It doesn’t really make sense to post this in this topic, but it’s the most relevant I could find.
Let’s say you have a nifty lightsaber composite node network setup. It may look really awesome from where the camera was when you were designing the network, but since blur values are pixel based, it will look really dumb if you change how far away the camera is by very much at all.
One way around this is to have several networks and use different ones depending on where the camera is, but this can be extremely time consuming and impractical. This is especially a big problem if the camera does much moving around during the animation.
This applies to many many effects applications, and a lot more scenarios that the one I gave. It’s also a very interesting problem when it comes to movie making.
Firstly, I’m wondering if anybody knows if a special kind of blur that blurs different amounts based on a different sized area to blur is even possible, and secondly, I’m curious as to what other peoples thoughts are on this.