Same Pose, Different Places
I created a series of photographs and illustrations based on a selfie that feet45paris posted on Instagram.
Using Stable Diffusion and SDXL, I generated around 500 of these images, putting him in different places in Paris, including the Paris Charles de Gaulle Airport, Gare du Nord train station, the Louvre, Versailles, the Eiffel Tower, the Nike store, the Lacoste store, on the street, at the supermarket, and several other urban locations.
The images share a similar pose by using the Depth control net. The clothes he wears also look similar because I have specifically prompted those in the prompts — with cotton shirt and ripped blue jeans.
I then assembled the images together inside Adobe After Effects into a 30fps video, playing a different image every frame to create a video portrait of him — with the same pose, wearing almost the same clothes, but in all different places.
Video Square Short
Video Vertical Short
Tech
- Steps: 50
- Sampler: DPM++ 3M SDE Karras
- CFG scale: 7
- Size: 1024x1024
- Model hash: 31e35c80fc
- Model: sd_xl_base_1.0
- VAE hash: 63aeecb90f
- VAE: sdxl_vae.safetensors
- Denoising strength: 0.75
- ControlNet 0: "Module: depth_midas, Model: sai_xl_depth_128lora f5ddde75, Weight: 1, Resize Mode: Crop and Resize, Low Vram: False, Processor Res: 1024, Guidance Start: 0, Guidance End: 1, Pixel Perfect: False, Control Mode: Balanced"
- Automatic1111 Version: v1.6.0
- Post-Upscale Topaz Gigapixel HQ 4x, 4096x4096
- Post-Processing: Adobe Lightroom
- Animation: Adobe After Effects
- Music: Bring iIt On by Movediz, licensed.
- Soundtrack: Ableton Live Suite