VAW ClipForge v1.2.2 - Local Performance Boost

VAW ClipForge v1.2.2 - Local Performance Boost

This release improves local video analysis and rendering performance, with hardware-aware fallbacks for Windows systems.

Highlights

• NVIDIA NVENC rendering: use supported NVIDIA hardware encoding for faster H.264 exports, with a safe software-encoding fallback.
• Whisper CPU optimization: use INT8 inference and a bounded five-thread configuration to keep resources available for Windows and the UI.
• CUDA runtime validation: detect incomplete CUDA/cuDNN environments and fall back to CPU instead of failing transcription.
• Faster scene analysis: reduce visual sampling from 30 FPS to 15 FPS to lower frame-decoding overhead.
• Improved activity diagnostics: expose analysis stages, transcription segment counts, processing timings and the selected encoder.

Windows download

Download VAW-ClipForge-v1.2.2-win-x64.exe below. It is a standalone, windowed Windows executable; Python does not need to be installed.

Notes

• NVIDIA GPU encoding requires a compatible NVIDIA GPU and driver. If NVENC is unavailable, ClipForge falls back to software encoding.
• Whisper currently uses CPU INT8 when a complete CUDA + cuDNN runtime is not available. CUDA transcription acceleration is not included in this release's tested configuration.
• First-time use may download the selected Whisper model.
• Windows may display a SmartScreen warning because this executable is not code-signed.

Verification

• Windows application build completed successfully.
• Local regression suite: 57 tests passed.
• Tested NVIDIA NVENC encoding on the local GTX 1660 SUPER system.

---

Vanitas Studio - Build. Design. Innovate.