How do they film scenes where there is 3d movement of the camera without people moving. Is this the same thing as second example where camera is moving in depth differently between things in foreground and background
Timestamp in video (but in general many parts of this music video) [https://www.youtube.com/watch?v=kTJczUoc26U&list=RDP1tI4VFMRmE&t=65s](https://www.youtube.com/watch?v=kTJczUoc26U&list=RDP1tI4VFMRmE&t=65s)
One way they do it, is to take many many cameras. The cameras are lined up in the path the camera takes in the shot. So like the matrix, when it pauses and the camera moves around the person, there are many cameras that are set up around the person. They take a photo all at the same time. Usually in front of a green screen. The put all the images together in post and animated it, adding in the background later.
There are different techniques in different situations. Sometimes (like the matrix bullet time) there are a bunch of different cameras in different locations near each other. They take one frame from one camera, another frame from another, and so on.
Other times, the people just move really slowly (or stand really still) and they use a camera with a fast frame rate and move it quickly. Then they play it back at a normal frame rate so it makes everything look slow.
In the links you shared, some of them were probably things like models of poured liquid that are actually plastic, or maybe 3D models made with CGI. The second link looked like it was just a static picture that slid past the camera, with a separate background image. You can see that the people’s faces don’t change. But if you move a camera past a real face, there would be different parts visible at different times.
At the start of the first video (Stay – The Kid LAROI) when he’s leaning back upside down, he’s actually on a green screen platform. The platform is digitally edited out in the final video to make it looks like he’s staying still.
Scenes like the one in the second video (Shadmehr Aghili – Halam Avaz Mishe) with the couple can be accomplished through a photo editing process where you take a bigger photo, cut out individual aspects of it (like the couple on the bed) and then use video editing software to pan across the cut out sections at a different rate than than the background. The white dots in the foreground are edited in.
Here is an example of the second effect: [skip to 17:55 for the final effect, the whole video discusses how it’s made from a still photo.](https://youtu.be/GZimVBFSGm0?t=17m55s)
There are a lot of different techniques which can be used. The simplest is just to make the actors stand still during the scene. If they maintain a stance which looks like they are in motion it will look convincing. Support things like clothes with wires to complete the illusion.
Another common thing is to break down different layers of the shot and film each layer at a different speed. Then when you compose them back together they look like they were all happening at the same time. Some of the layers may even be computer generated to complete the illusion.
But one very common almost comically bad technique is to take a still image and cut it into different layers and then change the zoom and pan of each layer to appear as if there is movement. This is very obvious though as the images does not change as the camera moves around so it almost looks as if the items in the foreground turns around to follow the camera. This is what is happening in the last link you show.
Many cameras in a rig all sinked up. Just take frame x from the cameras in order. Add some interpolation between the frames. And bang matrix bullet time. If time still needs to pass then take frame x form the first camera and then frame x plus one(or more) from the second.