Major concepts and features:

Vapor paint animations are constructed, much like 3-d animations, with an
editor interface, which I call the work screen.  The work screen has its
own built in renderer to give a schematic view of your finished animation
and allow you to fix it up.  This information can be used to drive a
"real" renderer which will finally produce the frame output.

The working space is able to show any two of 5 dimensions:  
X, Y, Z, R, and M.  Mouse moves are usually interpreted only in the shown
planes.

Actions with the mouse are broken into two phases:  the registry and
tracking of the mouse on the screen, and the conversion of mouse
information into vector and sizing information.

Certain transformations work by correlating the timestamps of a set of
6d-vectors.  Others correlate the normalized total distance.  The
interpretation of the vector at render time depends on a pen setting and
color setting.  Thus, certain dimensions may be ignored or combined.

All drawing gestures and settings are controlled by traversing a tree of
the actions.

The current render time (SMPTE time) is a continuous variable, to allow the
creation of motion blurred frames.

The gestures are ordered into sequences, which are like cell levels and
scenes, and within the sequences , key frames are linked together.  The
sequences specify the starting frame and each key frame specifies the
number of frames to the next key.  All linked key frames' elements must
correspond exactly.  (This is insured by the frame creation process)

<Other actions can be sequenced in the tree - Invocations, Linear
transformation, and Process transformation.  Invocation restarts the tree
at a named spot.  Linear transformations are the familiar scale,translate
and rotate .  Process transforms are open ended, loadable, operations on
other vectors and colors to create cross hatching, fills, or other
difficult effects>

The rendering view is also animated, as a series of view key frames.  The
view can be rotated and re-scaled, and has two view points for stereo
rendering.

Since the work screen can do a linear in-between of existing vectors in
time, this interpolation can be instantiated into a key frame.

Any element of the tree can be hidden from the working renderer, in order
to keep the confusion down.  It is also hidden from the "real" renderer
this way.

Any element of the tree can be selected for use in some action.

The rendering process traverses the graph, using current colors, pens and
transformations, and interpolations, to create a high color-depth "target"
plane.  This target is then postprocessed into the "correct" output format.
For example, the black and white renderer creates an 8-bit black and white
frame, which can be converted to a 4-bit Amiga IFF ILBM file.

<Sections of the tree can be dynamically invoked from a disk file, referred
to with transformations to save space.>

The tree is available for the user to edit in an iconic form.  Details
about a particular node in the tree are set with a corresponding
requester-like screen.

Large dialog-like menus are used for setting certain numbers, setting the
details of an element, <mapping and remapping commands, and navigating> .

The work screen can be animated to give a detailed "pencil test" of the
work in progress.  This animation can be saved as frames or played
internally frame by frame .

The Arexx port commands duplicate all of the physical actions used to
create <and animate> the animation structure to allow conversions and
"algorithmic work" from other programs.  AREXX macros can be invoked from
the keyboard.

A Help facility is provided which reads disk files full of documentation,
which can also be read in a normal editor..(you're using it now..)

Henry Lowengard, jhhl-at-panix.com / 324 Wall St. Apt 5 / Kingston NY 12401

---

© 1998-2022 Henry Lowengard