Hello, I have a suggestion: when I try to generate images with text using your AI models, I often encounter spelling errors in German. For example, I am currently trying to create a keychain with ChatGPT and would like it to generate a transparent image featuring the text “Lüneburg”. The chatbot should not translate the word; it should simply write L ü n e b u r g without errors. Perhaps this feedback will be helpful!
Kind regards,
Marcus Petersen-Clausen
A minimalist and elegant keychain design featuring the single word ‘Lüneburg’ in a clean, modern, and stylish font. The keychain has a sleek, abstract shape with subtle black-tone textures and patterns, creating a sophisticated appearance. The background is pristine black, ensuring the keychain stands out. The text ‘Lüneburg’ is designed in a refined metallic silver, making it suitable for engraving or printing. The overall composition is balanced, contemporary, and luxurious.
Great! How did you do that? What do I need to enter? I just want the text so I can have the keychain produced for private use.
First, shame on me 
because I forget the say “Welcome to the community!”
There was no extra effort on my side, maybe just coincidence or maybe divine alignment between words, who knows! 
As mentioned in OpenAI’s help page here:
https://help.openai.com/en/articles/6781228-how-can-i-generate-text-in-my-image
generating text in images can be tricky because DALL·E is not specifically designed for creating text. It’s more about generating artistic and realistic images based on your descriptions.
Mostly, text in images might not come out as expected, but luckily in this case, it worked out well!
Other examples are here, the English phrase is correct, but the German phrase is not. Interestingly, sometimes, long English text is correct. Probably due to the training data:
This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.
