Not only did I like the video I subscribed too he could be charging for this tutorial this guy deserves a new subscriber. Everyone should subscribe if this video helped you make better images.
Fascinating. Thank you. And yes, I would like to see an even more in-depth video on this type of prompting. I'm learning a lot by starting out with very simple, basic prompts and then adding the other things to see how they change the image. Sometimes it's a bust, but sometimes it's pretty amazing.
haha chaos 100 .. you must have seen some pretty wild stuff! I'm glad you like the longer video, I'll have more like this in the future. Thanks for the comment!
This was a great video for blending v. Image weighting etc. I found that MJ 5 prompts are different than MJ 4. I have been using longer more descriptive prompts for MJ5. When I use prompts that are similar to how I would describe something I get better images. I have also just used prompts that say "....depicted in the style of" and MJ5 get's it perfect everytime! Thanks for the tut and keep up the good work !
Blend the best results of previously blended images 🔥. Blending is my favourite MJ feature, it’s a little like those gambling slot machines, you never know what you’re going to get.
glad to see you're still having fun with 5pointz 😉😎 -- that graffiti design of yours @9:45 is so dope! I use it in so many of my prompts that I went & gave it a unique parameter ("--5pointz"), just so my lazy ass doesn't have to spell out a full sentence every time, hahaha 🤷♂
haha dude thanks for pointing that out. I've mentioned custom parameters before but I feel like I should make a new video about it because it's so convenient and helpful.
@@FutureTechPilot seriously...I only have them set up for like pretty pointless shit, like "5pointz" instead of writing 3 or 4 extra words, and another style similar to that lol. but you can def make some crazy parameters w/ image prompts or very specific design styles/unique characters/landscapes/etc that are a huge pain in the ass to dial-in more than just once...I just always forget that's there whenever I could really use & benefit from it 🤦♂ lol - would for sure save me a TON of time with like 75% of the prompts I use on MJ though...
5:16 👍 yup I noticed that immediately on the very first day of using v5 because I almost always promted with a --c14 in v4 and in v5, the results of c14 are much more dramatic in most cases. I like subtle chaos in most cases.
@@FutureTechPilot yup........ unless you have already worked out a laser focused prompt that is getting you the correct results...... then I don't use --chaos
Excellent! Thanks for the video. I didn't know you could take it up to 5. I'm really glad you mentioned the little thing like clicking on the '+4' when uploading images. 👍Things like that can be a real pain. To anyone reading this - If you don't want to use up your fast hours in 10 minutes, use chaos in relaxed mode. Chaos mode could end up presenting you with 2 awesome directions that appeal to you from the same prompt. Then you'll go further down the rabbit hole with _each_ of them. So I suggest using chaos mode as 'brainstorming' in relaxed mode. Question(s): when remixing, do the stylize values 'compound'? If I change the stylize value in remix from 100 to 300 for example, am I increasing the stylize value for the _original_ image, or applying it to the remixed iteration? Hope that makes sense. Great video!
Wow, what a good suggestion for exploring with chaos in relaxed mode, I wrote it down and I'll be bringing it up in a bunch of future videos lol And that's a really interesting question about the stylize values ... I just did some experiments and based on my understanding, you're increasing the value on the remixed iteration. It seems like anything you add in the remix gets added to the new prompt, and wouldn't compound with how the previous prompt was made. let's say you generated a picture with chaos 100 and then went to remix one of the photos, even if you leave chaos 100 in the prompt, the pictures all closely resemble the original. Let me know if you notice anything different!
Hi, I liked the picture of the old man. Or is there another name for it? Anyhow, my question is if you need a regular PC or a more poweful one to handle the pix ?
@@FutureTechPilot I am a new to this world, but I waited for the oppertunities to be creative with text and pictures. Thanks for the tutoral, its awesome !!
I am wondering if the comma that was between the image of the girl and the colourful batman may have influenced the final image .. but i have no idea why that would result in a dog-boy or the other results :) .. keep up the great videos
Great Video! My english is not very well. Can you please help me with the "--v 5". What does it mean? I looking already in Discord but i´m not realy understand.😩
--v 5 just means 'Version 5' which is the newest version of the model.. so you could write --v 4 and use the Version 4 algorithm, or --v 3 to use the Version 3 algorithm. Hope that helps
haha you're the second person to ever ask me that! ruclips.net/video/HnuMM8U-vEM/видео.html I explain it early on in that video but I'll give you a quick rundown: - you can add a permanent suffix to all of your prompts ... /prefer_suffix -> new value -> --ar 16:9 (so now I don't have to type in the aspect ratio every time) - but if you want something different than 16:9, you can just type it into your prompt like normal and it will show up before and take importance over the 16:9 Hope that helps
Soooo helpful, thank you 🙏 The one thing that I'm stuck on is I'm trying to use a picture of my 23-year-old son that's not really looking like him even with IW tweaking in my creations . It's a face shot and I also find most creation always features a big representation of his head. Any thoughts 💬 on this?
