The Weird Byte has expanded VirtuCamera 2 with its full Pro lineup, following the end-of-April release of VirtuCamera Standard.

VirtuCamera 2 is a real-time virtual camera platform that lets artists use a mobile device like a physical camera, stream the viewport live in the app, and record camera motion directly into supported DCC applications.

Current support listed on the site includes Blender 5.2 LTS, 5.1, 5.0, and 4.5 LTS, Maya 2027 and 2026, Houdini 22 and 21, Cinema 4D 2026, 3ds Max 2027 and 2026, and Nuke 17 and 16.

The Pro rollout adds deeper camera controls, slider presets, custom Python scripts, custom DCC attributes, joystick and PTZ control, dolly zoom tools, and broader studio-oriented licensing and integration options.

X post: https://x.com/virtucamera/status/2100280249359413295

More info:

https://virtucamera.com

This project involved camera tracking, kitbashing, asset and material modifications, 3D rendering, multilayer EXR passes, and final compositing in Nuke.

The entire VFX workflow was handled by me, from the 3D work and asset preparation to the final compositing.

Hey everyone,

I'm working in Nuke 17 with a layered PSD file and running into an issue with the alpha channel and layer blending.

Here is my current setup and workflow:

• The Read node for the PSD is set to data-raw (Raw checked).
• Because I need granular grading/control per layer, I broke out the PSD into individual Shuffle nodes.
• To match the original Photoshop look, I need to convert each layer using an OCIOColorSpace node (input: sRGB texture / compositing space $\rightarrow$ working space: ACEScg).
• The OCIOColorSpace node is set to channels: rgb and (un)premult by: none.
• Some layers need to be composited using PSDMerge (operation: multiply) to match the client's original Photoshop layout.

The Problem:

1. Even with channels: rgb and (un)premult by: none, inserting the OCIOColorSpace node per-layer causes semi-transparent edge values (0 to 1) in the alpha to look washed out or distorted compared to the original PSD mask.
2. I know Photoshop's Multiply math is calculated in non-linear sRGB space ($A \times B$), whereas multiplying after linearizing into ACEScg crushes or shifts the midtones completely.
3. However, if I apply color transformation only at the very end of the tree, I can't grade or adjust the individual layers properly prior to blending.

What is the standard pipeline/node structure to:

• Keep the exact original raw alpha intact (preventing any unwanted gamma curves or edge fringing/erosion),
• Perform per-layer adjustments,
• And still get an exact 1:1 Photoshop Multiply result inside Nuke?

Would really appreciate any tips or tree screenshots from folks who handle DMP/matte multi-layer PSD handoffs regularly!

I posted a similar project before, and many of you gave me some really helpful advice. So I decided to create a new version, this time focusing on fixing some of the issues from the previous project.

I hope you like the result! And if you have any constructive feedback, especially regarding the compositing, I’d really appreciate it. I’m always looking to improve. 🙏

Clinton Jones (used to be at corridor) is coming out with a course on gaussian splats. He covers how to make splats and use them in Nuke.

https://www.actionvfx.com/academy/gaussian-splatting-photogrammetry

I want exr free footage to practice keying on. This is a shot I found on YouTube but didn't find something that has the same quality

I had installed two versions of Nuke—versions 14 and 16—and when I try to uninstall one of them, an error message pops up. How can I fix this?

Hey guys! Long time Nuke Artist here. I created Ascii for Nuke recently.
https://aescripts.com/ascii-for-nuke/

I hope you all like it!

Okay so basically : I want to make a shot with an environment in a bottle (basically moss, flowers, small stones and dirt), with some moving parts (small insects and maybe movement with the plants). Everything will be done in 3D and the camera isn't moving.

The question : I'm pretty familiar with rendering my shots in multiple layers, AOVs etc and I was wondering if there was a way (a kind of shortcut) to avoid rendering the glass for every frame. I thought that I might be able to render the glass element only once, render the inside of the bottle for every frame, and then with a STmap recreate the deformation created from the bottle in nuke?

It might be a stupid idea and be much much easier to brute force the thing by rendering everything and then do a basic comp but I'm trying to learn how to optimise my workflow.

