In Part One we understood what The Foundry Nuke and NukeX is, and how to set it up for best results. In this part we’ll study the interface, workflow and import options.
The Nuke interface
This is what Nuke looks like:
This is called the Viewer pane. It displays the result of whatever mischief you’ve been up to. It can also display data per channel, so you can view your maps, passes, alpha and so on. Nuke supports a total of 1023 channels per script.
It has readily available buttons for color space, exposure, zebras, etc., so once you get the hang of it toggling between different modes becomes a breeze. Most of the time, you’ll be viewing one of the channels in an image, or another image entirely. You can create multiple viewers but only see one of them at a time. Each viewer gets its own tab in the pane.
This is where the action happens. It’s called the Node Graph. You can also use this space for the Curve Editor and Dope Sheet but most of the time it’ll be the node graph.
The node pane always starts with a Viewer node, so any ‘next step’ that you might take can be immediately visible. To see something, you need at least two nodes. E.g., if you have a ‘footage’ node and connect it to a viewer node, you’ll see your footage in the Viewer pane (A).
In the above image, there are three nodes in the node graph: the footage, an exposure effect and the viewer. This typically shows you the ‘flow’ in a node graph. Everything starts at the footage and ends in the viewer. Let’s say you add three effect nodes between the footage and the viewer node. The one that is closest to the footage is calculated first, and the one next to the viewer is calculated last.
Most nodes have inputs or outputs or both. Each output can be subdivided into channels (the ability to manipulate data channel-wise is Nuke’s superpower) so you can apply an effect to only one channel if you like. I’m going to stop here, but here’s a great video that shows how node-based compositing is done:
This is the Properties pane. Every node you put into the node graph will have properties associated with it (well, almost). This allows you to tweak effects, etc. on each node. You can clear your properties pane completely, or arrange which nodes you want to see and which to hide. It’s customizable.
This pane can also be the script editor.
This is the tool bar. Once you have a hang of the shortcuts, you won’t be using it much.
Like most software, you can change the layout of the panes as you see fit. It all depends on how you like your node graph to be. Some just tack on nodes without thinking about the overall effect. If one compositor falls sick and another has to take over, it becomes a nightmare to find nodes. It also is a nightmare for the same compositor if he/she has to revisit the shot at a later date.
For this reason, there are certain good habits one can develop in grouping the look of your node graph. This way, you can zoom out for a top level view, and quickly zoom back in to a particular zone to find a node or whatever. When you begin compositing a shot, you must already have an idea of how you’re going to group your nodes.
I like to have my node graph on the right, with as much space as possible. The properties pane comes to the bottom. You must experiment to find the best system for you. But whatever you do, don’t keep moving panes around like a noob. That’s wasting half your productive time as a compositor.
Here’s an example of what a well-ordered node graph should look like:
The first thing you’ll be doing after firing up Nuke is set your project settings. To access it go to Edit > Project Settings… or hit ‘S‘:
What Nuke wants to know is what the resolution of your project is (called the Full Size Format). All calculations will be based on this. If you import footage or images greater than this, it will be cropped by the edges (only visually, but not deleted, so you can move things around at any time and the hidden parts will still be there). What is the maximum resolution supported? According to the user guide:
The maximum image size the Nuke Viewer can display is 2^28 = 268,435,456 pixels. This is the same as 16k x 16k, 32k x 8k, 64k x 4k, or 128k x 2k.
You can also work in proxy mode for routine work. Above, you will see the LUT tab, in which you can define a LUT (color space) for the project as a whole. This won’t affect the viewer pane, only the project. If you have imported a shot, the frame count will show the number of frames in the shot, and the resolution will default to the resolution of the footage.
Full size vs Bounding Box
The full size format is always visible in the viewer by a white rectangle. However, because you can add images that are greater in size than the full size format, Nuke will also use the regions outside this for every calculation. Ouch.
This total maximum region of images is called the Bounding box. Some images are taller (4:3) than they are longer (2.35:1), so the bounding box is the smallest box that will fit every image you’ve thrown in your node graph.
As you can imagine, this is unnecessarily processor intensive. For this reason, if you’re absolutely sure you’ve positioned your large image correctly, it is better to crop it to fit the full size resolution. If all elements are contained within the full size resolution, the bounding box can even be smaller than the full size resolution!
You can change the project settings at any time. This way, you are not restricted to a particular way of working. This feature of Nuke is essential, because that’s how the VFX industry operates. A VFX facility is a highly organized place, serving unorganized filmmakers.
The Nuke workflow
To begin compositing you must first import footage. Unlike an NLE, you don’t import your entire source footage in one go. This is a shot-based compositing application, so the best way to go about it is just import the shot in question, and all the layers, maps, objects, etc., that are associated with it, one by one – and build it like a mosaic.
To import your footage, click the first icon on the tool bar (Image icon), and the first thing you’ll see is ‘Read (R)’. Click it or press R.
A browser will popup and you can select the shot/footage from your read/source drive. If it’s an image sequence, Nuke will detect it automatically. If it doesn’t, just check the sequence box. Click Open.
You’ll end up with something that resembles the first image – a thumbnail of your footage on the node graph. Of course, sometimes you won’t see anything except an empty rectangle so you’ll have to manually join the viewer to the thumbnail – the first time this is done is one of the most exhilarating moments in a compositor’s life. Nuke’s nodes are as smooth as butter.
If you’ve done this correctly, and have chosen the right project settings (assuming Nuke hasn’t covered your ass), you’ll see the footage as you know it. Now you start adding effects or whatever between the read node and the viewer node. Read page 53 onwards in the Getting Started Guide to see how to work with nodes. In the first few weeks or months of your learning, you will get acquainted with working with the node graph and navigating through it. I assure you, once you get the hang of it, you’ll hate layer-based compositing ever after.
Supported files for import
The Foundry Nuke works best with image sequences. There’s hardly anything you can’t throw at it. Here’s an incomplete list of supported codecs and file formats:
- AVI (Codec must be installed in your computer in Windows)
- Quicktime (Codec must be installed in your computer, both Windows and OS X, bit depth or gamma metadata does not show up, so the image might look ‘off’)
- Redcode R3D up to 16-bit
- YUV (resolution independent, Nuke assumes 720p) 8-bit
- Maya IFF
Still images and image sequences
- Arriraw 12-bit
- Cineon 10-bit log
- DPX up to 16-bit
- OpenEXR 1.0 and 2.0 16 and 32-bit
- TIFF up to 32-bit
- Targa up to 8-bit
- PNG up to 16-bit
- PSD up to 16-bit
- RAW files supported by dcraw
To know more about licensing for each codec, refer to page 774 of the Nuke User Guide.
What about audio?
You might be wondering: Should we composite in silence, or listen to our favorite music? No, you can bring in audio into Nuke – using the AudioRead node.
The AudioRead node allows you to import an audio file (WAV, AIFF, etc.). To learn how to use this node to actually listen to a synced audio file, read page 361 of the User Guide.
You can easily bring in multiple kinds of codecs or file formats. Nuke doesn’t care. Ultimately, all of them will be decoded and be ‘dealt with’ in 32-bit math. In a nutshell, you can think of Nuke as working a Rubik’s cube. There are many surfaces to play with, and you can turn and rotate in any direction you please – the end result is a perfect composite.
As long as you got there fast, neatly and without creating a scene, nobody cares.
In Part Three we’ll look at the rendering and export workflows in The Foundry Nuke and NukeX.