Playback Performance
 
 
 

The ability to play compositions and media in real time depends on both hardware configuration and software optimizations.

Internally, the two key features used to improve playback performance are tiling and the media cache. The media cache keeps images that have been loaded close at hand on the local drive for faster retrieval.

To improve playback performance, you can turn on hardware rendering (if you are compositing with Reaction) and/or define a region of interest. Each of these has trade-offs (usually between quality and speed) and therefore may or may not be feasible in a given situation.

Tiling

Images are automatically divided into tiles and processed. Instead of processing an entire frame, only those tiles that are part of the result of the operation need processing.

Each tile is 512 by 512 pixels. The number of tiles loaded in the Player at any point depends on the resolution of the image, the zoom level of the Player, and the region of interest, if one is defined—see Region of Interest:

Resolution

You can set the resolution that a Player view or the fullscreen Player uses—see Setting the Resolution for a Player. When full resolution is not required, working at less than full resolution can improve playback performance.

When you change the resolution of a Player, a new copy of the image is retrieved and added to the media cache at the new resolution and tiled for that resolution.

NoteWhen you zoom out in the Player, the resolution is automatically adjusted to the optimal one for that zoom level. This reduces the load on the computer.

Hardware Rendering

You can switch to hardware rendering when using Reaction to improve playback performance when the quality of the output is secondary to the speed of playback—see Turning Hardware Rendering On or Off.

NoteOnly Reaction nodes benefit from hardware rendering. For all other tool nodes, switching to hardware rendering has no effect.

Region of Interest

You can define a subsection of the frame processed for playback by defining a region of interest. This can improve playback by limiting the area of the frame that needs to be processed for the playback.

The region of interest has no effect on rendering an output. When you render an output, the entire frame is always rendered. The Region of Definition (ROD) is used to determine the area to render when you render an output.