Why this monitor?

The Asus ProArt PA27UCGE is a 27-inch 4K monitor aimed at editors, animators, designers, game developers and other people who spend too much of their lives looking at pixels. But why are we talking about that? Because, as you know, the state of the VFX industry is …. *gestures at the world* not great, so budgets for hardware are getting smaller. Let’s be thankful, the AI-Hyperscaler-datacenter-dolts aren’t onto that…. yet.

For 750€, it uses a 3840 x 2160 IPS panel, supports 10-bit color, covers 98 percent DCI-P3 and 100 percent sRGB according to Asus, and reaches up to 160 Hz. Asus also lists HDR10, HLG, VESA DisplayHDR 600, Calman Ready support, Light Illusion ColourSpace CMS integration and an integrated motorized colorimeter for self-calibration and auto-calibration.

At the time of writing, new units can be found around the 750 euro mark in Germany, while used, warehouse or buy-box offers sometimes appear in the roughly 525 to 580 euro range. At that point, the PA27UCGE becomes interesting for anyone who needs a decent everyday display or a secondary screen without spending reference-monitor money.

One warning before ordering: Asus’ ProArt line has several similar-looking models in 24, 27, and 32-inch sizes, plus higher-end variants with different resolutions, HDR specifications, and features. There is also a 6K version, if you are young, have very good eyes, or enjoy turning UI scaling into a lifestyle. Check the exact model number before clicking buy. Displays used to be “screen shows picture”. *sigh*

What Asus says it is

The PA27UCGE is specified as a 27-inch 4K IPS display with a 3840 x 2160 resolution, 178-degree viewing angle, 600 nits brightness, 10-bit color depth, 98 percent DCI-P3 and 100 percent sRGB coverage. Asus states that the monitor is factory calibrated to Delta E below 1. It also supports hardware calibration, storing calibration data in the monitor rather than only relying on an operating-system ICC profile.

The monitor includes an integrated motorized colorimeter. It supports standalone self-calibration via the monitor, ProArt hardware calibration, Calman and Light Illusion ColourSpace CMS. For multi-machine setups, that is useful because the calibration does not only live on one workstation profile, waiting to be ignored by the next application in the chain.

Unlike many calibrated displays that depend on software profiles stored and managed on the connected computer, the PA27UCGE stores its colour parameters internally in the monitor’s scaler chip. That means the calibration lives in the display, not as another small utility lurking in the background of the workstation, quietly taking up RAM. This may not fit every studio colour pipeline, especially where display management is centralised and standardised. But for smaller setups, freelancers, or the classic one-person production, internal hardware calibration is useful: fewer moving parts on the computer, fewer profile-management headaches, and less software debris between the monitor and the work.

The display also supports HDR10 and HLG, with HDR10 PQ curve options such as PQ Hard Clip, PQ Optimized and PQ Basic. In practical terms, that means HDR preview workflows are possible. It does not mean this is a HDR mastering monitor. Those are different animals, usually found in darker rooms with heavier budgets.

What we measured and saw

In our testing, Delta E stayed below 2. We did not measure further because for this use case, that was the relevant threshold. Chasing smaller numbers can be useful in specific environments, but for a general production monitor in this price range, below 2 is already where the discussion becomes less exciting than finishing actual work. The display did not produce one dramatic “behold the future” moment. That is not a criticism.

The integrated colorimeter worked well in our setup. During setup and the first few days, the motor noise of the colourimeter folding out will surprise you. Just sayin’.

Compared to an external colourimeter, it produced reasonable results. If Calman is already part of the workflow, the integrated unit may be enough for regular calibration checks. For critical work, we would still verify with an external probe. Trust is nice. Verification is what keeps production from turning into interpretive color science.

Asus also includes Light Sync, using an ambient light sensor and a backlight sensor to compensate for room and display changes. We did not test this meaningfully because our lighting setup is, technically speaking, unhinged. Any conclusion from that environment would be unfair.

HDR: present, useful, not magic

The PA27UCGE supports HDR10 and HLG, and Asus lists VESA DisplayHDR 600 certification. That makes it useful for HDR preview, editorial review, general video work and content production where HDR material needs to be checked without immediately entering high-end reference-display territory.

It should not be confused with a dedicated HDR grading monitor. For final HDR mastering, this is not the correct class of display. For editing, motion design, VFX review, game development, YouTube delivery checks, HLG preview and general HDR-aware production work, it is usable.

Refresh rate and gaming-adjacent matters

The PA27UCGE supports up to 160 Hz and a variable refresh rate. The feature worked in our testing, but we are not going to pretend to be e-sports authorities. We are not gamers, or possibly just too slow to admit it. Either way, no competitive claims will be made here.

For game development, viewport work, UI animation, fast scrolling, review work and general responsiveness, the higher refresh rate is nice – we set it at a comfortable 120Hz, and promptly forgot about it.

Connectivity and KVM

Connectivity is one of the monitor’s practical strengths. The PA27UCGE includes two HDMI 2.1 ports, DisplayPort 1.4, USB-C with DisplayPort Alt Mode and 96 W Power Delivery, a USB hub, two downstream USB-A ports, one downstream USB-C port, a headphone jack and integrated KVM switching.

The USB-C connection can carry display signal, data, audio and laptop charging over one cable. The USB hub can be used via USB-C or through a separate upstream connection when using HDMI or DisplayPort. For anyone using both a workstation and a laptop, which is nearly everyone now, this is getting rather important, because modern productivity means owning several machines and resenting them all. The integrated Auto KVM was one of the features we liked most in daily use. Once the cabling was understood, it switched reliably between workstation and laptop.

This is especially useful if the workstation is rendering, caching, conforming, exporting or otherwise doing its best impression of a very expensive space heater, while the laptop handles email, notes, admin or other survival tasks like catvideos and videocalls.

One small practical detail: the USB ports become active as soon as the monitor receives a signal and powers up. That means USB-powered accessories connected to the monitor, such as a monitor light, may also switch on immediately. This is not a bug. It is just something to know if you crawl toward the PC, undercaffeinated, and wonder why your desk has suddenly become an interrogation room. Just a random example.

Surface, reflections and physical handling

The PA27UCGE uses Asus’ LuxPixel AGLR surface, an anti-glare and low-reflection coating designed to reduce reflections without making the image look overly soft.

In use, the anti-glare behaviour is exactly where it should be. Reflections are well controlled enough that the screen remains usable even with shiny objects, windows, equipment, or other studio clutter in the background.

The coating does not make the image feel dull or smeared. That is the ideal result for anti-glare: visible in effect, invisible in irritation.

Another pleasant point is the physical handling. Despite the feature set, the monitor feels comparatively light and compact. We used it on a rather questionable monitor arm, and it still felt manageable. As usual with some VESA implementations, the mount area can feel tight depending on the arm or plate used. Not a dealbreaker, just another reminder that “standard” sometimes means “standard, but with tiny practical humiliations”.

Software and production features

Asus DisplayWidget Center allows settings to be adjusted from the desktop rather than through monitor buttons. The monitor also supports Picture-by-Picture and Picture-in-Picture modes, with selectable color modes for the PbP windows. Being able to switch quickly between modes such as Rec.709 and HLG is useful, not just in brochure workflows. The shortcut options are particularly practical. Assigning quick access to presets or safe areas makes sense for video and broadcast work.

DICOM calibration is also included. Asus states that grayscale levels are measured and adjusted to ensure DICOM compliance, which is apparently relevant to medical imaging review workflows. It may also be useful in adjacent technical visualisation contexts where grayscale behaviour matters. It does not magically turn the monitor into a certified diagnostic system, because sadly, context still exists, and we all take all the work we can, so if you go the industrial / medical route, this is helpful to have in the toolbox.

What we did not focus on

The monitor includes Low Blue Light modes. We did not test them in depth because this is not a review about surviving spreadsheets. For office work, they may be useful. For image-critical work, they are obviously not the point. The built-in speakers are fine for occasional YouTube playback, system sounds or emergency use when proper headphones are dead. Nobody should mix on them. Ever.

Verdict

If you need a Class A grading monitor, this is not it. If you need an everyday display for editing, VFX, and so on, the PA27UCGE is a very reasonable option. It does not do anything spectacularly wrong. For a reasonably priced monitor, that is the point.

Technical Notes for the Asus ProArt Display PA27UCGE

Panel: 27-inch IPS
Screen diagonal: 68.6 cm / 27 inches
Resolution: 3840 x 2160
HD type: 4K Ultra HD
Native aspect ratio: 16:9
Touchscreen: no
Display technology: LED
Panel type: IPS
Display surface: LuxPixel AGLR, anti-glare and low-reflection
Viewing angle: 178 degrees horizontal, 178 degrees vertical
Pixel pitch: 0.155 x 0.155 mm
Pixel density: 163 ppi
Visible display area: 59.7 cm horizontal, 33.6 cm vertical
Digital horizontal frequency: 31 to 255 kHz
Digital vertical frequency: 48 to 160 Hz

Brightness: 600 cd/m² typical, 600 cd/m² peak
Contrast ratio: 1000:1
Response time: 1 ms, Gray to Gray
Maximum refresh rate: 160 Hz
Variable refresh rate: yes, via VESA Adaptive Sync
NVIDIA G-SYNC: no
AMD FreeSync: no
Flicker-free technology: yes
Low Blue Light technology: yes

Color depth: 10-bit
Color-space standards: DCI-P3, sRGB
sRGB coverage: 100 percent
DCI-P3 coverage: 98 percent
Factory calibration: Delta E below 1 according to Asus
Measured in our testing: Delta E below 2

HDR support: yes
HDR technologies: HDR10, HLG Pro
Certification: VESA DisplayHDR 600

Calibration: integrated motorized colorimeter
Calibration support: self-calibration, auto-calibration, ProArt hardware calibration
Calibration software support: Calman, Light Illusion ColourSpace CMS
Light Sync: ambient light sensor and backlight sensor
DICOM calibration support: yes

Inputs and connectivity: 2x HDMI, 1x DisplayPort, USB-C with DisplayPort Alt Mode
HDMI version: 2.1
DisplayPort version: 1.4
Integrated USB hub: yes
USB hub version: USB 3.2 Gen 1
Downstream USB-A ports: 2
Downstream USB-C ports: 1
USB-C DisplayPort Alt Mode: yes
USB Power Delivery: yes
USB Power Delivery output: up to 96 W
Headphone output: yes
Headphone connector: 3.5 mm
Built-in KVM switch: yes
HDCP: yes
HDCP version: 2.3

Picture-in-Picture: yes
On-Screen Display: yes

Audio: built-in speakers
Number of speakers: 2
Speaker power: 4 W RMS
Integrated camera: no

Ergonomics: height adjustment, tilt, swivel, pivot
Height adjustment: up to 120 mm
Tilt: -5 to +23 degrees
Swivel: -30 to +30 degrees
Pivot: -90 to +90 degrees
VESA mount: yes, 100 x 100 mm
Cable lock slot: yes, Kensington

Dimensions with stand: 612.4 x 201.8 x 574.7 mm
Weight with stand: 6.4 kg
Dimensions without stand: 612.4 x 47.6 x 380.8 mm
Weight without stand: 4.5 kg

Energy efficiency class SDR: F
Energy efficiency class HDR: G
Energy consumption SDR per 1,000 hours: 26 kWh
Energy consumption HDR per 1,000 hours: 66 kWh
Power consumption in standard operation: 27 W
Power consumption standby: 0.4 W
Power consumption off: 0 W
Power consumption PowerSave: 0.5 W
AC input voltage: 100 to 240 V
AC input frequency: 50/60 Hz
Power supply type: internal
Energy efficiency scale: A to G
EPREL registration number: 2216353

Certifications: Energy Star, EPEAT Gold, TÜV Flicker-free, TÜV Low Blue Light, VESA DisplayHDR 600, Calman Ready, FSC MIX, Mac Compliance

Typical new pricing during our checks: around 749 to 750 euros
Observed used / buy-box offers: roughly 525 to 580 euros

Asus ProArt Display PA27UCGE product page

The post Asus ProArt PA27UCGE: A 4K Monitor That Mostly Refuses to Be Annoying first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

Remove goes local

Adobe Photoshop 27.7 adds an on-device AI model option for the Remove tool. After the model download, object removal can run locally without cloud processing. You can download, cancel, or remove the model whenever you want, and Photoshop checks hardware compatibility automatically before it starts. You can also switch between cloud and on-device processing.

In the on-device Remove workflow, Photoshop exposes a Device option for generative AI processing, and the model download is roughly 5 GB. If the Device option stays unavailable, the documented behavior is that the system does not meet the minimum requirements for the model. The net effect for artists is simple: you get a loacl option for Remove when you cannot, or do not want to, round-trip to the cloud.

So far no information if that still uses AI-credits in one way or another.

Boards, now with fewer copy-pastes

Photoshop 27.7 tightens integration with Firefly Boards in a few practical places. You can share any Photoshop document directly to Firefly Boards, either by creating a new board or adding to an existing one, from the Share panel and the Export menu.

Firefly Boards also appears as a destination inside the Variations panel for moodboarding and comparing AI-generated variations. When you generate image variations with Ideate in Firefly Boards, you can open those variations in a preconfigured Board layout for further exploration and iteration. New guided workflows target moodboards and visual concepts, branding and marketing assets, and social media content.

Credits and account controls move into the room

Photoshop 27.7 expands access to premium generative AI features for eligible subscribers. Existing generative credits can now be used across both standard and premium generative AI features. Photoshop also adds a Unified Account Menu in the Edit workspace, where you can view and manage account and subscription details, plus generative credits.

Small workflow fixes that add up

Save for Web gets a modernized interface intended to match the rest of Photoshop more closely. File handling and preferences get two quality-of-life toggles. You can open JPEG files as a regular layer instead of a locked background layer. You can also create new documents without a default background layer, starting with a transparent canvas.

Under the hood, the internal Metrics module is refactored to focus on heavy-use metrics, with the stated goal of reducing telemetry overhead and improving responsiveness during normal operation. Canvas HUD diagnostics move out of the canvas context area and into the View menu.

Fix-wise, Photoshop addresses a customer-reported issue where Hue and Saturation ranges did not update when selecting a prominent color preset from the Properties panel. On Windows, it also fixes an inability to launch versions 26.2 and 25.12.1 simultaneously. The Open JPEG without Background layer preference now defaults to off to avoid confusing existing users, and the Learn More URL in the generative credits quota exhausted notification is updated.