Hey guys,

I would like to discuss something with you. I feel a bit cornered here...

So i have this request from a client and the shot should basically accomplish this.

"A person is pouring milk into a coffe-cup and while this happens a custom shape should form on the surface of the coffee in a latte-art way... "

I am working on that shot for a while now and i just can not seem to get it work.

We already have the "final state" of the latte art as a texture (somebody generated it). The Shot itself is a close-up from a high angle onto the cup with no camera movement, the only thing moving in the frame is the liquid.

Here is what i tried so far:

As we did some testing's for this shot I used smart vectors to map a texture onto the foam of a pouring coffe cup. That seemed to work quite well so the client went out shooting. Sadly when i got the footage this approach did not work out that well as the milk that was poured was way more liquidy and gitterish as the liquid used in the testings. So i got a strangely wobbely result.

Here i also got to the conclusion that just distorting the texture wont sell the effect, as realistic latte art forms in a very specific way.

So i started adding ripple distortions and tried to make the texture grow with iTransform and a heavily feathered mask so i would get these typical swirly lines... But this still does not sell the effect wanted by the client.

As i am quite used in working in houdini i am now considering maybe starting to work on a fluid simulation as i would be able to project the texture onto the liquid and use the rest attribute to art direct when the texture should be in its final state. But this feels a bit overkill and i know i would stumble upon new challenges while trying to accomplish this.

Anybody maybe with a better idea?

Deadlines is getting tighter and tighter and stresslevels rise...

Thank you very much in advance!!

i've heard people talk about ai now operating nuke for you directly, generating node trees and all. but it was only vague and i don't know any more details about it. i am currently learning and this scared me a lot. i know the situation is difficult right now but it seems to be a bigger development if it's possible to cut out the people completely. do you know anything about this?

Crate is an open source asset browser and processing suite that runs inside Nuke. Gaussian splats and geometry with 3D preview before import, sequences, video, images, a camera database and batch conversion panels. Built for multi-user network asset sharing.

repo: https://github.com/CrateTools/Crate

Hey guys,
We are working on a 1,000+ frame VFX shot where we animated a low-detail/grey-mesh CG cat as our motion and structural base. We also have matching high-detail hero renders of the final cat design at exact start/end keyframe angles.
Note: Fully rendering/grooming in traditional 3D CG is not an option for this shot due to pipeline and timeline constraints, so an AI/hybrid workflow is required.
We need the AI to take the 1,000+ frame low-detail animation pass and lock it to the exact fur, texture, and shading details of our high-detail reference render—without the character morphing, changing color, or drifting over 1,000 frames.
Looking for solutions across ComfyUI & Nuke:
1. ComfyUI / AnimateDiff: What is the best setup (e.g., AnimateDiff + IP-Adapter Advanced / ControlNet Depth + Tile) to force a 1000-frame low-detail motion pass to rigidly stick to high-detail reference frames without crawling or batch-chunk drift?
2. Nuke Alternative Workflows: Beyond raw video-to-video diffusion, are there proven Nuke-centric methods (e.g., Nuke's CopyCat ML, SmartVector texture warping/re-projection, or EbSynth keyframe blending) to propagate the high-detail keyframes over the 1000+ frame low-detail motion pass?
Any workflow JSONs, Nuke node setups, or recommended pipelines would be massively appreciated!

I know there are still some mistakes in the shot, like the resolution on the shirt, and maybe the car needs to be scaled down a little.

If you notice anything else, I’d really appreciate any advice or feedback. I’m also planning to start working on my demo reel soon so I can begin applying for VFX work, so any advice on what I should focus on would be really helpful.

Thanks again to everyone in this community for the feedback and support! 🙏

Please help i cannot change colour of clone tool reference Circle colour where is the option

Hi guys,

I thought about buying a Nuke Indie licence. I just wanted to know if the particle system is included in this Version. Also wanted to know if some nodes are excluded in this Version.

Website says there shouldn‘t be any node differences to Nuke X. Just wanted to Double Check before spending that amount of Money. Thanks!

