AI Video 101: How to Master AI Videos (Beginner to Advanced)
Shifting from plain text prompting to image-start JSON frameworks allows AI filmmakers to eliminate random model cuts and enforce professional camera direction.
What this lesson covers
Youri van Hofwegen outlines a structured methodology for advancing from basic text-to-video generation to precise image-anchored JSON prompting in Higgsfield. By combining initial frame generation with structured parameters for camera, action, and lighting, filmmakers can maintain strict directoral control over shot pacing and visual continuity.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 01:05 ↗
Simple text-to-video prompts leave framing, lighting, and camera cuts to model randomness rather than deliberate direction.
- 03:39 ↗
Using a dedicated image model to design a start frame establishes composition and visual quality before video diffusion begins.
- 04:38 ↗
Higgsfield requires a one-time media eligibility check on custom initial frames before video generation can be triggered.
- 06:10 ↗
Kling 3 serves as a budget-friendly image-to-video alternative to Seedance 2 when animating stationary or single-frame setups.
- 06:38 ↗
JSON prompting categorizes video controls into distinct parameters (subject, action, camera, lighting, style) to enable isolated scene adjustments.
Workflow outlined in the video
- Create high-resolution opening frames in an image generator to lock shot composition before starting video animation.
- Upload source frames to Higgsfield and execute the single-time asset eligibility check.
- Use a dedicated prompt generator like videoprompt.studio to format creative ideas into structured JSON video prompts.
- Specify camera motion, subject action, and lighting parameters in separate JSON keys to avoid rewriting entire prompts during revisions.
- Select Kling 3 for cost-efficient image-to-video renders, or Seedance 2 for complex dynamic motion.
Before you use this workflow
These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.
ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.
Explore the tools
How to interpret the numbers
23,051 views captured 27 September 2026. ReelStack recommendation score: 17/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 45% performance: 0.38× views vs 7 other channel videos observed at a similar age. View-evidence factor 96% (views / (views + 1,000))
- 25% freshness: 5/100 with a 14-day half-life
- 20% momentum: 102 views/day vs 575 channel baseline (5 comparable uploads), with the same view-evidence factor
- 10% engagement: 5/100; likes + 4× comments, smoothed with a 500-view neutral prior