What to test before you bet a deadline on it

On-device Remove changes performance, storage, and review habits in a way that only shows up in your own shot types, textures, and delivery constraints. Test new builds and new AI paths before you roll them into production, especially if you need strict repeatability across machines and artists.

https://helpx.adobe.com/photoshop/desktop/whats-new/photoshop-on-desktop-release-notes.html

The post Adobe Photoshop 27.7 runs Remove on device first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the movie: Chrysalis is a surreal short film by Robin Lochmann blending live action with CG, built in Cinema 4D, rendered in Octane Render, and finished in After Effects, Premiere Pro, and DaVinci Resolve.

Who, what, and why this fog would not quit

DP: For the people in the back row exporting the wrong folder, who are you and what do you make?

Robin Lochmann: I am a Berlin-based filmmaker and 3D artist from Ireland. My work ranges from commercials to concert visuals, trailers and art projects and anything in between. My latest short film Chrysalis blends live action footage with digitally created characters and environments. You can find me at robinlochmann.com and linktr.ee/robin_lochmann.

DP: How did Chrysalis start, script first or vibes first?

Robin Lochmann: The film grew out of a series of visual experiments where I started to create characters and see what they looked like when integrated into real landscape footage that I shot. This is obviously the opposite of starting with a script and then creating the visuals. I was very motivated by combining the different techniques that I had learned and the resources that I had to create something on my own and push myself to make it look as seamless as possible.

DP: Where did you grab the real world plates, and what kicked it off on location?

Robin Lochmann: I was in the south of France with my Blackmagic Design camera and I just started shooting landscape shots. I find that being in a new place with a camera just creates options all of a sudden and is great for inspiration. The characters were the next step but the starting point was definitely the landscape. The project started in France back in 2023. I got a bunch of usable material before proper winter settled in. I came back about six months later, in the summertime, and did another round of shooting, this time with the drone. I chucked it into DaVinci Resolve for editing and colour correction, started organising everything by what might make sense together, and did a first pass at matching the colours of the two different cameras and seasons with very different lighting conditions.

Scope, time, and the luxury budget known as free time

DP: How long did you live with this project, and did you ever ghost it for a while?

Robin Lochmann: The project spanned over 3 years, but I did not work on it consistently. I still had to manage client work and regular life. At one point, there were about 10 months in a row where I did not open the project at all because I did not have any time. After taking longer breaks from it, it was always hard to pick it back up again, but I am stuck with it.

DP: Before the plot showed up, what did you actually know? Any hard rules, or just “please look slightly holy and slightly terrifying”?

Robin Lochmann: I did not have a script at this point, just a visual direction. I knew I wanted to integrate animated characters into the landscape, that they would be doing this ritual, and that there would be a procession.

DP: Where did the costume language come from?

Robin Lochmann: I started collecting images of religious and ceremonial robes and headpieces with a similar vibe. I wanted to contrast the religious feeling with leather straps and mix those opposing aesthetics.

DP: The main head shape feels oddly familiar, in a good way. What fed that silhouette?

Robin Lochmann: I love the look of the alchemist character in The Holy Mountain. That and a Catholic bishop’s hat, a miter, were the original inspiration for the main character’s head design. I wanted the shape to feel somewhat familiar but still unique, with a strong profile where the front is much longer than the back, almost like a shield.

DP: Any research rabbit holes you wish you could bill to a client?

Robin Lochmann: I looked at a bunch of other different religious practices, which was not particularly useful for the project in the end.

DP: What kept you from quietly deleting the folder and moving on?

Robin Lochmann: Honestly, I did not know it would take this long when I started. At some point, I had invested enough time into it that I thought I could not give up now. Maybe if I had really known how much was still ahead of me I would have given up. But I would always get motivated when I saw a new shot come together, and that would sustain me for a while.

DP: Money question, was there a budget, or just stubbornness?

Robin Lochmann: I did not have a budget, just my free time.

Tools, tracking, and the pipeline

DP: Give us the big picture workflow, from camera to final file, without pretending it was always tidy.

Robin Lochmann: Shoot real footage, first edit, 3D-track selected shots, build characters and environments in 3D, animate characters and simulate clothing, integrate 3D elements into tracked footage, render, and composite CGI with real footage.

DP: What apps did the heavy lifting, and which ones did you bounce between when nobody was watching?

Robin Lochmann: Cinema 4D was the main hub for everything, Octane Render handled all rendering, ZBrush sculpted the dog and bug characters, iClone handled base character animation and motion capture clean up, Rokoko provided motion capture, Marvelous Designer handled cloth simulation, and Substance 3D Painter handled texturing. I composited in After Effects, then finished editorial, colour grading and final delivery in Premiere Pro and DaVinci Resolve.

DP: Motion wise, how much was keyframed suffering versus mocap suffering?

Robin Lochmann: About half of the film is animated by hand in iClone. That was probably the most time consuming part of the whole project. About halfway through I got a Rokoko smart suit. It lets you record body movement and transfer it to any 3D character. It still needs a lot of cleanup, but it provides a great base and definitely speeds up the process.

DP: Clothing question for anyone who has ever rage-quit a sim cache. How did you build the garments?

Robin Lochmann: I used Marvelous Designer. I started by drawing the pattern in 2D and found it helpful to look at actual tailor patterns for reference. Then I positioned the pieces around the character, sewed it together, and ran the simulation. In some cases I added top stitching lines and simulated again so the garment movement followed the character animation. Different fabric types like canvas or silk drape very differently and give different results.

Rigging: stilts, metal, and non standard creature logic

DP: The creatures look like they refuse to behave in a normal rig. How did you build and rig them without losing your mind?

Robin Lochmann: Both the dog and the bug characters are built from scratch through a combo of modelling in Cinema 4D and sculpting in ZBrush. The bug was actually pretty easy to rig, but maybe not really a standard way of doing it. I am no expert and do not have a sculpting background, so I was winging it and pushing and scraping until it took shape. I painted it in Substance Painter, which feels similar to Photoshop with layers and masks. I added wings, bones for the legs, head sculpt variations to blend between, and a joint chain through the center of the head to make it wiggly. I animated the wings and added slight bends to the body. A shot with the bug trapped inside the jar ended up being one of my favorites.

For the dog, I used the quadruped rig in Cinema 4D as a starting point, but it was a little tricky because the dog is on stilts. Also, some of the metal bodies needed to stay rigid, so I constrained that separately.

DP: The robot dog is doing a lot. Where did that idea come from, and why the wonderfully impractical stilts?

Robin Lochmann: The initial inspiration came from a sculpture I saw in the Louvre in Paris. I liked the idea of a robot animal wearing a mask and the odd, impractical idea of having it on stilts. I wanted it to look cobbled together from scrap metal and discarded parts, barely held together by a rusty frame. I made a rig for it in Cinema 4D so I could animate it. I wanted the movement to feel janky and strenuous, like the rustiness holds it back a little bit.

World building, forests, and staying sane

DP: How much is plate based, and how much is full CG?

Robin Lochmann: Roughly half of Chrysalis uses real background footage and the other half is fully CG.

DP: When you go full CG, what does a scene build look like?

Robin Lochmann: I built the environments in Cinema 4D and made two main forest sets. Once they were finished, I could move a virtual camera around and frame like in the real world. I would do a previs first as a basic render for timing and camera movement, then refine and add details before final rendering with Octane.

DP: Lighting-wise, how did you get the France vibe into the CG shots?

Robin Lochmann: I made 360-degree HDRI images in France as a starting point to light the scene. That was rarely enough by itself, so I added extra area lights and a fog volume, basically a big box of fog that sits on top of the scene and breaks up how light particles shine through it.

DP: Any survival tips for reopening giant environment files without screaming?

Robin Lochmann: Try to stay organised in big environment scenes, so when you open the project again, you do not completely lose your mind.

DP: When it was time to hit render, what did you spit out for comp, and what passes mattered?

Robin Lochmann: I rendered everything in Octane. I used EXRs with beauty, cryptomate, volume, and Z-depth passes for each shot. That was usually enough for compositing. Because I was rendering locally on a 2080 and a 3090, I would usually just render overnight but at the very end I had about one week of nonstop rendering.

Then, everything came together for compositing and final adjustments in After Effects. Because I had the cryptomasks, I could select every object separately for small color or brightness tweaks and add CG elements to real footage. Edges are a dead giveaway, so I used light wrap to bleed background light into the foreground edges. I also added haze and sometimes more fog, took down digital sharpness, and added a bit of film grain while obsessing over small details to balance the image.

DP: What did you deliver in the end, and how wild were the render times shot to shot?

Robin Lochmann: 4k 24fps. Render times varied a lot. Some shots, where it was just one character that would later be composited onto real footage, rendered quickly while the fullCG scenes took much longer. Because I was rendering on the go, at night, it is really hard to say how long the total render times were.

Sound: music first, then realism on top

DP: Who made it sound expensive, and how did you brief them without writing a novel?

Robin Lochmann: Music by Robot Koch and sound design and mixing by Philippe Glandien. Throughout working on the project, I would listen to Robot Koch’s music while watching playbacks and I knew that I wanted him to do the music for the film. I sent him a locked edit and gave him some guidelines, but I already trusted his taste and style.

After just two revisions teh music was locked down. For the sound design, I gave Philippe Glandien the film with the final music. I wanted to aim for realism. The sound design should just support the visuals and make the film feel more tactile, while the music was the emotional heartbeat.

Updates mid project: the good, the bad, and the crashy

DP: Over three years, updates happen. Which upgrades were helpful, and which ones tried to eat your project?

Robin Lochmann: Because the project went on so long, I did have to update most tools along the way. The worst was After Effects. The number of crashes kept increasing, and by the end I could not even close the project without it crashing. Some updates in Cinema 4D were really useful, like their particle system and the updated cloth simulations.

I used these cloth simulations to tear the cocoon and the main character’s wings directly inside Cinema 4D. To control where cloth breaks, you can paint a vertex map, allowing tearing to occur precisely. For the cocoon emergence, I layered different versions so they pushed up from underneath and broke at various points, with layers reacting to each other and different physical properties.

DP: And the white webbing lines that creep over things? How did you do that?

Robin Lochmann: I shot particles at the object and traced the particle path to form a line. The particles split and form new particles. I used the same technique in a couple of shots where white webbing covers everything.

DP: If you rebuilt the character pipeline today, what would you streamline first?

Robin Lochmann: Some of the workflows were also not the most efficient *cough* in terms of bouncing between the different tools, especially for the characters. Bringing them from motion capture to cleanup, clothing, and then back into C4D to remove the base and attach the head involved too many steps. I am not a coder, so I had to do it all manually, but I think now I would try to use AI to write a script to automate it.

DP: Any deleted scenes you still mourn, or are they safely buried in a folder called OLD_FINAL_FOR_REAL_THIS_TIME?

Robin Lochmann: I made early test shots to see how the parts combined. None of those shots ended up in the film. In earlier versions there were other characters I built and then cut out, like a giant baby statue, an awful kid, and a mask. I even finished a couple of scenes with one character, then went to the cinema to see Dune 2 and saw characters so similar that I decided to remove that character completely.

DP: You skipped the festival waiting game at first. What was the thinking, and what happened after you hit publish?

Robin Lochmann: I decided to skip festivals at first and go straight online, mainly because I did not want to wait so long to share it. I am really glad I made that decision. Even with a decent festival run, you are lucky to have a couple of thousand people see your film. Two weeks after release, we had already gotten over 100k views on YouTube just organically. That was way more than I expected, since not many people want to watch short films. The feedback has been great. I am also submitting it to some smaller festivals now that do not disqualify films that are already publicly available.

DP: Last one. When did you stop tweaking and let it leave the nest, or the render farm, whichever hurts more?

Robin Lochmann: Sometimes all you need is a deadline. Right up until it was being screened, I was still tweaking details that nobody else would ever notice. Then, when the time was up, I was happy to let it go. I think only you can decide for yourself when it is finished; nobody else can make that decision for you.

https://www.robinlochmann.com/

The post Chrysalis first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the studio: Lux Aeterna delivers documentary and natural history VFX including CG environments, FX, and compositing, and it mixes scanned data with CG for shows made with broadcasters and streamers.

History likes receipts

Historical reconstruction only works when the images carry the same authority as the research. The studio leans on archaeological reference, field documentation, and specialist feedback to keep reconstructions grounded, especially when the brief asks for ancient humans that look plausible in stills and in sequence work.

Two recent projects show how far that approach goes. Secrets of the Neanderthals needed a cave interior rebuild and physical remains that read like real bones. Human pushed further, asking for early human avatars built from actor scans and refined against scientific consultation, plus maps and timelines driven by datasets.

Bones first, drama second

For the Neanderthal work, the team recreated the interior of Shanidar Cave in northern Iraq, the site of one of the largest ever discoveries of Neanderthal remains. It used a mix of reference drawings, photographs, and expert advice from professors at the University of Cambridge and Liverpool John Moores University to guide accuracy, including cases where remains had been damaged or moved from their original positions.

Skeleton placement used inverse kinematics so the bones could interact with the ground in a natural way. The pipeline also dealt with scan data arriving as a 3D point cloud, then converting it into polygons that could support a usable cave reconstruction. That conversion problem tends to separate demo reels from production, because bad meshing reads instantly, even when the camera does not linger.

Point clouds, polygons, and the cave that ate your schedule

The cave build started with LiDAR scans that came in as a 3D point cloud. The team developed algorithms to convert the points into polygons for the reconstruction, then used Houdini to isolate areas of scan data so the mesh could be assembled and recombined like a 3D jigsaw. That kind of procedure-heavy work sits squarely in the Houdini mindset: build repeatable ops, then iterate until the scan stops looking like scan data and starts looking like a place.

The result aims for a CG environment that matches the real location closely enough that it can carry the weight of the story, while also supporting the additional CG work needed for the remains and the shot context. In a documentary pipeline, that matters because the images often sit right next to photography, maps, and archive material, with no cinematic lighting trickery to hide behind.

Lux Aeterna delivered 60 shots for a single series, and that number matters because it hints at the production scale when planning resources. WE might e talking about geological and evolutionary tiespans in the story, but delivery is far more rapid.

Full bodies: photogrammetry, casting, and millimetre arguments

The later project asked for historically accurate avatars built from scans of real performers, refined with skeletal research and scientific input. It was a five-part series produced by BBC Studios Science Unit for the BBC and PBS NOVA, hosted by paleoanthropologist Ella Al-Shamahi, and it frames visual effects as a way to tell a story spanning hundreds of thousands of years without a written record.

The workflow combined ultra-high-resolution body scans with iterative model refinements guided by BBC Studios and scientific advisors. It worked with London-based scanning studio Sample & Hold, which describes custom head and body photogrammetry systems and a photogrammetry robot for objects. In practice, that puts photogrammetry into the same category as any other acquisition step: controlled capture, predictable output, and a pipeline that can survive the boring parts like data cleanup and repeat scans.

