These 4 Tricks Will Make Midjourney Listen to You

Поделиться
HTML-код
  • Опубликовано: 11 янв 2025

Комментарии • 85

  • @jackcotner8981
    @jackcotner8981 10 месяцев назад +16

    Another presentation with clear, concise, and valuable information. Thank you!

  • @crashdummyglory
    @crashdummyglory 10 месяцев назад +4

    After you use an image as a image prompt you will need to hit a vary command ( with remix on )as this lets mj do a better and more coherent blend of prompt and image prompt. For mysterious reasons image prompts work best after you hit vary ( with remix on )

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      I'll have to go do some experiments!

  • @moosekeating
    @moosekeating 7 месяцев назад +1

    Absolutely amazing. I'm relatively new to Midjourney and your videos are immensely helpful. Cheers!

  • @theultimateartist4153
    @theultimateartist4153 10 месяцев назад +6

    Cubism is also an African Artstyle, Batman the animated series takes allot from it because it is Jazz inspired

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +2

      That's a really good point and I can see that inspiration now that you mention it. The entire opening credits really fits that description lol love that show

  • @AG_before
    @AG_before 10 месяцев назад +1

    I think it's great that you're moving ahead with your videos but not skipping some of the basics new users might need to know. Excellent approach. It was only when I watched this that it crossed my mind - What if there was a way to weight 'style' vs 'style raw'. That would be awesome! Thanks for the video. 👍

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      haha I think there would be soo many cool possibilities if the Midjourney team wanted the bot to be more complex - but it seems like they're drawn to a simplified version of AI

    • @AG_before
      @AG_before 10 месяцев назад +1

      @@FutureTechPilot It's smart that they are I think. One way they could make the parameters even easier is keep the range of values the same for every one. Someone mentioned that in discord and I think it's simple and brilliant. I had completely forgotten the range for sref. So thank you for that. 👍 MJ is pretty much the nicest looking slot machine there is! You win most of the time, and when you win big, you win really big! Lol

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      @@AG_before lol it's soo weird that chaos goes from 0-100 and most other things go from 0 - 1000 but then weird goes from 0 - 3000 ... I've heard that changing the ranges isn't as simple as it sounds, but maybe they'll get around to it one day!

  • @wystemd
    @wystemd 10 месяцев назад +2

    thank you, cool information!

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      And thank you for taking the time to comment!

  • @LASartor
    @LASartor 10 месяцев назад +1

    So much to learn but you show us the way, thank you!

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      And thank you for taking the time to comment!

  • @downrock.creative
    @downrock.creative 10 месяцев назад +1

    Great ideas I hadn't heard of before. Thanks!

  • @MrB831
    @MrB831 10 месяцев назад +1

    Great tutorial, I really learnt a lot especially with the image upload. Peace!

  • @circelink
    @circelink 10 месяцев назад +3

    Nice guidance, thanks!

  • @antidiangelita
    @antidiangelita 10 месяцев назад +1

    Gracias Nolan, siempre aprendo algo nuevo contigo

  • @johnfretz1938
    @johnfretz1938 10 месяцев назад +1

    Super helpful. Thanks.

  • @paullenoue8173
    @paullenoue8173 10 месяцев назад +1

    He's a time traveller! I have proof! Ten minutes ago I took his survey and asked two questions and here he is now answering those questions eight hours in the past. How does he do it!

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      😂 that's hilarious

    • @miss69smart
      @miss69smart 9 месяцев назад

      I knew he was smart and talented but this is mind blowing!!!!!lol🤣

  • @GreenbellyStudios
    @GreenbellyStudios 10 месяцев назад +1

    Thank you for this, it is so helpful! For some reason I can’t accurately make the “dash dash” on my iPad Pro and have resorted to copying and pasting it in. Any suggestions?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      I think you can change that in your keyboard settings ... there is some setting that is changing your -- to an 'em dash' ...
      or maybe try hitting 'backspace' as soon as the dashes change ... that might work
      sorry I'm not more help!

