v2.1.10

• Layout rendering is now done with Metal if it is available
  • Metal device used for rendering is selectable in preferences
  • OpenGL based rendering is toggleable in perferences
• Fixed a crash that could sometimes occur when editing a compressor
• Added easy access recording of layouts, including live layout. Outputs now have a 'stream' and 'recording' toggle. The Stream and Record buttons in the main window will start their respective outputs. An output can be both a recording and stream output (it is started only once if that is the case)

v2.1.5

• Added support for Mojave dark mode
• Support forcing of light or dark theme, ignoring the system's global setting. You can also make CocoaSplit follow the system setting properly (which is the default)
• Improved plugin loader so it always prefers the highest version number if multiple plugins of the same type are found
• Various updates for minor changes in Mojave
• Update to newer Syphon framework
• Update to newer Sparkle framework

v2.1.3

• Added NewTek NDI input and output plugins
• Added a passthrough compressor: not actually a compressor
• Layouts now have user definable video filter chains
• Video sources are now 'pinnable'. A pinned source persists across layout changes/merges/removal
• CocoaSplit is now more aggressive about reusing underlying video source plugin instances. This means if you create multiple instances of a webcam, even if those instances are in different layouts, only one CaptureSource is used. This should improve performance and memory churn. Plugins can opt out of reuse if required
• There are now alternate source clone options that ignore the cache
• Added a user flag on video sources to keep them permanently cached. Even if all visible instances of the source are removed, the CaptureSource is maintained and reused in the future. This is useful for cameras or other devices that may have a small delay in coming online.
• Video sources can now create child audio sources if they have one. Only relevant for NDI and Movie sources for now
• Fixed: Movie source would not advance to the next item in the playlist
• Fixed: Drag and drop reordering in the source list was sometimes breaking if one of the sources was animated via javascript
• More performance improvements in the audio engine

2.1.1

Fixes save file naming issue.
Adds experimental HEVC encoder

v2.1.0

Merge remote-tracking branch 'origin/2.1'

v2.0.2beta

Fix some warnings, bump version to 2.0.2

v1.9.11

• New Feature: shape inputs. A few built in shapes are available. Additional shapes are drawable via python scripts that return CGPathRefs
• New Feature: Mask flag for inputs. White pixels become transparent; transparent (alpha channel 0) pixels make the pixels UNDER them transparent. Default shape input colors are setup for use as a mask.
• New Feature: EXPERIMENTAL. Some source types can provide timing for a layout's framerate. As an example: You might set the timing source to your video capture card. You still need to provide an FPS value for the layout and if the input's timing is off some frames may be dropped. Do not use with variable frame rate inputs, such as most USB webcams.
• New feature: The compositing filter for an input is configurable. You'll probably never really need this, but it can be used for some interesting effects.
• New feature: Movie source now has repeat settings
• A source can now be excluded from multi-source transitions
• Bugfix: fixed a crash when going live with active movie sources
• A bunch of bugfixes for the audio system. Changing sample rates is better behaved, and mixed sample rate audio sources work much better
• Miscellaneous internal changes: preview view rendering now uses a CALayer, python is now embedded directly instead of using py2app

v.1.9.10

• New Feature: Native Blackmagic support. Input signal auto detection on supported hardware.
• Vastly improved multi-channel audio inputs. All inputs are routed through a multi-channel mixer and volumes are individually adjustable per crosspoint
• New Feature: Export and import of layouts. The file format is an apple PLIST file. If the layout references any animations they are also exported and imported along with the layout.
• New Feature: MIDI control. Various actions are assignable to arbitrary MIDI controls. See View->Midi Manager and the Midi Manager contextual menu for inputs.
• Added Vaughnlive.TV output type
• Animations with the 'onLive' configuration value set to true are run whenever the layout goes live
• Animation functions now accept an array of values; the target property moves through these values in the time given via a Core Animation keyframe animation.

v1.9.9

• New Feature: Scriptable animations. A variety of source transforms and effects are animatable via python scripts.
• New Feature: Inputs can be attached to other inputs as sublayers.
• New Feature: Constraints relative to an input's parent. Allows you to pin an input to the parent's edges or center. See the 'Adjust' Input configuration tab to set these
• New(?) Feature: The glorious return of Core Image filters. Three types of filters: Background, Source and Input.
• New Feature: In addition to background colors, inputs also have gradient colors. The order of drawing is background color -> gradient -> source
• New Feature: View -> FullScreen menu item. You can set both the staging and live views to fullscreen on a specific display.
• New Feature: Adjust input configuration pane. Allows fine tuning of position and dimensions of the input.
• New Feature: Layout merging. If you shift-click a layout it merges that layout into the current one
• New Feature: x264 encoder is reconfigurable while live, with some limitations
• New Feature: Text sources have line wrap and alignment options
• New Feature: X and Y axis rotations for inputs
• The live view is not longer editable unless you are in live-only view mode
• 'Revert' button added for layouts: restores to the last saved state of the layout
• Changing to a new layout (even in staging view) saves the layout you are switching away from
• Updated to newer ffmpeg/x264
• Improved the performance of text sources
• Some source types have frame rendering options to control when frames are rendered into the layer.
• Both layout transitions and multi-source transitions now include Core Image transition filters.
• Multi source transitions have been converted to switching between sublayers
• Bugfix: Added proper glFlush() to layout renderer: ChromaKey and other opengl effects should work again

Version 1.9.8

• Layout rendering is now based on Core Animation layers
• Added new sources:
  • Now Playing (supports iTunes and Spotify)
  • Text File
  • Current date/time
  • Countdown until date/time
  • Elapsed time from date/time
  • Fixed time interval (countdown/stopwatch)
• Sources now have a border option, with rounded corners!
• Sources have a configurable background color
• Layout views are fullscreen-able for your editing pleasure. (Cmd-F)
• Removed custom filter configuration option. It may return in the future
• Added 'clone' contextual option for sources. Cloned sources share the same source instance as the parent.
• Webcam capture is now global across the entire app. This fixes new sources resetting resolution/framerate settings of existing sources and just behaves better in general
• Syphon inject app list updates in real time
• Syphon inject source supports non-bundled processes now
• x264 encoder propely fails if the avcodec setup fails. Fixes a crash when trying to go live with bad encoder settings
• Added option to select NO x264 profile
• x264 encoder now supports lossless (CBR off and CRF 0)
• Staging copy/delete fixed to operate on the selected staging layout
• Layout switcher plugin now restores saved configuration on launch instead of on first load if its configuration window
• Changed style of audio source on/off buttons
• Added rendering update options to Syphon and Webcam sources
• Don't allow framerate to change when outputs are active
• Fixed use of 10.10 only CMSampleBuffer creation function