<h2 id="professional-video-to-video-transformation-with-wan-vace-i7w088kze">🎬 Professional Video-to-Video Transformation with WAN VACE</h2>Transform your videos with professional quality using this comprehensive ComfyUI workflow for WAN VACE. This complete pipeline enables seamless video-to-video transformation of long-form videos with advanced features including seamless joining, upscaling, and frame interpolation. Break down lengthy videos into manageable segments, process them individually, and seamlessly combine them back into cohesive, high-quality output.<h2 id="key-features-nxirl0b5c">✨ Key Features</h2><ul><li>Long Video Processing: Handle extended video content by breaking into segments and seamlessly rejoining</li><li>Complete V2V Pipeline: Full video-to-video transformation workflow</li><li>Seamless Video Joining: Custom nodes for professional video concatenation without visible transitions</li><li>Multi-Step Process: Generate → Join → Combine → Upscale → Interpolate</li><li>Professional Quality: High-quality output with customizable settings</li><li>Memory Optimization: Low VRAM options for various GPU configurations</li><li>Batch Processing: Process multiple video segments efficiently</li><li>Scalable Architecture: Handle videos of any length through intelligent segmentation</li></ul><h2 id="requirements-oqtt92seg">📋 Requirements</h2><h3 id="essential-model-files-hhc5eyfx7">Essential Model Files</h3>🔴 WAN GGUF Models<ul><li>Download from: <a target="_blank" rel="ugc" href="https://huggingface.co/QuantStack/Wan2.1_T2V_14B_FusionX_VACE-GGUF/tree/main">QuantStack/Wan2.1_T2V_14B_FusionX_VACE-GGUF</a></li><li>Choose your preferred quantization (Q3_K_S, Q8_0, etc.)</li><li>Place in: ComfyUI/models/unet</li></ul>🟣 WAN VAE<ul><li>Download wan_2.1_vae.safetensors from: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors">Comfy-Org/Wan_2.1_ComfyUI_repackaged</a></li><li>Place in: ComfyUI/models/vae</li></ul>🟣 WAN Text Encoder<ul><li>Download GGUF text encoders from: <a target="_blank" rel="ugc" href="https://huggingface.co/city96/umt5-xxl-encoder-gguf/tree/main">city96/umt5-xxl-encoder-gguf</a></li><li>Place in: ComfyUI/models/text_encoders</li></ul>Required Custom Nodes⚠️ Important: Download these custom nodes from this page (not available in ComfyUI-Manager):<ul><li>seamless_join_video_<a target="_blank" rel="ugc" href="http://clips.py">clips.py</a></li><li>combine_video_<a target="_blank" rel="ugc" href="http://clips.py">clips.py</a></li><li>Place in: ComfyUI/custom_nodes/</li></ul>ComfyUI Extensions⚙️Install these custom notes using the ComfyUI-Manager.<ul><li>ComfyUI-GGUF</li><li>ComfyUI-VideoHelperSuite</li><li>ComfyUI-KJNodes</li><li>ComfyUI-ControlNet-Aux</li><li>ComfyUI-Frame-Interpolation</li><li>ComfyUI-Easy-Use</li></ul><h2 id="step-by-step-guide-vd4eohl43">📖 Step-by-Step Guide</h2>Initial Setup<ol><li>Configure Constants:<ul><li>Width/Height: 576x1024 (9:16 aspect ratio) or match your source video</li><li>Length: 81 frames per segment</li><li>Skip Frames: Start with 0</li><li>Filename Prefix: Set your output folder and prefix</li></ul></li><li>Load Source Materials:<ul><li>Load your source video for restyling</li><li>Load reference image (ensure similar pose to first video frame)</li><li>Use SDXL/FLUX with LoRA and ControlNet for best pose matching</li></ul></li></ol><h3 id="step-1:-generate-wan-videos-a69lysg2y">Step 1: Generate WAN Videos</h3><ol><li>Write Prompts:<ul><li>Describe subject, outfit, and background</li><li>Include action phrases for dynamic results</li></ul></li><li>Generate Video Segments:<ul><li>Click run to generate first 81-frame video segment</li><li>Increase skip frames by 81 to process next segment</li><li>Repeat for the entire length of your source video</li><li>Final segment can be shorter but may have lower quality</li><li>For long videos: Continue this process until you've covered the full duration</li></ul></li></ol><h3 id="step-2:-join-videos-seamlessly-unf8yt9qr">Step 2: Join Videos