This is the coolest Midjourney exploration I’ve seen in a long time, so many awesome results along the way! I might get sucked down this rabbit hole for a while. Amazing thumbnail btw
Oh man, thanks a lot!! Haha, it was about the 7th version of the thumbnail! MJ_Illus_v7_FINAL_rev_b.png We do drive ourselves crazy, don't we? Enjoy the rabbit hole, it turns out to be a Mobius loop! PS- big fan of the work you're doing on your channel!
@@TheoreticallyMedia Haha, I know the feeling for sure. I've definitely spent all day in Photoshop to end up with a thumbnail I hate, multiple times. Thanks! Yours is one of the few ai channels I watch on UA-cam. Usually I find stuff on Twitter then test it out myself. But you'll usually have tips or tricks I wouldn't think of myself. Like in this video, I've never even tried the stop command before.
Thank you so much!! That's something that I love about the current state of AI Tech-- everything this changing so quickly, we're ALL learning ALL the time! It's so invigorating!
Thank you for looking into this. I've been experimenting as well with this whole optical illusion style, nowhere near what you have been able to accomplish. Sharing your ups and downs is very helpful. I spent a couple of fast hours trying the spirals before setting them aside, going to try the other approach you mention for the non-spiral efforts. It's these serendipitous results, that seem to come out of nowhere, that I really do enjoy. I won't be surprised if someone comes up with a simple adaptation of something like the "stop" command, or some other parameter and revolutionize this Optical Illusion Style. Thanks again.
100%! And thanks for picking up on the ups and downs! I didn’t want to cherry pick the best results, since I think it tends to discourage people trying the technique out. It takes a LOT of playing to get something good, but the results are really worth it! The Spirals? Yeah, I’m going to hold off on that one, too much work!
Very detailed! I love how deep you went by adding optical illusion style imagery with the vary subtle method. There's so many great techniques yet to be discovered with midjourney I think! Killer video as always Tim :)
Thanks so much Wade!! Totally agree-- I love the fact that MJ is this endless rabbit hole, and each time they release a new feature, we have a whole new series of approaches and combinations with previous commands. It's funny, I was just thinking: One of the coolest things about covering AI on YT is that you're never really out of ideas. Things just keep evolving and changing. If anything, the challenge is keeping up!
I wanted to get to that in this video! But, I was short on time, and I have found that combining topics like MJ and Pika in one video tends to kill the retention. Oh well, good excuse for another Pika video!!
When I use the "Vary(Subtle)" button it does not prompt for me edit the original prompt. Instead it just gives me a slight variation of the original image. Am I doing something wrong here?
6:17 simple concept, but done so well. 👍 Great video. 💯 The creative rabbit hole can go so deep. 6:56 Was that a live realization? 😅I was blown away by the realization a few seconds before you mentioned it. Wonderful, beautiful 'accidents'! Have a great one! :)
Great Video. I was unable to a really good dancer/skulll look like you have even after trying subtle many times and with a very good painted skull if I do say so myself. Anything I might check or be missing?
So, the base image does play a pretty big part, so I’d try rolling up a variety of them. If you want, swing by the Discord and we can mess around with it. I’m curious what would happen if I passed you the same base skull image!
@@TheoreticallyMedia It’s funny we had a creativity Zoom call with DePaul University and people were afraid AI was going to end them. I think it will push creativity personally.
Tim's the Poster Boy for "When your free stuff is 10x better than the paid duds." Thanks for this, dude. ❤️ Also, you ever thought about creating a video with a finished project in mind (comic book, etc...)?
Great video! I've been messing around a lot with the vary subtle/stop methods a lot lately. It's a fun way to create stuff, although definitely finicky. Agreed about simple iconography. I've had the best luck using electric or neon colors on the base image, and prompting for black and white illustrations or mezzotint prints. My results with photos have been a little dicey. Also seems to help by adding "trompe l'oeil" to the prompt. Anyway, you've given me a lot more to work with, so thank you. This is one of the classiest AI channels around. Take care.
