Looks like some kind of ML version of a dissolve edit. There was also a shift in his position and no breath between sentences there. Like they cut out a big breath or interruption of some sort.

He’s way too still and uniform over the whole thing though. No 80 y/o can hold that still unless they’re napping in front of a TV. (And you usually end up checking for breathing when they do).

Good catch.

https://youtu.be/-q1N6Akpz9c?t=18

Yeah, I don't know how people can't realize that Trump in that video, is NOT what Trump actually looks like, even with the best makeup. He's old and tired. When you see pictures of him dressed in casual golf attire at maralago he looks like an old fat woman who's melting

Forget autopen, Trump has moved on to autospeech.

I know we just recently got into a bit of tussle elsewhere, but I think I mostly agree with you here.

The uh, the 'morph-cut', when his hands shift and kinda melt?

Look above his head when that happens.

The lighting and shadow on the trees in the background changes.

Its all one way before that happens, and its all another way after that happens.

... I think he literally stroked out, while they were filming this, thats why the morph cut is there, and if some body wanted to try to model the scene, the oval office, geopositioned right, with the sun, they could probably estimate how long it was between these two shots they merged together... or timeskipped between.

I don't think his face is completely fake, I think it may have some kind of... correction applied to it, to try to minimize the droopiness of ... what you can see in his face today, from the 9/11 memorial.

Sorta like how and instagram filter morphs your face... and a frame stabilization effect tries to keep objects centered and just boink the frame itself around...

Something like a face stabilization effect, if that makes any sense.

Maybe. If that's true, I still say it should be called out and it shouldn't be accepted as a practice.

His body barely moves at all until it randomly spasms into motion every like 30 seconds. That's not how people, especially Mr talks with his hands, communicates.

I've played a bit with frame interpretation using open source models. With those models, fast movement did seem to trigger some weird artifacts and glitches that were visible at lower speeds or paused.

I guess it's possible that a propietary model is using some other image generation techniques to make those artifacts less noticeable. I think that would still constitute something that should be called out.

If the issue is with interpretation, then we shouldn't see this visual glitch in the original footage.