Share

Flame Reactor Rendering and 32-bit Float

Flame Reactor now supports 32-bit float in addition to 8-bit and 16-bit float.

  • The 32-bit Graphics mode can be enabled in Batch Prefs / BFX Prefs -> Rendering. This enables 32-bit processing in Action, Text, GMask, and GMask Tracer. Other nodes process in 16-bit float. This setting is general to all Batch Groups and Batch FX, just like the Hardware anti-aliasing.
  • Some Tools have not been updated, but accept 32-bit float clips as input, but process in 16-bit float. Resulting media will be 32-bit float with decimation from the original content.
AreaSupport 32-bit float Input32-bit float Processing32-bit float OutputComment
Project ManagementN/AN/AN/A32-bit float projects can be created.
Colour ManagementYesYesYes 
2D TransformYesYesYes 
ActionYesConditionalYes
  • Option to output renders as 32-bit float added to previously existing 8-bit and 16-bit float. Data passes are output in 32-bit.
  • Processing is done at 16-bit float, or 32-bit float if the 32-bit Graphics option is enabled. Except for layer effects.
  • Layer effects (Blur, CC, CW, Comp, and MK) are always processed in 16-bit float. Resize is processed according to the bit depth of the layer.
Media Caching & RenderingYesYesYesRAW and OpenEXR media can be generated based on the Stone & Wire cfg file.
Media Import/ExportYes (OpenEXR only)YesYes (OpenEXR only) 
New SequenceYesYesYes 
Paint (Tools)YesNoYesProcessing done in 16-bit float.
Stabiizer (Tools)YesNoYesProcessing done in 16-bit float.
Text (Tools/Timeline FX/Batch)YesNoYesProcessing done in 16-bit float.
Batch NodesSupport 32-bit float Input32-bit float Processing32-bit float OutputComment
Motion AnalysisYesYesYes 
Motion BlurYesYesYes 
Motion ConvertYesYesYes 
PaintYesYesYes 
Difference MatteYesYesYes 
De-InterlaceYesYesYes 
InterlaceYesYesYes 
ResizeYesYesYes 
FlipYesYesYes 
TimewarpYesYesYes 
Colour Sources/Bars/GradientYesYesYes 
MatchboxYesYesYes 
Write File NodeYesYesYesRGB OpenEXR only.
Render NodeYesYesYes 
BFXYesYesYes 
LustreYesNoYesProcessing done in 16-bit float.

Was this information helpful?