The stated goal was four historically accurate avatars representing Jebel Irhoud, Neanderthals, Homo floresiensis, and Homo erectus, based on scientific data and real-world actors. Alongside the character work, the project also required prehistoric environments and landscapes to show migration and later disappearance of early human life across the globe. The work targets ancient humans, but it does not tolerate ancient guesswork.

Maps, timelines, and the part nobody notices until it breaks

A second strand of the work focused on animated maps and timelines that visualised migration, development, and disappearance across millennia. The graphics are being built using datasets provided by BBC Studios Science Unit researchers and scientific consultants, and not as speculative graphics.

One specific challenge involved changing coastlines and submerged landmasses, where the team created transitions and overlays to keep the visuals clear while staying scientifically accurate. That is the kind of “small” work that carries a lot of editorial risk: the map has to be readable, the timing has to serve the cut, and the choices have to stay consistent across episodes.

On the delivery scale, the project description states sixty crafted shots across five hour-long episodes, and it frames the approach as avoiding prosthetics and overly stylised effects while using a hybrid of photogrammetry, consultation, and minimalist Compositing.

What to take away if you do this for a living

If your job involves building reconstructions that sit next to real-world evidence, the takeaway is simple: acquisition quality and reference discipline decide how much time you spend “fixing” later. Scan data arrives noisy, research notes arrive detailed, and schedules arrive optimistic. The pipeline has to translate all three into images that read as plausible without distracting from the content. And as always, test any new workflow, plugin, or automation step on a throwaway shot before you bet a delivery on it. Just like with rare fossils, you don’t want to play loose with the materials.

https://www.lavfx.com/
https://www.lavfx.com/secrets-of-the-neanderthals
https://www.lavfx.com/blog/human

P.S. You have NO idea how big the challenge was not to replace “Ancient Humans” and “Prehistoric” with the users of one ore more DCC-packages….

The post Lux Aeterna rebuilds ancient humans first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: The Pixel Farm makes PFTrack, a node based matchmove and reconstruction tool that exports to DCCs like Autodesk Maya, Foundry Nuke, SideFX Houdini, Blender and Unreal Engine, with improved dense point clouds via OpenUSD.

Hero Cloud is a new reconstruction node

PFTrack Build 26.05.19 adds Hero Cloud, a node that generates a measured dense 3D point cloud from a single tracked plate. It sits in the Geometry category and runs downstream of a solved camera, turning an ordinary camera solve into spatial data you can actually measure against. Hero Cloud claims to be through-the-lens reconstruction, meaning it extracts depth from the camera movement in the shot rather than relying on a separate photogrammetry set or a LiDAR scan.

The node targets a familiar production reality: you get a brief that needs measured 3D, but the the on-set… people…. said “Fix it in Post, heheeheeh” and did not capture the extra data for cases like set extension, CG integration, lighting reference, forensic analysis, architectural measurement, and archival footage where you cannot go back and reshoot reference. If the plate is all you have, Hero Cloud aims to make that plate pull double duty.

Hero Cloud works on plates PFTrack can track and solve, and works footage sources that include cinema cameras, drones, body cams, dash cams, action cameras, mirrorless and DSLR cameras, CCTV, and archival film. Basically, every type of video. Camera metadata can help but it is not required, and the Camera Sensor Database can cover common cameras automatically, with the Camera Solver able to estimate parameters when metadata is missing.

There is a catch, and it is the same catch as every method that tries to infer depth from motion: the shot needs parallax. Locked off shots or shots with very little movement can produce sparse or unreliable results. Most production footage, including handheld stationary shots, can still have enough movement to produce usable reconstruction, while deliberate camera motion tends to produce richer results.

A cleaner route into splat training

PFTrack 26.05.19 adds a Postshot export option that writes Hero Cloud point clouds and associated tracked cameras in COLMAP format. In other words, matchmove and reconstruction can stay in the same app up to the point where you hand off to splat training.

If you already live in the Gaussian Splatting world, COLMAP is a common interchange format and COLMAP itself has both GUI and command line workflows. The practical value here is that you do not have to glue together separate tracking, reconstruction, and export steps before you ever open Jawset Postshot. You track the plate, generate the cloud, export in COLMAP format, then train the splat.

This is an artist friendly route into a workflow that has historically leaned on command line tooling. That is a reasonable and recommendable goal, but it still deserves the usual pipeline skepticism: try it on a real shot and verify your coordinate frames, scaling, and camera consistency before you let it anywhere near a deliverable.

USD point clouds now carry more than positions

The same build extends USD export to support dense point clouds with per point normal and colour data, because a point cloud that only carries XYZ positions often becomes a dumb scaffold the moment it leaves its origin app. Normals and colour help downstream tools treat the cloud as more than a vague spatial fog, and can make it easier to use for measurement, placement, modelling reference, and lookdev decisions.

Improved integration with USD aware tools including Unreal Engine, Maya, and Houdini. It also claims (!) improved FBX export reliability for very large dense point clouds, aimed at users who previously hit issues exporting big clouds via FBX.

If your pipeline already standardises on USD, this is the kind of unglamorous export detail that can decide whether a new feature becomes a daily tool or a demo-only party trick. Still, test it on your actual asset scale and on your actual DCC import settings. Export support is never a substitute for shot-based validation, and you only find the sharp edges when production hits them, much like furniture at night can only be found with toes and shins.

Photo Mesh gets a single shot input path

Studio and Enterprise editions receive an update to Photo Mesh that enables it to consume Hero Cloud point clouds as input. That creates an internal single-shot path from tracked plate to dense cloud to surfaced textured mesh, without leaving PFTrack. The Pixel Farm describes Photo Mesh as the highest fidelity route when you have a planned photo set, while Hero Cloud covers the case where you only have the hero plate.

In practical terms, this puts a reconstruction option closer to matchmove, and it reduces the number of handoffs in the middle of a cmaera tracking job. Whether it replaces any external reconstruction tooling depends on what you already trust for meshing, texturing, and scale fidelity, but the internal integration is at least a coherent story: generate the cloud, feed it into Photo Mesh, then export into the rest of your pipeline.

The viewport tweak you will actually use

26.05.19 also adds an Active Camera viewport option in viewer windows, with a snap to active clip toggle. It sounds minor, and it is, which is why it will get used: quick visual verification against the source plate tends to catch mistakes faster than hunting through node outputs. The release notes pitch it as useful for checking Hero Cloud reconstruction, but the feature applies anywhere you want a fast way to see what the tracked camera saw.

Trial and licensing notes

Solo trial mode now enables full export functionality for seven days. The release notes describe this as enough time to take a plate from track through Hero Cloud and out to Postshot or USD, rather than being limited to evaluating the toolset without export.

For Studio, the release notes state it is available on rent to buy for 5 or 30 days, or as a perpetual purchase. The PFTrack Creator Program page states PFTrack Solo licenses sell for £699. Treat that as a price reference for Solo licensing, not a promise of regional pricing or bundles.

What this changes in real work

Hero Cloud aims to shift PFTrack from a pure matchmove and survey friendly reconstruction hub toward a single shot scene reconstruction step that can be used even when the onset data plan fell apart. The strongest parts of the release are not flashy, they are connective tissue: COLMAP export for Postshot, denser USD point cloud payloads, and a meshing path that stays inside the same graph.

It is also a reminder that the industry has started to treat point clouds as a first class asset, not only a temporary byproduct. When point clouds export with colour and normals and survive the trip into USD aware DCCs, they become reusable poitn based reference that can support layout, set extension decisions, and reconstruction driven previs.

https://www.pftrack.com/post/pftrack-build-26-05-19-is-now-live

https://www.pftrack.com/post/hero-cloud-tutorial

The post PFTrack gets Hero Cloud first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: MoonRay is a production path tracer with a USD Hydra render delegate for tools using OpenUSD, positioned in rendering and lookdev next to ASWF staples like OpenColorIO.

The Academy Software Foundation has added MoonRay as a hosted project, keeping the renderer under foundation governance while opening contributions to a broader community. DreamWorks Animation developed MoonRay as an open source production path tracing renderer and has used it on every one of its feature films since 2019, including The Wild Robot and The Bad Guys 2. The project uses the Apache 2.0 license and is developed in C++.

MoonRay includes support for distributed rendering and a pixel matching XPU mode that processes bundles of rays on the GPU as well as the CPU. It also includes bundled path tracing and a USD Hydra render delegate for integration into content creation tools that support the OpenUSD standard, which should make interactive preview workflows less of a custom science project in teh average pipeline.

The ASWF angle

MoonRay now sits alongside other hosted efforts in the ASWF roster, including OpenEXR, OpenVDB, OpenTimelineIO, and Open Shading Language. If your facility already treats OpenColorIO as plumbing, this is the same neighborhood.

For contributors, the project points developers to openmoonray.org and to the ASWF Slack workspace for a MoonRay channel at slack.aswf.io. Documentation is published at docs.openmoonray.org, and the source repository lives under the OpenMoonRay organization on GitHub.

Keynote season meets SIGGRAPH season

ASWF Open Source Days runs in Los Angeles on July 19 to 20, 2026, and the schedule aligns with SIGGRAPH 2026 timing. The keynote comes from DreamWorks Animation CTO Bill Ballew, with a talk titled How to Train Your Renderer: MoonRay’s Journey from DreamWorks’ Dragons to the ASWF.

The event listing describes July 19 as the main program day in Los Angeles, followed by July 20 as a full day of open source Birds of a Feather sessions within SIGGRAPH. The Call for Proposals remains open through May 24, 2026, via the Open Source Days event site.

As always, test new tools and new governance setups before you bet a production schedule on them, even when the feautres list looks exactly like what your render wranglers asked for.

https://openmoonray.org/

https://www.aswf.io/

The post MoonRay Renderer joins ASWF as hosted project first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: Dash runs inside Unreal Engine as an asset browser plus world-building toolkit. The Fab build ships the Content Browser only, while the website installer can also unlock the wider toolset with a paid license.

Two installs, one idea

Polygonflow now offers Dash in two distinct ways: as a free plugin on Fab and as the full Dash installer from the Polygonflow website.

The Fab listing calls the package “Dash – Advanced Asset Manager” and in this version, it is “just” a content browser that lets you search and drag and drop local assets into any UE5 project. It includes tagging for your own assets with a stated limit of up to 1,000 assets tagged per month, plus a Smart Collection System for organizing assets.

The website installer route matters more, because the Dash Content Browser itself has been available for free since Dash version 1.9.0, but the free license is limited to the Content Browser side of the toolset. The free license includes the supported libraries, the tagging system, and the centralized browsing experience across UE5 projects, but not other Dash tools such as Easy Scattering, Procedural Meshes, Physics Tools, or Material and Atmospheric Tools.

What is actually different between Fab and the website build

The Fab version includes the Content Browser only, and explicitly points to the website download if you want the full plugin toolset. The website build splits into two experiences. The free license stays content-only, meaning you use the Content Browser, library integrations, tagging, and centralized asset browsing. A paid license can unlock the other Dash tools, Easy Scattering, Procedural Meshes, Physics Tools, and Material / Atmospheric Tools. So the practical difference is not the content browser UI, it is how far you want to go beyond asset browsing once you are inside the same plugin shell.

Included libraries, and why that matters more than the tagline

Both the Fab listing and the Dash docs have the same core set of integrated libraries: Quixel Megascans via Fab or Quixel Bridge downloads, Amazon Berkeley Objects with 6,900 models, ambientCG with 2,300 textures plus HDRIs and decals, Poly Haven with 1,900 assets plus textures and HDRIs, The Base Mesh with 1,400 models, and the IES Library with 1,300 lighting presets.

On the website side, the Dash Asset Marketplace lives inside the Content Browser, with the Dekogon library and a Sierra library entry that includes one free pack.

Polygonflow also supports browsing and importing content across multiple UE projects from inside the project you are currently working in. The docs add a warning on version compatibility for external browsing: UE5 projects are forward compatible but not backward, so you can browse content from the same or older UE5 versions, not newer.

Custom assets, right now

In the Project Library tab, you can currently add your own static meshes, textures, and blueprints. The tagging flow focuses on making assets searchable beyond filenames and folder names, and the content browser adds collections for grouping assets into reusable sets.

Pricing and the usual production warning

On Fab, Dash – Advanced Asset Manager is free. Commercial pricing for paid Dash licenses is 8€ for students, 20€ per Month for professionals below the 250.000$ barrier, and 95€ per month for bigger studios. Test any new browser, tagging, or library workflow on a non-critical project before it touches production depots and shared team setups.

https://www.fab.com/listings/865301e9-2d26-425d-8f3c-e0aadf35da61
https://docs.polygonflow.io/advanced/free-content-browser

The post PolygoFlow Dash asset manager goes free on Fab first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

With *Das szenische Portal RT*, Andreas Resch has developed a real-time system for live events that combines a physical stage, virtual scene spaces, camera feeds and musical inputs to create a hybrid performance environment. The approach is primarily aimed at live formats featuring improvising musicians and is conceived as a modular virtual stage extension rather than a mere background projection.

Andreas Resch works in a field where one is usually expected to choose between art and technology. He has evidently chosen both, and this is no disadvantage for this project. Classical music training on the piano and church organ, coupled with an early start in programming, later work on SGI Onyx-based training simulators in Munich, followed by stints in animation, simulation, tracking, projection mapping, virtual physics and mixed reality.

This combination of music, imagery and technical development has long been a recurring theme in Resch’s work. Since 2020, he has demonstrated in several projects how audiovisual staging, virtual spaces and live performance can be combined. Das szenische Portal V1.0 already relied on a form of live performance in which Resch played the piano whilst a surreal, distorted virtual reinterpretation of a physical painting exhibition was projected onto a large screen as an endless film. This was presented live on several occasions and ultimately served as the precursor to what is now being continued as a real-time system.

At present, Das szenische Portal RT is primarily an art project, a prototype and a technical development. However, according to Resch, project-based applications are conceivable: anyone requiring a suitable configuration for an event can, in principle, integrate him and his software. Enquiries? Questions? Idea? Sent them to innovation@virtual-paradises.org. Resch can also envisage further development with a partner from the product or events sector to expand the existing prototype into a robust production or rental solution.

Unreal Engine as the basis for the stage extension

With the scenic portal RT, Resch revisited the concept from 2024 onwards and completely redeveloped the system in Unreal Engine. The first working version was created in 2025. The aim is a virtual stage section that does not simply hang behind the real stage like a particularly ambitious backdrop, but forms a hybrid production together with the physical structure. The ambition is therefore greater than mere decorative projection. The virtual scene is intended to become part of the stage space.

The bigger, the better, of course. LEDs, a good projector etc. with a screen, … The prototype has so far only been tested on an 85-inch monitor. The size also depends on the actual stage dimensions to ensure a good overall design. Every event here is unique. 

