Skip to main content
J
jean-bernardm85233851
Known Participant
June 10, 2026
Answered

Les 2 dernières images ne sont pas finies : floues et incomplètes

  • June 10, 2026
  • 1 reply
  • 20 views

Les 2 dernières propositions sous forme d’image snt très floues et incomplètes

      Correct answer Hajar Toutsi

      Hi there, thanks again for sharing your example.

      I tried working with the same prompt you mentioned, and I noticed that using the AI Assistant within Firefly can actually make a big difference for cases like logo creation. Instead of generating a single result, the Assistant helped by:

      • Providing multiple initial variations
      • Analyzing the outputs and suggesting refinements
      • Iteratively improving the results to make the text more recognizable and aligned with the intended “T+P” design

      Text generation within AI models is a known limitation, and Firefly models (both internal and third-party) are no exception. Because of that, text is treated more like a visual element in generative AI, it can often appear distorted or inconsistent, especially in logos. The Assistant helps guide the process step by step, which tends to produce more structured and usable outcomes.
       

      It even suggested finalization steps, like refining the best version and then polishing it in a design tool, which is a great workflow for production-ready logos.

      If you haven’t already, I’d definitely recommend trying your prompt with the AI Assistant, it can help you iterate faster and get closer to a clean result.
       

      Let us know how it goes or if you’d like help refining the prompt further.

      Hajar

      1 reply

      Hajar Toutsi
      Community Manager
      Hajar ToutsiCommunity Manager
      Community Manager
      June 11, 2026

      Hi jean-bernardm85233851,
      Thank you for reaching out about this.

      To be able to help you better, please provide us with the following information:

      • The Full Prompt: The exact text you entered into the prompt box.
      • Settings and Model: The specific model used and any other settings (such as aspect ratio, content type, or style strength).
      • Reference Images: A copy of any reference images or keyframes used to guide the generation.
      • Visual Examples: Please upload a copy of the "bad" generation so we can see exactly what the output looks like and tell us what is wrong if it is not obvious.


      Once you have the details, let us know so we can be able to assist you efficiently.

      Hajar
      [Moved to questions]

      J
      jean-bernardm85233851Author
      Known Participant
      June 11, 2026

      Bonjour, merci de votre retour :
      PROMPT : 
       

      Create a premium vector logo for a tennis and padel club on the shores of Lake Bourget. Design a modern logo as a monogram "T+P". Subtly integrates a tennis racket string pattern and a padel racket perforation pattern. Elegant string pattern, expressing tradition, precision and performance. The modern padel racket with circular perforations, expressing conviviality, innovation and accessibility. The transition between the two racket must be seamless and natural, creating a single coherent symbol. Minimal geometric construction, strong visual balance, premium sports branding. Flat vector design, clean negative space, scalable icon. Inspired by contemporary sports club identities and outdoor brands. No text, no shield, no mascot, no gradients, no 3D effects, no clipart, no realistic rackets, no tennis ball.

      Bannière cinématique 8:1
      And i tried 4/1
       


      No reference

       

      Hajar Toutsi
      Community Manager
      Hajar ToutsiCommunity ManagerCorrect answer
      Community Manager
      June 11, 2026

      Hi there, thanks again for sharing your example.

      I tried working with the same prompt you mentioned, and I noticed that using the AI Assistant within Firefly can actually make a big difference for cases like logo creation. Instead of generating a single result, the Assistant helped by:

      • Providing multiple initial variations
      • Analyzing the outputs and suggesting refinements
      • Iteratively improving the results to make the text more recognizable and aligned with the intended “T+P” design

      Text generation within AI models is a known limitation, and Firefly models (both internal and third-party) are no exception. Because of that, text is treated more like a visual element in generative AI, it can often appear distorted or inconsistent, especially in logos. The Assistant helps guide the process step by step, which tends to produce more structured and usable outcomes.
       

      It even suggested finalization steps, like refining the best version and then polishing it in a design tool, which is a great workflow for production-ready logos.

      If you haven’t already, I’d definitely recommend trying your prompt with the AI Assistant, it can help you iterate faster and get closer to a clean result.
       

      Let us know how it goes or if you’d like help refining the prompt further.

      Hajar

      Terms & ConditionsAccessibility statement
      Using the community Terms of use Privacy Cookie preferences Do not sell or share my personal information ad choicesAdChoices