Midjourney v6 News: Turbo and New Describe for v6

Поділитися
Вставка
  • Опубліковано 30 вер 2024
  • In this video, I discuss the latest updates to Midjourney, including the new turbo mode for faster image generation and the revamped "describe" feature. I share my thoughts on using the new describe tool, compare it to the old version, and explore its strengths and limitations in generating useful prompts from images. I also provide a quick demo of how to use these new features.
    CORRECTION: It seems there is actually no way to use the old describe anymore.
    Music generated with Mubert: mubert.com/ren...
    All generated images created by Thaeyne with Midjourney: www.midjourney...
    Animation created with Kaiber: kaiber.ai?referrer_id=081b4b5f-1c61-423c-8e88-21c7dd445fbb
    Voiceover generated with Elevenlabs: elevenlabs.io/

КОМЕНТАРІ • 14

  • @gerhardbrunthaler4311
    @gerhardbrunthaler4311 6 місяців тому +3

    In general I do not feel very comfortable with V6, most time it does not what I expected, but it worked well with V5.2. To my opinion Version V5.2 was much more predictable.

  • @wd1n10
    @wd1n10 6 місяців тому

    🎯 Key Takeaways for quick navigation:
    00:13 *🚀 Turbo mode for Mid Journey V6 offers 3.5x speed at double the cost of fast hours, activated with "/turbo" or "SL settings".*
    01:21 *📸 New version of "describe" allows uploading images for descriptive sentences, with slight differences from the Discord version.*
    02:46 *😬 New "describe" version might offer harsh critiques on personal photos, following V6's conversational prompting style.*
    03:42 *🎨 The new "describe" recommends artist names less frequently but still suggests them based on stylistic similarities to famous artists' work.*
    04:38 *💡 Some users find the old version of "describe" more helpful for discovering impactful keywords, while the new version has longer prompts and less precise keyword identification.*
    Made with HARPA AI

  • @noahlloyd2034
    @noahlloyd2034 6 місяців тому +1

    Mixing this new describe feature with style reference gets me some pretty neat results.

  • @decoder33
    @decoder33 6 місяців тому +2

    they did you dirty with those selfie descriptions smh

    • @thaeyne
      @thaeyne  6 місяців тому

      Well I guess it's working like intended :D
      If it didn't do that, then it wouldn't be able to accurately translate it to the prompt language Midjourney needs.
      All the people that Midjourney generates are just so beautiful that an ordinary person obviously does not normally look like that. Hence it's describing a normal person more harshly so Midjourney would dial down with the beauty.

  • @erikaol1091
    @erikaol1091 6 місяців тому +1

    I changed to version 5 but I'm still getting the new describe only. How did you get to use the old one?

    • @thaeyne
      @thaeyne  6 місяців тому +1

      Oh dear, it seems that I must have been hallucinating or something. I could have sworn that there was a way to use the old describe still. But now when I go back, there is no mention of it. I don't know what happened there.

    • @erikaol1091
      @erikaol1091 6 місяців тому

      I hope they make it available somehow. Version 5 doesn't work very well with the prompts the new describe is generating.

  • @aipicsarts
    @aipicsarts 6 місяців тому

    Thank you, very helpful as always! ❤

  • @mounir792
    @mounir792 6 місяців тому

    Please I want 3D styles caricature

  • @JoshuaCatchur
    @JoshuaCatchur 6 місяців тому

    Does /shorten still work? I haven't used it for a long time, so I'm not sure. That might help with the wall of words prompts.

    • @thaeyne
      @thaeyne  6 місяців тому +1

      Yes, that I tested out just yesterday that it still works. It does significantly cut down on the amount of words that describe suggests.

  • @boerboelmummy5075
    @boerboelmummy5075 6 місяців тому

    Thank you