ComfyUI - Z Image Turbo 이미지 생성 최적화 워크플로우

This video introduces my main ComfyUI image generation workflow using both Z image Turbo and Z image Base Z image Turbo creates beautiful and detailed images but it can often produce similar compositions and moods In this workflow Z image Base is used first to create variation and composition then Z image Turbo adds cinematic detail and the final upscaling step improves the overall image quality 00:00 Intro 00:38 Models Used 01:00 Pros and Cons of Z image Turbo 03:39 FLUX1 Variation Comparison 05:04 Core Workflow Structure 05:46 Role of Z image Base 07:21 Draft Image with Z image Base 08:48 Detail Enhancement with Turbo 10:26 3MP Upscaling 11:10 Image Compare Result 12:05 Portrait Close-Up Test 15:01 How to Use LoRA 16:56 Reference Image Workflow 17:32 Qwen3 LLM Image Analysis 18:38 Prompt Extraction Result 19:40 Batch Generation and Variation 20:23 ControlNet Tips 20:43 Key Workflow Notes 21:47 Reddit Feedback and Quality 22:15 Additional Upscaling with SeedVR 22:37 sampler / scheduler Settings 23:10 Closing