Skip to main content
smiff1000
Participant
April 21, 2026
Open for Voting

reference image/video option for subject and background for image to video

  • April 21, 2026
  • 1 reply
  • 22 views

When using an image or first frame/last frame image reference for image to video, it would be very useful to add an option to upload a subject reference.

So in practice I have an environment image reference, and I want to add a specific effect to the space, or a specific character. or even a consistent effect/character for use in other generations. Currently its trial and error each time to try and find the correct prompt. Would additional references be possible? For instance a picture of the kind of character I want, or a video of the kind of effect?

Specific example: I rendered a 3d start and end frame image of a room interior. I wanted to animate the trees outside the window blowing in the breeze, or the sunlight casting shadows on the floor, or a person walking through the room.

It took a lot of trial and error to get the trees looking correct, but eventually managed it. I gave up on the sunlight effect since I was burning through credits, and I managed to get the character walking through the room, but on the next generation it was a different character even with the same prompt. I think video references for the effect might have helped, and image references for the character might give some consistency.

Additionally, I wonder if theres a way to specify which area of the image to edit in the video, by a grid ref over the image for example. Whilst attempting to add the shadows and moving trees I lost count of the amount of times I had tree branches and leaves flying through the room nowhere near the windows.

 

    1 reply

    Kartika Rawat
    Community Manager
    Community Manager
    April 21, 2026

    Hi smiff1000!

     

    Thanks for sharing this detailed feedback. I understand how frustrating the trial-and-error process can be, especially when it comes to credit usage.

    At the moment, Firefly’s image-to-video workflow supports only a single image reference, and additional inputs like character or effect references aren’t yet available. This can make consistency and precise control more challenging.

    Your suggestions, such as subject references, effect/style inputs, and region-specific controls, are very valuable and align with what we’re actively exploring for future improvements.

    In the meantime, using clear, consistent prompts and testing effects individually before combining them may help improve results. We’ll be sure to share your feedback with the product team.

     

    Let us know if you have any questions.

    Thanks,

    Kartika