I'm a professional photographer and programmer with 20 years of experience across a wide range of lenses. While I love working with AI models, the lack of genuine lens control has always frustrated me — so I'm releasing my first lens LoRA. If there's interest, I'll create more.
Note: The Wan and Z-Image versions use different trigger words — "telephoto" for Wan, and "600mm" for Z-Image.
This LoRA replicates the look of a long prime lens. It was trained exclusively on material shot with a 600mm lens, and with the right prompting, you can achieve that level of compression. At other times, it may produce the feel of a 400mm, 200mm, 150mm, or similar — all of which are significantly harder to replicate without this LoRA.
Tips: Prompting your background, middle ground, and foreground makes a real difference.
To get the full long lens aesthetic, prompting for a full-length shot of a person helps considerably if your image is vertical — this will regularly deliver authentic long focal length perspective and full 600mm looks can be achieved.
For subjects where a 600mm is rarely used, like in side a building, achieving the effect is trickier — you may end up with something closer to 150mm. It's still possible, but you'll need to refine your prompt to push as far as you can to the 600mm look.
Z-image: between 0.5 and1.3 strength