Null Object for Applying Layer to Tracking Data - AE 2020

[Faizan Shaikh]

Hi Everyone,

I have been using AE 2020 for some days now and I am just working on tracking and applying layers to tracked data. I was wondering if there is any way that a null object can be avoided and the layer can be applied directly to tracked data in AE 2020. I a video where the user did not use null objects for applying layers. I tried the same but couldn't get same output as the user.

The reason why I am asking this is because I worked on a 6 mins clip where I had to create 17 null objects for final output.

[Mylenium]

The contextual functions inside the 3D tracker are pre-defined and can only do so much. however nothing stops you from copy&pasting position data to regular layers or manually placing them in the scene once the camera is solved.

Mylenium

[Rick Gerard]

You can apply motion tracking data to any layer. It does not have to be a null layer. The advantage of using null layers in your comp is that nulls are never considered when rendering pixels so they don't slow down renders or previews, the anchor point is at the upper left corner (0, 0) making calculations easier, you can parent any other layer to a null and then adjust that layer's position to make it be exactly where you want it to be. 

If you apply tracking data to a layer the Anchor Point of that layer will be attached to the point of interest. This does not guarantee that the layer will end up in the exact position you want it to be. It can be very difficult to reposition a layer properly later. That's why you use nulls.

If you are using Camera Tracking I would suggest that you not use nulls, but use solids because you can't see how well a null is sticking to a surface when you run a preview. You can easily move another 3D layer to the same position and orientation of any other 3D layer by simply holding down the Shift key when you parent the new layer to the target. That is almost always the way I work with camera tracking. My shot ends up with several solids that I either set to guide layers, or add grids to, or change the blend mode so that I can see how well I have lined up the solids with the surfaces in the shot. I then add my other elements and shift + parent them to the appropriate solid. Once everything is set up and working fine I'll either make all of the solids guide layers, turn them off, or in some cases, just delete them before I render. 

6 minutes is a very long comp. What kinds of things are you tracking? Do they stay on the screen for the whole 6 minutes? More than 90% of my comps are under 6 or 7 seconds and a single shot. Some of these shots have hundreds of layers, but I never try and edit in AE. It would really help us help you if we knew exactly what you were trying to do. A detailed project description would help. You might also consider including a screenshot that shows the layers or even the flow-chart for your project if you have a lot of nested comps (pre-comps).

[Faizan Shaikh]

Thank you for a detailed response, Rick. I have a clip from a movie which is six minutes long. There are two actors and the camera keeps on changing focus to each of them in small intervals. Like, it is a conversation between both of them and camera keeps focusing each of them when they speak. My task is to replace their faces with two different faces. Whenever focus is shifted to another actor, I hit ctrl+shift+D to split the layer. This gives me more than 17 layers to work on. My laptop is not that big, so it becomes difficult for me to have 17 layers and 17 nulls. There might be another way of working efficiently using all these data which I am not aware of.

[Rick Gerard]

If your second actor is not still in the frame make each shot a different comp. I know this sounds like a lot of work but each time you split a layer, pre-compose it adjusting the length of the new comp to the in and out points of the layer. Do your tracking by opening up the nested comps. If you are replacing faces I would suggest that you try face tracking instead of motion tracking. You can take the face tracking data and apply that to other layers. 

If both actors stay in the frame then you have bigger problems. To give you the most efficient workflow we need to see screenshots of your comp and the footage. If you can't show all of the footage several sample frames with an accurate description of the camera movement will suffice. Even if both actors stay in the frame I would make separate comps for each section of the six-minute movie, render them, then edit the pieces back together in Premiere Pro. The chance of a 6-minute render failing is very high. The resources required to work on a 6-minute comp is also high. You'll be better off breaking the movie up into smaller pieces.

[Faizan Shaikh]

Thanks again Rick. I trimmed down my video to 59 seconds removing all unwanted parts. That idea to work in pieces is perfect but I had a concern if that will, for any reason, hamper the quality of final output. I will definitely give a try to that and will use face tracker to track the faces of actors.

[Rick Gerard]

As long as you render to a visually lossless production format you will not lose any quality even if you go down many generations.