Ha, that one is an odd duck. It's a 1980s era Peavey Horizon 2. Swamp Ash, and kinda stupid heavy for a strat. Fairly unique pickup situation: If you crank the tone knobs to 10, it's a single coil, but if you lower them to 7, they turn into humbuckers. Such an odd guitar. Picked it up off a random facebook marketplace ad from a guy who collected guitars but never learned to play. Strange fella. Pretty great guitar though!
@@TheoreticallyMedia Thanks for the info! Yeah strange to collect guitars but not even play them. Peavey Horizon, I don’t need anymore guitars but just one more wouldn’t hurt right?
How do you think it would take a photo prompt with and image that is 50% opacity of one photo and 50% of another? It could get the bot ideas for how to combine them and where
Very useful video, I have subscribed. I have a question , can midjourney create the same character in different settings/poses /emotions for adding to a video with different scenes involving the same lead characters . Is that possible ? Or each character generated will always be different ?
Oh, I’m so glad you found this video!! It’s such a cool and rewarding technique. I still think about that moth to light moment often! It really did blow my mind!
Interesting video thanks. Oh,and that image really evoked Rodney Matthews, so yes, 70s (and 80s) rock covers. Roger Dean and Matthews do have some similarities and thats a great Yes album:)
Matthews is so good! It’s funny, a lot of early AI Art always reminded me of his work. Not in a trained way, more in a weird retro color palette and shapes sort of thing, I’m glad I went with Yes and not Gentle Giant!
@@TheoreticallyMedia I had a Mathews poster on my wall at university back in the 80s... love his art. I think many of the illustrators from the 70s and 80s were way more interesting than the ones we have post digital, and contrary to what a lot of people are saying I think its actually digital that has done a lot of damage to art and AI may invigorate it after decades of stagnation. Im saying that as a digital illustrator lol, but I started trying to get into the field with a trad portfolio pre digital and it was a much tighter community with far higher standards and expectations back then and a lot harder to break into, digital lowered the bar and along with the ease of finding cheap illustrators online flooded the markets with unimaginative art that all started to look the same. There are good digital ones but they kinda get drowned out by all the noise. While obviously there is huge potential for AI to be more of that plus some, it might also shake people up and cause a rethink. Im hoping it will be a rethink!
Tales from Topographic Oceans. Subscribed! (Does Midjourney recognize Roger Dean or Hugh Syme?🤔) I’m trying to pick up Midjourney and other AI tools for my music video channel. (Editing video is a necessary evil.) I’m happy to have recently discovered your channel. It’s is invaluable.
Thank you so much!! And if I ever hold up a copy of Tormato, please feel free to unsub! We should try Roger and Hugh as keywords! Now I'm curious myself!
Now THAT’s a ten cent word! Gotta work that in a few times this week! But yeah, this does seem to solve that problem. There are some real issues in language that AI just doesn’t “get”- Vary Subtle seems to solve a few of them. MJ still won’t do a cyclops though!
Not gonna lie, just pulling out Tales for the video made me give it a spin! Then…I had to listen to The Yes Album…and then I figured it was time to reorganize my record collection. And then it was 4 days later!
Hay Anyone help me with midjourney subscription. In India only card method available and my card always decline, It's seems 2d and 3d is not accepted in INDIA. Please someone help me. If anyone help me i send money immediately 🙏
I think you'll have better luck checking in with Midjourney Support/Billing. Given the amount of scams (not saying you are one of them) going around these day, most people will be wary.
@@TheoreticallyMedia yeah I contacted midjourney support and they say 2d and 3d authenticate not accepted in INDIA ,and I'm not scam anyone I just need subscribtion,i can pay extra if anyone help 🙏
I don't know, I'm just kind of done with how AI fucks up human figures. Like the woman with 4 hands at 6:48. I just don't want to look at that kind of stuff anymore
I don’t see it that way. I see it as either speeding up a working artists workflow, or creating more value for original art. Traditional artists have always had a tough time making a living solely through art. AI might increase competition, but I don’t really think so. In the commercial space, artists will fold these tools into their workflow. In the fine art space, it’ll just increase the value of a piece created with paint and canvas.
Thanks
Wow! Thank YOU so much! Hope you enjoy the technique!!
When I click the vary button, there isn't a window for any new prompts.. how do I fix that?