  • @lunazamoracreative
    @lunazamoracreative 10 месяцев назад +1

    In MJ a web you can just drag the image up. I can’t figure out how to use -iw. When I include 2 images. Thankyou. You are soooo helpful

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      I don't think the website is optimized for that yet. That's one of the reasons I choose to generate on Discord still ... but please let me know if you figure out how to do it!

    • @lunazamoracreative
      @lunazamoracreative 10 месяцев назад

      @@FutureTechPilot ok.

  • @Deaflephant
    @Deaflephant 10 месяцев назад +1

    what times would you want to AVOID using --style raw?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      When you write shorter prompts. As a general rule, shorter prompts are better in the default

  • @cg2135
    @cg2135 9 месяцев назад +1

    Awesome videos 👊🏼👊🏼👊🏼

  • @TENKIDGAMESLLC
    @TENKIDGAMESLLC 10 месяцев назад +1

    Thank you

  • @lunazamoracreative
    @lunazamoracreative 10 месяцев назад +1

    How to do the -iw on mj alpha? With 3 images ?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      I actually don't know! Please let me know if you figure it out

  • @kieren7763
    @kieren7763 10 месяцев назад +1

    Im a custom car builder and ive been trying out different car designs the problem im having is i will type in a specific car and if its not common it wont be the right car any ideas ?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      Image prompting, but if Midjourney doesn't know the car, it's going to have trouble creating it

  • @metrosoft
    @metrosoft 10 месяцев назад +1

    MidJourney 6 does not listen to me with the camera angles. For example: "/imagine a warrior from a low angle shot" dont work in 6.0, but in 5.2 and previous works fine. you have not had this happen to you? (sorry for my bad english).

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      Camera angles are soo frustrating in Version 6. I probably should have mentioned that in this video... all I can say is that these tips aren't going to guarantee success, but they're exactly what I would do if I was having trouble

  • @MegaStoneyMahoney
    @MegaStoneyMahoney 10 месяцев назад +1

    Love your videos. They have been helping me out a lot. I've run into an issue. I've been trying various different ways to blend characters into kind of Amalgam-Versions. But it doesn't work most of the time, any tips?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      You could straight up try /blend? Not saying it will be perfect - and maybe you've already tried it - but that's what I would do

    • @MegaStoneyMahoney
      @MegaStoneyMahoney 10 месяцев назад +1

      @TechPilot The results of /blend are not at all satifactory. But I think I maybe found a way, to get good results. First the text prompt, then the image prompts (one for each character), then I use an image as a style-reference and finally --style raw. At least up until now, that has given me the best results. Still working on it though.
      EDIT: I forgot to mention, I give the first two images a IW of .5.

  • @NizamulMulk.Hassan
    @NizamulMulk.Hassan 10 месяцев назад +1

    Actually Style Raw give a realistic and real life look to the image. Am I right?

    • @NizamulMulk.Hassan
      @NizamulMulk.Hassan 10 месяцев назад

      But it doesn't work as good in niji

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      Yah it does for sure, but that's a result of it being less opinionated and listening to your words more - so when you don't pick an art style, it definitely defaults to realistic!

  • @dianimartinez1477
    @dianimartinez1477 8 месяцев назад +1

    I can't create full body characters, it always cuts them off somewhere, any tips?

    • @FutureTechPilot
      @FutureTechPilot  8 месяцев назад

      Vertical Aspect Ratio --ar 2:3 / and mention their shoes!

  • @gtSpleen
    @gtSpleen 10 месяцев назад +1

    One problem that is really frustrating me. I am trying to create a rabbit with three ears. Midjourney and also DALE are completely ignoring any try implementing this idea. Even if I use an example... any ideas or is this a hard border for AI at the moment?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      I would definitely try image prompting with a reference, but maybe you can also try in-painting? Vary (Region) underneath an upscale, and then maybe select where you want the third ear to go? that could work

    • @gtSpleen
      @gtSpleen 10 месяцев назад +1

      @@FutureTechPilot Thanks a lot, I will try our ideas!