I don't think there is a perfect way to get what you're looking for. Although you can try adding in more pictures of him, maybe 3 or 4 ... and maybe you could try describing his appearance. Hope that helps a little
@@wonkawarrior Yeah exactly, the more references you give MJ, the closer it might get to a resemblence ... but it's not an exact science. I personally don't think it works well but that is your best chance
@@FutureTechPilot It's a total waste of time to just combine images with image prompting UNLESS you're using more than 5 images which is the current image limit with Blend. It's MUCH quicker to Blend than the combine images with image prompting........ at least on a tablet or Smart phone
you don't have to use the blend mode...all that can be done with the imagine command. Even for no text prompt. Plus with imagine you also have the option to change aspect ratio, stylize value and even chaos.
Please more image prompting. Been testing doing corgis. Someone sends me a picture of their corgi and I try to infuse the corgi into a neat watercolor or so image. Another thing I’d love to see, is image promoting for Pokémon. Extremely difficult to do, near impossible for me just doesn’t work well. Midjourney only knows pikachu
hmm, infusing a subject is tough! I'm not sure if there is a fool proof way of doing it yet, but I'll keep a look out for you. And for pokemon? I think it's just a dataset issue ... as in, if MJ wasn't trained on certain pokemon, image prompting won't be able to help!
@@23SonicBoom haha I tried with Officer Jenny and it didn't work out at all ... maybe you could try image prompting like 3 or 4 different pics of the pokemon and then include a brief description. maybe that will get it closer!
@@FutureTechPilot I’ll do more practicing like you showed in the video. Pretty interesting seeing the results when playing around with prompts and settings. Thanks for posting the videos you do!
haha I appreciate the comment. www.midjourney.com/app/users/991837135667728504/ - that is a link to my gallery but I generate on private and don't update it often. So all of those pictures are from a while ago. I'll try to remember to publish new pictures more frequently
@@FutureTechPilot ah, gotcha. I appreciate your insight. Maybe for certain videos where you do comparisons/tips etc. publish some of those for the newbs learning? Just a thought. Thanks for sharing.
Why not show the prompt in all these photos man. Also by image prompting do you mean the imagine prompt? Technically /remix and /blend are also image prompts but they're not the imagine instruction. I got confused a couple of times following thru but thanks for the insight
I'm not sure what you mean though, I don't think I've ever cut off the prompt in any of my videos. They're always above the picture. And 'image prompting' is when you actually include the link to another photo in the prompt itself. Sorry if it was a bit confusing. Thanks for taking the time to comment!
To get the web address, you need to upload it to discord first - Just paste or drag the picture into a conversation. And to delete an image, you can 'react' to it with the red X emoji.
If you use /blend on desktop, you can upload images directly from your computer. It’s easy and super fun. However that method doesn’t allow you to add text prompts. For that you’d need to click on the + button next to the message box, upload the image, once uploaded right click on it and select copy link from menu , and copy that link in your new prompt after /imagine. If uploading more than one image to the same prompt, don’t forget to add a space between image links. Happy blending 😊
I was playing around with the Blend-Function. Combining 5 different characters, just to see what it looks like. And on one of those, the bot told me I used an NSFW pic, but I didn't. However, a while after it happened again, when I was high and combined Wrestlers to basically create ones. I'm guessing, it's because they don't wear shirts. Problem being, in both cases, it told me, that ny account could be banned because of it. And I can't get a hold of anybody from Midjourney, to rectifiy this. So basically, if I keep blending pics, which I liked to do just to see what the mashed-up characters look like, I could end up banned. Anybody has an Idea where I can get a hold of someone? Because the bot doesn't even tell me, wich one of the 5 pics, is the one causing the issue, as to just easily replace it.