The effect is based on a geometrically structured overall concept that closely interlinks physical and virtual image elements. Resch describes this as a form of hybrid stage, in which objects from the virtual projection appear to extend into the physical stage space. This approach becomes particularly relevant where improvising musicians are involved. Audio signals, whether from digital or acoustic instruments, can spontaneously trigger events in the virtual scene, thereby directly influencing the visual narrative. The stage, in other words, listens in. Something that rarely happens even in everyday life.

Live input, triggers and virtual events

For the prototype, Resch first developed several three-dimensional stage sets in Unreal Engine. He then implemented real-time communication between the computer and a Korg KRONOS workstation, which serves as an external trigger source for scene events.

In the test configuration, three live webcams and the KRONOS were continuously fed into a specially developed Blueprint system. This interprets audio and video data and translates it into dynamic events within the virtual scene.

The Blueprints

Nevertheless, this allows for artistic, surreal constellations to be achieved with the video
projection surfaces, such as embedding a performer into a picture frame, which, as shown here, begins to float alongside the concert grand piano. A triggered Niagara system could then be attached to the picture frame, for example, which reacts to the music and scatters sheet music around the room, etc.

The live video of the performance on the keyboard could, for example, be superimposed over the virtual grand piano’s keyboard to make the artist’s hands visible in the virtual representation.

These include object transformations, cloth in the wind, collisions, Niagara effects, camera control, choreography, scene changes and other VFX actions. Central control is handled via Unreal Engine’s sequencer, in which specific sequences are defined in an endless loop, such as camera paths through the scene. Technically speaking, this is the point where performance, event logic and scene organisation converge.

What goes into the pipeline?

A pipeline that integrates external tracking processes can run on anything that communicates with Unreal Engine. Tools such as VooCAT, 3ds Max, C4D, Houdini, Blender, Datasmith, and Unreal Engine itself form the basis for tracked image sequences (or AI-generated image spaces as virtual environments), for example, via a camera backplate or a projection map. An imported 3D point cloud serves not only as a spatial reference but also provides coordinates for the placement of additional events.

The real space is not surveyed, as there are no parallactic references, as in a virtual production. Ideally, however, the real space is photographed so that the virtual worlds can be constructed to ‘inherit’ surfaces and geometric features from the real space. The real stage can then also be furnished with haptically appropriate objects to create visual harmony.

Unfortunately, a comprehensive “field test” of the system’s behaviour is still pending. The Blueprint system must always be adjusted to the target instrument’s requirements and signal strength using appropriate parameters.

A key structural feature is the separation between internal and external triggers. Within the 3D scene, an audio source selected by the artist can be statically positioned and function as playback. These internal signals trigger defined processes within the scene, such as Niagara effects or other pre-prepared reactions. External audio signals from microphones or digital stage signals, on the other hand, activate other types of events.

Two Niagara systems were present at the demonstration, triggered internally via a stored scene sound. With two additional webcam inputs and an external signal from the KRONOS workstation, the system ran stably on the aforementioned “small computer”. The latency times were “noticeably” correct. Geometrically, this took place in a 3D model of the Colosseum, after all.

In this way, an internal beat can underpin the basic structure of a sequence, whilst external impulses additionally influence, shift or disrupt the scene. This makes sense for live formats because it mediates between predictability and reaction – that is, between show control and what musicians simply love to do: something unexpected.

Modular structure and live video

The modular blueprint structure is designed to allow the appearance and behaviour of the virtual stage section to be freely adapted to the artist, song or event format. The project is therefore not conceived as a rigid show package, but as a flexible framework within which specific logic for individual pieces or client requirements can also be implemented.

Ideally, a large LED wall is provided on the physical side, though projection via a projector is also possible in principle. The prototype was tested on a Windows PC with 16 GB RAM and an RTX 3060, running Unreal Engine 5.5.4. Depending on the complexity of the scene, two to four webcams can currently be integrated in real time. Their live images appear on defined projection surfaces within the virtual stage section. In this way, both artists and audience become part of the virtual world, not as a post-production effect, but as an ongoing component of the production.

It is precisely this integration of live video that is one of the more interesting aspects of the concept. When the audience appears, for example, as a background projection within the virtual stage section, the viewer may get the impression that the real stage opens up at the back into another room. This is more than an animated background and less than a completely virtual production. It is precisely in this intermediate zone that the appeal of the project lies.

A ‘field test’ is still needed here. During the prototype demonstration, the room was quite dark. The pianist was locally lit and clearly visible. The camera image of the audience was “normalised” using a blue screen, i.e. the brightness was adjusted so that the people on the projection screen could be clearly recognised. A test in a larger, much better-lit hall is still pending.

First presentation on a small scale

An initial live presentation took place in a small, private setting with selected guests. It was deliberately not aimed at a specialist audience, but at interested visitors, including visual artists. A photorealistic 3D model of the Colosseum served as the simulated stage set, displayed on an 85-inch monitor in Full HD at 70 fps. Resch himself took on the role of the live-performing artist at the KRONOS workstation, simultaneously controlling the behaviour of the virtual world.

Two live cameras (Full HD Rapoo webcams, i.e. very affordable equipment) captured him from the side and from above, with a view of the keyboard; additionally, the audience was integrated into the event-driven 3D world as a video projection. 

For Resch, this presentation is proof that traditional stage technology and a responsive virtual universe can be meaningfully combined. The audience reacted with clear amazement. This is not yet a technical acceptance test nor a production benchmark, but at least a useful test of whether the basic idea works on a perceptual level. That, too, is not entirely unimportant in the live sector.

Between prototype and production tool

The scenic portal RT is continually evolving. With current gaming hardware, 4K output and higher frame rates should also be possible in the future. In parallel, Resch is working on a musical project that is being developed partly using AI platforms. He defines the musical structures himself, whilst AI is used for arrangement and instrumentation.

For performers, narration and singing in fantastical worlds, he cites HIGGSFIELD AI. Whilst this is a separate project, it points the way forward: real-time, virtual staging, music, simulation and AI-assisted content creation are increasingly merging here. To what extent does the availability of online AI content generators depend on the ‘current state of things’? Which platforms will still be available by the end of the year (see SORA) and which ones have working APIs, must be checked again before each performance. The checklist is getting longer.

The bottom line is that the RT scenic portal is neither a classic media server product nor a ready-made event graphics package. It is a custom-developed real-time system based on Unreal Engine, trigger logic, live input and modular scene control. Whether this will become a widely applicable tool for the events sector ultimately depends, as is so often the case, not on the idea itself, but on robustness, workflow suitability and scalability. The idea itself is certainly technically interesting. Not every stage needs a surreal real-time parallel universe. But it would also be a shame if it were always just an LED wall with wave animations.

Want to know more? virtual-paradises.com and here is the associated YouTube channel: https://youtube.com/@VirtualParadises

The post Das szenische Portal RT: When the Stage Starts Listening to the Music first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

The ProArt Display PA16USV centers its pitch on 12G-SDI. One BNC input handles uncompressed 4K at 60 Hz and targets direct camera-device monitoring with ultra-low latency, using the SDI family standardized by SMPTE. That matters when you want the same signal type your carts and villages already trust, without detouring through converters.

The panel is a 15.6-inch IPS at 3840 by 2160 with 282 PPI. It ships with an anti-glare surface, a typical contrast ratio of 1200 to 1, and 170-degree viewing angles. Response time is rated at 5 ms gray to gray with a maximum refresh rate of 60 Hz. You also get the expected alternates: one HDMI 2.0 port, plus a USB-C port that supports DisplayPort Alt Mode. A USB hub port is in there as well, with USB 3.2 Gen 1 Type-C, and USB-C power delivery is at 7.5 W. Not enough for a laptop, but enough for most phones.

Color specs, presets, and the reality of 400 nits

Color coverage is listed as 100 percent sRGB and 100 percent Rec.709. sRGB ties back to the sRGB standard published by the IEC, and Rec.709 is the broadcast workhorse from the ITU. Factory calibration targets Delta E under 2, with Calman Verified status via Calman, and the box also includes a colour pre-calibration report.

Preset modes include Native, sRGB, Adobe RGB, DCI-P3, DICOM, Rec.709, HDR, and two user modes. HDR support is HDR10, which the industry commonly references through work from the Consumer Technology Association and related standards bodies. The monitor also lists multiple HDR10 curve options: PQ Hard Clip, PQ Optimized, and PQ Basic.

Brightness is typical at 400 cd per square meter. That number can work fine for studio carts and shaded setups, but it puts outdoor reality checks firmly on you, not the spec sheet. If you plan to use this as your primary village screen, run your own midday tests before you show up with confidence and a single C-stand.

This is also where good process still beats good marketing: even with a calibrated panel, you should validate your full monitoring chain in your own environment, including signal path, LUT handling, and who gets to touch the settings. The best surprise on set is no surprise.

Mounting, handling, and the box you keep

On the physical side, the monitor supports VESA 75 by 75 mm mounting. It also has a 1/4-inch tripod socket, tilt from plus 20 degrees to plus 70 degrees, a Kensington lock, and an ambient light sensor. Net weight is at 1.4 kg including the stand, and 1.3 kgs without the stand.

Bundled accessories vary by region, but the list includes a mini stand and an HDMI cable, plus USB-C cable, power cord, and power adapter. Warranty details on the tech specs page list LCD ZBD warranty as yes, 3 years. Power consumption is specified as under 15 W, with power saving mode under 0.5 W and power off mode under 0.3 W.

There is also a clear nudge to keep the packaging: the carton can be reused as a monitor hood. The packaging hood comes with a caveat that its condition will naturally deteriorate over time and is not covered by the warranty.

If you have ever hacked together a hood out of whatever the rental house cardboard gods provided, this detail will either make you smile or make you wonder how many folds it takes before it looks like a crushed matte box.

Picture tools and desktop control

The display supports PiP and PbP. It also supports DisplayWidget Center, a monitor management app from ASUS that lets you adjust settings with a mouse instead of the monitor buttons.

In practical terms, that combination points at two workflows: live camera feed over SDI for set decisions, and laptop-based review over HDMI or USB-C for DIT and post-adjacent tasks. The PiP and PbP options can help when you need to compare sources quickly, but you still need to confirm how your specific devices negotiate formats and framerates in the field.

If you live in the world of tight Rec.709 handoffs and client screens, the real test is not the spec list. It is whether the monitor holds up under real cables, real power, and real people.

Pricing, warranty, and the parts that need a footnote

The monitor costs 850€ in Germany, with a similar Amazon Price, and used samples landing somewhere between 500€ and 600€. As always, treat new monitoring gear like any other new link in the chain: test it before you put it in front of a director, a client, or a colorist who can spot a gamma wobble at twenty paces.

https://www.asus.com/de/displays-desktops/monitors/proart/proart-display-pa16usv/

P.S. The PA16USV now ships in Germany, after a longer wait than many expected. At FMX, conversations with two on-set VFX-people pointed at the gap for this screen: they want 12G-SDI monitoring in a smaller, more affordable package than the usual set-village screens. This is also why we pulled this one back on the list, since we have hands-on testing lined up for this panel very soon. Keep your eyes peeled.

The post ProArt goes 12G-SDI portable first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: Unreal Engine is a realtime 3D engine and editor used for games, VFX, and virtual production, with tools like MetaHuman, Nanite, and procedural systems like PCG built in.

The preview nobody should install on Friday

Unreal Engine 5.8 Preview is available now to download via the Epic Games Launcher, GitHub, or for Linux. Epic Games positions this Preview around performance advancements and systems meant to feel more reliable, scalable, and intuitive.

That Preview label still means what it always means: The build is not quality tested, it remains under active development, and it can and propably will be unstable until the final release. Make a copy of your projects for testing rather than converting your mainline work.

Mesh Terrain goes full 3D, experimentally

The headline feature in 5.8 Preview is Mesh Terrain, flagged as Experimental. It uses a 3D mesh based architecture for building massive environments and it works natively with the Procedural Content Generation Framework. If your pipeline already leans on procedural layout, the idea is straightforward: terrain becomes part of the same graph driven worldbuilding conversation, instead of a separate special case that artists tiptoe around.

The practical impact depends on what you need out of terrain today. The announcement calls it a “novel architecture” (Scaaaary words) and keeps it at the overview level, so details like feature completeness, interchange expectations, and performance characteristics remain a guesswork.

PCG gets edits without losing the recipe

5.8 Preview also updates the PCG framework to allow manual edits directly on procedural output while keeping the underlying procedural logic intact. That matters because it targets a classic production pain: the moment an artist has to hand fix something and the whole setup turns into a one way bake.

In plain terms, you can push results around without instantly severing the link back to the graph. The sources describe this as an improvement alongside other iterative changes, but they do not specify which nodes, asset types, or editor workflows get the most love in this Preview.

Vegetation: more biology, more Nanite, more questions

The Procedural Vegetation Editor receives significant enhancements in 5.8 Preview. The goal here is building high quality, biologically correct, Nanite ready vegetation assets directly inside the editor. The sources also say you can import meshes from external DCC tools, which points at less time spent rebuilding base shapes just to satisfy an in engine authoring workflow.

Currently, we are about to get the following features:

• Growth Node
• Seed Generation
• Skeletal Extraction from 3D Mesh or 2D image/Atlas
• Foliage Grafting
• Object Avoidance and Carving
• Trimming- branch loops, displacement tessellation
• Mesher Improvements
• Visualization Modes
• Trunk Material Compositor

Real world artists will appreciate that this tool stays honest about what it is trying to do. It aims at fast iteration and authoring inside the editor, not at replacing dedicated plant tools overnight. The community replies in the same thread also report issues and broken behaviors in specific cases, which fits the Preview warning label. Treat it as a sandbox build, not as a deadline build.

MetaHumans scale up, and they get less picky about input meshes

Crowd work gets a bump through the MetaHuman Crowd plugin. The sources describe scaling from tens to thousands of characters for populating realtime worlds. If you have ever tried to sell a crowd shot without a crowd system, you already know why that matters.

On the creation side, the Preview also expands the ability to transform almost any human mesh into a MetaHuman, now with simultaneous head and body conforming. Epic Games calls this as a workflow improvement, but the current sources do not spell out constraints, quality expectations, or edge cases for stylized characters.

Lights, rigs, faces, and a grab bag of everything else

MegaLights reaches full production readiness in 5.8 Preview. Control Rig Physics moves to Beta, and the face rigging and morph target editing tools expand to support in editor sculpt driven facial workflows and shot sculpting. The sources describe this as strengthening blend shape authoring for stylized characters, custom skeletal meshes, and MetaHumans, plus pose corrections.

Rigging also gets Direct Mesh Controls, an Experimental system that lets Control Rig controls sit directly on sections of a skeletal mesh. If you like your controls where your hands are, that sentence probably reads like a small miracle.

