A new build of the RTX Render open beta is now available: Beta2_0806. This release adds iClone Personal support, experimental volumetric spotlights, and fixes several render-related crashes. Full changelog below.
How to update
This build does not update in place. To move to Beta2_0806:
Remove your existing RTX Render beta files (the previous build).
Download the new build using the links in your beta email.
Unzip and run as before. No installation needed.
If you can’t find your download email, sign up again through the beta form and the links will be re-sent.
RTX Render Open Beta 2 – Custom Character Import Issue Appears Fixed
Just wanted to report some positive feedback after installing RTX Render Open Beta 2.
In the previous beta, I was experiencing problems importing my own custom-made characters into iClone. During my testing today, I loaded one of my custom characters (TEST 001) by dragging it into the default iClone scene, and I’m pleased to report that it imported successfully with no issues.
I repeated the test using my normal workflow, and the character loaded correctly without any of the problems I experienced in the previous beta.
So, from my testing, it appears that the custom character import issue I reported previously has now been resolved in Beta 2.
A big thank you to the Reallusion team for the rapid turnaround between beta releases and for listening to community feedback. It’s great to see issues being addressed so quickly.
Looking forward to continuing testing and reporting back if I find anything else.
Eventhough ,I still cannot import video as texture of a screen or a palne inside neither Iclone8 nor CC5 in rtx beta2.I tried many formats of video that are mentioned in reallusion web,but no success
I’m glad we got the problem sorted. I’ve been having fun putting the RTX renderer through its paces since then — the results have been great.
I’m looking forward to doing a lot more testing and experimentation with it. Thanks again for your help, and for all the work you’re putting into the beta.
Thank you for your reply,
is there any alternative way to display imported video on a big screen or plane inside rtx beta? (for exemple image sequence)
@HansRL - Been putting the RTX rendering beta through its paces and I have to say, the path-traced results are great. While I’ve been testing, one question keeps coming up for me, has anyone else been thinking about where DLSS 5 fits into this picture longer term? DLSS 5 dropped September 3rd and it’s a fundamentally different animal from 4.5, we’re talking three real-time neural rendering models with per-object masking and Structural/Tone Intensity controls. For character animation specifically, the ability to fine-tune photorealistic lighting at the object level inside a real-time pipeline could be genuinely transformative for CC5/iClone workflows. Curious whether anyone in the beta community has heard any internal discussion around DLSS 5 being explored as a future phase, or if the current beta scope is firmly 4.5 for now. Not looking to jump ahead of the current testing, I just want to see if this is a conversation others are having too.
Would love to hear what the community and the Reallusion team think.
I have been testing ways to use it with iclone using obs not the best approach but it let’s me see how it looks slightly on my Iclone work. I posted here and will post more stuff when I figure about way to upload or share videos with out doing it via YouTube. Testing DLSS 5 with CC5 Characters and iClone Footage
We see DLSS 5 more as a rendering filter, similar in concept to AI Studio, where the 3D render serves as the source and AI is used to enhance the final visual result. So, we don’t see DLSS 5 as something that would replace our RTX development. Rather, it could be treated as another local AI-powered rendering filter that complements the existing RTX pipeline.
That makes complete sense, and honestly mirrors how I was thinking about it. RTX as the foundation, DLSS 5 as a local AI enhancement pass downstream. Clean architecture, no cloud dependency. The AI Studio parallel is a useful reference point too. Looking forward to seeing how it develops as the beta progresses. Thanks for clarifying.
Following up on the DLSS 5 thread. There’s a sanctioned dev path: NVIDIA Streamline plus a UE5 plugin.
The SDK exposes selectable AI models, Structure and Tone Intensity, and per-object semantic masking, that last one matters for character work, since you can hold back on faces while boosting the environment.
Worth noting: output quality scales with input. Path-traced source gives noticeably better results than rasterized.
I’m currently testing this through the iClone/CC5 to Unreal Live Link plugin, pushing animated CC5 characters into UE5 and running the neural pass there. Too early for conclusions, especially around character consistency across shots, but the licensed route looks far more promising than the community workarounds going around.
Will report back. Curious if anyone else has gone down the Streamline path.
