Copy link to clipboard
Hello there, I am very new to After Effects (started last week) and for my first project, I've been using the 3D camera tracker to track text onto the stage of a dance performace. To make it easier for the tracker to track the video, I split the video into multiple layers for the zooms and fixed views. The video tracks fine, but when I create the text and camera, I have multiple layers of text. I have to then, by hand, match up the last frame of the first layer, to the first frame of the next layer so the different layers of text are in the same position for the whole video.
I have also tried creating one text layer and after the tracker is finished, I will just create the camera under the effects of the current layer but that will cause the text on the other layers to jump around a lot.
Is there an easy way to do this? Am I using the tracker wrong? This really slows down my workflow so I am hoping to find a way to speed it up. Thanks.
Copy link to clipboard
Camera tracking a dance performance can be quite difficult, especially if you are overlapping shots or editing shots in a single comp. Here's some basic advice.
For more detailed instructions I need to see a timeline and have a lot better understanding of the design goals and how the piece is assembled.
After removing the tracking points that are on the dancers, the text is tracked smoothly in both layers, but if I try to resize and move the text to the correct position in the second layer, it's in the wrong place in the first layer. If I split the text layer to match the length of the different layers, I can resize the text freely without messing up the other layer but then I will have to match the size, orientation and location of the text on the second layer to match the last frame of the first layer so that it looks like one smooth text layer with not cuts.
I'm unsure if that makes any sense, if you need me to clarify something, then let me know. Here's my timeline currently if that will be of any help. "Snakes and Ladders" is the name of the dance.
Copy link to clipboard
Layer 3, 4, and 5 are all nested comps of what I assume to be the footage of the dancers. I don't know what is going on in those shots, or whether the layers have been scaled or moved, but there are 3 layers in your comp that are nested comps with the same name so there need to be 3 cameras, one for each layer that is camera tracked, and 3 text layers that match all 3 shots. You will have to establish an origin and ground plane, add a camera and a track solid (reference solid named Ground and set to a guide layer with a layer marker on the frame you used to set up the origin) using the same points you used for the Origin and Ground plane, then you need to add reference solids or text layers to the appropriate geometry in the scene. You'll have to keep track of the location and orientation of each piece of geometry you use in each scene so that the cuts work.
I hope this helps. You would be better off using a separate comp for each shot in this 3 shot sequence, then editing them in Premiere Pro. Since it is only 3 shots you can probably get away with just these 3 shots, but if you have more than a few seconds and shots in your sequence it is usually better to cut them up into short sequences.
I did a shot for a short film last year. The shot was about 145 frames long, it all had to be camera tracked, but an actor crossed in front of the camera 4 times blocking most of the background and all of the part of the scene that needed fixing in that single shot from the Premiere Pro sequence. I started with a comp that contained just the original shot. I duplicated the footage layer then proceeded to split the layer every time the actor blocked the camera. Each of the duplicates was trimmed so that just the part of the shot that needed fixing was retained. There were holes in the split layers, but the original footage on the bottom layer was there to fill the holes.
All 5 edited layers were pre-composed trimming the new comp to layer length. This was one time I just kept everything in one comp. I then ran camera tracking on each nested comp, added the reference solid and the replacement element for the scene, and set the nested comp as a guide layer because I only needed it for camera tracking. My shortest camera tracked shot was about 8 frames. My longest was about 40 frames. The completed comp had 5 cameras, 5 origin guide layers, 5 copies of the repair, and 5 placeholders. That is how you set up a comp with multiple cuts of the same shot. This is what it looked like:
A layer that is camera tracked cannot be moved, scaled, or rotated. It must be a 2D layer. I always add these simple expressions to a layer that I'm going to camera track. I have it saved as an animation preset and I apply it to the layer or nested comp that I am going to camera track.
// for Anchor Point [width, height] / 2; // For Position [thisComp.width, thisComp.height] / 2; // For Scale [100, 100] // For Rotation 0
Here's the preset. Save it to your user presets and apply it to every shot or comp you camera track.
What will the preset do?
After I have everything tracked, when I copy the orientation over from the first layer to the second layer, the second one isn't really close to the orienation of the first one, even though the values are the same. Did I miss a step?
The preset locks the original footage layer in position so you don't accidentally move it.
You don't copy anything. You have to make sure that you are using the same surface between shots. It's like every shot is a separate comp. If you don't pick the same surface each time you run the camera tracker the layer you want to insert (my sign replacement layer) won't line up between shots, but if there is a cut between shots, it should be close enough.
It would really help if I could see a preview of your edited shots. I don't think that you are completely grasping the concept of camera tracking.
I have finished the first layer and have almost finished the second but the text isn't lined up yet. I have the preset on both too. I've set the tracking points on the left scaffolding as the Ground plane and Origin but the text still isn't right. I'm probably missing something but I'm not sure what it is.
What does the sequence look like if you move to 7 seconds so we can see the second shot? Is everything coming from the same camera and the same performance? Is the Snakes and Ladders text layer the only thing you are inserting into the scene?
If this is just one camera and a couple of takes at different focal lengths, and the camera move is mostly a tripod pan, I would just use Motion Tracking or motion stabilizing to insert the text layer. It would take a lot less time and be a lot easier to handle.
The whole comp is just one short part of a 2 minute dance. I need to add the tracked text into a few different dances similar to this one. This dance and all the other dances consist of 1 or 2 different angles of tripod pans and zooms. In this comp there is just the one angle.
I will try the Motion Tracking. It will probably be easier for this project. When I was trying to find something that could add tracked text, the 3D camera tracker was the first couple things that came up and I assumed there wasn't anything else in After Effects that would be more effective.
It would really help me help you if you could show me the 3 different shots or even render a low-resolution copy of what you have created so far with this comp. If I could just see the last frame of the first shot (layer 5) and the first frame of the second shot (layer 6) it would help a lot. It's awfully hard to give somebody a good road map if you don't know where they want to go.