C-LARA

Here’s a recent interaction with the new OpenAI model o1-preview, which I at least found rather impressive:

mannyrayner:

I have an unusual task you may be able to assist with. I’m currently reading Jon Fosse’s novel “Septologien”, where as you may know a painting is central to the story. The painting is described like this on the first page:

Og eg ser meg stå og sjå mot biletet med dei to strekkane, ein lilla og ein brun, som kryssar einannan på midten, eit avlangt bilete, og eg ser at eg ha måla strekade langsamt og med tjukk oljemåling, og ho har runne, og der den brune og den lilla linje kryssar einannan blandar fargen seg vakert og renn nedetter og eg tenkjer at dette ikkje er noko bilete. men samstundes er bilete slik det ska vera, det er ferdig, det er ikkje meir at gjera med

Later in the story, it is described as an “andreaskross” and indeed receives that title. Here is the definition from Wikipedia:

Andreaskors, eller X-kors (engelsk: saltire eller St. Andrew’s cross) er et kors med diagonale armer formet som en X. Det har fått sitt navn etter apostelen Andreas, som ifølge legenden døde som martyr på et slikt kors. Andreaskors forekommer som skrifttegn i mange kulturer, i kunst og dekorasjoner, i heraldikk, flaggvesen, bumerker, logoer og andre kjennetegn.

Andreaskors inngår også som en del av den synlige konstruksjonen (avstiving) i norske stavkirker.

The painting’s shape is a rectangle, considerably broader than it is tall.

I tried to persuade gpt-4o to create a photorealistic image of the painting sitting on an easel in the painter’s studio, without success. It can get quite close, but something is always off.

I wonder if you can create detailed instructions that I can relay to gpt-4o, which might resolve the issue?

o1-preview:

Certainly! Here’s a detailed prompt you can relay to GPT-4 to help generate the image you’re envisioning:

---

Prompt:

“Create a photorealistic image depicting a painting resting on a wooden easel in an artist’s studio.

Painting Details:

• Canvas Shape: The painting is on a rectangular canvas that is significantly wider than it is tall—an oblong shape.
• Background: The canvas has a neutral off-white background.
• Main Elements: On the canvas, there are two thick, diagonal strokes of oil paint:
  • Colors: One stroke is a deep purple, and the other is a rich brown.
  • Positioning: The strokes cross each other exactly at the center of the canvas, forming an ‘X’ shape, also known as a St. Andrew’s Cross.
  • Technique: The strokes were applied slowly with thick oil paint, resulting in a textured appearance. The paint is thick enough that it has started to run and drip downward.
  • Color Blending: At the point where the purple and brown strokes intersect, the colors blend beautifully, and this mixed color also drips downward, creating a visually striking effect.
• Texture: Emphasize the texture of the thick oil paint, with visible brush strokes and drips running down the canvas.

Studio Details:

• Lighting: The studio is illuminated with natural, soft light coming from a window, casting gentle shadows and highlighting the textures in the painting.
• Easel: The painting rests on a sturdy wooden easel that supports the wide canvas.
• Surroundings: Around the easel, there are various artist materials:
  • Paint brushes of different sizes.
  • Tubes of paint, including ones with purple and brown labels.
  • A palette with mixed paints.
  • Possibly a stool or a small table with additional art supplies.
• Ambiance: The overall atmosphere is quiet and contemplative, reflecting an artist’s personal workspace. The studio may have paint-splattered floors or walls, adding to the authentic feel.

Additional Notes:

• Focus on realism and attention to detail, ensuring the textures, lighting, and materials look lifelike.
• The composition should draw the viewer’s eye to the painting on the easel, with the surrounding studio elements complementing the central focus.
• Avoid adding any extra elements not described in this prompt.”

---

By providing these detailed instructions, GPT-4 should be better equipped to generate the accurate and vivid image you’re looking for. Let me know if you need further assistance or adjustments to the prompt!

I gave the instructions to gpt-4o, and it immediately produced this:

Hardcore Fosse fans will no doubt complain that it isn’t powerful enough to be the central image in a Nobel Prize winning novel, but let’s see what the next version but one can do.