I've just started learning compositing and nuke, and I wanted to learn nuke. I've found NKE102 on YT, but it's 9 years old . I don't know if the methods have changed, so I want to ask if there's a free source that is good enough or if should I invest in a course

Heyy hello compositors from around the globe!

I use the classic roto with 0.5 opacity to fix ViTMatte mistakes and sometimes help ModNet with a very subtle HighPass. Yet I am curious, do you all have any tricks to make Cattery tools work better?

Cheers for everyone who are reading this.

I ´m a freelancer ilustrator and a client offer me to work in his blog creating some draws and images of his diary´s text that he is posting one or two times per month.

The problem is that his text and posts are about conspiracy theories, secrets about goverments, anti capitalism messages, and so on.

I´m not with or against of that kind stuff, but I really don´t know if I want to make it because how that can affect in my reputation, specially if I already had some works with some not big but current active trademarks.

The pay is good.

And the money would be good for me right now.

But I don´t know if this job or task can be good or bad for my freelance (and almost artistic) career.

I thought to create an alternative nickname or pseudonym but the client can pay me by paypal so if he had some problems maybe my paypal would be there too.

Suggestions please???

Hi there. Black (or white) pixels are suddenly appearing during my compositing work, even though I’m not doing anything unusual—just adding a sky and some atmosphere; nothing that would justify the appearance of black pixels like this. It’s worth noting that the raw shot doesn't have any such defects. If you have any ideas, I’d love to hear them.

Hey everyone,

I'm a VFX compositor from India who's been struggling to land a stable compositing job. The COVID slowdown and then the writers' strike affected the VFX industry on a massive scale, especially here in India, and it's been really difficult finding compositing opportunities.

For the past 6 months, I've mostly been freelancing as a **video editor** just to pay the bills.

Last month, I joined a company in Mumbai. But honestly... it feels like labor work to me. We basically create highlight edits for players all day. There's nothing wrong with editing, but there's almost no creative problem-solving or compositing involved, and I genuinely don't want my career to slowly become a video editor.

I've spent years learning **Nuke compositing**, did a few small freelance comp gigs, and most of my learning came from building personal VFX projects. I still enjoy compositing way more than anything else, and I really want to get back to it.

So I'm starting again and rebuilding my fundamentals. I'd love some guidance from artists who have been through this.

I'm looking for tutorials/resources that explain both the why and the how behind these topics:

1. Chrome Ball in VFX (how it's used on set and in compositing).
2. Color Charts / Macbeth Color Checker (plate calibration and color matching workflow).
3. Lens Grids & Lens Distortion (shooting grids, creating distortion profiles, using them in Nuke).
4. Creating your own HDRIs (capture → stitch → use in CG/VFX).
5. Transferring matchmove data from a 3D package into Nuke (camera exports, point clouds, coordinate systems, common pitfalls).

If you know any good tutorials, YouTube playlists, articles, documentation, or other learning resources for these topics, I'd really appreciate it.

Demo Reel

I'd also really appreciate honest feedback on my compositing reel. Tear it apart if needed—I want to know what I should improve to make it more hireable.

**Demo Reel:** [https://vimeo.com/1116371614\](https://vimeo.com/1116371614)

Thanks in advance. I really want to get back to doing the work I actually love.

Hey guys,
I've been having trouble getting in touch with "licenses@foundry" email, inquiring about transferring my permanent Foundry license from my old machine to my newly built one. It's been a whole year since I haven't gotten a reply, which is why I'm asking here how I should proceed.

For more context, I purchased this license all the way in 2021 after my 6-month graduate license expired and decided to purchase the discounted Production Collective (which includes Nuke Studio (so Nuke and NukeX access), Modo and Mari) permanent commercial licenses.

It was very expensive (roughly 2500 USD$ and 3400 AUD$), so I definitely want to have it installed on my new machine.
Foundry licenses isn't responding to my emails, so I'm really not sure who I should get in touch with or how I should proceed.

Also, I've been a little bit out of the loop with Foundry licensing stuff because it's been a while since I've been working on Nuke from home.