VGens Overview

Extension

Video Oscillators

There are only a few test oscillators implemented for now, with plans to add many more. One key conceptual difference between SuperCollider audio oscillators and Scintillator video oscillators to bear in mind is that video signals are constrained to values between [0, 1], unlike audio signals, which operate normally between [-1, 1].

Image Sampling

VGens for reading from ScinImageBuffer objects.

Tween Sampling

The ScinTween class describes a series of simple curves which the server can populate into a table, and then sample with interpolation with VGens at runtime. Scintillator provides VTweenGen VGens for sampling the tween curves over time, and VTweenSampler VGens for time-invariant tween sampling.

Fragment Position

Scintillator offers a few different means to determine the position of the current fragment shader relative to the geometry being rendered, or the onscreen pixel dimensions. The VTexPos VGen is in the Image Sampling section.

Vector Manipulation

These utility VGens allow the grouping of one-dimensional elements into vectors, access to individual elements within vectors, and lastly a convenience function to repeat (or splat) a single element across all elements within a vector.

Vector Building

Some VGens require inputs that are higher-dimensional vectors. To construct those inputs from single-dimensional components, Scintillator provides the VecN classes.

Element Access

To break out a single-dimensional signal from a higher-dimensional vector, use the VX and related classes. These follow the computer graphics naming conventions for elements within the vector, where the names x, y, z, w are used to indicate the first through fourth element respectively.

Color Output

Note that any vec4 output is considered valid ScinthDef output, but these can help with code clarity or with grouping of single-channel inputs into a vec4, sometimes with that fourth alpha component hard-coded to 1.0. In the future the server may perform opacity calling, so knowing an output is entirely opaque will allow VRGBOut and VBWOut to take advantage of that culling automatically.

Mathematical Operations

Scintillator offers per-element (or piecewise) operations as well as some more traditional vector mathematical operations such as dot and cross products. Many unary and binary operations are offerred with the same names as their analog counterparts,

Vector Operations

Built-In Unary Operations

The SuperCollider UGens support a broad variety of unary, binary, and n-ary mathematical operations. These are typically expressed as operators within the SynthDef flow, and are transformed by the SynthDef programming into one of the subclasss of BasicOpUGen before being sent to the server. Scintillator implements parallel classes for unary and binary operations, with the BasicOpVGen as the base class and UnaryOpVGen and BinaryOpVGen derived classes handling a subset of the operators handled by the UGen programming.

For documentation of these operators in base SuperCollider see the Operators overview or the SimpleNumber class documentation.

The following tables detail the current supported operations along with the ones that are not yet supported, the name of the VGen sent to the server to realize the operation, a brief explanation of their function and conceptual mathematical code. Since all numbers within a Scintillator ScinthDef must be floating-point, several of the integer operations like bit manipulation are marked as not applicable. If support is planned, but not yet implemented, the function is marked as not yet implemented.

For higher-dimensional signals all operations happen piecewise, meaning the operator is applied to each component of the signal independently. For example if b = a.neg and both a and b are VVec4 objects then:b = VVec4.fr( VX.fr(a).neg, VY.fr(a).neg, VZ.fr(a).neg, VW.fr(a).neg);

Built-In Binary Operations

Binary operations also happen piecewise, meaning that the binary operator is applied to each individual component separately. Conceptually if c = a * b and both a and b are VVec4s then:c = VVec4.fr( VX.fr(a) * VX.fr(b), VY.fr(a) * VY.fr(b), VZ.fr(a) * VZ.fr(b), VW.fr(a) * VW.fr(b));