Feature Requests

Reference image is very useful, but couldn't it be even more useful if we had some sliders or checkboxes to specify how much influence the reference image has on generations?   Color, for instance. Reference image greatly influences colors. What if I don't want that? What if I just want the ref image to influence shape/form/detail, and not determine coloration? How about an overall influence strength slider?

Option to control generated results that consist of just a single isolated object image, on either a solid color or even rendered on alpha transparency.   _________________   Despite being call Generative "Fill", it's not always desirable to get results with complete filled backgrounds that look like complete scenes.   Often you want just iterations of object designs or character designs and do not want a fully detailed background complicating the results. When I'm lucky I can get results with minimal or no background, but it's very tricky, basically random.   Futuristic vehicle design ideas, for instance. I tend to get complete scenes with fully colored/rendered city backgrounds, etc. My intent is to simply generate concepts for futuristic vehicles to work from, not fully illustrated or photographic scenes/panoramas. 

Favorites list for Generataive Fill effects.   _________________   Concerning Generative Fill, it's important which "Effects" you have selected, and so far I tend to keep using mostly the same few effects when fine-tuning text prompts to achieve the results I'm after.   It feels like a missing feature that there's no way to quickly select recently used effects (like how the Brushes panel keeps a list of recently used brushes settings) or to be able to "star" an effect so it's saved in a "favorites" list I can quickly access and apply. Instead, I must tediously navigate to each needed effect, even though I only ever use ~2 of the available effects (which I expect is common use case).   (Sidenote: it's rather awkward that if I click on a generation variation thumbnail, my selected effects are  arbitrarily overwritten. I understand why the ui works this way, but it's not always desired. Not sure how to best address this.)

Generative fill is a great tool for patching 360 photos, typically to remove the tripod at nadir.  But it currently can't be applied directly to an image in equirectangular format.  If I use the prompt 'equirectangular 360' it does seem to get the equirectangular perspective sort of right.  But it clearly does not understand how pixels are distorted as you approach the pole.  So the result has bogus detail in that area, and  lots of 'cat butt' distortion. If this coud be fixed, it would dramatically speed up patching 360 photos, compared to the method I use now: extract a rectilinear image with a 3rd party tool; patch that in Photoshop; use the external too to re-insert the patch  and re-render the equirectangular image.

In FF2, we could save the links after we generated an image. Having a link could be useful.

Upon opening a new file, you are presented with a bar that says "Image Import" and "Generate Image." When you click to generate an image, a very organize and visually appealing dialog box comes up with everything needed to  to generate (and looks very much like Adobe Firefly's options). After you generate an image, the original import / generate goes away and turns into the older style input field and a button next to it for reference. Even when adding a new layer, the bar doesnt switch back to the import / generate bar. I think the dialog box should pop up everytime and should also be per layer (In cases when I want to generate and review multiple images in one document)    

Hi, I noticed that when using the clone tool, brush smoothing is unavailable. I would like to see brush smoothing available when using the clone tool. A small amount of smoothing helps to keep your brush going where you need it to. Please consider this! Thanks!

I love this tool; however, I get occasional blotching in the resampled version if the brush head is any sort of proportionally large size. This seems like something that could be improved but I don't want to be picking nits either.    

For those of us who use similar prompts repeadedly, is it possible to have PS remember our last five or ten choices and give us a drop down menu in the prompt bar to re-choose those prompts?

It would be great to have a similar effect of krea or magnífic where you can just improve a base image with AI adding extra details to the images.

The new v3 AI model in the beta is worse for digital artists. It seems to only want to give me photorealistic results and when trying to prompt for art, it just gives me really bad unusable results.Please give us an option to switch between v2 and v3 models so this doesn't disrupt our current workflow.

I would like a "reference layer" feature in the desktop version of Photoshop. Fresco already has it and it works great, but I would also like this feature in the desktop version of Photoshop. Photoshop is my main illustration tool, which I mainly work with every day. Adobe fresco has a feature where you can change the drawing's line layer to Reference layer. After the change, you can drop colors onto the new layer, but the colors remain behind the line and on their own layer. You can work with lines on their own layer and colors on their own layer. Fresco - Reference layer - https://helpx.adobe.com/fresco/using/layers.html

This just occurred to me when responding to a post in THIS THREAD.   The search bar in the right click brush panel is a total no brainer with so many presets available, (700 just in Kyle's Mega Pack).  Something that confuses and annoys a lot of posters to the Photoshop forums is selecting a new preset, and finding themselves using a different tool.  Would it be useful if there were tool type filters, like at the top of the Layers panel?  That sounds like a lot of work, so how about being able to search by tool type in the search bar, and with similar flexability to the Type tool?  i.e.  br spat (brush spatter), bl ro (blur round), sm ha (smudge hard) etc.   If filters for each tool would be too much work, how about a single switch that only displays current tool presets?  

As a photographer and videographer of people I would love the idea of selecting a face and removing the acne in one click and the glare on glasses. It would speed up my workflow immensley and outsourcing as well. 

