Wrap4D 2025.9 introduces sweeping new features. Instead, Faceform has delivered dozens of corrections and small refinements that directly target reliability in production use. For facilities processing large batches of scans or capture data, reduced crash frequency and faster project opening may prove as valuable as new headline tools.

Test before you trust

As always, Faceform’s documentation lists what should now work, not what will always work. Production artists are advised to test thoroughly in non-critical contexts before deploying them in live pipelines.

The post Wrap 2025.9 and Wrap4D 2025.9: Faceform updates first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

Topaz Labs confirmed that from 3 October 2025, its software will no longer be sold under perpetual licences. Current perpetual customers keep their licences, but no new perpetual purchases will be possible after this date. Products affected are Topaz Gigapixel, Topaz Photo, and Topaz Video (formerly Gigapixel AI, Photo AI, and Video AI). Until the October deadline, perpetual licences remain available at $99, $199, and $299 respectively.

Subscriptions only

After the deadline, the only way to access future versions is via subscription. Topaz calls out “founding customers” with active upgrade plans. They will receive discounts on Studio subscriptions and additional access privileges. The company has not specified the size of the discounts.
Pricing announced:

• Topaz Gigapixel: $29/month or $149/year, Pro plan $499/year
• Topaz Photo: $39/month or $199/year, Pro plan $599/year
• Topaz Video: $59/month or $299/year, Pro plan $699/year

The shift mirrors the launch of Topaz Studio, a subscription-only bundle that combines all three applications.

What Studio brings

According to the official Topaz Labs forum statement, Studio includes seven products and access to over 100 AI models. Local and cloud-based image processing is bundled without extra cost. New AI models debuting inside Studio include Wonder, Standard MAX, Starlight Sharp, and Nyx XL. Models in development include Face Realism, Subject Preservation, and Mosaic.

What this means in production

For postproduction artists, the switch means one-off investments are replaced by recurring costs. Existing perpetual licences continue to function, but users will need a subscription for updates and new features.

The post Topaz drops perpetual licences, bets on subscriptions with Studio launch first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

The September 2025 update of Adobe Photoshop (version 26.11) is modest in scope but introduces a structural change: cross-app cloud projects. Alongside that, Adobe has added Firefly asset import and vector rendering for CJK text. The typography and Firefly updates are clear feature upgrades. The cloud projects, however, touch the way teams manage and share assets across applications.

From Libraries to Projects: what changed?

Creative Cloud has offered “Libraries” for years: shared collections of colours, brushes, symbols, and logos. Libraries are lightweight, designed to pass assets around quickly. But they do not handle project-level structure or permissions.

Photoshop 26.11 introduces cloud projects as containers for multiple file types and assets across Illustrator, Photoshop, and Adobe Express. Unlike Libraries, cloud projects can: Group full files (not just swatches or elements) into a single unit. They also can Assign user roles: editors can change content, while commenters can only review and annotate. So to Keep project-wide organisation, rather than loose collections of items.

In essence, a cloud project is Adobe’s attempt at a collaborative folder, integrated directly into Creative Cloud apps.

Collaboration: practical gains and limits

For freelancers and small studios, cloud projects could simplify version control. Instead of shuttling Illustrator vectors, Photoshop comps, and Express exports back and forth via email or cloud drives, everything lives in one Adobe-managed container.

For larger pipelines, the situation is less straightforward. Many VFX, post, and design studios already rely on dedicated asset management (ShotGrid, ftrack, or in-house DAMs). These systems support pathing, metadata, and automation in ways Creative Cloud has historically struggled with.

Questions remain open:

• How are conflicts handled if two editors open the same file?
• Does Adobe keep a file history or only the latest save?
• How are large binary files (e.g. layered PSDs >1 GB) treated inside projects?

Adobe has not provided detailed technical documentation on these points. Until clarified, projects should be considered convenience features rather than replacements for established pipeline tools.

Permissions: useful, but basic

Role-based access in cloud projects is limited to two tiers: edit and comment. Unlike enterprise DAM systems, there is no granular control over layers, linked assets, or export rights. For example, a user with “comment” rights cannot modify a PSD, but they can still see full-resolution assets. This may raise concerns in projects with strict confidentiality requirements. For teams needing more controlled distribution (for instance, advertising agencies under NDA), traditional controlled file-sharing may remain preferable.