This is the coolest Midjourney exploration I’ve seen in a long time, so many awesome results along the way! I might get sucked down this rabbit hole for a while.
Amazing thumbnail btw
Oh man, thanks a lot!! Haha, it was about the 7th version of the thumbnail!
MJ_Illus_v7_FINAL_rev_b.png
We do drive ourselves crazy, don't we?
Enjoy the rabbit hole, it turns out to be a Mobius loop!
PS- big fan of the work you're doing on your channel!
@@TheoreticallyMedia Haha, I know the feeling for sure. I've definitely spent all day in Photoshop to end up with a thumbnail I hate, multiple times.
Thanks! Yours is one of the few ai channels I watch on UA-cam. Usually I find stuff on Twitter then test it out myself. But you'll usually have tips or tricks I wouldn't think of myself. Like in this video, I've never even tried the stop command before.
I love how when I think I've seen it all I get proven dramatically wrong. Excellent video.
Thank you so much!! That's something that I love about the current state of AI Tech-- everything this changing so quickly, we're ALL learning ALL the time! It's so invigorating!
Thank you for looking into this. I've been experimenting as well with this whole optical illusion style, nowhere near what you have been able to accomplish. Sharing your ups and downs is very helpful. I spent a couple of fast hours trying the spirals before setting them aside, going to try the other approach you mention for the non-spiral efforts. It's these serendipitous results, that seem to come out of nowhere, that I really do enjoy. I won't be surprised if someone comes up with a simple adaptation of something like the "stop" command, or some other parameter and revolutionize this Optical Illusion Style. Thanks again.
100%! And thanks for picking up on the ups and downs! I didn’t want to cherry pick the best results, since I think it tends to discourage people trying the technique out. It takes a LOT of playing to get something good, but the results are really worth it!
The Spirals? Yeah, I’m going to hold off on that one, too much work!
Very detailed! I love how deep you went by adding optical illusion style imagery with the vary subtle method. There's so many great techniques yet to be discovered with midjourney I think! Killer video as always Tim :)
Thanks so much Wade!! Totally agree-- I love the fact that MJ is this endless rabbit hole, and each time they release a new feature, we have a whole new series of approaches and combinations with previous commands.
It's funny, I was just thinking: One of the coolest things about covering AI on YT is that you're never really out of ideas. Things just keep evolving and changing. If anything, the challenge is keeping up!
@@TheoreticallyMedia Agree 100% mate!
Thanks for sharing your awesome technique! Loved that woman lightbulb pic 😍
That whole lightbulb sequence still amazes me. I think that’s the first time I’ve ever caught myself on camera having my mind blown!
Pika Labs has a system like the spiral example you showed. You can use 2 tone 'templates' and text to influence the images now.
I wanted to get to that in this video! But, I was short on time, and I have found that combining topics like MJ and Pika in one video tends to kill the retention. Oh well, good excuse for another Pika video!!
Great work as always! I appreciate the unique workflows that really through new ideas and concepts into the mix :)
Thank you!! And yeah, I love this idea, namely since it invites so much personal exploration and interpretation!
Very cool experiments! I’ve been using Vary Subtle more often lately. Solid tool!
Vary Subtile is 100% the Jan Brady of the Vary buttons- but it does have a lot to offer!!
loved your video about the illusions but have one question how do i get the very subtle button to appear because i can't seem to get that
Go into /settings and make sure the Remix button is turned on. That should get you there!
When I use the "Vary(Subtle)" button it does not prompt for me edit the original prompt. Instead it just gives me a slight variation of the original image. Am I doing something wrong here?
So, go into /settings and make sure you have the Remix button turned on. That should clear you up!
Also make sure you're upscaling one of your images first. Otherwise, you won't see Vary (Strong) and Vary (Subtle)
6:17 simple concept, but done so well. 👍
Great video. 💯 The creative rabbit hole can go so deep.
6:56 Was that a live realization? 😅I was blown away by the realization a few seconds before you mentioned it. Wonderful, beautiful 'accidents'! Have a great one! :)
Total Live reaction! That the EXACT look of my mind being blown!
