Feature Requests

In order to improve the user experience, I have encountered a problem that I really don't know how to solve.When you generate a video with any model, whether it be Firefly, Luma, Pika, etc., it often happens that the videos are longer than 5 seconds, even if it is only 5.001 seconds longer. The Runway Aleph model does not accept them because they are longer than 5 seconds. This is annoying for the user because they have to download the video, go to a video editor, cut it, upload it back to the board, and repeat the whole process.I do not know if it is possible to leave videos at 6 seconds when using Runway Aleph, which would solve the problem.That's all! Thank you!

It would be amazing if the generated images previews could get way bigger. I am having a hard time checking if I want to keep one or not, because I cannot see any detail before adding it to my board. 

I need to colorize old sports photos that are black and white. The problem arises when I can only use a prompt to instruct the AI on where colors should be placed in the photo. The photos usually consist of two teams with different clothing kits to colorize. My suggestion is simple. Provide a color wheel and a pen stroke ability to show AI what parts of the photo should get what color.

Would be great to start having 3:2 as an aspect ratio option for image generations.

Hey Guys, it woild be lovely to add a zoom feature just like in illustratot or phtoshop. I keep pressing Z to get the zoom tool and zoom doing a marquee or just zooming in or out. 

By having the AI learn the style of the character you are drawing, you could streamline the process by having the AI draw a character that fits the situation. And since it learns your own drawing style, it could clear up copyright issues and be easily recognised as an AI for creators.

Hello. I am a trainer in visual communication and visual artist. I am a user of Midjourney and I have tested Dall e.I notice that no AI can manage correctly the composition in the workspace.I suggest you to work on this dimension of the design which in my opinion would give Firefly a significant advantage over other visual generative AI.It seems to me that integrating the golden ratio and or the Fibonacci sequence in the algorithm would be judicious. Also I suggest you to think about how to integrate the theory of the good shape (gestalt) as well as elements allowing artists to use tools of the theory of colors and contrasts.Or at least the rule of thirds.I dream of an AI that can recognize the periods of art history, the artists and their aesthetic and technical characteristics.I thank you and am at your disposal.Sincerely.

This layout for displaying a close-up of one generated image and the thumbnails of the quartet of generated images would be dramatically improved if the new display appeared over a solid background instead of appearing over the original screen. It is difficult to judge images with so much visual clutter surrounding them.

At the moment when generating people Firefly cuts off a portion of the face or body. There needs to be an option for fully body shots where the entire person is in frame and for portraits for the entire face including hair and everything to be in frame. Currently it wants to cut off the top 10% of a persons head/face or just gives 3/4s of a body. 

It is hoped that adding input text in a future version will generate generative model and bone actions

Hi, one of the hardest things to produce in text-to-image is the right type of view, My idea is to add orthographic views options (front, top, side) on the options, It would help a lot. 

Hi, for the reference image slider, it would be helpful to have a numerical indicator or % Then we can know more precisely how much we need to change the slider consistently across images to obtain more or less of the reference image.

Add an ability to remoove from my favorites

unable to edit the pictures, remove background, a mini version of the capabilites that adobe photoshop has would be great 

It would be very usefull to add a feature that allowed you to write what elements of a particular image you would like to change. For example you have a picture of a woman in a red dress created by Firefly and you want the same picture but you want the dress to be green.

I actually like the Firefly beta so far. What would be even better, is if there was a release date for the "In Exploration" stuff so that I know when it comes out. I am really looking forward to trying all the "In Exploration" features out. Thanks.

It would be nice a brush tool to be able to point the areas that you want to be regenerated by the AI

Hello, Dear Adobe Firefly team I am very glad to say that about Adobe Firefly it very nice text to image creation, so humble request nto you please add Aspect ration 9:16, and one more thing some time i given command but wrong image creation, so please improve,  ThanksSACHIN KUMAR SHARMAAdobe Firefly user 

Turn blurry, low-resolution images into crisp detailed images, using AI.  Such as Remini or stable defusion. I'm sure you have already been working on this. Just wanted you to know we want an Adobe version.

Im working with advertisers and would like to have a way to add their specific product to a generative AI image

I would love to be able have the option of importing product images, such as clothing, accesssories, etc. The A.I. would then generate a custom background based on specific conditions and realistically incorporates my image with this background.

I created really good textures with the AI and then was trying to get the AI to apply it to a suit. It would be interesting to have options on how the Reference image is used with your prompts; such as texture, background, subject, ect... 

Would be wonderful to allow users to upload a refference image to start from it

Based on an image, I would like to generate a website theme based on an image.For example, I have a videogame with a main character and I would like to generate a website layout / design based on the image of that character, extracting colors, art style, elements, etc. applying it to a cool modern website. 

an option where any text generated by text to image can be edited to say what you want