Past the hero items, the Preview notes further work across production rendering and virtual production workflows, plus updates to Dataflow and Chaos Physics.
https://forums.unrealengine.com/t/unreal-engine-5-8-preview/2721597

The post Unreal Engine 5.8 Preview rolls in first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

What exactly is Device Preview Plus supposed to do once the calibration itself is finished? And how far can a monitor tool stretch before it starts trying to become part of the wider image pipeline? In this interview, Heath Barber speaks about Datacolor’s current direction for SpyderPro, including display simulation, custom device profiles, C2PA content credentials, and the role of calibration in workflows that no longer end at a single screen.

Heath Barber (LinkedIn) is Senior Product Director, Consumer Solutions at Datacolor and has spent nearly 21 years with the company in a series of senior software, product and imaging roles. His previous positions including Senior Director, Software Technology; Global Market & Technology Manager, Imaging; Global Software Technology Manager; and Director of Development, Imaging Color Solutions. That long run across Datacolor’s software and imaging side makes him a useful guide for a conversation.

From monitor calibration to device simulation

DP: SpyderPro now includes Device Preview. On the box, that sounds simple enough. In practice, what does it do?

Heath Barber: Device Preview is really a way of simulating how your image or media might look on other output devices without you having to physically move the file across ten different screens and start guessing what changed where.

Crawly for Videoscreens: SpyderPro

The classic comparison is print preview. People in print workflows have done this for years. You work on one display, but you still want a reasonable idea of how the final output will behave on paper. Device Preview takes that idea and extends it to screens. Instead of asking only, “How will this print?”, we are asking, “How will this look on a phone, on a tablet, on another class of display, or on another type of viewing device?”

That matters more than people sometimes admit. A lot of media today is first consumed on a phone. My wife is a wedding photographer, and like many photographers, one of the first things she wants to understand is how an image will look when a client opens it on a mobile device. That first impression often happens on an iPhone, an iPad or some Samsung tablet, not on a calibrated studio monitor. So what we did was build custom profiles that simulate those device classes as closely as possible on your own screen.

In other words, you are still looking at your calibrated monitor, but the software is applying a transformation that approximates how the image would appear on another device. It is not magic, and it is not the same as physically holding that exact phone in your hand, but it gives you a much more informed preview than simply hoping for the best.

Custom targets and ICC profiles

DP: There are already quite a few screen presets in there. Can users add their own?

Heath Barber: Yes, and that is where things get interesting. If you actually have access to the screen you want to simulate, you can create your own profile for it. The underlying idea here is the ICC profile, which is basically a standardised description of how a device represents colour.

If you calibrate that display with SpyderPro, create an ICC profile for it, and place that profile where Device Preview expects it, the software can use it as another target. So you are not limited to the built-in presets forever. You can extend the system yourself if you have the hardware and you are willing to profile it properly.

That is actually a useful kind of crowdsourcing. It means users can build simulation targets for screens we may not have added officially yet. Otherwise, the alternative is simply waiting for us to issue an update with more presets. This way, if you need a very specific display and you have access to it, you are not stuck.

In practice, this matters because not all displays fail in the same way. Some are cooler, some warmer, some over-saturate, some clip shadow detail, some compress highlights in ugly ways. An ICC profile gives you a measured description of that behaviour. Once you have that, you can simulate it much more intelligently.

„For Those Who Dare“

DP: Can Device Preview also be useful in a more classic post-production or grading pipeline? For example, can it simulate ACES or broadcast-style viewing targets?

Heath Barber: Yes, absolutely. Some of those things are already there in one form or another (RAW, JPG, PNG, TIFF, obviously, which is common for screenshots and frame exports), and one of the goals of Device Preview Plus is to support a much broader set of output types and colour spaces than the earlier beta did.

That is important because once you move beyond stills and basic display calibration, you run straight into colour-management questions. ACES, for example, is not just a buzzword people throw around to sound serious. It is a standardised colour pipeline used in film and post to help move images consistently between cameras, applications, VFX, grading and final delivery. Broadcast standards do something similar in their own context. They define how media should behave so that a signal is interpreted predictably.

What Device Preview Plus is trying to do is give users a way to preview those kinds of transformations and output conditions within a simpler interface. It is not trying to replace a full finishing pipeline, but it does let you think more concretely about where your image is going and how it might look once it gets there.

That matters because color correction is never done in a vacuum. You are always correcting toward some destination, even if people do not always say it that way.

DP: So in the long run, could this also help compare different source looks or make camera matching easier before you are deep into grading?

Heath Barber: That is part of the bigger ambition, yes. I would phrase it less as “this replaces grading” and more as “this gives you more context earlier in the workflow.” A lot of users do not only want calibration. They want to understand what their material is going to do when it moves from capture through editing and eventually toward delivery. So one goal of this relaunch was to build a stronger foundation for that kind of ecosystem thinking.

Part of that was hardware support, including better support for multiple monitors. Part of it was product structure, so users can start at one level and move upward without feeling that they bought into a dead end. But the big conceptual shift was image processing. That is what starts moving SpyderPro out of the little utility corner and into actual day-to-day workflow relevance.

If you can preview, compare and process images more meaningfully before the very end of the chain, that is valuable. People do not always need a giant facility pipeline to benefit from that. A photographer, editor, videographer or small studio can still gain a lot by seeing likely output behavior earlier and more clearly.

What C2PA actually is

DP: Let’s say somebody changes an image in the pipeline and then blames someone else. How do I know who broke it?

Heath Barber: That is where C2PA comes into the conversation. C2PA is essentially a standard for content credentials, which has been driven by Adobe and a few other companies, including Meta and Google. The simplest way to explain it is that it works like a verified provenance layer for media, which is supplied by approved partners and products. It is a way of attaching information to an image, video or audio file so that the history of that asset can be preserved and inspected.

Now, it is important not to oversell it. This is not an unbreakable security vault. It is not meant to function like heavy-duty encryption where the whole system collapses if one person touches the file. The goal is different. The goal is to create a reliable chain of information around authorship, origin and edits, so that platforms, applications and viewers can read that history and present it meaningfully.

If somebody strips that information out, that itself can become part of the story. So the point is not only to preserve history, but also to make tampering or removal more visible. Important to note is: Not everybody can supply the data, because a “C2PA” approval stamp needs a verified supplier to integrate it. Stripping the data away is easy, as anybody who has had an intern delete sidecar files before knows.

But adding that information requires a vetted, approved software company, like Adobe or Datacolor. Quite a few other big names in our space are currently in the process of that verification, by the way – we don’t know if this system will be the one that solves the problems, but as far as we can guess today, it might stick.

From our perspective, that becomes interesting because we sit so early in the image workflow. Calibration and device preview happen near the beginning of a chain. If you can begin attaching useful metadata there, and if that metadata continues downstream, then the image starts carrying some of its own history with it.

Tracking edits through the chain

DP: So if somebody in a studio changes an image and sends it on, that history could travel with the file?

Heath Barber: Yes, that is the idea, assuming the applications in the chain support reading, preserving and writing that metadata. If a file moves into Photoshop, for example, and Photoshop adds edits while preserving the content credentials, then a later viewer or platform that supports C2PA can show that there was an earlier version, then another change, then another change, and so on. That is why I often describe it more like versioning or historical tracking than some kind of hard lock. You are building a timeline of the asset.

And that is valuable because a surprising amount of confusion in media workflows is not about whether an image was changed. Of course it was changed. The question is how, where, by whom, and in which order. That is where provenance becomes useful. It turns vague suspicion into something much more inspectable.

Why that matters for media professionals is obvious. If you are a photographer, editor, colorist or VFX artist, so-called technical metadata is not boring background noise. It is often the only trail of breadcrumbs you get when something looks wrong and you need to work backwards.

DP: Could C2PA become a foundation for more reliable metadata in professional media workflows?

Heath Barber: Yes, I think it could. Right now, the public conversation around C2PA is often dominated by AI detection, content authenticity and platform labeling. Those are real issues, but for media professionals there is another side to it, and that side is workflow visibility.

A lot of tools and viewers today do not really surface the kind of information a colorist, editor or photographer would care about most. To them, that data may just look like clutter. To us, it can be gold. It can tell you what settings were used, what was embedded, what transformations were applied, and how the file evolved.

That is why I think there is a major opportunity here that goes beyond headlines about “real versus fake.” For working professionals, the value may be that you can finally see what happened to an asset in a way that is consistent and transportable.

Camera metadata, edit history and proof

DP: I assume camera makers already contribute a lot of that information.

Heath Barber: Yes, they do. A camera can capture all kinds of useful metadata: model, lens, focal length, exposure settings, stop values, capture conditions and more. That information is already part of the professional world. The big question is not whether metadata exists. It does. The question is whether that metadata survives and whether it remains visible and trustworthy as the file moves through the pipeline.

If you can preserve camera-side information, then add edit-side information, then preserve output and transformation information, suddenly you have something much more powerful than isolated metadata fragments. You have a chain.

DP: So, for a professional production, that could eventually help prove what is captured, what is altered, and what is generated?

Datacolor’s SpyderExpress: Colour Calibration on a Budget, Now with Growth Path Options

Heath Barber: That is certainly one of the ambitions around this whole space. It is tricky because media workflows are messy, and the real world is rarely as neat as standards documents would like it to be. But yes, the direction is toward being able to say: this is where the asset originated, this is what was done to it, and this is what remains attributable, and it is traceable. That is useful not only for protecting creators, but also for protecting the workflow itself. When something goes wrong, being able to inspect the chain matters.

DP: Ideally, you would like every screen in the workflow to have a little Spyder hanging from it.

Heath Barber: Sure, but the larger point is that I want to push us more into software. Spyder, as a hardware product, is well established. People know what a Spyder does, and the current platform supports most of the display technology our customers are working with today and what we see coming in the next several years. That is exactly what frees us to put our energy into software, because that is where we can deliver the most value the fastest. Workflow integration, format support, analysis tools, environmental adaptation, and content credentials. Those are software problems, and software is how we solve them best.

So Spyder remains the foundation, but the surrounding software layer is where we can build more capability and more relevance. Things like device simulation, image processing, and metadata let us move beyond “put sensor on screen, click calibrate, done.” That workflow still matters, but it does not have to be the whole story. And honestly, we are not done innovating on display calibration or colour-critical display analysis. 

DP: You recently updated the Spyder software. What is next?

Heath Barber: There is a lot in the pipeline, but the part I can talk about today is the ecosystem side. We are working closely with key vendors across both video and photography to make sure the experience is as seamless as possible from capture through delivery. The goal is to give creators the critical tools they need to do their best work and to genuinely improve their productivity, not to add another piece of software they have to babysit.

Imaging is a team sport. People move between cameras, software, displays, tablets, phones, edit systems, grading tools, review tools, and delivery platforms all the time. So if you want to be relevant, you cannot act like calibration lives on its own little island. It has to connect into the broader ecosystem. That does not mean every user needs a huge enterprise pipeline. It does mean the tools should feel aware of how people actually work.

The market has also changed. A lot of users today are not interested in spending endless hours tweaking every parameter inside complicated full-featured suites that are harder to use than the editing and DCC tools they are already trying to keep up with. They want tools that do most of the job well, clearly and efficiently. That is not laziness. That is reality. So part of the job is figuring out where in the workflow we can add value without making the tool feel like homework.

Light-meter integration and the viewing environment

DP: What about the light meter side? Is that becoming a bigger part of the system too?

Heath Barber: Yes, very much so. That is one of the more exciting areas for us. Built-in ambient light sensing is useful up to a point, but it is broad and limited. A dedicated light meter is much more accurate, and once you have that, you can begin describing the viewing environment in a much more meaningful way.

Right now, we already think in terms of multiple zones around the viewing setup. What is the light behind the screen, what is the light facing the viewer, what is happening around the workspace. The current system uses that information in a relatively constrained way, but the longer-term plan is to expand that into a much richer map of the environment, because your monitor does not exist in a vacuum either. The room changes what you perceive. Ambient light, direction, intensity and colour temperature all influence how you see the image.

So if you can measure the environment more accurately, you can calibrate more intelligently and also preview output conditions more realistically. In other words, better environmental data improves both the monitor correction side and the device preview side.

From three zones to room mapping

Getting Colors Right with the Datacolor LightColor Meter

DP: So eventually, instead of just one or two measurements, you could map the room.

Heath Barber: Exactly. Think of it less as a single reading and more as building a spatial picture of the environment. If you know what the room is doing around the display, then you can make much smarter decisions.

And that leads to some interesting possibilities. If you have multiple measuring points, or eventually multiple devices monitoring the environment, you can start blending those readings into a more complete model. That could matter for studio setups, mobile work or any workflow where lighting conditions are inconsistent.

Once you start seeing it that way, calibration becomes less about a single screen measurement and more about understanding the conditions in which that screen is actually being used.

DP: So the real shift here is that calibration is no longer treated as an isolated task.

Heath Barber: Exactly. That is really the core of it. Calibration still matters, obviously. It is the foundation. But once that foundation is in place, the next logical step is to ask what else can be built on top of it. That is where device simulation, image processing, metadata, content credentials and environmental measurement all come in. We are trying to move from a narrow utility toward something that participates more fully in the image workflow.

That does not mean it suddenly becomes a complete post-production platform. It does mean we are trying to make it more useful to the way people actually create, review and deliver media today.

The post SpyderPro learns new tricks: where colour goes wrong first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

The latest update to Nuke Stage stays with the preview: real-time playback of photoreal environments on LED walls, plus live compositing tools that shoulkd feel familiar to artists. The tool targets virtual production and in camera visual effects (ICVFX), with a workflow designed to keep VFX artists in creative control on set.

The core stays high-resolution playback for 2D and 2.5D environments, with EXR as the playback format and USD used for 3D geometry. Live compositing comes from a set of Nuke nodes rewritten for real-time performance, so operators can blend the on-set environment with the physical build and manage the relationship between camera and screen.

The practical takeaway: this is not a game engine project that happens to output to a wall. It is a VFX-flavoured playback and live comp system that tries to keep your pipeline language intact.

The parts you actually run on set

Foundry built a three part architecture: an editor where the operator works, a headless networking relay, and render nodes that run on each machine driving the wall. A launcher defines where those processes run and can start them across a larger stage, then the editor connects to the relay and render nodes and reports cluster health.

A typical topology puts the editor and relay on the control machine, distributes renders to the render nodes, then sends images to the wall. The setup flow walks through networking first, then hardware outputs and physical displays, then display mapping, and then cameras and tracking. Configurations can be exported and imported so teams do not have to repeat the same stage setup every day. Unless they want to. But in general, that is the unglamorous part of virtual production that still decides whether you shoot before lunch or after dinner.