I agree it's a frustrating problem because sometimes it's a 'false' positive. My understanding is that it is rare and has become even more rare with V5, so it sucks that you ran into it. I also agree with your assessment that it might have flagged their shirts being off, but I've got same message when absolutely nothing was nsfw ... anyways, the first offense is just like a temporary ban so they can take a closer look at your account and see if you actually did anything wrong. The only chance for you to talk to someone is going to be on the 'member support' channel of the Discord server. They can give you a better answer on how to find the problem image.
midjourney is brutal, thanks for the tips, yours are the freshest
haha cheers buddy!
Not only did I like the video I subscribed too he could be charging for this tutorial this guy deserves a new subscriber. Everyone should subscribe if this video helped you make better images.
hahah thanks so much for the comment. I really appreciate you taking the time to share your thoughts and I'm glad I could help out.
Cheers!
Fascinating. Thank you. And yes, I would like to see an even more in-depth video on this type of prompting. I'm learning a lot by starting out with very simple, basic prompts and then adding the other things to see how they change the image. Sometimes it's a bust, but sometimes it's pretty amazing.
You've pretty much nailed what I would teach! But yes thank you for the suggestion, I'll try to bump that up the priority list!
Thank you for the video! Very interesting and useful content.
Happy to help! Thanks for the feedback :)
Hey FTP - you gotta good vibe dude, rock on!! 😀🙏🏽✌
I appreciate the comment! Cheers :)
It is producing some really cool images for me!
Awesome! Let me know if you find any other cool tricks
Thanks for sharing this!! I tend to be an over kill and have been going to -c 100 and now it’s all making sense. Favorite video from you!
haha chaos 100 .. you must have seen some pretty wild stuff! I'm glad you like the longer video, I'll have more like this in the future. Thanks for the comment!
Just another fabulous in-depth video that sparks my creativity to new levels. Keep the videos coming! Thank you. 😎✌️
And thanks so much for the feedback! Glad I could help and I'm glad you enjoy the content
great video as always
I'm glad you enjoyed it!
This was a great video for blending v. Image weighting etc. I found that MJ 5 prompts are different than MJ 4. I have been using longer more descriptive prompts for MJ5. When I use prompts that are similar to how I would describe something I get better images. I have also just used prompts that say "....depicted in the style of" and MJ5 get's it perfect everytime! Thanks for the tut and keep up the good work !
'depicted in the style of' eh? Thanks for the tip! I've been adjusting how I speak to the bot as well and it's difficult to find what works best
Blend the best results of previously blended images 🔥. Blending is my favourite MJ feature, it’s a little like those gambling slot machines, you never know what you’re going to get.
hahha now that's a rabbit hole I should go down sometime
Using the 40% image to blend with another image is a technique I've never seen before, very interesting!
Yeah haha I was surprised by that one too ... let me know if you learn any other wacky tips like that
glad to see you're still having fun with 5pointz 😉😎 -- that graffiti design of yours @9:45 is so dope!
I use it in so many of my prompts that I went & gave it a unique parameter ("--5pointz"), just so my lazy ass doesn't have to spell out a full sentence every time, hahaha 🤷♂
haha dude thanks for pointing that out. I've mentioned custom parameters before but I feel like I should make a new video about it because it's so convenient and helpful.
@@FutureTechPilot seriously...I only have them set up for like pretty pointless shit, like "5pointz" instead of writing 3 or 4 extra words, and another style similar to that lol. but you can def make some crazy parameters w/ image prompts or very specific design styles/unique characters/landscapes/etc that are a huge pain in the ass to dial-in more than just once...I just always forget that's there whenever I could really use & benefit from it 🤦♂ lol - would for sure save me a TON of time with like 75% of the prompts I use on MJ though...
@@mkm1856 saaame, always forget about it, but it would be so useful lol
Hey thanks Nolan that was awesome 🔥
Glad you like the longer video! I might have to make more of these in the future
Wow .. very amazing what u can do with just a few words.. thanx
I completely agree! Glad you enjoyed it
5:16
👍 yup
I noticed that immediately on the very first day of using v5 because I almost always promted with a --c14 in v4 and in v5, the results of c14 are much more dramatic in most cases.
I like subtle chaos in most cases.
Having no chaos is almost boring in V5
@@FutureTechPilot yup........ unless you have already worked out a laser focused prompt that is getting you the correct results...... then I don't use --chaos
Excellent! Thanks for the video. I didn't know you could take it up to 5. I'm really glad you mentioned the little thing like clicking on the '+4' when uploading images. 👍Things like that can be a real pain.
To anyone reading this - If you don't want to use up your fast hours in 10 minutes, use chaos in relaxed mode. Chaos mode could end up presenting you with 2 awesome directions that appeal to you from the same prompt. Then you'll go further down the rabbit hole with _each_ of them. So I suggest using chaos mode as 'brainstorming' in relaxed mode.