Seamlessly</h3><ol><li>Configure Joining:<ul><li>Set folder path to your generated videos</li><li>Set filename prefix matching your generated files</li><li>Start with filename suffix = 1</li><li>Use same prompt from Step 2</li></ul></li><li>Join Process:<ul><li>Run to join first and second videos</li><li>Increase filename suffix by 1</li><li>Run to join second and third videos</li><li>Repeat until all segments are joined</li></ul></li></ol><h3 id="step-3:-combine-upscale-and-interpolate-02jk8bkz3">Step 3: Combine, Upscale, and Interpolate</h3><ol><li>Final Processing Setup:<ul><li>Set folder path to joined videos</li><li>Keep filename suffix = 1 (constant)</li><li>Set combine filename for final output</li><li>Set upscale filename for enhanced version</li></ul></li><li>Execute Final Pipeline:<ul><li>Combine all joined videos</li><li>Upscale using RealESRGAN (2x scale)</li><li>Interpolate frames using FILM VFI (2x frame rate)</li></ul></li></ol><hr /><h3 id="advanced-settings-1w1umz9u7">⚙️ Advanced Settings</h3>Low VRAM Configuration<ul><li>Use the UnetLoaderGGUFDisTorchMultiGPU node for memory optimization</li><li>Set virtual_vram_gb to 2.0-4.0 for 12GB and lower GPUs</li><li>Enable use_other_vram for additional memory fallback</li></ul>Performance Optimization<ul><li>Bypass PathchSageAttentionKJ and ModelPatchTorchSettings if you don't have Triton</li><li>Adjust batch sizes based on your GPU memory</li><li>Use appropriate quantization levels for your hardware</li></ul><hr /><h3 id="tips-for-best-results-efa1xg682">🎯 Tips for Best Results</h3><ol><li>Long Video Strategy: Plan your segmentation approach - 81 frames per segment ensures smooth transitions while maintaining manageable processing chunks</li><li>Reference Image Quality: Use high-quality reference images with poses similar to your source video's first frame</li><li>Prompt Engineering: Be specific about subject details, clothing, and background elements</li><li>Segment Planning: Plan your video segments to maintain narrative continuity across the entire video length</li><li>Hardware Considerations: Adjust settings based on your GPU capabilities - longer videos benefit from optimized VRAM settings</li><li>Consistency Maintenance: Keep prompts consistent across all segments to ensure visual coherence in the final long video</li></ol><hr /><h3 id="troubleshooting-7wgjo1z84">🩺 Troubleshooting</h3><ul><li>OOM Errors: Increase virtual_vram_gb or reduce batch sizes</li><li>Missing Nodes: Ensure all custom nodes are properly installed</li><li>Quality Issues: Check reference image alignment and prompt specificity</li><li>Processing Slow: Consider using lower quantization models for faster generation</li></ul><hr /><h2 id="custom-nodes-parameter-guide-4t2vajiij">🔧 Custom Nodes Parameter Guide</h2><h3 id="wanvideovaceseamlessjoin-node-vkls7bxr9">WanVideoVaceSeamlessJoin Node</h3>This custom node seamlessly joins two video clips with intelligent masking for smooth transitions.Parameters:<ul><li><code>mask_last_frames</code> (INT): Number of frames to mask at the end of the first video<ul><li>Default: 0</li><li>Range: 0-20</li><li>Use 0 for no masking, 5-10 for subtle blending</li></ul></li><li><code>mask_first_frames</code> (INT): Number of frames to mask at the beginning of the second video<ul><li>Default: 10</li><li>Range: 0-20</li><li>Recommended: 10 frames for smooth transitions</li></ul></li><li><code>frame_load_cap</code> (INT): Maximum number of frames to load from each video<ul><li>Default: 81</li><li>Range: 1-1000</li><li>Should match your segment length (typically 81)</li></ul></li><li><code>first_video_path</code> (STRING): Full path to the first video file<ul><li>Format: <code>"C:\path\to\video1.mp4"</code></li><li>Use absolute paths for reliability</li></ul></li><li><code>second_video_path</code> (STRING): Full path to the second video file<ul><li>Format: <code>"C:\path\to\video2.mp4"</code></li><li>Ensure file exists and is accessible</li></ul></li></ul>Outputs:<ul><li><code>image</code>: