In this post I want to talk about the differences between the Blur and Defocus Nodes. I will talk in details about the ZDefocus in the next post. These two node do something similar and the difference might not seem obvious at first. But they have different uses. 

BLUR 

You can split width and height by pressing on the "2" button on the right side of the value slider.

Let's start with the Blur node. It's the simplest of all blur operations in Nuke. It's also the less expensive to use as it doesn't require nearly as much processing power as the other ones. 

The Blur is a very basic node. It works by mixing each pixel's RGB value with its surrounding neighbors. If we enter a value of 1, it will mix the value of each pixel with its immediate surrounding neighbors. If we enter 2 it will mix the value of each pixel with its next two neighbors, and so on. As we up the value it will start mixing with more and more surrounding pixels and will get you a blurrier image. 

Watch out for massive bounding boxes as they will slow down rendering time considerably.

You can split up the size in Width and Height and that's pretty much all you need to know about it. I don't think I've ever changed the filtering in a Blur node. As usual there's a mask input if you want to blur only parts of the image. And you can sometimes get interesting result by playing with the mix slider. Uncheck crop to format to keep your bounding boxes in when working with lens distortion but be mindful that it might cause the bounding box to become massive. 

Most of the time I use Blur to soften matte elements and roto, or to slightly blur something but it's not the best choice when working on the beauty of the CG render or a live action plate or elements. For a nicer and more realistic effect on an image, Defocus is much better. 

 

DEFOCUS 

Defocus is different than blur since emulates the way an out of focus image would look through a lens. It's slower to run but it looks a lot prettier than a basic Blur. 

Blur vs Defocus

Right away we can see that something more is going here. As with a real lens, brighter areas are more visible and we get what we call bokeh. I'll get into more details about bokeh in the next video. But a good example is the classic out of focus shot of a city at night where each light source becomes a big blurry circle. 

With Defocus you can't separate Width and Height like you can with the Blur Node, but you can change the Aspect Ratio. This is very useful when working on films shot with an anamorphic lens which will change the shape of the bokeh from round to an elongated oval shape. You can emulate this by entering a value of 0.5. 

The scaling slider is not very useful in my opinion, it's just a multiplier of the values of the Defocus and Aspect ratio slider. I personally, have never used it.  Quality is usually fine left at the default value. I don't think I ever needed to change it. 

Defocus is the node of choice when working on live action plates, 2D elements such as environments, etc. Anything that doesn't have a Depth pass. To work with a depth pass we’ll use the ZDefocus node which is the subject of my next post. 

 

Comment