    • @gtSpleen
      @gtSpleen 10 месяцев назад +1

      @@FutureTechPilot tried the whole evening, it seems as if Midjourney (and DALE) are told not to show stuff as a third ear. Possibly to be more realistic in most cases. Even giving the system pictures of rabbits with three ears, it just corrects them back to two (saying it has three)). Four was possible again.
      Also defining an area and telling to add another ear did not work.
      If anyone finds out how to get this going (breaking barriers seem to be hard implemented) I would be very glad to know!
      Anyway thank you a lot for your great videos! I am just starting with KI as a normal person, not really having a connection to my daily work, but understanding, that THIS will impact our whole society.

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      @@gtSpleen Very well said! Hopefully some of these frustrations get ironed out in future releases

    • @gtSpleen
      @gtSpleen 10 месяцев назад +1

      @@FutureTechPilot Made it, create the picture first, then "Vary (Region)" for half the head. And in some cases (10-25%) the system does not detect, that there are already two ears and adds a third one. Telling the system still does not work, but like a genetic defect some random changes did it.

  • @KingZero69
    @KingZero69 10 месяцев назад +2

    i saw that Mac Finder... i see you joined team Apple 😂

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      👀 it was either pay a giant amount of money for a new PC or pay a lot of money for an amazing laptop haha I don't know if I'll ever get an Iphone though

    • @KingZero69
      @KingZero69 10 месяцев назад +1

      @@FutureTechPilot respect, i use a mac too, it’s great for making videos 💪

  • @ArielBrito82
    @ArielBrito82 10 месяцев назад +1

    Great!

  • @lunazamoracreative
    @lunazamoracreative 9 месяцев назад +1

    I wonder how it knows what beauty is

  • @eurossocial
    @eurossocial 9 месяцев назад +1

    great tutorial. Thank you for these Great tips. 👍 Personally i never use the "Niji", because i hate manga cartoons. 😊 I don't like that style, that's all.

    • @FutureTechPilot
      @FutureTechPilot  9 месяцев назад

      Honestly you should try it and negative prompt like --no manga, cartoon haha you'd be surprised, maybe you'll like the results!

  • @discoarcade_
    @discoarcade_ 10 месяцев назад +1

    Nice video - can you apply both --iw and --sref in one command to build the composition but also reference the style?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      Oh yah! I'll have a video about that idea soon

  • @cliftonperry4528
    @cliftonperry4528 10 месяцев назад

    None of your free PDF's ever arrive in my inbox or spam when I subscribe

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      really?? dang. You can shoot me an email and I'll send you them if you want. ftp @ futuretechpilot .com

  • @LouisGedo
    @LouisGedo 10 месяцев назад +1

    What's the deal with some sort of collaboration between Midjourney and Twitter I've heard being mentioned?

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад +1

      From what I can see, it's a combination of three separate pieces of info.
      1. David has mentioned that he has been friends with Elon for more than a few years
      2. Elon said something about allowing other ai's on Twitter
      3. David recently mentioned an openness to collaborating with other AI labs this year
      I think it's a stretch to say they will be working together, and I think it's irresponsible for people to act like they're a super detective who can read between the lines. But, if there is a collaboration - like allowing Midjourney images to be prompted from Twitter - these would be the clues that led to that

    • @LouisGedo
      @LouisGedo 10 месяцев назад +1

      @@FutureTechPilot
      Thanks for the info....... I didn't know anything about this until in passing I heard a quick soundbite on the radio I believe

    • @FutureTechPilot
      @FutureTechPilot  10 месяцев назад

      @@LouisGedo the radio?? haha that's crazy. Maybe I missed some bigger clues!

    • @LouisGedo
      @LouisGedo 10 месяцев назад +1

      @@FutureTechPilot
      I don't remember exactly where I "heard" it but it was on audio.
      But because I listen to a lot of YT videos, it may have been a video I was hearing while doing other tasks!
      👍

  • @Coollsmalls4kTechArtist
    @Coollsmalls4kTechArtist 10 месяцев назад +1

    You should trying my art style .

  • @eona
    @eona 10 месяцев назад +1

    I'm spending my days on MJ and I still can't keep up with all the new info! Drag the image? After the prompt?? And what's --sw ?! Oh god....