- Have the option for the item generated to be cutout by default e.g. If I have ask for a flying pig, have it cutout so I can place it on any background/ move it around, etc. - Have a 'flip' button for horizontal and verticle orientation so after the image is generated the user can quickly flip it (this wouldn't be need as much if the image was cutout like suggested in the first point.) - Have a slider to adjust the blur/focus of the generated image so it blends in more acurately upon creation (again, maybe not such an issue if image was cutout already.) - Have a feature where the user can bring in their own additional image and have it easily blend in with the existing image as the AI generation would have done e.g. If I want to place a hat on a gorilla, I can select my own hat image and the AI positions, blends, resizes, lighting, etc onto the gorilla as the AI may have done (I say 'may have done' as I'm sure I read Adobe claims these things get considered when it's AI places their g

What I need from photoshop with AI is a pattern removal tool to where you can use a lasso or brush and brush or lasso a sample area to remove throughout the image. For example, a fence. AI needs to be used for complex masks with hair or fur, especially. Even in areas with impossible masks due to hair against a low contrast background with similar colors. AI would be able to create in those areas based on the rest of the selection. AI generative fill still needs a lot of work to make it look more believable. AI Generative fill to where you have options to transform the current pic into an artistic finish, such a pop art, matte, splash art, etc...

With every number change in the settings panel the program initiates the process. I want to change several parameters without each change running the filter. An apply button would eliminate this problem!

Issue 1: Generating a fill as a transparent overlay. If I want magical sparkles overlayed, that's what I want. I don't want it to alter the background or blend it - just add it. Issue 2: Generating a fill as a transparent overlay with 3-dimensional object recognition so it doesn't alter any of the other layers and blends it with shadowing AND OR doesn't alter the other layers but blends shadowing from the add: Why, with all of Adobe's amazing software and power is this still not a thing? In order to render images and cast the shadowing correctly, the AI HAS to recognize the 3-dimensional placement of objects. So, if it can analyze the photo: house, person, front, back, etc., recognize light placement, shadows, then when I add a transparent object (a tree, rocks, magical sparkles, a light bulb, etc.), the AI should be able to render that into the image. This should extend to fonts in Illustrator, where I can take a core font and use the AI to break it up, add to it, etc. Issue 3: Adobe

The more I use the Contextual Task Bar the more I like it, but what would make it into a truly killer feature would be if we could edit it like we can with the Toolbar.  I would need to be savable with custom workspaces like we can with Toolbar and Shortcuts.     Give this a like if you'd like to see this?

When is Adobe going to build a version that allows presets and customizations to carry forward to new versions?  Since I started using PS in 1998 things have gotten better but many presets still don't carry forward during an update. For example, custom crops get deleted with every update. This is the folder that needs to carry forward.C:\Users\username\AppData\Roaming\Adobe\Adobe Photoshop (Beta)\Presets\ToolsGo one better and add custom crops for for all of the common social media image sizes.  .

 Photoshop Beta 25.1 I love the new generative ai feature and I use it alot. Sometimes I will generate 12 to 15 versions. As I sort thru the different versions to see which one(s) I like the best, I find it difficult to  remember which ones are the best. My idea is to have some rating system above each thumbnail. It could be like Bridge, maybe use 3 stars. I would LOVE to have this has a feature. Pam Hawkins

It would be realy nice if the text effects (like in Firefly) could be used in PS, any chance that it will be implemented?

I'm continually impressed with the innovations and improvements you bring to the platform. I would like to share an idea that I believe could further enhance the capabilities of the Generative Fill feature. I have found that there is room for improvement in terms of efficiency and creative flexibility. Specifically, I would like to propose the inclusion of a feature that enables the AI to remember and reproduce specific elements or objects in different scenarios. The concept is simple yet incredibly useful. It would allow users to designate certain elements, such as a main character or a related environment, and have the AI remember and replicate them in various situations. Here's how it could work: Selection and Designation: Users would be able to select an element within their canvas, whether it's a character, an object, or a background. They can then designate this as a "persistent element" by activating the feature. Scenario Variations: Once an element is design

Please add a feature to the generative fill to remember what it has done before, so it can create something similar/identical. for me, the use case is image sequences, either frame-by-frame patches, or keyframes for i.e. Mocha remove. giving the generative fill a memory function would help tremendously to keep the work inside image sequences more consistant.

Hi Team,   I use Photoshop actually to uncrop images to match their x or y resolution. I want to have quadratic images and for this use case is content aware and generative fill perfect. Unfortunately, there is no real function in Photoshop to automate the canvas extension for this task. There is a workaround, but it works sometime and sometime not.   If there were some kind of option to extend the canvas to x or y of image resolution, I would save a lot of time, due to the possibility of automation of the process.   There s a workaround, but this process takes time (select, clone image, rotate and so on). I think if it is coded it can be really fast - extend canvas to longest side of image or so. Or is there already such a feature and I didn't saw it?   Thank you!