If the Camera is selected in a second view, moving the camera moves both the Camera and the Point of Interest. After Effects has always worked this way.    If the Camera is not selected, you will notice a different cursor when you click and drag on the camera body. There is a little white square around the Camera body, and you can now move the camera around without moving the Point of Interest. When working on a project like yours, I almost always add a 3D POI null to the stage (Comp) and use a simple PickWhip expression to tie the Camera's point of interest to the Null.  temp = thisComp.layer("POI Null").transform.position; [temp[0], temp[1], temp[1]] This allows me to see the path that the Point of Interest will follow in the Composition Panel, which makes fine-tuning that path much easier. That is the only way to visually check the path of the point of interest. It makes moving the camera and adjusting the framing much easier and more accurate. ... View more

You do not need the array in the last line: a = effect("Trace Path")("Progress"); x = easeOut(a, 0, 100, 0.3, 100); [x, x]; ... View more

You cannot animate a layer's position, rotation, or scale when applying Puppet Pin. The Mesh does not follow the layer movement, so everything breaks. You must complete the puppet pin animation on the layer, then pre-compose or add the comp to a Master comp before you can animate the layer's position. This makes complex character animation difficult. The fingers would have to move independently of the rest of the hand (on a separate layer, then be pre-composed, then the hand and fingers combined in another comp and attached to an arm. If the aArm is also animated using a puppet pin, then the Arm and hand must be pre-composed and connected to the body so you can move it.   Puppet Pin is very handy for simple movements, but complex animations like fingers snappinwilld waving as your character is much more efficiently. Ifkly.   Character Animator or a 3D app like , I would explore those optionsBlender or C4D allows you to rig a character much more efficiently. I would explore those options if you have a lot of this kind of work to do. ... View more

Start by choosing the Selection tool (v) and then digging down in the shape layer contents to the correct path. You should then be able to move the pointer close to the point you want to use as the first vertex, see a change in the pointer's shape, and drag a selection around the vertex you want to set as the First Vertex. This is much easier to do with the Selection Tool than the Pen tool, but you have to have the path selected in the timeline. ... View more

The best advice I can give you is to break your eleven-minute video into scenes that are only a few seconds long. Each scene would be a separate comp. The scenes would include the start and end of the move through the web page from one point of interest, button, headline, or section to the next one.   The average shot (time between edits) in a modern action movie is 4 seconds; in a horror film, it is a little less than 16 seconds. Some directors like to do incredibly long takes and then use single shots in their live-action movies, but that is rare, and most of these shots are not longer than a minute or two.    It takes years to create and months to render all the shots in a Pixar movie, and hundreds of animators work on the project. None of the animators do all the work on a single shot by themselves. It takes several shots make a scene. A bunch of scenes make an act, and there are usually at least three acts in a movie.    Pick points where you want the character to make changes in direction, click a button, or emphasize a point, then cut the video. When you get it all cut up, create a new comp from each cut, then animate your character working their way through the shot. When the character successfully follows the action by discovering a new button or feature on the web page, that shot is done. Render that comp and stick it in a bin. Save a frame of the last position of the character and use that for the first frame of the continuing journey through the shot to keep the action consistent. That's how they made the film 1917 to look like a single shot. That's the only workflow that makes any sense. Spending days, weeks, or even months animating an 11-minute video, then trying to render it in one go, is not only incredibly inefficient, it's risky.    Working on individual shots that are only a few seconds long and then rendering them will allow you the freedom to make sure that animation works and the story is told well. More importantly, it will give you the opportunity to correct the few frames of one of the actions that just does not work well instead of re-rendering the entire 11-minute movie.   I hope this helps.  ... View more

When you crop to the Region of Interest, and you have 3D layers, the relative position of the camera to the 3D layers changes. Your 3D layers are not exactly in the center of the composition, so the new default Active Camera is now in the center of the new comp, so the position of the layers relative to the camera changes. If you add a comp camera and carefully line things up or stick to only 2D layers, the problem will be easy to solve or completely disappear. ... View more

If the background is constantly moving, but the red reflection is always in the same place in the video. In that case, you may be able to create a feathered subtraction mask over the reflection, so there is just a hole in the video that removes the reflection. Then, you can try duplicating the footage layer, shifting the position of the bottom copy to the background to fill the hole, and maybe adding a little blur to the bottom layer to help hide the edges.    Another approach would be to timeshift the bottom copy of the footage and animate the opacity of the mask so that the mask makes a hole in the video when the light is on. Because the blinking tally light is running at a constant rate, a simple loopOut() expression with three keyframes would animate the opacity of the blinking reflection for the entire shot. Again, you may have to blur the bottom copy to hide the time difference.   Another option would be to create some other reflection for the frame using Photoshop, which does not obviously look like a flashing tally light, and drop it in the shot.   If those options do not work because the reflection moves around on the windshield or the background changes too much, you are probably stuck with Content-Aware Fill. Without seeing a few seconds of the shot, that's the best I can come up with. ... View more