Stream to 3D Release 6.1: Detailed User Feedback
Long-term Stream to 3D user shares detailed feedback after watching several films with the new AI-assisted Depth Guided conversion in Release 6.1.
Okeb purchased Stream to 3D on 30 January 2026 and had already been using it actively for several months before the introduction of AI-assisted Depth Guided conversion in Release 6.
He has previously provided detailed feedback on Stream to 3D, including feedback that helped improve the product documentation.
When Depth Guided was first released, Okeb also reported performance issues with the new AI conversion on his system. Release 6.1 subsequently introduced NVIDIA TensorRT optimisation for compatible RTX systems.
On 14 September 2026, after spending considerably more time with Release 6.1 and watching several films using Depth Guided, Okeb posted the following detailed feedback in the Stream to 3D Discord community.
The feedback below is reproduced verbatim. Only paragraph formatting has been changed for presentation. Okeb has explicitly given permission for it to be shared, quoted and published.
Okeb's feedback
I wanted to leave some proper feedback after spending a lot more time with Depth Guided and watching several movies with it.
I had already been using Stream to 3D quite actively before the big AI Depth Guided update, so I was already very familiar with what the software could do. I liked it before and had already watched a lot of content in 3D with it.
When Depth Guided first arrived, though, it immediately felt like something much bigger.
At first I had some performance and optimization problems with the new AI method. I sent some feedback to the developer, and honestly I did not expect the main issue to be solved so quickly. Release 6.1 arrived with NVIDIA TensorRT for RTX support, and that update basically solved the performance problem for me.
After that, everything really came together.
Since then I have been using Depth Guided almost exclusively. I have already tested it with several different movies, and recently I watched The 13th Warrior, one of my favourite films, completely in Depth Guided Quality.
And honestly, I’m blown away by the result.
What impresses me most is not simply that there is “more depth”. It is how coherent the whole image feels.
The AI seems to understand what is happening in the scene rather than simply producing an isolated depth effect for one frame and then doing the same again for the next one.
I don’t know exactly how the internal system works technically, so I don’t want to make assumptions about the implementation, but from a viewer’s perspective it feels like the scenes are connected properly over time.
When the camera moves quickly, when it shakes, when something suddenly moves across the frame, or when the movie cuts to a completely different shot, the depth remains remarkably smooth and stable.
The 13th Warrior was actually a great test for this because it is an older movie and the camera is not always perfectly stable. There are handheld shots, shaking, fast action, sudden camera movements and quick scene changes.
I expected moments like that to expose weaknesses in the AI conversion, but instead I was surprised by how well it handled them.
The depth remains clear, the image stays comfortable for the eyes, and scene transitions feel very natural. Even when the camera suddenly changes direction, I do not get the feeling that the 3D structure is falling apart and then being rebuilt again.
It simply remains coherent.
Details also look excellent. Faces, people standing at different distances, objects in the foreground, interiors and distant backgrounds all have a convincing sense of separation and space.
One thing I found especially interesting is how little I now think about stereo stabilization.
In the original functionality there is a Stabilize Eyes View option, which can be useful for making the stereo image more comfortable and stable.
With Depth Guided, I basically stopped thinking about it.
The AI seems to produce such a stable result by itself that for me it has become much closer to simply selecting Depth Guided, pressing Play or Convert, and letting it do its job.
And it does that job extremely well.
That simplicity is a big part of why I like it so much.
After watching for a while, I stop analysing the conversion completely. I am no longer looking for depth errors, checking whether the next scene will work, or thinking about what the software is doing.
I am just watching the movie in 3D.
For me, that is probably the strongest compliment I can give it.
I also tested full Quality conversions at the maximum H.265 settings. One source movie was only around 1.5 GB, while the final converted 3D version ended up around 14 GB.
At first that file size surprised me a lot. 😁
But after watching the final result, I was completely happy with it. The image is extremely clean, the stereoscopic result holds up throughout the movie, and it actually feels like the extra data is preserving something worthwhile.
Before Depth Guided I used Stream to 3D regularly and already enjoyed what it could do. But after this AI update, and especially after the optimization in 6.1, my usage has genuinely changed.
Now I basically want to use Depth Guided for everything.
Once you have seen a result that looks this convincing, it becomes very difficult to forget it and go back. Your reference point changes.
And after testing more movies since the update, that impression has only become stronger for me.
I know there are more major ideas and features planned for the future, and I have no idea how all of that will work from the technical side. But what has already happened with Depth Guided and the 6.1 optimization is seriously impressive.
The first release showed me how good the AI conversion could look. Then 6.1 solved the main performance problem surprisingly quickly and made it genuinely practical to use on my RTX system.
Now that I can actually sit down and watch complete movies in Depth Guided Quality, the whole feature makes much more sense.
I’m extremely happy with how this has turned out.
Thank you for the work on this update, and especially for listening to feedback and improving it so quickly. What Depth Guided can do now is honestly incredible.
Okeb followed his detailed feedback with:
It’s hard to list everything - as you can see above, the buffer limit doesn’t really let me say it all. But in short, I’m absolutely blown away. This is fucking incredible.
He then added:
That’s basically my full review above 😁 I fully stand by everything I wrote, so feel free to share it, quote it, or use it anywhere you like. There was honestly so much more I wanted to say, but the message limit wouldn’t let me fit everything in, like I said before.