Cómo usar
- Copia el comando y revisa el paquete antes de instalarlo en tu entorno de agente.
- Consulta las herramientas, credenciales y entornos necesarios en el documento original.
- Describe las entradas y el resultado deseado; contrasta la salida con las fuentes.
npx skills add https://github.com/nphu0811/pmg-skills --skill pmg-camera-parallaxAntes de empezar
- La instalación no incluye suscripción al modelo, cuota API ni permisos sobre tus archivos.
- Las estrellas e instalaciones reflejan actividad, no una prueba independiente de calidad o seguridad.
Contenido del repositorio de origen con títulos, ejemplos, código, tablas, enlaces e imágenes preservados.
PMG Camera & Parallax Engine (pmg-camera-parallax)
Responsibility and dependencies
Shared camera transforms, paths, pan/zoom, depth coordinates, parallax, camera-relative placement and continuity. 3D/model composition must reuse this system. Canonical rig consolidation is staged in v1.7.0; retain working legacy APIs meanwhile.
Dependencies: pmg-engine; core geometry; Morph/native motion compilers. All generated candidates must pass PMGOrchestrator before delivery; direct save/COM snippets below are lower-level composition examples, not independent production entrypoints.
A presentation should not feel like flat slides changing; it must feel like a continuous 3D space viewed through a virtual cinematic camera.
1. Multi-Plane Parallax Architecture
To simulate genuine physical camera depth, elements are segregated into 3 distinct spatial planes moving at differentiated velocities during Morph transitions:
+---------------------------------------------------------------------------------+
| PLANE 1: BACKGROUND GRID / HORIZON (Shift: 0.7x) |
| Moves slowly across slides to create infinite horizon depth. |
| Identifier: !!BG_Studio, !!BG_Grid |
| |
| PLANE 2: 3D MODEL / MACHINE MIDGROUND (Shift: 1.0x, Zoom: 100% -> 240%) |
| Primary focal anchor. Undergoes camera push-in and lateral panning. |
| Identifier: !!Hero3D_Focus, !!Midground_Core |
| |
| PLANE 3: FOREGROUND PARTICLES / MIST (Shift: 1.4x) |
| Floats close to the virtual lens, moving faster to exaggerate depth. |
| Identifier: !!FG_Dust, !!FG_LightStreak |
+---------------------------------------------------------------------------------+Parallax Mathematics
When the virtual camera pans by $\Delta X{\text{camera}}$: $$\Delta X{\text{BG}} = 0.7 \times \Delta X{\text{camera}}$$ $$\Delta X{\text{MID}} = 1.0 \times \Delta X{\text{camera}}$$ $$\Delta X{\text{FG}} = 1.4 \times \Delta X_{\text{camera}}$$
2. Virtual Camera Move Types
A. Wide Establishing Shot (Zoom = 100%)
- Purpose: Orients the audience by displaying the complete system or facility.
- Coordinates:
x = 0.0", y = 0.0", w = 13.333", h = 7.5".
B. Macro Push-In Zoom (Zoom = 180% – 240%)
- Purpose: Zooms deep inside a specific subsystem (e.g., combustion chamber, turbine blades, microchip core).
- Coordinate Formula:
$$W{\text{zoomed}} = W{\text{base}} \times Z$$ $$H{\text{zoomed}} = H{\text{base}} \times Z$$ $$X{\text{zoomed}} = X{\text{screen\target}} - (X{\text{subsystem\rel}} \times Z)$$ $$Y{\text{zoomed}} = Y{\text{screen\target}} - (Y{\text{subsystem\rel}} \times Z)$$
C. 3D Tracking Pan
- Purpose: Camera slides laterally along an energy conduit or pipeline, maintaining focus on the moving energy packet.
D. Pull-Back Reveal
- Purpose: Smoothly decelerates from a macro detail back to the global overview, illuminating the entire synchronized system.
3. Python Camera Control Pattern
from scripts.forge_lib import Cinematic3DDeck
deck = Cinematic3DDeck(title="AI Neural Core")
# Slide 1: Wide Overview (Zoom 1.0x)
s1 = deck.add_cinematic_slide(
scene_idx=1,
camera_zoom=1.0,
camera_focus_center=(6.66, 3.75),
header_title="NEURAL SUPERCOMPUTER",
callout_title="CLUSTER OVERVIEW",
callout_desc="High-density parallel processing architecture."
)
# Slide 2: Macro Push-In into Tensor Core (Zoom 2.2x)
s2 = deck.add_cinematic_slide(
scene_idx=2,
camera_zoom=2.2,
camera_focus_center=(4.2, 3.8), # Targets tensor processor coordinates
header_title="TENSOR PROCESSING UNIT",
callout_title="FP8 MATRIX CORE",
callout_desc="100 petaflops sparse mathematical execution."
)
