Feature Requests

I have used ChatGPT, Bard, Bing, and Pi to discuss writing ideas and they have all proved helpful. What would make Firefly a world better than any other image generation AI, is if you could integrate what you are doing with a language model like the ones I mentioned. Then we could kick around ideas, rather than simply directing Firefly. We could ask questions, like "What colors would go well with this shade of red in the dress?" or "What style of painting might work better for this subject?" and the AI could actually responde based on a knowledge of Art History. Potentially, it could even go further, making suggestions and asking questions of it's own, like "Are you wanting to invoke a clearly sad emotion or something more ambigulous?" or "Those two colors clash and may prove distracting for many viewers."

Wäre sehr praktisch, wenn es auch eine Eingabe für Negative Promts gäbe.

Text Better text support Add own text to picture and use ai to wrap it.    Picture Recolour sections of a picture easily (yellow cup rather than white) Better support for everyday objects like a blimp, plane, stationaries... Better logo/symbols (fake is ok) Better particle effect like steam or bubbles.  Ability to have half a person, for example in the water or in sinking sand.  Ability to make a comic strip

Would be amazing especially for the text effect if we had the ability to not only save it as a .jpg but as a vector file as well so we can use for big print files!

suppose I create a flower drawing, and I want exactly the same style of drawing but I just want to change the subject, from a rose to a daisy. it would be nice to have a button that can retain the exact same line and amount of detail so that the drawing looks like it is from the same artist/family

This ai Photoshop editing on my phone is powerful but what if we had our brushes we could use to mask out specific shapes??:). I think that would be a great idea. Plus the system crashes super easily:(

text to image has a lot of things you can do. Model version, Aspect ratio, content type, visual intensity, match, etc. all of these options are replaced when you edit an image. Once we are left with an image we can alter we are left with add, subtract, settings, background, invert. it would be amazing if we could import our photos and tweak them with the initial things we could use like, aspect ratio, content type, visual intensity, match, etc.  if we could import say a picture of a log cabin door and lower the quality of the image with visual intensity and then rework the image with text to image to alter it to give it a certain look that would be Gold. You wouldn't need other ai generators because this would do it all:)

Improve the accuracy of the image generation. Adobe Firefly is still under development, and it can sometimes generate images that are inaccurate or unrealistic. One way to improve the accuracy of the image generation would be to train the AI on a larger dataset of images. This would help the AI to learn the different features of images and to generate more realistic images.Add more image styles. Adobe Firefly currently offers a limited number of image styles. Adding more image styles would give users more options for creating images. This would also make Adobe Firefly more versatile and useful for a wider range of creative projects.Make it easier to use. Adobe Firefly can be a bit difficult to use, especially for users who are not familiar with AI image generation. One way to make Adobe Firefly easier to use would be to provide more detailed instructions and tutorials. This would help users to understand how to use the app and to create the images they want.Add more feat

I would find the usefulness of the Generative Tools far more useful if it could work with match color. Select layer then layer to match and it use Ai to change lighting and color to match target. i can achieve something like this already by individually selecting clothing, skin, eyes, etc and generating each piece to something similar. If the algorithm could basically do the same thing at once it would make photoshop retouching, the main reason for photoshop even better 

I am using Stable Diffusion right now, which allows me to create generative art with my own archive of photographs and illustrations, processed completely on my macbook. It is a brilliant way to re-think my own creativity, from the past into the future. Will Adobe offer an option for users to do the same with their Firefly model sometime in the future?

I am a graphic designer and for a long time I have been faced with the difficulty of drawing wheels in perspective, building buildings in perspective and there are very few lines of direction at an angle, as in an illustrator. Now we have to draw it in separate layers and make it transparent. It would be very cool if these lines could be twisted, build a horizon point and they would be magnetized during drawing.      

is there any APIs availabe of Adobe Firefly to integrate into our websites ?

Because AI places great importance on learning. However, learning begins with accuracy. So, fundamentally, it should include an additional feature to redraw a copy and have accuracy calculated in percentage from 1% to 100%. Additionally, there will be mechanisms to ensure the integration of various artistic styles and techniques of artists. Lastly, there should be a section for training AI for each individual to self-train their AI in their desired manner.

Y'all should use DALLE-3 like Bing AI bc a lot of the generative fill looks like AI from the original ai generator thing

for video  editing bettter to add some features kike removing bag and add whatever the user want to change that ,as like green screen graphics changing .such like that better impcove this kind of tool 

As a disabled person myself, I'm very desapointed of how your AI don't understand our reality. I've tried over and over to get a motorized wheelchair in my image in English and in French (my native language) and all I get is the fu***ing blue hospital wheelchair that absolutely no disabled people never NEVER use. There's a whole variety of wheelchair, I mean literally hundreds of different wheelchairs. Why it's always the same that prompts? You should really add images of different wheelchairs in your database cause it's chocking that we can't represent our reality! I just wanted to I fun in a reality that is my life. For you information my original prompt was "motorized wheelchair chair with reactor on a car racing circuit" in French then I tried many synonyms and a switch to English and I always got the hospital wheelchair with a reactor IN it... thank you for your understanding.

Here's some suggestion for featuresAllow to clean up the texture of the ai generated image, allow selectively change the quality of the image, like smoothing out lines, correcting glitches.  Allow for increasing the image resolution 

Wechsel von Eingangsbild zum Ergebnis und zurück.Protokoll wie in Photoshop.

Would be nice to do some seconds of special effects on images

Generative Neufärbung mit den eigenen Farbbibliotheken 

I need the ability in Firefly to 1) go back to the previous screen, and 2) recover a history of the image generations and prompts I have worked on. This is extremely frustrating as I try to learn the tool. I click on one button that doesn't have any tool tips or directions, and then I end up on a screen that doesn't allow me to go back to the previous screen. I have lost a lot of work and don't see why I would keep using this tool if it's so easy to lose all of that work. If there is a way to see my history or go back, it should be more prominent and easy to find. You also need a help button with context for the current screen. The UI is way too minimal.

it would be very interesting to be able to propose an image of a person (not famous) and ask him to represent it with a pictorial style, perhaps suggesting the style with another image.It can't be done now right?

Having AI understand how to create Cubic Environment Maps, Reflection Maps and Textures for 3D application would be great. Textures seem to have moderate sucess already, but it would be good to have a way to define scale of the texture. I think "Seamless Tile" already works, but if not, that is helpful too.

Currently, defining a subject in a scene randomly places the subject so there is no way to control placement and composition.   Please add a control for Distance to Subject so we can retain composition while changing lens settings. This is important for authentically simulating camera effects.   Please add a Placement control as we can consistently place the subject within the frame.

Automatic Background Removal: Remove the background of an image and generate a transparent PNG.Smart Cropping: Use AI to automatically crop images based on the content within them, such as faces or objects.Style Transfer: Apply the style of one image to another with a text prompt.Image Captioning: Automatically generate captions for images based on their content.Image Recognition: Identify objects within images and provide information about them with a simple text prompt.Color Correction: Use AI to automatically adjust the colors of images to achieve a desired look.Interactive Image Editing: Allow users to interactively edit images using natural language commands.Animation Creation: Generate animated graphics and videos from a detailed text description, including the ability to customize the animation style, length, and content.Custom Font Creation: Generate custom fonts based on a user's handwriting or a detailed text prompt.Photo Enhancement: Use AI to enhance the quality of photos,