A digital brain, created by Stable Diffusion from the prompt "a digital brain imagining a future film"
Support ProVideo Coalition
Shop with Filmtools Logo

What Do You Think? Let Us Know.

Notify of
Newest Most Voted
Inline Feedbacks
View all comments

I like the building of the thread and the conclusion. Thank you for the overview of what is happening and what is coming, and especially for your common sense assessment of the situation.


Just as an FYI: Keyper is total *CRAP* compared to FCP’s Scene Removal Mask. Very surprised you wouldn’t have used it instead as an example.


Clearly, you have _not_ tested both side by side. I have. Extensively. And the _only_ time Keyper comes out ahead is with a moving camera, yes. But even with a _static_ camera, Keyper can’t hold a candle to the SR filter. And if the camera is moving the results are unusable more often than not without major touchups, effectively rendering it pointless.
Check out Mark Spencer’s video on the topic and see for yourself. https://youtu.be/oyuhTmO6gHY The SR would have done a FAR better job with every static example he shows. The others, as I said, are mostly unusable, so how is that an advantage?
Never mind that Keyper it is only useful _exclusively_(!) for humans, nothing else. And even then it’s a crap shoot.
To suggest the SR’s little more than a difference matte makes me think you don’t know what that is or what that looks like. At best the SR uses a difference matte for reference (which is demonstrably the better way to go) to then use ML (not AI, those are two different things!) to create a near-flawless matte compared to Keyper. Even with backgrounds that, according to Apple, shouldn’t even work. It also does far better with hair detail.
I also never said that either were meant as a “full green screen replacement” nor would I ever.

You Might Also Like