I suggest giving the user the OPTION TO CHOOSE between (A) fx generation for the WHOLE WORD, understanding the whole word as ONE OBJECT and then apply the visual effect OR (B) fx generation for SINGLE LETTERS, as available today
Benefit: in some projects, e.g. a jungle scene, many elements such as leafs, snakes, tropical birds, etc. are repeated again and again on each individual letter which sometimes is just too much. Instead, I'd prefer to have e.g. lots of green leaves across the complete word, with only a FEW tropical animals on just some of the letters. In this case the COMPLETE WORD would be recognized as ONE OBJECT
I could imagine prompts like "dark tropical jungle with huge leafs in different green tones and 3 red parrots, 2 green snakes and 1 alligator", for example - everything related to the word as a whole
I think this offers more variations and better harmonization with the overall layout
Thanks for considering!