Metadata Vault: your future self says thanks

The headline feature is metadata capture via a Vault that records on-set decisions for handoff into post. The Nuke Stage supports logging scene data, camera tracking, lens metadata, timecode, scene settings, and colour decisions. The Vault documentation also describes two capture workflows: Snapshots for a point-in-time save, and Take Record, which creates a camera USD file between defined start and end points.

If you have ever tried to reproduce a wall look from a vague note like “match take three but slightly warmer with some *Ooompf* “, this aims to make the record concrete. It will not fix bad habits, but it can at least preserve what you actually did on set.

NotchLC and EXR playback, plus a faster sequencer loop

On the playback side, the feature list now includes native support for NotchLC and B44 EXR for high-resolution background playback. There is a drag-and-drop sequencer panel, one-click keyframing, and editing and interpolation tools for timing changes.

In the operator view, the track-based sequencer remains central. Content created in Nuke can be staged over time, keyframed, and adjusted live for collaborative iteration with the on-set team.

USD scenes and splats, now aimed at the wall

Nuke Stage supports importing OpenUSD scenes that can be edited and overridden for real-time blending of virtual and physical sets. For file and colour standards, the tool lists OpenEXR and OpenColorIO alongside ACEScg support via ACES. It also claims support for HDR and a color pipeline intended to carry creative decisions from set into post.

For scene representations, the tool now highlights importing and controlling Gaussian Splats for high fidelity 3D scenes, alongside standard USD geometry. In the broader ecosystem, Foundry also shipped Gaussian Splat support in Nuke and built a USD based 3D system in there too, which positions splats and USD as shared data types across stage and post.

Standard hardware, scaling, and the usual fine print

The product page repeats the goal of running on commodity hardware without specialist media servers or bespoke equipment, and scaling across render node clusters. It frames this as cost and setup-time control and as a way to make stages repeatable.

Pricing is not specified in the sources. The product page offers a register your interest contact flow.

New tools and innovations should be tested before use in production, especially anything that touches playback codecs, cluster timing, genlock, tracking inputs, and colour consistency. Run your own worst-case plate, your own camera tracking feed, and your own handoff into post before you bet a shoot day on it.

https://www.foundry.com/products/nuke-stage

The post Foundry expands Nuke Stage for LED walls first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: KitBash3D sells themed 3D asset kits and delivers them through Cargo, with downloads for DCCs like Blender and engines like Unreal Engine.

A set dressing kit that actually knows set dressing

KitBash3D just shipped Lights, Camera, Action!, a 3D asset kit built around the behind the scenes world of filmmkaing. You get sound stages, grip gear, light stands, sandbags, and monitors intended to sell the idea that something got shot here, even if it was just your previz. The kit has 139 models, 112 PBR materials, and a total poly count of 6.3 million. If your scene still feels empty, you can now add more stuff that looks expensive, without having to explain why your hero prop budget went to folding tables.

Here’s part of the set:

• Apple Box: The ancient on-set unit of measure. Also the quickest way to “fix” an eyeline without touching your camera solve. Actor ego not included.
• Dark Apple Box: Same apple box, but in “this has seen things” edition.
• Bathroom Toilet: For those hero shots where the realism budget includes plumbing. Also useful for sim reference when compositing a bad day.
• Boom Microphone: The object your roto artist will definitely have to paint out, no matter how many times someone promised “it’s out of frame.”
• Box Truss Scaffolding: Because every virtual set needs at least one structure that screams “temporary,” even when it will live forever in a master scene.
• Camera Tripod High: The “we swear it’s locked off” rig that still mysteriously jitters just enough to ruin your matchmove.

• Chroma Stage: Your favourite shade of green, now available in “why is the spill that colour” format.
• Clapperboard: Classic clapperboards, which the onset personnel always mark in unintelligible handwriting, and VFX always tries to read like it’s a 20-times AI-translated-OCR-enhanced ancient manuscript.
• Coffee Cup: Continuity’s greatest enemy and the comp team’s most common set extension.
• Coffee Trailer: The real production hub. Put it in your scene and everyone instantly believes there was crew there.
• Dolly Rails: For that smooth move you will still stabilize because somebody stepped near the track.

• Extended Camera Crane: When the shot needs to feel expensive, even if the environment is still a greybox.
• Film Light: A two-lamp unit that screams “stage,” and whispers “watch your speculars.”
• Foam Core Boards: The poor person’s bounce, now available as a rich person’s render asset.
• Frosted Donut: The snack that disappears between cuts. Perfect for continuity tests you will fail on purpose.

• Frosted Donut B: A second donut, because the first one was “too pink in the grade.”
• Gray and Chrome Ball Rig: VFX’s emotional support spheres. Put them anywhere and suddenly you’re doing “proper” on-set reference.
• Macbeth Chart: The colour chart that politely reminds everyone that “we’ll fix it in post” still starts on set.
• Production Slate: The clean slate version, for when you want the prop without the handwriting crimes.

Unreal gets the game ready treatment

For Unreal Engine, the kit gets a Gameplay Ready variant with full Nanite support, custom collisions, Packed Level Instancing, and modular runtime ready assets. If you have been building virtual sets and thinking, this needs more realistic cable chaos, congratulations, you have found your people.

In practical terms, it is also a tidy addition to any kitbash workflow where you want believable production context around your hero environment, not just more neon signs and wet asphalt.

Formats, DCCs, and the inevitable renderer bingo

The supported software list includes Blender, Unity, Cinema 4D, Maya, 3ds Max, and Houdini. On the interchange side it calls out FBX, OpenUSD, and MaterialX. Renderer logos shown include Arnold, Cycles, Karma MatX on Houdini, OctaneRender, Redshift, and V-Ray.

Pricing, licensing, and the fine print you should still read

The product page lists a price of $95 and a current price of $47.50. It also shows three license tiers: Individual License for self-employed artists, Small Business License for companies with 50 or fewer employees, and an Enterprise License tier for companies with 51 or more employees. So far, we can see now discounts when you own some of the stuff in real life.

As always, test any new tools and assets in your pipeline before you let them anywhere near a real delivery, especially if your project depends on strict naming, instancing, and performance budgets.
https://kitbash3d.com/products/lights-camera-action

The post FilmSet kit lands in KitBash3D first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

Fallout Season 2 asked for more scale and less forgiveness. Rise delivered 244 shots across seven episodes, with the heaviest work landing early, and kept two sites moving as one crew. The playbook stayed familiar, but the demands went feral: closer cameras, heavier mechanical interaction, city-wide procedural destruction, and sand-driven reveals that punished every shortcut.

Fallout!

Andreas Giesen is a Rise VFX Supervisor with a Houdini-heavy background and a track record on FX-intensive shows. He is credited by the Television Academy for a 2024 Primetime Emmy nomination for Fallout in “Outstanding Special Visual Effects in a Season or a Movie”. His Rise FX credits include serving as the studio VFX Supervisor on The Matrix Resurrections, as well as work on has worked on X-Men, The Matrix, The Last of Us and the various entries in the Marvel universe, among others.

DP: When we last spoke about Fallout Season 1, you worked on a show that combined large-scale environments, destruction, Vertibirds, Power Armour, and complex integration work. On Season 2, where did the challenge shift?

Andreas Giesen: The challenge shifted purely to massive scale and extreme scrutiny. In Season 1, we were establishing the rules. In Season 2, we took those rules and stress-tested the pipeline. We had a huge jump in scale, like procedurally destroying the entirety of Los Angeles or dealing with the multi-layered sand reveals of Area 51, but we were operating with roughly the same core team size. Brute forcing was not an option. We had to lean heavily on our procedural setups and the USD foundation we built in Season 1 just to survive the shot count.

DP: Did Season 2 feel like a continuation of an established visual language, or more like a technical reset because the demands had changed so much?

Andreas Giesen: Artistically, it was a continuation. We already had that shorthand with Jay Worth and Jonathan Nolan, so we knew what that tactile, grounded Fallout look felt like.

Technically, though, it was a massive upgrade. We did not throw everything away, but we had to rebuild or upgrade many assets because the camera was suddenly much closer or the interactions were much more complex. The Vertibird docking with the Caswennan, for instance, required a whole new level of mechanical detailing that just was not necessary in Season 1.

DP: RISE delivered work across several episodes, with particularly dense work in some of the early ones. How did that affect the structure of the production on your side?

Andreas Giesen: We delivered 244 shots spread across seven episodes, but the bulk of the heavy lifting hit right in episodes 2 and 3. To keep the chaos in check across our Berlin and Stuttgart offices, we operated as one unified team. We ran single VFX dailies so it never felt like two separate facilities. We relied heavily on our in-house production management tool, risebase 3, to keep the sites synced.

Honestly, a massive shoutout to our VFX Producer Annelie Dangel , Production Manager Scott Kunysz, and the coordination team. They were the real firewall keeping the production structured while the artists dealt with the pixel pushing.

DP: You mentioned that the work was divided into more specialised teams. How exactly did that division work in practice?

Andreas Giesen: Instead of splitting the work purely by department, where an asset just gets tossed over the wall, we built small, cross-functional strike teams dedicated to specific sequences. For example, we had a dedicated Area 51 team and a completely separate Caswennan team.

Each of these pods basically consisted of one layout artist, one lighting artist, and an FX artist if the sequence required it. This meant the artists became absolute experts on the continuity, lighting setups, and specific technical hurdles of their assigned sequence, rather than constantly context switching between different environments.

DP: The Caswennan ship has become even more important in Season 2. At what point does a hero asset like that stop being “just a vehicle”?

Andreas Giesen: When you get that close, it absolutely stops being just a vehicle and becomes a massive flying set. The layout has to treat it like a static environment just to block the shots, especially when we are tracking complex character animation and heavy mechanical interactions, like the Vertibird docking, directly onto the hull.

DP: What were the rebuild decisions under the hood?

Andreas Giesen: It was actually a really interesting round-trip process. Since the flight deck scenes were shot on a virtual stage, we received the Unreal Engine asset from production as our starting point. Funny enough, that Unreal asset was originally based on our own Season 1 dirigible model!

So, our CG team, led by Jenny Leupold , took the updated proportions from the virtual stage, brought the asset back into our pipeline, and completely re-engineered all the high-res cinematic details and shading quality.

DP: How did you approach balancing a highly detailed close-up asset with practicality for layout, lighting, and shot iteration?

Andreas Giesen: We relied heavily on our USD workflows and instancing. We had to design functional heavy hydraulics and complex mechanical joints so the lifting arm physically made sense up close, but we could strip that back or use lighter variants for wider layout tasks so the scene would not grind to a halt in the viewport.

DP: In Season 2, there seems to have been a stronger overlap between physical production, virtual production, scans, and final CG. How did that affect the pipeline in concrete terms?

Andreas Giesen: It meant we lived and breathed LIDAR. For the Area 51 reveal and the hangar sequences, production gave us incredible plates and highly accurate scan data. Having that precise spatial data was the only way to guarantee a seamless transition from the practical sets into our massive digital extensions. Once that data enters production, it locks in your scale, distance, and speed, which were crucial for making the flight dynamics feel grounded in reality.

DP: There was also mention of Unreal-based virtual production material related to the airship. How did you manage the relationship between real-time assets, scanned set data, and final cinematic assets?

Andreas Giesen: Again, USD was the bridge. We took the Unreal assets, which gave us the exact scale and spatial relationships from the LED volume, and used them as the spatial framework to build our final pixel cinematic assets over the top. It ensured that when we integrated the final CG back into the live action plates, all the contact points and lighting references matched the physical set pieces perfectly.

DP: Was the biggest benefit from USD technical elegance, or simply that it stopped artists from stepping on each other’s work?

Andreas Giesen: Ha, a little bit of both, but honestly, stopping the file overwrites is worth its weight in gold. On a more technical level, the real elegance was how flawlessly it let us bridge the gap between Maya and Houdini.

For the Shady Sands environment, our Layout Lead Karin Wolf and her team built the pristine, non-destroyed city parts in Maya. We then brought that via USD straight into Houdini to scatter all the granular details and endless cornfields. Being able to mix and match layouts without friction meant artists could just stay in the tool they were fastest in.

DP: The Vertibird work in Fallout has always been interesting because it needs to feel stylised and believable at the same time. What were the challenges for it in Season 2?

Andreas Giesen: The challenge is that it is a massive, lumbering vehicle, but it still needs personality. For Xander Harkness arriving at Area 51, he is a fantastic, slightly cocky pilot, so we had to inject this aggressive precision and specific swagger into the flight path and mechanical braking.

The integration challenge, especially during the Caswennan docking, was finding that fine balance of heavy momentum. We had to ease the mass and simulate procedural micro vibrations to show it buffeting in the air currents. The hardest part was nailing the exact frame where the mechanical arm takes over the full weight and the rotors power down.

DP: How much of selling a vehicle like that comes from animation, and how much from camera language, editorial rhythm, and integration into atmosphere and environment?

Andreas Giesen: Animation gets you the weight and the performance, but the environment integration is what sells the lie. In the hangar landing, the plate was dramatically backlit. Our lighters had to introduce specific light sources to shape the geometry nicely and match the foreground. Then the FX rotor wash tied everything together. We simulated heavy CG dust and debris blowing right past the lens, matching the practical wind from the live-action plate to physically anchor the digital ship into the real set.

DP: For sequences involving scanned practical sets, such as hangar or landing environments, what are the biggest pitfalls once the scan data actually enters production?

Andreas Giesen: The biggest pitfall is that the scan data and the live action plate lock you into absolute physical reality. If the plate has a specific wind direction or interactive lighting, absolutely every piece of CG vegetation, dust, or vehicle integration has to physically react to that exact same environmental energy, or the illusion breaks immediately.

DP: Fallout still has this very specific balance: it is heightened and absurd in places, but the world still needs physical credibility. How does that affect VFX decisions in review?

Andreas Giesen: You have to ground the absurdity in absolute, boring physical reality. For the dramatic dirigible crash, which is inherently a heightened, spectacular moment, we went straight to historical footage of the Hindenburg. We ran intense, clustered simulations starting with plastic deformation for the metal bending, triggering RBD sims for the hull ripping apart, and feeding that into multiple layers of pyro for flames, smoke, and contact sand. If the physics are incredibly strict, the audience will buy into the universe’s stylised madness.

DP: On a show like this, where does proceduralism genuinely save time, and where does it still need a heavy layer of manual art direction to become production-ready?

Andreas Giesen: Proceduralism saves you on the sheer volume, but art direction saves the shot. For the destroyed Los Angeles, we started with basic map data, literally just cubes and planes. Our brilliant TD, Kilian Baur , built a Houdini tool that procedurally converted those cubes into heavily detailed, destroyed buildings. It generated massive amounts of geometry. But you cannot just hit simulate and walk away. We had to manually paint the degree of destruction across different zones for compositional reasons. The tool does the heavy lifting, but the artist still has to compose the frame.

