Nano Banana Tutorial: Removing Lens Flare with Negative Prompts

When generating images with strong light sources, AI tools often introduce optical imperfections known as lens flare. These bright streaks or circular artifacts can distract from the main subject and ruin the composition. While some users might want this effect for artistic reasons, many need a clean image where the lighting enhances the scene without creating visual noise. This tutorial explains how to use the Nano Banana interface to specifically target and remove these unwanted elements using negative prompts.

Understanding Negative Prompts for Flare Removal

Negative prompts are instructions that tell the AI what not to include in the final output. In the context of Nano Banana, which supports both text-to-image and image-to-image workflows, these prompts act as a filter against specific visual errors. When you describe a scene with a bright sun or a spotlight, the model might interpret the high contrast as an opportunity to generate lens flares automatically. By explicitly stating what you do not want, you guide the generation process toward a cleaner result.

It is important to note that prompt instructions describe desired outcomes but do not guarantee identity, label, object, or typography preservation. Therefore, when removing artifacts like lens flare, the goal is to refine the image quality rather than alter the core subject matter. The technique relies on clear communication between the user and the generator to distinguish between intentional lighting effects and accidental digital artifacts.

Step-by-Step Guide to Cleaning Your Image

To effectively remove lens flare using Nano Banana, follow these structured steps within the tool's interface:

1. Access the Generator: Navigate to the Nano Banana 2 product page at /nanobanana2. Select either the text-to-image or image-to-image mode depending on whether you are starting from scratch or editing an existing file.
2. Define Your Scene: Enter your primary positive prompt describing the subject and the desired lighting. For example, "A portrait of a person standing in a golden hour field with soft sunlight." Ensure you mention the lighting style you want so the AI knows what to keep.
3. Locate the Negative Prompt Field: Find the input area designated for negative prompts. This section is crucial for excluding specific unwanted features.
4. Input Exclusion Terms: Type specific terms related to the artifact you wish to remove. Common terms include "lens flare," "optical flare," "light streaks," or "glare." You can combine these based on the severity of the issue in your initial attempt.
5. Generate and Review: Run the generation process. If the first result still contains minor artifacts, refine your negative prompt by adding more descriptive terms or adjusting the weight if the interface allows.
6. Iterate: Use the results to fine-tune your approach. Since this is an AI tool, multiple attempts may be necessary to achieve the perfect balance between natural lighting and artifact removal.

A Usable Prompt Example

Below is an example of how you might structure your inputs. Please remember that these are examples of prompt structures and do not guarantee specific identity or object preservation in every generation.

Positive Prompt: "A cinematic shot of a vintage car parked under a streetlamp at night, dramatic shadows, realistic texture, 8k resolution"

Negative Prompt: "lens flare, light streaks, glare, optical distortion, overexposed highlights, artificial bloom"

This combination tells the AI to create a dark, moody scene with specific lighting while actively suppressing the bright artifacts that often appear around light sources. You can adapt these terms to fit different scenarios, such as outdoor photography or studio setups.

How to Judge Results and Fix Issues

Evaluating the success of your negative prompt strategy involves checking the areas around light sources. Look for any remaining streaks, halos, or circular patterns that were present in previous attempts. If the image looks too flat or the lighting seems unnatural, it may indicate that the negative prompt was too aggressive, removing legitimate atmospheric effects along with the flare.

If the lens flare persists despite using negative prompts, consider the following fixes:

• Refine Terminology: Try synonyms like "bloom" or "ghosting" alongside "lens flare."
• Adjust Positive Descriptions: Be more specific about the type of lighting you want, such as "soft diffusion" instead of just "bright light."
• Use Image-to-Image: If starting from a generated image with heavy flare, upload it to the image-to-image workflow and apply the negative prompt to guide the re-generation.

By understanding how to communicate clearly with the AI, you can maintain the intended atmosphere of your image while ensuring technical cleanliness. For those ready to experiment with these techniques, Try Nano Banana offers the platform to put these strategies into practice immediately.

Remember that while Nano Banana provides powerful tools for image manipulation, the results depend heavily on the clarity of your prompts. There are no guarantees of specific outcomes, but consistent refinement of your negative prompts will yield progressively better control over unwanted artifacts.