Source Values

Devices

Mouse

The Mouse node provides the current x and y coordinates of the mouse in the view plane (screen space) within the Scene Viewer. Both values are in the range 0-1. The button output is 1 while the mouse button is held down in the Scene Viewer, and 0 otherwise.

To compare 3D positions with view plane coordinates, use the View Plane Position node.

Keyboard

The Keyboard node monitors the keyboard for a particular key press. It matches the current key against the input value (as an ASCII value). When it matches, it returns 1; otherwise 0.

MIDI Control

The MIDI Control node outputs where a knob, fader, touch strip or pitch bend on a MIDI controller is, from 0 to 1 (a pitch bend rests at 0.5). Choose one of the project’s sixteen controls; what each one means is set - and named - in the project settings, under Music, MIDI (click Learn and move the knob). Wire it to anything numeric: a color’s brightness, a position, a noise’s scale. It works like the MIDI Control node of an instrument, so the same control can shape a sound and a visual.

Until the control has been moved, the node gives when not live.

MIDI Note

The MIDI Note node watches one note (0-127, middle C is 60) on a MIDI controller: gate is 1 while the key or pad is held and 0 otherwise, and velocity is how hard it was hit, from 0 to 1. Pads send notes too (36 and up on many controllers), so a pad can switch a color or an effect on and off.

Texture

Texture Value

The Texture Value node returns the RGBA value from the texture at position uv. To see how you can map 3D coordinates to 2D UV values, see Texture Mapping.

Texture Value LOD

Use the Texture Value LOD node to avoid texture edge artifacts when, e.g., wrapping a texture around spheres or applying an equirectangular mapping to a cube.

Start with the Texture Value node for regular texturing. Switch to Texture Value LOD when you see seams or artifacts at texture boundaries, or moiré patterns or aliasing.

Texture Value LOD Zero

The Texture Value LOD Zero is used to extract texture values in vertex (deformation) shaders.

Feedback Depth Value

The Feedback Depth Value node returns the raw depth and linear depth values from the output of the previous frame at view plane position uv. You can use View Plane Position to derive the current (or other) position in view plane coordinates. The linear calculation uses current camera near and far clipping plane values, but these can be overridden. The linear value is normalized to 0.0 - 1.0. Calculation: linear = (2.0 * near * far) / (far + near - ((raw * 2)-1) * (far - near)), normalized with (linear - near)/(far - near).

Sound

Frequency Volume

The Frequency Volume node looks at incoming sound data (from a Microphone or Audio Track node), and turns it into a floating point value that can be used in other nodes.

When you click on the node, you can select which frequencies you want it to listen on:

• Overall
• Low Frequencies
• Mid Frequencies
• High Frequencies

The audio spectrum is converted into logarithmic frequency bands that match human hearing perception, and for each A-weighting is applied to compensate for how humans naturally perceive different frequencies (i.e. adjusting for how our ears are less sensitive to very low and very high pitches).

Frequency Split 5

The Frequency Split 5 node makes it easy to pick up the intensity in any of the frequency bands based on audio data (from a Microphone or Audio Track node).

The audio spectrum is converted into logarithmic frequency bands that match human hearing perception, and for each band A-weighting is applied to compensate for how humans naturally perceive different frequencies (i.e. adjusting for how our ears are less sensitive to very low and very high pitches).

Frequency Split 10

The Frequency Split 10 node further subdivides the frequency bands based on audio data (from a Microphone or Audio Track node).

The audio spectrum is converted into logarithmic frequency bands that match human hearing perception, and for each band A-weighting is applied to compensate for how humans naturally perceive different frequencies (i.e. adjusting for how our ears are less sensitive to very low and very high pitches).

Frequency Split 256

The *Frequency Split 256 node divides the frequency bands based on audio data (from a Microphone or Audio Track node).

The audio spectrum is linearly divided into 256 segments, and for each segment A-weighting is applied to compensate for how humans naturally perceive different frequencies (i.e. adjusting for how our ears are less sensitive to very low and very high pitches).

Texture Frequency Sample

The Texture Frequency Sample node reads a single frequency band out of a Audio Track (Texture) (or Microphone (Texture)/Mix (Texture)/Runsheet Track (Texture)) node, the same texture a raw texture(...) lookup would use, but averaged into perceptually-even bands instead of the texture’s own linear-Hz pixels. It reads the texture’s actual width at compile time, so it also works unchanged with an older, 256-wide Audio Track (Texture) (OLD)/Microphone (Texture) (OLD) node from a project saved before the 512-wide versions existed.

position (0-1) is not a raw pixel coordinate - it follows the same standard ISO octave-band curve as the frequency buckets of Frequency Volume and Frequency Split 10, calibrated so sampling it at evenly-spaced label positions (e.g. 0.1, 0.2, 0.3, … 1.0 for a 10-point profile - the natural “i/N” loop) lands on musically-even bands, low frequencies near 0 and high frequencies near 1, without you having to work out where any given band actually sits in the texture. position=0 lands at the very bottom of the spectrum and position=1 at the very top.

samples controls how many texture reads are averaged together for each call - the averaging window automatically widens as position increases, mirroring how a real octave band gets proportionally wider at higher frequencies. Leave samples unconnected (a plain number, not wired from another node) so it compiles to a fixed loop count rather than a per-frame value; 8-32 is a reasonable range depending on how smooth vs. responsive you want the result.