DP: Was there a particular sequence in Season 2 where the real constraint was neither the creative idea nor the client feedback, but simply scene complexity, cache weight, or iteration cost?

Andreas Giesen: The dirigible crash and the Area 51 reveal were absolute beasts for cache weight. For Area 51, we were blending volumetric dust, suspended sand particles, and contact sand actively swirling off the geometry. The dirigible had even more interactive layers, expertly managed by FX TD Arnaud Malherbe under Jens Martensson .

We survived by heavily clustering the sims for faster turnarounds and by relying on deep compositing workflows in Nuke. Our Compositing Supervisor Michael Lankes , Comp Lead Kader Bagli , and Comp Lead Jannik Walzer did incredible work taking those massive datasets and balancing them without bringing the farm to its knees.

DP: What became easier on Season 2?

Andreas Giesen: Things like lookdev for the energy signatures became easier. We reused the core effect for the cold fusion and diode glows from Season 1. But because we had to scale it up to so many more shots, we had to build highly efficient setups in Nuke involving heavy rotomation and complex relighting, since that blue glow affects every shiny metal surface around it. The look was established, but the volume and the necessary integration were significantly harder.

DP: Was there a shot or sequence this season that looks deceptively simple in the final cut, but was technically far more difficult than most viewers would ever suspect?

Andreas Giesen: It has to be the Power Armor enhancements. Production has this absolutely incredible practical suit that a stunt performer wears on set. However, because it is a physical, wearable suit, sometimes individual parts will naturally jiggle or bounce a bit too much during heavy action, which breaks the illusion of solid, heavy metal armor.

In the shot where the massive dirigible is coming down and exploding in the background, we actually had to completely replace the shoulders of the Power Armor while the character is running away. Everyone in the audience is obviously staring at the massive fiery airship crash taking up the sky, but there we are, meticulously tracking and doing invisible CG replacement on shoulder pads just to make sure the suit feels appropriately heavy.

DP: And conversely, was there a shot that looked terrifying at first but became manageable once the right structure or workflow was in place?

Andreas Giesen: Definitely the procedural destruction of Los Angeles. The idea of destroying a city of that scale is terrifying. But once Kilian’s tool was in place, and we utilised our USD pipeline to block the shots with simple geometry while the high-res destruction updated seamlessly in the background, it actually became a highly manageable, iterative process. We could instance the endless piles of rubble and rusted cars, which kept the massive scale from becoming unrenderable.

DP: What is a technical lesson from Fallout Season 2 that would transfer well to other large episodic shows?

Andreas Giesen: The absolute golden rule is to keep everything as procedural as humanly possible so you can actually survive rapid client changes. That, combined with establishing those small expert strike teams we talked about, is how you manage the chaos. But the real lifesaver that transfers to any heavy show is aggressive pipeline automation. For sequences with massive shot counts, we utilised our proprietary riseflow tools to completely automate the heavy lifting.

We could submit CG renders from one principal master scene across 50 different shots before leaving the studio. By the next morning, the pipeline had automatically generated 50 temporary composites, already perfectly combining the new CG with the live-action plates using deep learning. Waking up to 50 ready to review temp comps instead of a massive rendering bottleneck is the only way to deliver a show of this scale.

DP: What is the next project you are working on, insofar as you can already talk about it, and is there anything from Fallout Season 2 that you are taking with you into that next production?

Andreas Giesen: I am currently working on something very exciting, but unfortunately, it is strictly under wraps right now, so I cannot speak about it just yet. But I am absolutely taking the USD workflows and the procedural pipeline strategies we refined on Fallout straight into this next challenge!

The post RISE: Fallout Destroys LA, Politely first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: OptiDraw is an in editor mesh painting extension for Unity that aims to keep vertex, mask, and splat style painting close to your materials, shaders, and scene context.

A mesh painter that stays in engine

OptiDraw positions itself as a framework for texturing work inside the Unity Editor, with a focus on centralizing painting and data handling rather than sending artists on a DCC roundtrip for every tweak. Marketing claims call out a zero lag painting experience on static and animated meshes, plus workflows designed to feel familiar to artists coming from Adobe Substance 3D.

The core pitch: paint directly on meshes in the editor, store paint data in reusable assets, and keep the iteration loop short enough that you actually use it. If your pipeline already leans on mesh painting to drive blends, wear, grime, decals, or procedural masks, the idea here is simple. Do the edits where you preview the final shading.

Setup, shortcuts, and the first five minutes

OptiDraw claims (!) a one click setup flow. You open the Mesh Painter window from a top menu entry labeled R2G then OptiDraw then Mesh Painter, dock it where you want, and select an object in the scene.

On selection, the tool describes an auto initialization step that prepares a MonoBehaviour component, adds a Mesh Collider to the selected GameObject, and links a ScriptableObject asset to store painting data. The tool also bringsa Draw Mode toggle you can hit via the B key or a Draw Mode button to enter paint mode, with a brush gizmo shown in the Scene view.

There is also a compatibility check concept baked into the workflow. The documentation describes checks for mesh and shader compatibility, and it describes UI states where Draw Mode becomes non interactive if no compatible object is selected.

Data management, not just brushes

OptiDraw leans hard on a ScriptableObject driven data model, with an OptidrawData asset as the persistent storage for vertex colors and mask data, and an OptiDraw Mesh component as the bridge that connects that data to a GameObject.

The tool has a non destructive auto copy approach that keeps the original mesh untouched while managing the data needed for painting. It also includes a folder strategy split between an Editor folder setup intended to keep final build size smaller by excluding raw data, and a Runtime or Resources approach intended for cases where you need to modify paint data via C# during gameplay for things like dynamic damage or snow.

This matters more than it sounds. A painter that creates mystery folders and fragile references becomes a liability fast. OptiDraw makes claims about clean data management and efficient linking. In practice, you still want to verify how it behaves with prefab variants, source control, and multi artist merges before you bet a production on it.

Two engines, two very different failure modes

OptiDraw exposes two headline painting modes: Vertex Painting and Splat Painting.

Vertex Painting stores data directly in mesh vertices. The documentation frames it as useful for variation, blending, and environmental detailing without high resolution texture overhead. It also calls out the obvious constraint: resolution depends on mesh density, and topology changes after painting can break the relationship between vertex IDs and the stored data. If you plan to repaint after mesh edits, you will want a repeatable process and a stable asset handoff.

Splat Painting stores data in a texture map via UV coordinates, framed as a way to get pixel precise masks for complex blends and weathering effects. The documentation also lists the expected tradeoffs: texture memory, a need for clean non overlapping UV layouts, and some performance overhead from shader sampling compared to reading raw vertex data. If you ship on constrained hardware, this becomes a budgeting conversation, not a feature checkbox.

The tool also describes channel targeting across RGBA, paint operations like add, replace, and erase, a normal limit control to reduce bleed across sharp angles, and a seamless paint option to minimize seams across UV islands, texture borders, or submeshes. Around one third into a production, this is also where you will appreciate having robust undo and redo instead of a single irreversible mistake that haunts your repo forever. Marketing claims say undo and redo is robust. I did not test that.

Animated meshes and the unglamorous armpit problem

OptiDraw supports Mesh Renderers and Skinned Mesh Renderers. For skinned meshes, the documentation describes an Animation Draw tab that appears automatically when a Skinned Mesh Renderer is detected.

Animation Draw Mode is described as a specialized mode that lets you paint directly on meshes while they are in motion or at a specific animation frame, with a single Animation Clip slot you drag and drop into the UI. The UI is described with play and pause controls, a scrubber, frame stepping, and a clip duration readout.

The documentation describes the workflow advantage in plain terms: you can pose a character to access difficult regions and paint fixes in context, rather than guessing in T pose. It also claims deformation consistency by being aware of bones and weights, with the goal of avoiding texture sliding or warping issues. That is a marketing claim, and it is exactly the kind of thing you should stress test on your own rigs, since skinning setups vary wildly.

If your project sits inside game development production reality, you already know the rule: always test on the worst character in the build, not the best demo mesh.

UV channels, UDIM tiles, and why your mesh can still be the bottleneck

OptiDraw has support for up to four UV channels, UV0 through UV3, and it also describes UDIM aware splat painting for UV islands that extend beyond the standard 0 to 1 space. The documentation frames this as important for splat painting setups where textures use clamp for performance and where other tools might force remapping.

It also describes how UDIM support gets surfaced in the UI. You link an OptidrawData asset, inspect mesh info to see which UV channels support UDIM and how many submeshes exist, then navigate UV tiles through a viewer that shows tile coordinates. You then input those coordinates into the OptiDraw Mesh component per submesh, with the note that optimized workflows typically separate tiles into different submeshes for material needs.

Mesh compatibility still matters. The documentation describes an incompatible mesh case where the active UV channel contains overlapping islands or islands that cross tile boundaries, which can cause incorrect painting results and unexpected behavior. It also describes a compatible mesh case where at least one UV channel contains islands properly contained within a single tile, even if other channels have cross tile UVs.

It also describes a Unity side fix for missing or invalid UV data using the model importer option to generate lightmap UVs, producing a secondary UV channel where islands pack into a single tile. That is a pragmatic fallback, but it does not magically fix every projection requirement. Asset quality still sets the ceiling for how well any painter behaves.

Layers that look like a familiar UI, with an unfamiliar performance bill

OptiDraw includes a Layer System similar to modern image editing software and also similar to Unity Terrain Layers in how you select and paint a specific layer. There is also an explicit performance warning: the system is designed to support four PBR texture layers as a sweet spot. The documentation says the system can support eight layers or more on a single object but advises against eight layers on one mesh due to performance bottlenecks from texture sampling and memory overhead. That warning is more useful than most feature lists, because it tells you where the tool expects you to stop.

If your shader stack already sits in URP, you will still want to validate how many texture samples you can afford in your target platforms, and how many variants you want to maintain once layers enter the mix.

Shader integration: template route or build your own

OptiDraw describes shader requirements for proper data communication between the painter and the mesh, and it frames shader integration as either plug and play with templates or a custom setup.

The documentation says the package includes 11 template shaders located at Assets then Run2Go Studio then OptiDraw by ProfX then Shaders, and that you can find them in the shader picker under a category labeled R2G then LayeredLit. It also says that for developers using Unity 6.3, the tool is compatible with an official Sample Texture Terrain Shadergraph with minor tweaks. For custom setups, the documentation describes Shader Graph custom support via 21 subgraphs and custom HLSL scripts for blending and vertex modifications, described as modular so you can include or exclude functions.

Utility tools: channel packing with guard rails

OptiDraw includes a Texture Channel Packer described as a free utility that combines up to four grayscale textures into a single RGBA texture. The documentation frames this as a standard channel packing workflow for PBR maps to reduce texture samplers and memory footprint, and it says the tool supports custom channel mappings including common conventions like ORM.

Export formats are TGA and PNG, and the channel packer window cannot be docked as a tab within the Unity Editor.

Price and release details

On the Unity Asset Store, OptiDraw is listed as an “Extension Asset” and the latest version 3.2.2 and an original Unity version of 6000.0.67. The render pipeline compatibility section shows compatibility for Built in, URP, and HDRP at Unity version 6000.0.67f1.

The Asset Store price is €36.79.

As always, test any new tool before you let it touch production assets. Verify the data it generates, the shader hooks it expects, and the performance profile on your target hardware before you scale it across a project.

Link Citation Block

https://assetstore.unity.com/packages/tools/painting/optidraw-advanced-mesh-painting-tool-361248?clickref=1110l3SaYMtQ

The post OptiDraw paints animated meshes inside Unity first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

What Project Violet is

Project Violet is a tutorial series made in Houdini 21, of the “Intermediate” Level, by Sam Taylor. The series shows new character features and workflows in Houdini 21 while combining the production readiness of Epic’s MetaHuman system with Houdini rigging and simulation workflows to build an approachable full human workflow.

That matters because it narrows the promise. This is not a inspiration reel, but a character pipeline exercise, with the character coming from MetaHuman and the heavy lifting happening in Houdini. If your day job sits anywhere near character setup, deformation, or character-adjacent sims, Project Violet wants to show you how to reuse, get a decent iteration speed, and getting a believable human through a procedural toolset without constantly falling back to manual rig babysitting.

The pipeline pitch: MetaHuman base, Houdini control

The main idea: take MetaHuman as the starting point and use Houdini for rigging and simulation versatility. It calls MetaHuman “production-ready” (we could discuss that, yet, the character is still in intensive care), and it presents that as the baseline you inherit rather than something you rebuild. Everything after that becomes a question of how far you can push Houdini’s character toolset once you stop spending weeks on initial asset authoring.

Houdini 21: Like good wine (Part1, VFX & Geo)

A separate SideFX-side (hehehe) for Project Violet describes it as an evolution of an earlier character-centric project, and it wants to focus on greater realism and expanded capability of the character, to showcase Houdini rigging and animation capabilities updated for new features and efficiencies in Houdini 21, and demonstrate a character pipeline that emphasises proceduralism while using MetaHumans as a base for the workflow.

That language is still a pitch, but it dosn’t widen the scope. The project is about a workflow, not a single solver trick. It aims to demonstrate procedural character work in Houdini 21 with a MetaHuman starting point.

One practical caveat: none of these claims your exact studio constraints, problems and idiosyncrasies of the colleagues get solved. Your skeleton naming rules, mesh structure, and export requirements still exist. Project Violet shows a route, not a universal treaty.

Houdini character tooling under the hood: APEX Autorig Builder

If Project Violet is the show, the APEX Autorig Builder is one of the stagehands. The APEX Autorig Builder SOP is a tool to interactively apply and configure pre-built and custom rig components through the viewport, which aims to make character setup more intuitive and lets you quickly build a complete rig.

SideFX draws a line between the APEX Autorig Builder SOP and the APEX Autorig Component SOP. The Builder can add multiple rig components, supports interactive viewport rigging, and allows modifying a limited set of component parameters in the viewport. The Component SOP adds one rig component at a time, can be used to create custom components, and exposes the full set of parameters in the parameter editor.

Back to Project Violet, because it glues the workflow in Houdini’s modular rig component approach rather than a one-off rig graph built for a single hero asset. Even if you do not adopt the exact tutorial approach, those constraints and mechanics tell you what Houdini 21 expects when you try to scale character setup.

Unreal and MetaHuman side of the bridge

Project Violet ties into Unreal. Epic’s MetaHuman are tools to assemble a MetaHuman character or create MetaHuman-compatible grooms in SideFX Houdini.

That tells you two things. First, the pipeline is not only about ex/im-porting Alembic and hoping for the best. It involves a toolset that lives in the Unreal ecosystem but targets Houdini as the working environment for certain tasks. Second, the workflow depends on version compatibility and installation details that can change across engine builds, pipeline images, and studio IT lockdowns.

