Copy link to clipboard
Copied
I am running current vesion of Firefly. I am working with TEXT to IMAGE application. I realize that when I insert text and select styles, effects, themes, etc. , I am basically providing "seeds" and I can regenerate until I get something I like. I then can add and subtract to the text and the generated image on regenerate basically has the same look" or "style." At least that is what I experience.
I then change the text to add , say, a log cabin in the woodland scene. The regenerated scene has the same look and feel with changes and , of course, some cabin options to choose from. I then cycle thru the various scenarios (text) for the illustrations for my story all in one sitting.
But, later I wish to further manipulate the images by adding/modifying the text. UhOh! Once I leave that session , I can no longer get back to the look and feel of the previous session images. If I record my "setting" of style, etc. and paste in the same text, I will get a whole new look. (of course, I will) Is there any way to force Firefly to return to the look I like when starting a brand new session?
Copy link to clipboard
Copied
@mckinney3 So, if I understand your situation correctly, you've generated a few images that match the look and feel you're looking for. You then close your browser and come back later and would like to be able to keep the look and feel from before. Even with different prompts but the same styles selected?
If so, try downloading one of the images you generated and liked then upload them as a refrence image. Here is some more inforamtion this new features: https://helpx.adobe.com/firefly/using/whats-new/2024.html#style-match
Copy link to clipboard
Copied
Thank you for the reply. I did as you recommended and obtained reasonble results. I found that I was better off . though. just reinserting the original text with no styles or filters and then after the first render, applying the group of filters that I used to create images that I liked after the first Firefly rendering. Once I got a panel of 4 options that I liked, I started modifying the text by adding one thought at a time along the timeline of the story I am creating and was rewarded with similar illustrations but with the modifications I had added included.
I also found that I was better off adding text to change the point of view instead of picking, for example, a new point of view from the styles options. I prepared. Kinda like a story board progresion so I entered the session with a series of story modifications and kept making appropriate illustration choices for saving. after letting the story line progress in the text entry. Again, I did not make a new one line statement but kept adding new text to the story line.
For me this is an ideal work flow for creating the images to accompany my short story books. I then took the entry text used for Firefly and used that to create a condensed story line which I then entered into ChatGPT along with some requests on story length, writing style, Gunning Fog index for age appropriateness , etc. The ChatGPT story has developments I never thought off which enrich my story board tale. 😃
An example of this story board text and the resulting images. There are gaps in the story board as I am only attaching 3 examples in the progression.
Text 1: A scene in the fairy forest which is the secret habitat of fairies and small woodland creatures at night. The trees are pine. A glimpse of the moon can be seen. The view is from very high upon a mountain trail looking down into the forest in the valley below. A small light can be seen very far below. A winged fairy appears and beckons the travelers to follow the trail.
Text 2: A scene in the fairy forest which is the secret habitat of fairies and small woodland creatures at night. The trees are pine. A glimpse of the moon can be seen. The view is from very high upon a mountain trail looking down into the forest in the valley below. A small light can be seen very far below. The weather changes and wind blows hard and snow turns into a blizzard covering the ground. The winged fairy can no longer fly and asks a long legged woodland creature to be the guide through the deep snow.
Text 3: A scene in the fairy forest which is the secret habitat of fairies and small woodland creatures at night. The trees are pine. A glimpse of the moon can be seen. The view is from very high upon a mountain trail looking down into the forest in the valley below. A small light can be seen very far below. The weather changes and wind blows hard and snow turns into a blizzard covering the ground. The winged fairy can no longer fly and asks a long legged woodland creature to be the guide through the deep snow. They finally get to the valley floor and the snow is deep. A log cabin is seen ahead offering shelter from the storm. They enter and a guardian fairy watches over them as they rest.
I then go to work on editing the verboseness and other weird writing artifacts that come with AI creations.
Voila, I have a small illustrated story for binding into small limited editon, hard bound , hand made books. 😃
What a wonderful set of tools for use by a creative person who is hampered with poor painting skills and limited writing skills. I still feel full ownership of the creation and the creative process.
Copy link to clipboard
Copied
@mckinney3 you might be interested in this idea then: https://community.adobe.com/t5/adobe-firefly-ideas/p-maintaining-a-scene-or-character-across-differe...
Copy link to clipboard
Copied
Again, thank you! But the video "teaser" as well as comments in the thread do not get at solving the issue of "locking -in" a certain look as the story board progesses with different text prompts. Each re-generation or refresh just produces another "brand -new" rendering which may or not be compatible with the previous rendering.
I LOVE Firefly and have quickly outstripped , maybe, its intended purpose. But I am thirsty for more. Specifically , the abiity to maintain continuity in creating a story board or an illustrated book.