In this case:
I used an animation sequence with a fixed camera
But the result wasn’t good. I couldn’t test if another model would have worked better…
I think this workflow is best for these animations:
What was the model you used? I couldn’t read it from the screen capture. Not every model is suitable for this. I can’t access this at the moment but in my AI experiments (not involving AI Studio), I’ve used a model that gave good results.
I think the problem with AI Studio currently is that there are quite a number of models and there is no model that is universally good for everything. Also, new models are released regularly and it takes experimenting and thus credits to find the most suitable one in a given situation.
The default one: Z Scail Un3C… The biggest problem is the limited testing that can be done due to the high cost of consumption.
I’ve never heard of that one. It is not one of the more established models, like Seedance or Kling.
So I decided to Google it (searching for “SCAIL”) and apparently it is specifically designed for AI generation with a video reference.
Hi,
From the video you provided, I think the issue might be related to an inconsistency between your first frame and the 3D video. I’m not completely sure because the thumbnail in the video is quite small, but they don’t seem to match very well.
The best approach is to first render the opening frame using the 3D to Image workflow. Once you’re happy with the result, move on to the 3D to Video workflow. The first frame will automatically be set to the image you created in 3D to Image.
This helps keep the action aligned with the video and can greatly reduce weird rendering artifacts or unexpected results. I think Reallusion has a thorough guide in their tutorial.
And as always, Seedance 2.0 tends to produce the most stunning results. It can be a bit expensive, but in my opinion it’s worth it because it saves a lot of time that would otherwise be spent regenerating videos.
Hope this helps!