Great Video. I was unable to a really good dancer/skulll look like you have even after trying subtle many times and with a very good painted skull if I do say so myself. Anything I might check or be missing?
So, the base image does play a pretty big part, so I’d try rolling up a variety of them.
If you want, swing by the Discord and we can mess around with it.
I’m curious what would happen if I passed you the same base skull image!
Keep up the great work. I truly appreciate your work and information.
Thank you so much! This one was a real joy to do! Plus...y'know, MJ broke my mind in this one!
@@TheoreticallyMedia It’s funny we had a creativity Zoom call with DePaul University and people were afraid AI was going to end them. I think it will push creativity personally.
Tim's the Poster Boy for "When your free stuff is 10x better than the paid duds." Thanks for this, dude. ❤️ Also, you ever thought about creating a video with a finished project in mind (comic book, etc...)?
Great video! I've been messing around a lot with the vary subtle/stop methods a lot lately. It's a fun way to create stuff, although definitely finicky. Agreed about simple iconography. I've had the best luck using electric or neon colors on the base image, and prompting for black and white illustrations or mezzotint prints. My results with photos have been a little dicey. Also seems to help by adding "trompe l'oeil" to the prompt.
Anyway, you've given me a lot more to work with, so thank you. This is one of the classiest AI channels around. Take care.
I'll give that trompe l'oeil prompt a shot! At least...as we learned in this video, I don't have to pronounce it out loud!
Thanks for another insightful MJ take. I didn't realise you were in the band, "Yes!"
Vary cool.
Thank you very much! And yup, but only the 70s era stuff! I’ve got no time for Owner of a Lonely heart!
@@TheoreticallyMedia When I ran the Google-check... "Who was in Yes," it came back with... "No they weren't."
Unfortunately when I click vary I don't have the option to remix, it automatically vary the photo. Do you know how to solve it?
I do! So, issue the command /settings and make sure the "remix" option is turned on. That should solve your problem!
Thank you! it worked :) @@TheoreticallyMedia
What’s the maple coloured 3 pick up guitar behind you? Looks really really nice
Ha, that one is an odd duck. It's a 1980s era Peavey Horizon 2. Swamp Ash, and kinda stupid heavy for a strat. Fairly unique pickup situation: If you crank the tone knobs to 10, it's a single coil, but if you lower them to 7, they turn into humbuckers. Such an odd guitar. Picked it up off a random facebook marketplace ad from a guy who collected guitars but never learned to play. Strange fella.
Pretty great guitar though!
@@TheoreticallyMedia Thanks for the info! Yeah strange to collect guitars but not even play them.
Peavey Horizon, I don’t need anymore guitars but just one more wouldn’t hurt right?
How do you think it would take a photo prompt with and image that is 50% opacity of one photo and 50% of another? It could get the bot ideas for how to combine them and where
THAT....would be really interesting. Gonna have to give that a shot!! Let's do it in tandem and report back!
Very useful video, I have subscribed. I have a question , can midjourney create the same character in different settings/poses /emotions for adding to a video with different scenes involving the same lead characters . Is that possible ? Or each character generated will always be different ?
Excellent 💡
Haha, now I cant’t UNSEE it in a lightbulb!
Thank you for this💯🎉👍🏾
Excellent to Hear! Enjoy those illusions!
Great video!
Thank you so much!
whoa, i was wondering how to do pareidolia in MJ, here it is. you're a GIN-YISS!
Oh, I’m so glad you found this video!! It’s such a cool and rewarding technique. I still think about that moth to light moment often! It really did blow my mind!
When I click Vary Subtle it just varies by itself
Issue the command /settings and see if “remix” is turned on. That should solve your problem!
@@TheoreticallyMediathank you!
I was definitely going to say it looked like a Rush or Yes cover, for sure! LOL
Those 70s prog bands had the best album covers! Super trippy and intricate details! For sure a lost art!
Interesting video thanks. Oh,and that image really evoked Rodney Matthews, so yes, 70s (and 80s) rock covers. Roger Dean and Matthews do have some similarities and thats a great Yes album:)
Matthews is so good! It’s funny, a lot of early AI Art always reminded me of his work. Not in a trained way, more in a weird retro color palette and shapes sort of thing,
I’m glad I went with Yes and not Gentle Giant!
