I need a specialist who can take my mixed set of photos and short video clips and generate highly realistic full-body and face swaps. The output will be used in a research / educational context, so authenticity and reproducibility matter more than showy effects. You’ll receive organised folders containing both still images and 4-K video footage of the “source” and “target” subjects. From these, I expect swapped renders that hold up to frame-by-frame inspection: consistent skin tone, natural lighting integration, flawless lip sync, and no tell-tale artefacts around hairlines or clothing edges. Tools such as DeepFaceLab, FaceFusion, Stable Diffusion (video extension) or comparable custom Python pipelines are all welcome as long as the final quality is comparable to studio-grade deepfake demonstrations in published papers. Please include a short note on: • your preferred workflow and hardware (GPU specs, libraries, model versions) • the typical turnaround time per minute of processed footage Deliverables 1. Two sample stills and a 5-second proof clip for validation 2. Final swapped images in lossless PNG and videos in ProRes or high-bit-rate H.264, matched to the original resolution and frame rate 3. A concise setup summary so the experiment can be reproduced by other researchers I’m happy to answer technical questions quickly so we can keep the iteration cycle tight. Looking forward to seeing your approach and previous examples of high-fidelity swaps.