Best practices for video mask lifecycle

[diffidentfactor]

I'd like to talk about the best practices for saving and loading videos with an associated mask. I'm raising this in "ideas" because it doesn't seem like there are currently good practices for this. I've been doing it "by hand" with separate nodes, but it seems like the most obvious candidates would be:

1. Create a convention of linking a video named fname.suffix with a mask video file called fname_mask.suffix2 (not necessarily the same format), put the mask in the red channel of the latter, or in all channels of the latter. When you load fname.suffix, check for files named fname_mask.[anything relevant] and also load that file, providing that stream through a mask output parameter.
2. Use one of the video formats with an alpha channel for this purpose, again providing the mask stream though a mask output parameter.

The second option seems handy, even if it limits you to storing the video in one of those formats, but also potentially problematic given #135.

I would be interested in contributing to a solution that can exist within VHS, as that would be convenient for everyone, but some guidance as to what that solution should be would be appreciated.

Background: I'm looking into some making some documentation, and maybe some helper nodes, for creating WAN VACE control videos and masks. Although it seems like massive everything-happens-at-once ComfyUI workflows are the norm, that's not how I tend to work and I'd like users to have the option of breaking their projects up into different stages, maybe making their control video in one stage and then loading it in the next one.