Point Tracking, Corner Pinning, and Planar Tracking in Fusion
In this video Insight we’re adding to your Fusion skills. You learn the several different modes and types of trackers in Fusion.
There a four main sections to this Insight and they cover:
- How to use a single point tracker to do a sky replacement
- How to track features which get covered/occluded by other objects during part of a shot
- How to use four trackers to do Corner Pinning
- Using the Planar Tracker for Corner Pinning
- A technique to stabilize and then un-stabilize a shot, allowing multiple effects to be applied in between without the need for any further tracking.
I don’t go into too much detail about fine-tuning the composite here, that’s a whole other subject in its own right. I just give few examples of how each tracker can be used.
Tracking Method 1: Point tracker + Append
The big thing about learning a new bit of software is translating the terminology from software you already know. A common problem is tracking a visual feature in a shot and then need to ‘offset’ the tracking point when that visual gets covered up – or occluded (as the VFX pros say) – by something which stops the tracking. The key is to find another feature which has similar motion to the occluded visual. In Fusion handing off the tracker from one visual feature to another is called ‘Append’. (For the record: This isn’t the terminology that I was looking for when recording this Insight! It was only with the help of the Blackmagic trainers, user forums, and websites like Mixing Light, that helped me figure this one out myself!)
As I show in the video, the main thing to consider when choosing a feature to append a track to is that the feature should be moving at the same speed in x and y as the original feature. If you pick something that’s in the foreground then all its movement will be faster in both x and y due to parallax (slower if the new visual feature is far in the background relative to the original tracking point).
Tracking Method 2: 4-Point Tracker for Corner Pinning
To do corner pinning, allowing you to place a new image in a fixed-sized frame or a screen, you need to track all four corners of the screen/picture frame. A corner pin is simply four individual point trackers, working together. Fusion lets you track them all at once. You can have more trackers if you need to apply them to something else like mask points, for instance.
In my experience, getting a good 4-point track for corner pinning is really difficult. There’s often some jitter or noise that throws it out and it needs lots of tweaking afterward. However, now that the Fusion trackers are so much faster than trackers I’ve used before, I found that experimenting with increasing the Target area, increasing the Target shape, and increasing the Search area can get good results without taking ages and ages to process.
Tracking Method 3: Planar Tracker for Corner Pinning
The Hybrid Point/Area tracker in Fusion is generally a much better option for corner pinning. It’s a really good tracker. To be clear: Mocha Pro is the gold-standard for planar tracking. I’ve only used Mocha a few times and it has many extra features compared to Fusion’s planar tracker. But I’m impressed that Fusion in Resolve has this great tool and it’s very good in its own right.
One thing to note: If you have a shot with a very shiny screen but clearly visible corners then you’d be better using the 4-point tracking since the reflection could throw the planar tracker off. But for tracking most other flat shapes the planar tracker tends to be the best option. Features on a face, for instance, can sometimes be tracked very well with the planar tracker.
Tracking Technique: The ‘Steady – Invert Steady’ Transformation
The idea of steadying a shot so that you can work on it and then applying the inverse of the steady movement afterward is a really powerful technique that feels like a sneaky trick. This is another one of those things where the terminology or methodology is a bit different in various other software packages and there are different ways to achieve it. While you can execute this technique in Fusion similarly to other software (pick-whip link parameters to each other and invert the values with maths expressions), don’t bother. Fusion makes it easier for you since all Transform tools have an Invert Transform option. You simply need to paste a copy of the tracker and invert its movement using the invert checkbox!
Transform tools in Fusion will ‘concatenate‘ in that the values will link together and the final transformation value will be the combination of all the previous values. This means that if you scale down in one Transform and scale up in the next, you won’t lose image quality. You can even move the image out of frame in one Transform and move it back again in the next. That is unless you break the concatenation with tools in between (or choose the ‘Flatten Transform’ option). A Merge tool won’t break this concatenation but most other tools will.
In this Insight you’ll learn the ‘steady-unsteady’ method using a simple corner pin example. You can corner pin a face, or a portion of a face if the angle and the movement are suitable. But all tracking techniques and tools in this Insight completely depend on the nature of the shot you’re working on, so you have to choose (and maybe experiment with) the right tool for each shot.