The Firefly angle

The new Firefly integration is deliberately narrow. Firefly-generated images can be imported into Photoshop for refinement, but Firefly remains a standalone app. Edits are non-destructive, meaning the Firefly original is preserved while Photoshop creates a modified copy. This workflow may fit well with cloud projects. Teams could keep Firefly outputs and Photoshop refinements in the same shared project container, reducing file duplication across drives and emails. Still, without in-app generation, the workflow remains split across two applications.

CJK text: better previews, fewer headaches

Typography in Chinese, Japanese, and Korean (CJK) has long been one of Photoshop’s weaker areas. Raster-based previews made text appear jagged and inconsistent, especially when shared across machines. Photoshop 26.11 replaces this with vector-based rendering. On-screen previews should now match Illustrator’s fidelity, giving artists sharper results and reducing misalignment in multilingual layouts. In cloud projects, where documents may be passed between different Creative Cloud versions, this could cut down on errors.

Adoption strategy: cautious roll-out

Photoshop 26.11 is available via Creative Cloud subscriptions at no extra cost. Version rollout may vary by OS and region. For individual users, cloud projects and Firefly import may improve daily efficiency. For studios, adoption should be cautious. Cross-app projects may not play well with custom pipeline scripts or local render automation.

As always: test before deploying to production.

The post Photoshop 26.11: Firefly hops in, CJK text sharpens up first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

Kyno has returned with version 1.9, the first significant update in years. Now under the ownership of Signiant, the software shows that one of the most practical video workflow tools is still maintained.

Apple Silicon done properly

The most obvious change is support for Apple Silicon processors. Kyno now runs natively on M1 and M2 hardware, avoiding the translation overhead of Rosetta. However, users must download the dedicated installer from the vendor’s site, since the auto-updater still delivers only the Intel build.

A darker, sharper interface

The update also refreshes the interface. Kyno now uses a darker look with higher contrast in selection colours, a larger system font for better readability, and smoother scrolling on macOS. Buttons have moved: the drilldown function is now placed in the general toolbar, while activity progress is linked directly from the status bar. These are minor shifts, but they help when navigating large collections of footage under time pressure.

Staying current with RAW

For camera workflows, the Blackmagic RAW and RED RAW SDKs have been updated to match the latest camera models and firmware. For editors and DITs handling recent cameras, this alone makes the update worth considering.

Metadata and shared storage

The new release introduces volume settings that allow users to store technical metadata and thumbnails directly within a shared file system. Indexing behaviour can now be optimised for shared storage environments, which should reduce inconsistencies when multiple workstations access the same material.

Background analysis and smarter proxies

Pre-analysis of media now runs in the background without blocking the interface. Proxy workflows gain smarter presets that automatically create batches of proxies with scaling. For teams managing large projects, these refinements mean fewer interruptions and more predictable file handling.

Small details that matter

Some changes are less visible but still useful. A new context action, “Show in Enclosing Folder,” helps locate media quickly when working in drilldown mode. The refresh button has been improved; holding Shift while clicking forces a hard refresh, useful if the interface shows inconsistent file states. These small adjustments address long-standing workflow irritations. Not all features survived the update. Integration with Frame.io and Archiware P5 has been removed.

Fixes and maintenance

Security updates and a series of bug fixes round off the release. Copy-paste shortcuts work again, macOS users can now send to Adobe Premiere Pro 2025 without issue, and Windows users benefit from a more reliable update installer. Stability fixes cover problems ranging from broken keyboard shortcuts to folders being incorrectly marked as offline after macOS upgrades.

Still a favourite, cautiously

Kyno remains a favourite among professionals for previewing, transcoding and metadata management. Version 1.9 is no revolution, but it proves the tool is alive and slowly evolving. For users who have invested in Kyno as part of their daily workflow, that alone is significant.

We will be speaking with the developers at IBC and will report back on their plans for the software. Until then, consider this release a maintenance update with a few useful surprises. As always, test before deploying in production.

The post Kyno Update: Still Alive, Still Useful first appeared on DIGITAL PRODUCTION and was written by Bela Beier.