Create Consistent Ai Character Videos using a Single Image! - Hailuo Update - Subject Reference

Поделиться
HTML-код
  • Опубликовано: 14 янв 2025

Комментарии • 35

  • @jessykazann9742
    @jessykazann9742 13 часов назад

    Oh ! I didn't know I could do that with Hailuo ! I get used to make videos with a pic, but I didn't know we could make consistent characters !!!!! Thanks a lot ! ❤

    • @taoprompts
      @taoprompts  9 часов назад

      It's a new feature they just added that in!

  • @FilmSpook
    @FilmSpook День назад +1

    Thanks Brother! Happy New Year to you and your family. It's a dope update. I'm really hoping they soon add generations that are 10 seconds long, which they announced back in October. That would really help filmmakers like me out, since 6-second clips is barely enough to edit together for narrative filmmaking.

    • @robertmacku3059
      @robertmacku3059 17 часов назад

      For me as narrative filmmaker is more important to have possibility to get image on the end of the clip, same like Runway and Kling has. It is must have for me...

    • @taoprompts
      @taoprompts  9 часов назад

      Happy New Year man! I'm not sure why they haven't added in longer videos yet, it's one of the most requested features

  • @ronnyalves7
    @ronnyalves7 14 часов назад

    Great explanation. Following from Brazil.

    • @taoprompts
      @taoprompts  8 часов назад

      Thanks! Awesome to hear from Brazil 👍

  • @Otchengazoom
    @Otchengazoom 4 часа назад

    1:56 Are these videos created in Hailuo? The output quality is pretty cool.

  • @ReveRecordLabels
    @ReveRecordLabels 20 часов назад

    wow ❤❤❤❤❤

  • @pdjay8912
    @pdjay8912 22 часа назад

    저두 써봤는데 장단점이 있는 것 같아요. 그래도 학습할 필요 없이 곧바로 영상을 만들 수 있는 장점은 진짜 좋은 것 같아요. 좋은 영상 감사합니다:)

    • @taoprompts
      @taoprompts  8 часов назад +1

      Yeah, it's easier than ever now! It's really fast to create character videos using this type of feature

  • @PTHZEROAI-oz8ns
    @PTHZEROAI-oz8ns 19 часов назад

    Thank you ! So ... Do you know any AI that can make multiple characters consistent in one frame?

    • @taoprompts
      @taoprompts  9 часов назад

      you need to use masking, it's a bit more complicated. I have a tutorial here: ruclips.net/video/h-7SIFL6gP4/видео.html

  • @apongaming5132
    @apongaming5132 3 часа назад

    I want to know that does our information safe and secure in this website?

  • @Sean43322
    @Sean43322 4 часа назад

    "Thanks for the great tutorial! I'm still looking forward to one that shows us how to create an AI character holding one of our products, like a chocolate bar. Is consistent character design with a specific product possible?"

  • @TheIllusionRecords
    @TheIllusionRecords День назад

    any strategies for creating two consistent characters in the same image

    • @taoprompts
      @taoprompts  9 часов назад

      I have a guide here, it takes some more work ruclips.net/video/h-7SIFL6gP4/видео.html

  • @UNDEADARTE
    @UNDEADARTE День назад

    Hi man, In my last video I tried to make a consistent character - SISO😀after your videos about this, it seems to have worked a little.😊

    • @UNDEADARTE
      @UNDEADARTE День назад

      I used Midjourney, Kling 1.6 with lip sync and for shooting scenes VIDU

    • @taoprompts
      @taoprompts  9 часов назад

      You're video looks awesome! You added in so many different characters 👍

  • @KNWENT
    @KNWENT 14 часов назад

    I dont have S2V option. Only I2V-01 and I2V-live
    why?

    • @taoprompts
      @taoprompts  9 часов назад

      I think they just changed the name to "subject reference"

  • @MichaelBaynana
    @MichaelBaynana 11 часов назад

    doesnt work for me with a samurai mask character : (

    • @taoprompts
      @taoprompts  9 часов назад +1

      I think it needs visible faces to work

  • @ivan_12345
    @ivan_12345 22 часа назад +1

    LOL, just tested and got completely another person😄

    • @PTHZEROAI-oz8ns
      @PTHZEROAI-oz8ns 19 часов назад

      really ?

    • @ivan_12345
      @ivan_12345 18 часов назад

      @@PTHZEROAI-oz8ns yep, and what's your result?

    • @taoprompts
      @taoprompts  9 часов назад

      I've found it works pretty well for photorealistic images of people, the more stylized the more problems it will have

  • @MabelYolanda-c9i
    @MabelYolanda-c9i 19 часов назад +2

    I will also advice to use always the same seed for clothing consistency. However, this new model is really bad with hairstyles, specially long hair or anything that goes out of the cropping zone. To me this is a step backwards rather than solving a problem.

    • @taoprompts
      @taoprompts  9 часов назад

      I didn't know you could set the seed in hailuo? How are you doing that

  • @ivan_12345
    @ivan_12345 День назад

    image to video will give more predictable result, IMHO

    • @taoprompts
      @taoprompts  9 часов назад

      yes, image to video is still the best method

  • @dansv3872
    @dansv3872 22 часа назад

    Great Update ! I'm still waiting for the era of object consistency 🥲🥲🥲

    • @taoprompts
      @taoprompts  9 часов назад

      they'll probably add that in if there's more demand, I think people mostly want to create characters atm