Introduction
Synfig, like most every other competent graphics program, breaks down
individual elements of a Canvas into Layers. However, it differs from other
programs in two major ways:
An individual layer in Synfig usually represents a single "Primitive".
ie: A single region, an outline of a region, an imported JPEG, etc... This
allows you to have a great deal of flexibility and control. It is not
uncommon for a composition to have hundreds of layers(organized into a
hierarchy for artist sanity of course).
A layer can not only composite information on top of the image below it,
but also distort and/or modify it in some other way. In this sense, Synfig
Layers act much like filters do in Adobe Photoshop or the GIMP. For example,
we have a Blur Layer, Radial Blur Layer, Spherical Distortion Layer,
color-correct layer, bevel layer, etc...
Each layer has a set of parameters which determine how it behaves. When
you click on a layer (either in the canvas window, or in the Layer Dialog),
you will see its parameters in the Params Dialog.
Synfig Studio has an autorecover feature. If it crashes, even if the
current file has not been saved, it will not lose more than 5 minutes of work.
At restart it will automatically prompt the user to recover the unsaved changes.
Unfortunately history isn't recovered yet. That feature comes later.
One thing you may notice is that Synfig Studio is SLOW, making it
practically unusable on hardware that is over 3 years old. The biggest reason
for this is that all of the color calculations are done in floating point
because Synfig Studio was built from the ground up with High-Dynamic-Range
Imaging in mind. HOWEVER, this will not be the case forever.
darco has some fairly major re-implementations and optimizations that
he plans to implement that should quite dramatically improve the performance
of Synfig on all platforms. The goal is not a 200% speed increase, it is at
least a 2000% speed increase. With the optimizations that are planned to be
implemented, we will be able to pipeline operations in such a way that this
performance improvement can be realized. It should also pave the way to hardware
acceleration using todays powerful graphics processors, which should yield
further performance improvements measurable in orders of magnitude.