Cdaprod

"""
Move each finished recording to a UNC share.

OBS ▶ Tools ▸ Scripts ▸ + ▸ select this file
---------------------------------------------------------
•   Works with Simple or Advanced output mode.
•   No change to your regular Recording Path is required;
    in fact, keeping it on a local SSD makes the initial write faster.
---------------------------------------------------------
Tested with Python 3.11 + OBS 30.

Cdaprod

Updating Drivers for Ubuntu on Raspberry Pi 5 to support the VideoCore VII GPU

✅ GPU Support Overview

1. Mesa / V3D (OpenGL ES & Vulkan) drivers included, thanks to upstream support merged since Mesa 23.3 .
2. Initial desktop installs (e.g., Ubuntu 24.04/24.10) sometimes exhibit slow graphics, GNOME animation lag, even green-screen artifacts—likely due to incomplete or mismatched Mesa stacks .
3. Community feedback highlights that while OpenGL ES/Vulkan are available, you may need to install or update Mesa packages manually (mesa-vulkan-drivers, etc.) .

To Update Drivers:

Cdaprod

docker run --rm -i -v "$PWD":/w linuxserver/blender:4.4.3 \
  blender --factory-startup -b --python - <<'PY'
import bpy, math, os
vid = "/w/phone_screen.mp4"                 # <<< your vertical video
assert os.path.exists(vid), "Video missing"

# ------------------------------------------------------ build objects
bpy.ops.mesh.primitive_plane_add(size=2)
plane = bpy.context.object

Cdaprod

# /scripts/two_cam_auto_vfx.py
"""
Two-Camera VFX Auto-Rig -- Audio-Sync, Insta360-aware Edition
─────────────────────────────────────────────────────────────────────────────
•  Imports hero + witness plates
•  Automatically aligns them via audio-waveform cross-correlation
•  Auto-detects FPS, resolution & 360⇆rectilinear
•  Injects missing XMP → always pano-aware
•  Tracks witness plate on a seeded grid, iteratively cleans & re-solves
•  Builds   RigRoot ▶ WitnessCam ▶ HeroCam hierarchy

Cdaprod

NDI Stream to MoviePy Professional Video Processing

A comprehensive system for capturing NDI streams, recording them, and processing with MoviePy for professional video production workflows.

📦 Cell 1: Install Dependencies

# Install required packages
!pip install ndi-python opencv-python ipywidgets moviepy numpy pillow
# Note: Ensure NDI SDK/Runtime is installed on your system

Cdaprod

docker run --gpus all --restart unless-stopped -d `
  -v "B:\Models:/models" `
  -p 8000:8000 `
  ghcr.io/ggml-org/llama.cpp:server-cuda `
  -m /models/llama-2-7b-chat.Q4_K_M.gguf `
  --port 8000 --host 0.0.0.0 -n 512 --n-gpu-layers 35

Cdaprod

🎬 CDAProd — NodeVideo Expressions Guide

⚙️ Built by David Cannan — DevOps Engineer & AI Solutions Architect
💡 Self-made | #devopsdad | #tripletdad | #hacktheplanet

Cdaprod

Simple Cinematic Example for Text

fadeInDuration = 1.5
delay = (index - 1) * 0.2
t = time - (thisLayer.startTime + delay)

progress = clamp(t / fadeInDuration, 0, 1)
easeT = ease(progress, 0, 1)

Cdaprod

import os
import cv2
import xml.etree.ElementTree as ET

def get_video_technical_info(path):
    cap = cv2.VideoCapture(path)
    fps = cap.get(cv2.CAP_PROP_FPS)
    frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))

Cdaprod

TL;DR • It’s fine that both XQD cards individually mount as D:. • Differentiate them by VolumeLabel, not by drive letter. • Modify your script so the SMB share name = VolumeLabel (e.g. XQD_Card1, XQD_Card2). • Register your task once via PowerShell—after that, “hot-plug” automatically shares whichever card you inserted under the correct name.

Feel free to reach out if you want help auto-removing the share on eject, adding logging to see exactly when each card fires, or any other fine-tuning!

Explanation of Key Changes