This is where the series may appeal to teams that want to get to motion and deformation problems quickly, or to studios that already standardised on MetaHuman for certain classes of characters and now want deeper procedural control on top.

Either way, run a controlled wokrflow test before you bet a show schedule on it.

https://www.sidefx.com/tutorials/project-violet/

https://www.sidefx.com/docs/houdini/character/kinefx/autorigbuilder.html

https://dev.epicgames.com/documentation/metahuman/metahuman-for-houdini

The post Project Violet links Houdini and MetaHuman first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

The pitch: graphs, not layers

ArcBrush builds images in a node graph. You wire inputs into operations and wire the result into exports. The app focuses on non-destructive editing, with parameters staying live so you can adjust a pipeline long after the first export. The core editor is free to download and use, and it does not require an account for non-AI work (Oh, blessed 2026, when this is a mention-worthy feature of a software….) . It runs as a native desktop app written in C++ and works on Windows and macOS. Linux coming soon.

If you live in node land already, the concept will feel familiar. The difference here is that the output target is mostly 2D deliverables: icons, sprites, UI variants, texture maps, and piles of formats that someone will ask you to regenerate five minutes before lock.

Node count, node types, and the day job stuff

The app has 75 composable node types and type-safe connections that prevent invalid wiring. It also brings five pin types, separating images, masks, palettes, and variants so the graph can validate what you connect.

A lot of the feature list reads like it came from someone who has rebuilt the same export stack too many times. There is auto save with crash recovery. There is a minimap, node search, undo and redo, node groups, per node notes, and sticky notes.

The preview tooling targets production rather than marketing . You can lock the preview to any node output, zoom deep with a pixel grid, inspect RGBA values, isolate channels, and view a live histogram overlay. For monitoring consistency, the app reads your monitor ICC profile and applies correction so the display matches exports via display color management using the monitor profile from the International Color Consortium. Test that on your own display chain, especially if you work across multiple calibrated panels, because colour surprises are lovely… .

On the output side, there are single-image exports and batch exports. Export Batch can produce named files per palette variant, or pack variants into a single sprite sheet PNG while also writing a TexturePacker compatible JSON manifest. Export nodes also support relative paths, and the app can create directories automatically when exporting.

Pixel art and texture helpers, without pretending to be a full DCC

The feature goes for game art and texture workflows, but it stays in 2D territory. There are pixel art oriented nodes like Pixelate, Outline, Dither, Quantize, Palette Remap, and sprite sheet export. Dither lists Floyd Steinberg, ordered, and blue noise options.

For texture work, the app includes Seamless Blend to make images tileable, Noise and Pattern nodes for procedural bases, Tile Preview to check repetition, and a Normal Map node that generates lighting maps from grayscale. This is not a replacement for a full material authoring suite, but it covers the everyday prep-and-iteration loop where you need quick variants, quick tiling fixes, and quick exports without babysitting layers.

If you spend time in texture painting or material prep, the most interesting claim is the pipeline idea: build the graph once, then regenerate every variant and export target when the source changes. That includes a Watch File toggle on input nodes that reloads when the source file changes on disk. The page even spells out a simple case: edit in Photoshop and see the result live in ArcBrush. That makes it less of a walled garden and more of a graph-driven post station for assets.

OpenCV processes images on CPU and GPU, and OpenGL renders previews. For SVG import, it rasterizes via ThorVG. None of that guarantees performance on your particular workstation, but it does clarify that the app aims for native processing rather than another browser wrapper.

Optional AI nodes: credits, cloud, and knobs you can count

AI features are optional. The app has AI cloud based nodes that only reach out when you use them, while the rest of the graph runs locally. There are nodes for text to image, image edit with text instructions, background removal, and 4x upscale powered by AuraSR. It also has differnet models, including FLUX variants and Nano Banana 2 for text to image and editing.

Pricing for AI is a one time credit packs with no subscription and credits that never expire. The packs listed are Starter at $10 for 350 credits, Plus at $20 for 800 credits, and Pro at $40 for 1800 credits. The site also states 10 free credits on signup. Credit costs vary per model and task, and at standard resolution up to 1024 by 1024.

That setup makes it easy to budget AI usage as a production expense, but it also adds a network dependency for those nodes. Treat it like any other external service: check availability, latency, and whether cloud processing fits your client and studio rules. Or if the bubble already burst, and you need to rewire things to the little Docker Server you have in your Cupboard that replaces all the Hype…. If you only want the node graph editor, you can ignore AI completely and still use the full non AI toolset.

Where it fits if you already have node tools

Node based workflows already dominate compositing and lookdev, so the real question is whether a dedicated 2D node graph editor reduces friction compared to existing tools. ArcBrush positions itself as a build once and export everything machine: variants, palettes, batch exports, sprite sheets, and texture prep steps chained together so you can re run them on demand.

First, when you need to ship a lot of consistent variants, like tiered item icons, UI colorways, or platform specific asset sets. Second, when you need to keep a process editable after the first deliverable, because late changes rarely arrive in a neat package.

Reality check before you bet a deliverable on it

ArcBrush looks designed for artists who want node graph control for 2D assets and textures without living inside a heavyweight suite. The free core editor and the explicit credit pricing for AI lower the barrier to testing (Which we’ll do after FMX) , and the node count and export features cover a lot of common production chores.

Still, new tools and innovations should be tested before use . Put it on a few real assets, run it through your team review loop, and see whether teh graph model actually reduces mistakes compared to your current wokrflow.

ArcBrush is available as a free download from the product site, with documentation hosted on the docs portal and community links including Discord.

https://arcbrush.com/

The post ArcBrush turns image editing into a node graph first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the topic: The Visual Effects Society published the VES On-Set VFX Data Collection and Usage Guide to align on set capture and handoffs between production-side VFX, production, facilities, and technology teams. Specific DCC application support is not specified in the sources.

Webinar first, PDF later

If you would rather hear a walkthrough than skim a document, a public webinar runs May 12, 11:00 am to 12:00 pm PDT, with registration via Zoom. The session explores the guide with some of its creators and keeps the conversation in the open, where on set data questions belong instead of buried in a frantic email chain. Join Here

The guide’s job: stop the “what did we shoot” spiral

The VES On-Set VFX Data Collection and Usage Guide is a practical on-set resource that maps key data sets and capture workflows. Its goal is a shared playbook for productions, vendors, and technology teams to use and capture on-set data more effectively.

It aims to establish a common language between on-set VFX, production, VFX facilities, and technology teams, with the intended outcome of clearer communication, smoother handoffs, and better-aligned expectations across departments. In teh real world, that translates to fewer “we thought you had it” moments when the truck already left.

What it actually contains

The guide explains major on-set data sets, their capture methods, their practical applications, and intended stakeholders. It also documents both current and emerging on set data capture workflows and the choices that can impact production pipelines, timelines, and budgets. Included is a sample structure for directories, Reference Docs, ToDo Lists, Scope Definitions and many more usefull bits.

It also frames modern filmmaking as a hybrid of real-time virtual production, AI, and traditional pipelines, and the VFX department these days as responsible not only for post, but also for on-set data capture, continuity, and asset integrity from pre-production through final delivery. It seems the VFX department has become the primary custodian for management and tracking of on-set data across a project’s lifecycle. You wouldn’t trust actors with that, would you?

The boring parts that save your schedule

But on-set data has value beyond VFX. It states that data can support better informed creative and operational decisions and argues for open access and visibility so that teams across a production can engage with and benefit from shared knowledge.

One example is the cross-department logic: LiDAR scans created by VFX can be repurposed for set construction, stunt planning, and other production needs. Obvious, you’d think, but… That is the wrokflow reality check that makes the guide feel like a crew document rather than a glossary, and now we just have to get everybody to read and, ideally, understand it.

Credits, licensing, and the fine print you can actually use

The guide was developed over the past year by the VES Technology Committee. Its lead author is Sheena Duggal, with contributions from Sam Richards, Jim Geduldick, and Jake Morrison, and technical support from Jean-Francois Panisset.

The guide is licensed under the Creative Commons Attribution 4.0 International license, which permits redistribution and adaptation in any medium or format, including for commercial purposes, provided appropriate credit is given. So, no excuses for “Oh, we don’t have that, so we couldn’t follow it!”

Get the PDF here: On-Set VFX Data Collection and Usage Guide

And browse the Interactive Version here: On-set Data Specs

https://vesglobal.org/event/webinar-introduction-to-ves-on-set-vfx-data-collection-and-usage-guide-online/

https://ves-on-set-data.org/

The post VES releases On-Set VFX Data Guide first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: World Creator is a GPU terrain tool from BiteTheBytes with bridge plugins for Blender, Cinema 4D, Godot, Houdini, Unity, and Unreal Engine.

The big headline is a small textbox

World Creator 2026.4 lets you type mathematical expressions into numerical fields. That sounds like a nerdy feature, because it is. It also means fewer hand-calculated offsets, fewer copy-and-paste accidents, and more repeatable setups when building layered terrain that needs to stay art-directable. The same release adds full decal support. Decals are often the difference between a terrain that looks generated and one that looks dressed. They also help when you need localised detail without turning every change into another texture bake cycle.

Seams get quieter

Terrain normal blending wants to improve object-to-terrain transitions, so reduce the obvious cut line where a hero object meets the ground. If you build environments where a single rock edge can pull focus in a wide shot, this is the kind of “small” addition you end up using all day. World Creator 2026.4 also introduces configurable object shadow settings. That gives artists another lever for matching scene lighting behavior when objects sit on terrain, especially in lookdev and layout passes where readability beats physical purity.

Performance fixes that actually sound like production

VRAM scaling links the Max Object Count to available GPU VRAM. In practice, this ties a key scene limit to the hardware ceiling instead of a fixed guess, which can make scenes behave more predictably across machines.

There is another performance change that targets daily workflow friction. Shape Design Mode no longer triggers automatic material recomputes.

Bug fixes: import, UI, syncing, crashes

World Creator 2026.4 fixes an importing issue where Raw and R16 formats were hidden during import. It also corrects compass centering in the UI. On the data side, the update fixes missing Baseshape Color and Emissive toggles in the sync.xml file, resolves issues involving empty model objects, and fixes Stochastic Distribution failing when used with height blending. The release notes also state that several crashes were resolved.

Community Edition and licensing tiers

BiteTheBytes now offers a free Community Edition, positioned as a way to learn the realtime terrain workflow before upgrading to a commercial license. The export is disabled in Community Edition, while paid tiers enable export, with Indie listed as up to 8k and higher tiers listed as unlimited.

Yearly subscription pricing on the buy page lists Indie Yearly at $49 per year for revenue under $100,000, Professional Yearly at $99 per year for revenue up to $200,000, Studio Yearly at $199 per year for revenue up to $1,000,000, and Enterprise Yearly at $799 per year for revenue above $1,000,000.

Perpetual pricing includes Indie Perpetual at $99 and Professional Perpetual at $199. Enterprise Perpetual includes 12 months of updates with an optional renewal listed at $449 per year. Lifetime pricing shown includes Indie Lifetime at $249, Professional Lifetime at $499, Studio Lifetime at $999, and Enterprise Lifetime at $3,999.

https://docs.world-creator.com/release-notes/version-2026.x/world-creator-2026.4

https://www.world-creator.com

The post World Creator 2026.4 adds expressions and decals first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: Mavis Studio is a touch-first live switcher for iPad that ingests cameras over NDI or a companion camera app, mixes audio, overlays HTML5 graphics, and outputs to stream, record, HDMI, or NDI.

The iPad gets a vision mixer job

Mavis ships a live production app for iPad that runs a Program and Preview workflow with an interactive multiview, source tiles, and a transition section that can take cuts or mixes between Preview and Program. The multiview is configurable, can be locked to hide setup controls, and can bank to a second multiview setup.

A show saves as a document that keeps the show assets and settings together, including imported media that gets stored inside the show document for faster playback.

Sources in, including remote phones and local NDI

You can bring in camera sources over NDI or by connecting up to four iPhones running the companion camera app. The app supports up to four NDI inputs, including NDI HX3 or full bandwidth NDI, and it includes an integrated tally system.

Camera discovery can happen through a cloud hub that helps devices find each other, or by finding nearby devices for local discovery. Nearby discovery uses the same underlying idea as AirDrop for bringing devices close to discover them, then letting the connection route over the most appropriate network path.

External cameras can also enter the chain via adapters that feed HDMI or SDI into a phone running the camera app, including Atomos hardware like Ninja Phone and Accsoon capture devices in the SeeMo line. That keeps the phone as the ingest point while still using larger cameras.

Players, layouts, and web graphics as a first class citizen

The built in Media Player can play videos, and the Image Player can show stills, with the option to auto play media when it appears on prgoram. Media control includes a jog wheel for frame accurate control.

Graphics can come from a Web Player that loads a URL and can be used as a keyed overlay, with HTML5 based graphics overlay support. The app also supports scene layouts, including preset and custom layouts, and layouts can use up to four layers plus a background. You can move, resize, and crop live video inside layouts, then cut layouts directly like any other source.

iPad web graphics come with platform constraints. iPadOS does not allow video playback inside these embedded web views, and heavy JavaScript pages can consume more system resources.

Audio is a mixer, not a checkbox

Studio includes a 16 channel audio desk with channel strips and controls that include faders, pan, gain, EQ, grouping, and master output control. Sources can present audio metering per tile, and grouping can act like VCA style control for managing multiple channels quickly during a show.

Outputs: stream, record, HDMI, and a flexible NDI send

Streaming can target YouTube and Twitch or other destinations using RTMP or SRT. Local recording writes the output directly to disk for archive or edit.

For hardware, the app can output over wired HDMI from the iPad. NDI output works as a send that can transmit a selectable source from the internal video matrix, rather than being locked to Program only. That allows workflows like sending Program, a layout, a media player, or a clean feed to another device on the local network while still running your main show.

If you rely on NDI, validate network stability, device thermals, and end to end latency before you bet a paid gig on it. streaimg is a harsh teacher.

Pricing, platform requirements, and what is still vague

The app is listed as a free download on the App Store with in app purchases. The listing shows four purchase options titled Remove Watermark at $24.99 and $79.99, and Remove Watermark + NDI at $39.99 and $129.99. The App Store listing also states that some features require an app subscription, but it does not specify durations for those price points.

The App Store listing states iPadOS 26.0 or later is required. It also lists compatibility for macOS 26.0 or later on Apple silicon Macs, and visionOS 26.0 or later.

https://mavis.cloud/studio

Suggested tags:

The post Mavis Studio turns iPad into a live switcher first appeared on DIGITAL PRODUCTION and was written by Bela Beier.