Question(s): when remixing, do the stylize values 'compound'? If I change the stylize value in remix from 100 to 300 for example, am I increasing the stylize value for the _original_ image, or applying it to the remixed iteration?
Hope that makes sense. Great video!
Wow, what a good suggestion for exploring with chaos in relaxed mode, I wrote it down and I'll be bringing it up in a bunch of future videos lol
And that's a really interesting question about the stylize values ... I just did some experiments and based on my understanding, you're increasing the value on the remixed iteration.
It seems like anything you add in the remix gets added to the new prompt, and wouldn't compound with how the previous prompt was made.
let's say you generated a picture with chaos 100 and then went to remix one of the photos, even if you leave chaos 100 in the prompt, the pictures all closely resemble the original.
Let me know if you notice anything different!
@@FutureTechPilot Thank you for replying! :)
I'm experimenting with just this in mind right now.
7:59
😆 🤣 😂 that image 1 is a riot!!!
It's burned into my nightmares lol
@@FutureTechPilot 😆 🤣 😂
Hi, I liked the picture of the old man. Or is there another name for it? Anyhow, my question is if you need a regular PC or a more poweful one to handle the pix ?
To use Midjourney? You can use your phone lol it's all through discord and their own gpus. Nothing from yours
@@FutureTechPilot I am a new to this world, but I waited for the oppertunities to be creative with text and pictures. Thanks for the tutoral, its awesome !!
6:45
Image 4 is really sweet!
haha dude I had a hard time choosing the thumbnail for this video .. that #4 was a top choice for sure
@@FutureTechPilot
That #4 is a happy accident! 😉
More comic book artists and styles please 👍
That's a good suggestion! Thanks
I am wondering if the comma that was between the image of the girl and the colourful batman may have influenced the final image .. but i have no idea why that would result in a dog-boy or the other results :) .. keep up the great videos
😂 only god knows where that dog boy came from
Great Video! My english is not very well. Can you please help me with the "--v 5". What does it mean? I looking already in Discord but i´m not realy understand.😩
--v 5 just means 'Version 5' which is the newest version of the model.. so you could write --v 4 and use the Version 4 algorithm, or --v 3 to use the Version 3 algorithm.
Hope that helps
I have a question: I’ve noticed you’ll have 2 aspect ratios in some of your prompts (ie - -ar 4:3 - - 16:9) What affect does this have?
haha you're the second person to ever ask me that! ruclips.net/video/HnuMM8U-vEM/видео.html I explain it early on in that video but I'll give you a quick rundown:
- you can add a permanent suffix to all of your prompts ... /prefer_suffix -> new value -> --ar 16:9 (so now I don't have to type in the aspect ratio every time)
- but if you want something different than 16:9, you can just type it into your prompt like normal and it will show up before and take importance over the 16:9
Hope that helps
I actually like --c 80. Ha. Great tute. Again! Thanks FTP.
🤣 I knew there would be someone! haha I'm sure you've found a lot of gems with a high chaos
Soooo helpful, thank you 🙏 The one thing that I'm stuck on is I'm trying to use a picture of my 23-year-old son that's not really looking like him even with IW tweaking in my creations . It's a face shot and I also find most creation always features a big representation of his head. Any thoughts 💬 on this?
I don't think there is a perfect way to get what you're looking for. Although you can try adding in more pictures of him, maybe 3 or 4 ... and maybe you could try describing his appearance.
Hope that helps a little
Oh, so you just add pictures of the same person in your prompt and then MidJourney will interpret them together as one?
@@wonkawarrior Yeah exactly, the more references you give MJ, the closer it might get to a resemblence ... but it's not an exact science. I personally don't think it works well but that is your best chance
@@FutureTechPilotWhen all else fails, there's always Photoshop 🤣 I'll give it a try tonight and report back, thanks again for all your help 🙏
@@wonkawarrior 😂 good luck!
Image prompting is significantly superior/ better IMO since I always use text prompt with it
Yah honestly they're 2 different things because of that. You can blend with image prompting but you can also do sooo much more
@@FutureTechPilot
It's a total waste of time to just combine images with image prompting UNLESS you're using more than 5 images which is the current image limit with Blend. It's MUCH quicker to Blend than the combine images with image prompting........ at least on a tablet or Smart phone
you don't have to use the blend mode...all that can be done with the imagine command. Even for no text prompt. Plus with imagine you also have the option to change aspect ratio, stylize value and even chaos.
