At the moment when you make a Merge transition between a source and its virtual input with different Position values ​​the transition smoothly changes the Position parameters of one to the other, however, if you enable Flatten Layers on both inputs the Merge turns into a regular crossfade. It would be great if there was an option to make a Merge with a Flatten Layered input and its virtual input.

I've been saying for a while now that, if there was ever a use for AI in vMix, it'd be image analysis-based merge transitions.

So I'm going +1 this, but I feel like tracking a particular image through mixes is going to be a tough ask, and the image analysis idea is probably too resource intensive.


The simple answer is that, in the best case scenario, it's a boondoggle. You're doubling inputs just to get a clean merge transition.

In a worst case scenario, it's even more of a boondoggle.

I have a Zoom split screen setup, for example.

People aren't always centered on their webcam. This isn't always obvious with the full image, but it is pretty obvious when they're put in a smaller square/rectangle on the split screen. And there's only so much maneuvering of them or their camera that can be done sometimes.

So as part of the onboarding, I have an input where I discretely realign them so they're centered. That's the input I use for the split screen.

But because the crop and positioning of the image on the split screen needs to be independent of the center positioning, it needs to be flattened.

I can't merge to the actual layer, because the adjustment means there's a variable sized chunk missing from the side of the frame.

So the only option is to switch from layer flattening to a Mix input (which I'm running up against the limit on as it is)

Then add another input and layer the Mix input under the full frame input & merge to that.

For each person.

Is it possible? Sure.

But it's enough of a hassle that I just use cuts.