For a plain linear read of the texture with no octave curve and a narrow, fixed averaging window - the better choice for following a single pure tone - use Texture Frequency.

Texture Frequency

The Texture Frequency node reads a frequency out of a Audio Track (Texture) (or Microphone (Texture)/Mix (Texture)/Runsheet Track (Texture)) node exactly where it sits in the texture - a plain, linear read, with no logarithmic curve applied. It works with the 512-wide and the older 256-wide spectrum textures alike.

position (0-1) runs straight across the spectrum: 0 is the lowest frequency, 1 is the top (half the sample rate, around 22-24 kHz), so the frequency read is position × half the sample rate. Every step along the way is the same number of Hz - which means the low notes of music sit close together near 0 (the whole range from 25 Hz to 1 kHz takes up only the first 4-5% of the way), and a tone sweeping evenly from low to high crosses the full width at an even pace. For a perceptually even, octave-based reading, use Texture Frequency Sample instead.

width (0-1) is how much of the spectrum is averaged around position, as a fraction of the whole width (0.01 is about 5 pixels of a 512-wide texture, roughly 240 Hz). Keep it small to follow a narrow peak - such as a single pure tone - and widen it to smooth the result. samples is how many reads are spread across that window and averaged (1 reads just the single pixel at position). Leave samples unconnected (a plain number, not wired from another node) so it compiles to a fixed loop count.

Gestures

Hand Gesture

The Hand Gesture node captures a particular gesture from a hand based on texture input (usually video from the camera).

These gesture property accepts the following values:

• None
• Closed Fist
• Open Palm
• Pointing Up
• Thumb Down
• Thumb Up
• Victory
• I Love You

The hand property determines if it’s the first or second hand present in the video stream. If only one hand is present, it should be set to 0.

Gesture Landmark 2D

The Gesture Landmark 2D node captures the 2D view plane (screen space) coordinates of a specific gesture landmark from a hand based on the gestureData input provided by the Gesture Data node.

To compare 3D positions with view plane coordinates, use the View Plane Position node.

Gesture Landmark 3D

The Gesture Landmark 3D node captures a specific gesture landmark from a hand based on the gestureData input provided by the Gesture Data node. It’s similar to Gesture Landmark 2D, but also provides depth (z-coordinate). NB. The value of the z-coordinate is small, as is the depth relative to the wrist landmark.

Pose Landmark 2D

The Pose Landmark 2D node captures the 2D view plane (screen space) coordinates of a specific pose landmark from a figure based on the poseData input provided by the Pose Data node.

To compare 3D positions with view plane coordinates, use the View Plane Position node.

Mapping

Planar Mapping XY

Maps 3D coordinates to a 2D surface in one plane only, either XY (front/back), XZ (top/bottom), or ZY (left/right).

The example below shows planar mapping in the XY plane.

Planar mapping does not flip the back image, so when viewed from the back, it will appear mirrored.

Planar Mapping XZ

Maps 3D coordinates to a 2D surface in one direction, either XY (front/back), XZ (top/bottom), or ZY (left/right). For more information see Planar Mapping XY.

Planar Mapping ZY

Maps 3D coordinates to a 2D surface in one direction, either XY (front/back), XZ (top/bottom), or ZY (left/right). For more information see Planar Mapping XY.

Box Mapping

Maps a texture onto all six sides of a box. Orients all textures outwards, so that no side is mirrored, e.g., a text in the texture can be read on all sides.

Spherical Mapping

Maps a texture onto a sphere. As the texture wraps around, it is best to use Texture Value LOD to avoid artifacts at the seam.

If you map an equirectangular image using this method, it will seamlessly wrap around the sphere (as if the sphere reflected the surroundings).

Cylindrical Mapping

Maps a texture onto a cylinder (in the Y direction).

View Plane Mapping

Maps an image in the view plane (i.e. the screen plane). This means the 3D properties of the object becomes irrelevant.

Box Atlas Mapping

The Box Atlas Mapping is used to map six different images onto the six sides of a cube. The six images must be part of a composite atlas image containing 2x3 smaller images (see below).

Equirectangular Mapping

The Equirectangular Mapping node is used to map an Equirectangular Image onto a cube.

If the cube is seen from inside, it creates a seamless 360-degree background, just as when Equirectangular Images are used for the overall scene Background.

nodeake sure to set Material Sides to Back or Both, and disable shadows.

Stretched Mapping

The Stretched Mapping node creates a UV mapping in the XY plane that wraps around the edge, taking into account the Z-coordinate to a degree determined by the zFactor parameter.

The example below shows Stretched Mapping with a zFactor of 0.5.

Box Map Index

The Box Map Index node returns the index of the side of a box to which the 3D coordinates would be mapped. This should be used together with the RGBA Selector nodes to map different RGBA streams to the different sides of a cube. If you want to map different images to the different sides, make sure to use Box Mapping for the uv input to Texture Value .