Viewer nodes, unlike process nodes, don’t alter data in any way; rather, they act as windows on it. Each Viewer node displays the render output of any connected process nodes in the Viewer panel. Viewer nodes are essential for quickly assigning the right values to parameters because they allow you to edit in context - that is, edit a given node’s parameters upstream in a script while viewing the effect of those changes downstream.
You can place as many Viewer nodes in a script as you wish, which allows you to simultaneously view multiple outputs. You can also pipe the output from up to ten process nodes into single Viewer node, and then cycle through the various displays. This allows you to easily compare an image before and after processing by a given effect.
NOTE: The maximum image size the Viewer can display is 64k x 64k (or the equivalent number of total pixels at other resolutions). Make sure though, that you have sufficient RAM memory available if you want to use the maximum image size.
Panning and Zooming the Viewer Window
Improving playback performance
Channel set and channel dropdown menus
Superimposing an image’s alpha channel over its RGB channels
Lowering the display resolution of individual Viewers
Using the Viewer composite display modes
2D / 3D toggle and camera controls
Input Process and Viewer Process controls
Hiding and Showing Viewer Toolbars