When masking a camera or vMix call input to a non 16-9 shape, it's important to be able to reposition the image in order to keep the subject centred in the image. With calls or remote camera setups - where we can't adjust the camera framing - it would be nice to have a simple way to do this in vMix.

In their current form, Key/Fill inputs do not allow the mask or source to be repositioned. Changing the position of the Key input does not affect the masking, and repositioning the Fill input moves the mask along with it. It seems as though the Key is applied before any layer transforms. It would be great if there were an option for Key/Fill inputs to behave like using a matte in compositing software and allow the Fill layer to be repositioned while the Key layer's position remains unchanged.

This topic has come up before in the forums - https://forums.vmix.com/...2-Key-Fill-Mask-Control. As an example, a client has just asked me to prepare a multiview with 3x side-by-side 9:16 crops of subjects who will be standing in front of fixed cameras. The 3 subjects are not professional presenters, and so I'd like to have the possibility to manually recentre the subject in case they stray too far from their marks. The 3 windows are too close together to be able to position them behind a foreground layer with cutouts, without them overlapping when repositioned. Unless I'm missing something, the only way to achieve this seems to be to create a Virtual Set and use UV maps to sample the desired part of the image. Which of course works, but adds more complexity to the process of setting up a show and requires external graphics software.

If it were possible to add an option to disassociate Key/Fill positioning and allow them to be positioned independently while keeping the masking effect, it would make it possible to achieve the effect without leaving vMix. Thanks!