Integrate Khmer Complex Script Layout Engine (HarfBuzz/Tokenization) into Video Generation Backend

Title: Feature Request: Integrate Khmer Complex Script Layout Engine (HarfBuzz/Tokenization) into Video Generation Backend

Component/Context:
Google Vids / Gemini Veo Video Generation Pipeline (Server-side Video Rendering Engine)

Issue Description:
Khmer Unicode script renders perfectly on the client-side (storyboard/browser interface) due to local OS layout engines. However, during the final video generation process (server-side baking/rasterization), the Khmer text layer breaks. Sub-consonants (coeng/legs) and vowels fail to stack correctly, detaching into isolated, misaligned glyphs or empty boxes.

The backend font rasterization and video encoding nodes are missing a complex text layout engine capable of processing Khmer script grammar rules.

Proposed Technical Solution:

  1. Backend Layout Engine Upgrade: Integrate a complex text layout engine like HarfBuzz or an equivalent open-source library into the server-side text-to-video rasterization pipeline.
  2. Tokenization & Segmentation: Implement a Khmer syllable-level tokenizer/segmentation library (such as khmer-nginx, KhmerNLP, or ICU Khmer word breaker) to correctly handle cluster boundaries before the font layers are baked into the video frames.
  3. Font Fallback: Ensure the rendering engine defaults to web-safe Khmer Unicode fonts (e.g., Noto Sans Khmer or Battambang) during server-side composition.

Impact:
Enabling correct Khmer text rendering will unlock multi-language video generation, automated AI captioning, and local business communication features for millions of Khmer speakers utilizing Google Workspace and Gemini AI tools globally.