Combined video frames as image sequence</li><li><code>mask</code>: Generated mask for the transition area</li></ul><hr /><h3 id="combinevideoclips-node-a72c99hg6">CombineVideoClips Node</h3>This node combines multiple video clips into a single continuous sequence with advanced masking options.Parameters:<ul><li><code>frame_load_cap</code> (INT): Maximum frames to load per video<ul><li>Default: 81</li><li>Range: 1-1000</li><li>Should match your segment frame count</li></ul></li><li><code>mask_last_frames</code> (INT): Frames to mask at the end of each video (except last)<ul><li>Default: 0</li><li>Range: 0-20</li><li>Use 0 for clean cuts, 5-10 for fade effects</li></ul></li><li><code>mask_first_frames</code> (INT): Frames to mask at the beginning of each video (except first)<ul><li>Default: 10</li><li>Range: 0-20</li><li>Recommended: 10 for smooth transitions</li></ul></li><li><code>first_video_path</code> (STRING): Path to the first video in sequence<ul><li>Base video - typically your original generated video</li></ul></li><li><code>first_joined_video_path</code> (STRING): Path to first seamlessly joined video<ul><li>Result from first WanVideoVaceSeamlessJoin operation</li></ul></li><li><code>second_joined_video_path</code> (STRING): Path to second seamlessly joined video<ul><li>Result from second WanVideoVaceSeamlessJoin operation</li></ul></li><li><code>third_joined_video_path</code> (STRING): Path to third seamlessly joined video<ul><li>Continue pattern for additional segments</li></ul></li><li><code>fourth_joined_video_path</code> (STRING): Path to fourth seamlessly joined video<ul><li>Optional - use if you have this many segments</li></ul></li><li><code>fifth_joined_video_path</code> (STRING): Path to fifth seamlessly joined video<ul><li>Optional - maximum supported segments</li></ul></li><li><code>last_video_path</code> (STRING): Path to the final video in sequence<ul><li>The last generated video segment</li></ul></li></ul>Output:<ul><li><code>image</code>: Combined video sequence as image frames ready for final processing</li></ul><hr /><h3 id="parameter-optimization-tips:-4mjlqztwe">Parameter Optimization Tips:</h3>For Seamless Joining:<ul><li>Short transitions: <code>mask_first_frames = 5</code>, <code>mask_last_frames = 0</code></li><li>Smooth blending: <code>mask_first_frames = 10</code>, <code>mask_last_frames = 5</code></li><li>Long crossfades: <code>mask_first_frames = 15</code>, <code>mask_last_frames = 10</code></li></ul>For File Paths:<ul><li>Ensure all video files exist before running</li><li>Use consistent naming conventions for easier batch processing</li></ul>Frame Count Considerations:<ul><li>Set <code>frame_load_cap</code> to match your segment length (usually 81)</li><li>Smaller values may truncate longer segments</li></ul><hr />This workflow provides professional-grade video transformation capabilities with comprehensive control over the entire pipeline from generation to final output.

WAN VACE Professional Video-to-Video Complete Workflow

About this versionNew Features &amp; EnhancementsWan 2.2 Animate Model<ul><li>Introduces the Wan 2.2 Animate Model to capture pose dynamics from source video and apply them to a reference image, combined with Wan Vace technology to create seamless joins for extended animated video sequences.</li></ul>Installation &amp; Update InstructionsRequired File Updates:<ol><li>Navigate to your <code>ComfyUI\custom_nodes</code> folder</li><li>Remove the following legacy files:<ul><li><code>seamless_join_video_clips.py</code></li><li><code>combine_video_clips.py</code></li></ul></li><li>Install the updated files from the zip package:<ul><li><code>seamless_join_video_clips.py</code> (updated version)</li><li><code>combine_video_clips_extended.py</code> (new extended functionality)</li></ul></li></ol>Important: Make sure to completely remove the old files before installing the new ones to prevent conflicts.

Files with SHA256: fb715178b153aee7449ec6d9caffd7fc027c8fd6246231621f67a2f44101c605