@@TheoreticallyMedia I had a Mathews poster on my wall at university back in the 80s... love his art. I think many of the illustrators from the 70s and 80s were way more interesting than the ones we have post digital, and contrary to what a lot of people are saying I think its actually digital that has done a lot of damage to art and AI may invigorate it after decades of stagnation. Im saying that as a digital illustrator lol, but I started trying to get into the field with a trad portfolio pre digital and it was a much tighter community with far higher standards and expectations back then and a lot harder to break into, digital lowered the bar and along with the ease of finding cheap illustrators online flooded the markets with unimaginative art that all started to look the same. There are good digital ones but they kinda get drowned out by all the noise.
While obviously there is huge potential for AI to be more of that plus some, it might also shake people up and cause a rethink. Im hoping it will be a rethink!
Great video Tim!
Thanks Jeff! It was nice to touch base back with MJ. I know we’re off in Video Land a lot, but I still love MJ as the “engine room”
Great stuff!!
Thank you, Alex!
Tales from Topographic Oceans. Subscribed!
(Does Midjourney recognize Roger Dean or Hugh Syme?🤔)
I’m trying to pick up Midjourney and other AI tools for my music video channel. (Editing video is a necessary evil.)
I’m happy to have recently discovered your channel. It’s is invaluable.
Thank you so much!! And if I ever hold up a copy of Tormato, please feel free to unsub!
We should try Roger and Hugh as keywords! Now I'm curious myself!
Nice! I've been wondering about AI pareidolia for a while. Thank you!
Now THAT’s a ten cent word! Gotta work that in a few times this week!
But yeah, this does seem to solve that problem. There are some real issues in language that AI just doesn’t “get”- Vary Subtle seems to solve a few of them.
MJ still won’t do a cyclops though!
@@TheoreticallyMedia Yeah, I also had my cyclops phase 😅
YES!
Yes!’n
I was so excited, I put an extra n there!
@@TheoreticallyMedia 😆
How do you write the name of the french artist you butchered?
Haha, solidly butchered! So, the artist name is unknown, but the piece is called L'Amour de Pierrot.
On the plus side, at least I didn't say Pie-Rot.
Great vid! (Prog people unite)
Not gonna lie, just pulling out Tales for the video made me give it a spin! Then…I had to listen to The Yes Album…and then I figured it was time to reorganize my record collection. And then it was 4 days later!
Of course, but in what order?! #highfidelity
@@TheoreticallyMedia
Favorite prog rock album cover?
Im mean, I gotta go with Brain Salad Surgery, because: HR Giger.
Hay
Anyone help me with midjourney subscription.
In India only card method available and my card always decline,
It's seems 2d and 3d is not accepted in INDIA.
Please someone help me.
If anyone help me i send money immediately 🙏
I think you'll have better luck checking in with Midjourney Support/Billing. Given the amount of scams (not saying you are one of them) going around these day, most people will be wary.
@@TheoreticallyMedia yeah I contacted midjourney support and they say 2d and 3d authenticate not accepted in INDIA ,and I'm not scam anyone I just need subscribtion,i can pay extra if anyone help 🙏
@@TheoreticallyMediasir can you buy my midjourney subscription
I pay first if you say
"How Very!"
Vary Subtle is the Jan Brady of Varies!
Yeah honestly on dalle 3 bing app ypu could def get more egat u were intending 😉
I don't know, I'm just kind of done with how AI fucks up human figures. Like the woman with 4 hands at 6:48. I just don't want to look at that kind of stuff anymore
Yeah, that was a pretty weird one. It's mostly gotten better about it, but every once in awhile, you get something odd like that. Very Cronenberg
@@TheoreticallyMedia Maybe it is getting better with the new versions coming out. Like DALL-E 3
This is killing the livelihood of REAL artists!
I don’t see it that way. I see it as either speeding up a working artists workflow, or creating more value for original art.
Traditional artists have always had a tough time making a living solely through art. AI might increase competition, but I don’t really think so.
In the commercial space, artists will fold these tools into their workflow. In the fine art space, it’ll just increase the value of a piece created with paint and canvas.