The Chair No. 2, after Elias Avramidis
After Series
- 5The Fall of Man, after Titian, c.1550
- 6Modern Day Myth and Mythology, after Ben Fink, 2023
- 7Wave (SOTT), after Tom of Finland, 1974
- 10The Fall of Man, after Titian, c.1550
I’ve done a few series of works based on other artists, but mostly old masters that are long gone — The Birth of Ken, After The Birth of Venus by Botticelli, and Vitruvian Man, After Leonardo Da Vinci, c. 1480. But never with a contemporary artist. This would be the first.
Recently, I began following an Instagram account @elias_alpha_ through suggestions from Instagram Explore. I was initially drawn to Elias’ self-portraits, which have a delightful blend of seriousness and playfulness. I find many of them to be quite humorous and have developed an appreciation for his artworks.
Yesterday, I got really inspired and I made a series of paintings based on a drawing he posted. Titled “The Chair No. 2” (color pencil on textured paper), it depicts a male figure interacting / pointing at a yellow chair.
My re-interpretation mostly picks up on the pose of the figure, the yellow chair, and the vertical blue and white stripes on the walls. The link might not be very apparent, but hopefully the references are noticeable.
Images
Elias Avramidis
About Elias Avramidis, quoted from his website:
Elias Avramidis is a visual artist based in East London. His work often deals with the human figure within an aquatic geometry of colours. Thematically, he has been dealing with imaginary islands and contemporary mysticism through icons and hagiography. His practice involves colour pencils and crayons on textured paper. The human figures are based on live models which he photographs, keeping them uncoloured so that they appear as alien and unintegrated as possible in the aquatic environment. This is both an ecological metaphor and a way of expressing the anthropocenic peculiarity of our planetary emplacement.
Tech
Elias’ original work is rectangular, so first I used Photoshop AI to expand the physical space into a composition in square. Then in Stable Diffusion, I employed two control nets: OpenPose for the figure, and Reference Only mostly for the yellow chair and wallpaper.
Before using Reference Only, I tried using Canny and Soft Edge, but they suffer from limiting the overall composition too much. Reference Only is a good control net to use because it picks up visual ideas (yellow chair, blue and white textured walls) without making them look the same.
I rendered these with Airfuck’s Brute Mix, and Clip Skip 2 because it has the best balance of imagination and a photorealistic painterly quality that I prefer.
Text prompts in Stable Diffusion txt2img, with control nets.
- Airfuck’s Brute Mix, Clip Skip 2
- Euler a, 20 steps, 512x512
- vae-ft-mse-840000-ema-pruned
- Denoising: 0.5
- Hires upscale: 2, 1024x1024, 10 steps, 8x_NMKD-Superscale_150000_G
- Control Net 0: OpenPose Full. Weight 1, Start 0, End 1, Balanced
- Control Net 1: Reference Only, Weight 0.5, Threshold A: 0.5, Start 0, End 1, Balanced
- Composition expansion: Adobe Photoshop AI
- Post Upscale: Topaz Gigapixel HQ 4x
- Color Correction: Adobe Lightroom
- Inpainting: Adobe Photoshop AI
Also see full technical parameters.
Studies
Here are a few sketches made during the process. These are provided to show the range of compositional ideas using Reference Only at 0.5 weight. But you’ll note that I had a very hard time picking up on the pointing finger in most of them.
Images
Additional Studies
Not posted to Instagram, but made during the same time.
Images
Earlier Studies
At first, I tried using photographic renders, using the Virile Realty v3 Beta checkpoint. Some of these turned out ok but they lack the emotional quality that I like in paintings, so I switched to using Airfuck’s Brute Mix instead.
Images
Image Index
Technical Parameters
- Steps: 20
- Sampler: Euler a
- CFG scale: 5
- Seed: 4141041130
- Size: 512x512
- Model hash: 8ff8347bb0
- Model: airfucksBruteMix_v10
- VAE hash: 235745af8d
- VAE: vae-ft-mse-840000-ema-pruned.vae.ckpt
- Denoising strength: 0.5
- Clip skip: 2
- ControlNet 0:
- Module: openpose_full
- Model: control_v11p_sd15_openpose (cab727d4)
- Weight: 1
- Resize Mode: Crop and Resize
- Low Vram: False
- Processor Res: 512
- Guidance Start: 0
- Guidance End: 1
- Pixel Perfect: False
- Control Mode: Balanced
- ControlNet 1:
- Module: reference_only
- Model: None
- Weight: 0.5
- Resize Mode: Crop and Resize
- Low Vram: False
- Threshold A: 0.5
- Guidance Start: 0
- Guidance End: 1
- Pixel Perfect: False
- Control Mode: Balanced
- Hires upscale: 2
- Hires steps: 10
- Hires upscaler: 8x_NMKD-Superscale_150000_G
- Lora hashes:
- bulgerk-dickprint: b1934ecdac9a
- TI hashes:
- negative_hand-neg: 73b524a2da12
- NegfeetV2: df90b1ff666d
- Asian-Less-Neg: 22d2f003e76f
- GS-DeFeminize-neg: 26bd0d842cc5
- face_editor_enabled: True
- face_editor_face_margin: 1.6
- face_editor_confidence: 0.97
- face_editor_strength1: 0.4
- face_editor_strength2: 0
- face_editor_max_face_count: 20
- face_editor_mask_size: 0
- face_editor_mask_blur: 12
- face_editor_prompt_for_face: handsome greek face
- face_editor_apply_inside_mask_only: True
- face_editor_apply_scripts_to_faces: False
- face_editor_face_size: 512
- face_editor_use_minimal_area: False
- face_editor_ignore_larger_faces: True
- face_editor_affected_areas: Face
- face_editor_upscaler: None
- face_editor_tilt_adjustment_threshold: 40
- Automatic1111 Version: v1.6.0
- Post Processing
- Composition Expansion: Adobe Photoshop AI
- Color correction: Adobe Lightroom
- Limb fix / inpainting: Adobe Photoshop AI Generative Layers