Feature Requests

Let's say you generate an AI background for an image and you want to build upon/layer upon that image- say add a subject or objects , it'd be nice if the newly generated subject/objects could match the perspective, lighting, camera angle, style etc. of the background image and if it was able to figure out depth of field etc.  Output wise, it'd be great to have a layered psd document.

If there is one truly "silver bullet" problem that gen AI should be solving -- and specifically Firefly should strive towards -- for creatives, especially those in archviz, interior design, environment concepting, etc. is to make Firefly brilliant at understanding how to populate an image, render, photo of a space and fill it with accurate crowd imagery, and provide controls on how dense to populate the scene.To accurately account for lighting, perspective, have controls for the demographs (general public? VIP event?) and how dense the scene should be filled, and intuitively know to put a few people in the front of the scene, a few in the back, to understand how to generate crowds to interact with the elements of the image, like seated guests on couch furniture, or ordering a drink at a bar element.

Hello, everyone,maybe I haven't found a way to get there, but as Firefly presents itself to me, I find myself on the page with all the generated drawings and underneath I have the prompt that forces me to type in what I need to generate, without being able to choose anything, i.e. format, quality, whether photo or drawing, thus making me waste a Generation from my monthly credit every time.If there isn't a way already, it would be fairer and more logical to have direct access in the first instance to the panel for choosing options for Generation.Thank you! --> https://prnt.sc/fMn4-Ro51ILp 

Warum können Bilder, die eine CC-Lizenz haben, nicht als Referenz bei Firefly genutzt werden? z.B. https://de.wikipedia.org/wiki/Die_Erschaffung_Adams#/media/Datei:Creaci%C3%B3n_de_Ad%C3%A1m.jpg

MDFANTASIR 

It would be great, and possibly would allow for more accurate "reading of text" if there was an option button to choose fonts from an adobe typeface. I know this is a tall ask requiring a deep look at the underlying code, but it would also likely solve the general AI issue of being poor at generating text. Basically, if the option button for text is checked, a subroutine would be inacted that ran text recognition and then replaced the recognized text with real text from a font database.....Having it run only when a user selects a radio button would prevent other elements from being recognized as text and help it to learn. just a thought. 🙂 

save to local pc the set up - all settings and images used as ref / function to import 'job' with all these setting and inputs files 

I really don't like that a credit gets used to see the larger upscaled image. It is extremley hard to see the details until it is "upscaled" and alot of times when I do waste a credit and view the upscaled image it looks strange and I dont end up dowloading it... then a credit is lost.... I also dont see a point of offering the smaller images to users either because we can just size the larger ones down. The small images are WAY too hard to see the details of the image. I also think if we are paying for the entire adobe suite, that we shouldn't have to use "credits" for using ai.... or at the very very minimum not until we download the image...I feel like its just another way to take advantage of loyal customers....

It would be nice if you were able to access favorites while IN firefly, without having to go back or have it saved as a bookmark.

Introducing new dimensions for the aspect ratio feature in Adobe Firefly would be a great improvement. Another suggestion would be addressing the issue where previous designs become inaccessible after a certain period in the history section

Experimentando con el relleno generativo desde photoshop v25.12 observo que cuando se dan instrucciones (prompt) hace cosas distintas a lo que uno literalmente necesita. Para poder manejar el resultado he tenido que optar por fraccionar la selección de las cosas que quiero eliminar o modificar hasta que en una de las versiones de resultado me pueda dar lo que necesito o algo aproximado.

Perhaps could you add a Slider? For us to to see it the whole history of the pictures

como guardar un recorte después de haber eliminado en fondo. Después de eliminar el fondo me quedo con una figura que puedo guardar para luego insertarla en otra foto. Me obliga a que genere un fondo. 

可以加入比較不同種類的動漫風格,例如二次元風,或是像美少女戰士或柯南風,不用太立體

When creating AI graphics in 672x384 resolution nothing is visible at least 1000x500. No details and we lose valuable tokens.

I've been waiting for feedback for a long time and Adobe doesn't respond or give a position on a simple download button of all the generated images and in the browser's favorites. I have an Adobe CC subscription but I don't want to save one image at a time. Please listen to your payers and subscribers. The price is not cheap!!! Developers do not respond or give fair feedback on this. I'm tired!

• A search option for gallery.To sort required category based showcase.• A History of generated images.Come back and grab another one from the past. Thank you

Foi uma péssima ideia a que tiveram hoje, dia 3/10. Perde-se mais tempo e custa mais créditos.

Thank you for the very useful Ai image generator. But it's annoying that all the styles are Chinese customs, Chinese images, and Chinese paintings.

- entourrer dans l'image de référence des zones pour indiquer a l'IA de quoi il s'agit, un peu à la facon photoshop.- avoir des zones dynamiques dans l'image générée (selection dynamique de photoshop) pour faire varier la sélection selon le prompt

How can I tell Firefly to explicitly NOT do something? How can I distinguish between a description of what I want to see and the description how I want it to be?When will there be simple and basic functionalities like "zoom out/ zoom in" "transparent PNG". When will it finally be useful for something and not just produce random stuff?!

Aadobeエンジニアへ合成用に頭から足までの画像をアップロードしました。プロンプトには全身立ち姿と書いた。ただし生成されるのは腹から膝までの画像。が、プロンプトに表示される画像は遠慮されている。見本画像がアスペクト比に応じて切り捨てられるのはおかしい。過去、フォーラムで手書きでも何でもいいから参照用画像をアップする様にアドバイスを頂きましたが、それがこのように切り切れるのでは意味が無い。

I raised this issue in a Roundtable Discussion but I thought I would include it here for posterity. (Also to see if we could get any upvotes.)It would be great if Firefly had tight(er) integration with Photoshop on the Web. I realize there are options to Copy and Save to library in the Share Menu. However, I think it would be ideal to have an "Open in Photoshop" option. When clicked it would open a new tab and push the Firefly image into the Ps Web GUI (creating an appropriately sized canvas). Please see attached mockup.  I know there were some concerns about retaining the generative capabilities when moving the image to a new surface. However, I'm not sure that is as important. Just being able to move the image into Photoshop in one click would be useful. Then it could be conveniently generated, opened in Ps, edited, saved and shared for feedback with ease. Keeping this workflow within the Adobe Web ecosystem seems more convenient and retains Content Credentials along

I think i could be fun for a lot of families with small children to have the ability to transform familyphotos og holidayphotos etc into colorbook-pages. 

Would love to input my images and have Firefly generate in the same style. For instance, if I created a cartoon character named Freddy Fish, I would like to input current images of Freddy Fish and have Firefly generate based on those so that the output images all match the character.