[text to video] Seed doesn't establish Storyboard Consistency

[themewesetech]

I imagine the goal isn't to produce 5 second videos. If the goal is to use firefly video to create shorts, commercials, educational videos, or full-length videos, we need a storyboard function. Currently I am not sure what the "seed" does as it doesn't provide and generation consistency.

I divided the complete story into 5 second prompts buAnd taking a screenshot of the character generated in the first video and an image upload to control the output generates an error. I would love to create a complete scene and a storyboard that locks in the main character/s output would be amazing. For now I am doing it old school. 

[themewesetech]

Yes please! I can't imagine the ultimate goal is to generate only 5 second clips. A storyboard is so necessary. And it would be easy to add. The space already exist with the thumbnail video history. Just make that a storyboard. Make the seed actually mean something by locking in the main character. And if a portrait image is uploaded as a guide, autogenerate the body, using the image as a guide not the entire video. Thank you for requesting this. Let's keep our fingers crossed.

[droopydog500]

Hello @themewesetech,

Thank you for your message. I am sorry you are having this problem. 

Using the last frame of a clip to start the next five second clip is probably the best consistency technique. Should also use the same seed.  If you are getting an error with image to video, please share a screenshot of the error, the reference image, and your prompt.

Thanks,
    droopy

[droopydog500]

[Moderator note: This question is about text to video rather than text to image, as this original thread is, so I am moving it.

    droopy]