💯
Wow, great channel. Didnt know --iw came back in mj5. Great stuff ;)
Shame that you cant weight each upload image. Look forward to that day.
Yeah that would be pretty cool! Maybe one day lol
In my images, going over chaos 40 is when the strange figures in the background start showing up. There creepy usually
hahah don't fall down the rabbit hole ...
Please more image prompting. Been testing doing corgis. Someone sends me a picture of their corgi and I try to infuse the corgi into a neat watercolor or so image.
Another thing I’d love to see, is image promoting for Pokémon. Extremely difficult to do, near impossible for me just doesn’t work well. Midjourney only knows pikachu
hmm, infusing a subject is tough! I'm not sure if there is a fool proof way of doing it yet, but I'll keep a look out for you. And for pokemon? I think it's just a dataset issue ... as in, if MJ wasn't trained on certain pokemon, image prompting won't be able to help!
@@FutureTechPilot have you tried using an image for reference for Pokémon? It definitely tries but not able to actually get it similar enough.
@@23SonicBoom haha I tried with Officer Jenny and it didn't work out at all ... maybe you could try image prompting like 3 or 4 different pics of the pokemon and then include a brief description. maybe that will get it closer!
@@FutureTechPilot I’ll do more practicing like you showed in the video. Pretty interesting seeing the results when playing around with prompts and settings. Thanks for posting the videos you do!
@@23SonicBoom
Great videos. Is your MJ profile public? I’d love to follow you. Thanks!
haha I appreciate the comment. www.midjourney.com/app/users/991837135667728504/ - that is a link to my gallery but I generate on private and don't update it often. So all of those pictures are from a while ago. I'll try to remember to publish new pictures more frequently
@@FutureTechPilot ah, gotcha. I appreciate your insight. Maybe for certain videos where you do comparisons/tips etc. publish some of those for the newbs learning? Just a thought. Thanks for sharing.
6:59
😆 🤣 😂
I'm Team Chaos > 35
@@FutureTechPilot 💥
Why not show the prompt in all these photos man. Also by image prompting do you mean the imagine prompt? Technically /remix and /blend are also image prompts but they're not the imagine instruction. I got confused a couple of times following thru but thanks for the insight
I'm not sure what you mean though, I don't think I've ever cut off the prompt in any of my videos. They're always above the picture.
And 'image prompting' is when you actually include the link to another photo in the prompt itself.
Sorry if it was a bit confusing. Thanks for taking the time to comment!
Probably a dumb questions: How do you make a web address from an image to blend? And, how do delete images? With my lack of skills, some are ugly! 😢
To get the web address, you need to upload it to discord first - Just paste or drag the picture into a conversation.
And to delete an image, you can 'react' to it with the red X emoji.
If you use /blend on desktop, you can upload images directly from your computer. It’s easy and super fun. However that method doesn’t allow you to add text prompts. For that you’d need to click on the + button next to the message box, upload the image, once uploaded right click on it and select copy link from menu , and copy that link in your new prompt after /imagine. If uploading more than one image to the same prompt, don’t forget to add a space between image links. Happy blending 😊
You look very beautiful in the graphic novel...
🤣 I'm glad someone thinks so
No, you're actually beautiful.
I was playing around with the Blend-Function. Combining 5 different characters, just to see what it looks like. And on one of those, the bot told me I used an NSFW pic, but I didn't. However, a while after it happened again, when I was high and combined Wrestlers to basically create ones. I'm guessing, it's because they don't wear shirts. Problem being, in both cases, it told me, that ny account could be banned because of it. And I can't get a hold of anybody from Midjourney, to rectifiy this. So basically, if I keep blending pics, which I liked to do just to see what the mashed-up characters look like, I could end up banned. Anybody has an Idea where I can get a hold of someone? Because the bot doesn't even tell me, wich one of the 5 pics, is the one causing the issue, as to just easily replace it.
I agree it's a frustrating problem because sometimes it's a 'false' positive. My understanding is that it is rare and has become even more rare with V5, so it sucks that you ran into it.
I also agree with your assessment that it might have flagged their shirts being off, but I've got same message when absolutely nothing was nsfw ...
anyways, the first offense is just like a temporary ban so they can take a closer look at your account and see if you actually did anything wrong.
The only chance for you to talk to someone is going to be on the 'member support' channel of the Discord server. They can give you a better answer on how to find the problem image.
👋
🤝