Stereoscopic in Studio

I'm, starting to work with Stereoscopic footage in Fusion Studio, and was wondering if anyone had thoughts on how to best set up Stereo footage, when integrating CGI renders. I've been successful setting up the L and R eyes as projections in 3D space and adding elements that way. And I've also used 2D merges after Render3D to integrate the footage. Anyone have any general workflow practices that would be helpful or one way is better than the other?
Thanks,

I'm assuming you have 3D tracks? Then yes, at the very least get both cameras in. You can project independently onto the same meshes/materials. You may not necessarily render from those imported cameras, depending on what stereo corrections you might need to do, but you'll need their projections. It's a good sanity check.

Thanks Chad, Yes I have separate 3D tracks for both cameras. I've tried solving for 1 camera and manually adjusting the 3D camera for CGI work, but I think the production cameras had convergence rather than in parallel.

I felt like I have a lot more control when working with the projections than trying to use transforms in 2D after the Render3D.

I'm trying to find a better workflow for convergence of the 3D renders that are coming from outside of Fusion. I've tried creating projections onto geometry in Fusion that is displaced by the position pass from CGI, but run into small tearing artifacts when I start to converge the Fusion stereo camera. I can render in stereo for the 3D renders, but not sure how to match convergence to the footage before getting into Fusion.

A lot of this is testing.... so is it better to shoot parallel and render parallel CGI and converge both the footage and the CGI renders in Fusion separately and have them all Merge together at the end?

Thanks Chad, Yes I have separate 3D tracks for both cameras. I've tried solving for 1 camera and manually adjusting the 3D camera for CGI work, but I think the production cameras had convergence rather than in parallel.

Doesn't matter, really, the production cameras and rigs have tolerances that at even 2K will throw off the images. There's just too many moving parts and something is going to be off. Just hope it isn't the focus or shutter sync.

I've tried creating projections onto geometry in Fusion that is displaced by the position pass from CGI, but run into small tearing artifacts when I start to converge the Fusion stereo camera.

One displaced mesh or one per eye? Might need to have 2 meshes.

Thanks again Chad,

One displaced mesh or one per eye? Might need to have 2 meshes.

I'm only using one displaced mesh, I'll try setting it up one per eye. Does this require two Merge3D scenes with instanced cameras between the two, or can I separately project both meshes and have each eye only see the assigned projection?
Thanks.

Does this require two Merge3D scenes with instanced cameras between the two, or can I separately project both meshes and have each eye only see the assigned projection?
Thanks.

You can do it either way.