Exploring storytelling and creative videos has been my focus lately. VideoGPT appeared, and it's like having this secret weapon that gives my content that extra professional flair.
The consistent character creation in Midjourney is truly innovative. I've longed for such a feature since before. However, once SORA comes out, it feels like everything will just come to an end. I wonder if there will be many drawbacks, though?😔🙂
@@WayOfTheZombie. I also primarily used Pika and Runway, but seeing how Sora is far superior, I hadn't considered whether they could keep up with its capabilities. Anyway, it would be great if everyone could show their own distinctive advancements
What I don't like about Pika is the fact that it's a toss of the dice. If the render does not match the prompt, or if the render is just terribly warped out of shape, too bad... so sad you just tossed credits out the window. But with others like Runway, at least you get the free option to choose an appropriate image to use.
TIP. You can select two or three icons at the same time. Just hold down SHIFT while selecting. Then the image is at the same time a character, a style and a hint to the image.
Hey Caleb, is there a reason I don't have the option to copy link? I only seem to have the option to copy message link, which doesn't allow for me to use this cref command.
To answer my own question: this option is only available if you're using the Discord desktop app, otherwise you have to enlarge the image and select "copy image address".
Correct! In Discord, you need to make sure the image is already inserted into discord, then click on it to expand, then right click to get the copy link option .
why is pika adding new features (that are terrible anyways) before getting the lipsync feature to a decent level? it is already a very unique feature still in the field. if they get it right, it can sky rocket them ahead of competition.
@@curiousrefuge is there any way I can add two characters into one photo? For instance, I have an image of a girl and a photo of a cat. How can I use Midjourney to merge these two into one image while ensuring they fit the script? For example, if I want the girl and the cat eating rice.
I’d really like to know what you think about the future of filmmaking… if AI takes over, won’t the original works dry up leaving AI constantly remixing and regurgitating the same source material? I honestly can’t see how this can end well, seeing as studios and financiers are always going to opt for the cheapest solutions. Snake eating its tail comes to mind. It’s going to decimate the film industry unless the paying public choose real people and real footage with their wallets. Personally I can’t think of anything more boring and pointless than watching a film made by AI.
Well, there is room for both (not a zerosum game) right?. For example, like animation, some people believe generative AI will be a subsection. Or, it could be a mix like LiveGen (live action + live generation). We believe ART in all forms will continue to live on and evolve.
to Midjourney Consistence Character - My question ? What actually is a character ? A human being, an animal or just a simple object ?? What I want to say is, that in an AI based project even non-living things like a ship or a building need a certain consistency. Any experience with Objects or something else in conjunction with --Cref ?
@@curiousrefugeThank you. I have tested it in relation to objects and of course it doesn't work. It probably can't work at all yet. But in the end this has to work to let a character move in an environment and interact with things.
Do you know if an AI photo/video generator that can take consumer product imagery as inputs and reimagine them like MJ does with characters? It seems like what's out there doesn't exactly replicate the product/logo.
They CAN be replaced but that doesn’t mean they will be. I’m really hoping the public will always want authentic real people in their media space. Human to human connection is undervalued.
When I use the consistent character option using a selfie of me, it doesn't seem to work as well as it did for Caleb because it generates images of characters that don't look like me. I mean, am I really that ugly jeez haha😂 not sure if anyone had the same experience?
It’s so weird because some peoples are more easily replicated. My gut tells me that Midjourney is creating its own image of you, then it creates the clones.
@@curiousrefuge With V6 I've used sref, weights "::3" and ce 0-3 with multiple variations of blend. Ive lost count. Ive tried different things on and off for a month. Maybe I'm too picky.
@@curiousrefuge It is a bit tiring that many apps talk about consistency, when the morphology changes from a little to a lot in each generation. Let's not talk about the clothes. Semi-consistent would be an acceptable expression, at least it wouldn't be a blatant lie.
Er...herm. None of the MJ character outfits you showcased were the same as the source image. Fine for concept art, maybe, but let's not get carried away with hype just yet.
OLD MAN YELLS AT CLOUD: I do not like your Game of the Week. And you never give the results. You should just give the results at the end of the video (forget the free swag gimmick) or forget the game and just tell us which is which so we can make an informed comparison. Yes, people love free crap, but no one is bringing their friends here because of the GotW. If you really want a give-away (in exchange for comments), just tell people to make a specific comment and you'll pick a random winner, instead of creating stressful secrets for all the rest of your fans. I do otherwise love your channel and Film of the Week and appreciate all your reporting!
We appreciate your engagement and comment. We'll definitely take your feedback and reconsider how we do things. We certainly don't want this to be 'stressful'. Thanks so much for the feedback and wathcing!
the idots at hollywood signed a contract to not use ai. But independents will smash that and make good ai films that aren't constrained by the hollywood strike contract. All actors need to be completely replaced by ai. period..
holly wood is bad. that is all you have to know. And if you pretend yo udon't know, then you are an apologist for hollywood, which means you are summarily dismissed from all forms of credibility. take your pick. either way, you are exposed.@@jamesr6562
We tend to agree that Hollywood not taking advantage of this is a likely not a smart and long term move. If they embrace it they can create some amazing things!
![[TH] JFT vs RLF - VCT Ascension Pacific - Day 2](http://i.ytimg.com/vi/PuIH7w_jO6c/mqdefault.jpg)
1. Eleven Labs, 2. Artlist, 3. PIka
Congrats! You are the winner! Reach out to us at hello@curiousrefuge.com and we will give you next steps for claiming your prize!
Hey, thanks for including a couple of my Midjourney generations at the start of your video. 😃
Exploring storytelling and creative videos has been my focus lately. VideoGPT appeared, and it's like having this secret weapon that gives my content that extra professional flair.
This is so interesting to watch, learnt a thing from it, thank you
Outstanding information. Happy new subscriber.
1. Pika, 2. Artlist, 3. Eleven Labs, another great video!
1. Eleven labs 2. pika 3. artlist
Thanks for the news Ai, there are really good and big developments.
1 - Pika
2 - Eleven Labs
3 - Artlist
1.Artlist 2. Pika 3. Eleven Labs... Just found this channel and I am super excited to see what you guys are up to over here x
Im really sorry to say this cause Im a comic book and manga reader... but RIP to all of the artists :(
Remember, existing artists can use these tools too to create awesome, consistent stories very quickly.
Nice way to explain and share!
Glad you liked it!
A like before watching as always! The best AI channel on YT
09:30 it's called a duck sound
Haha good call!
The consistent character creation in Midjourney is truly innovative. I've longed for such a feature since before. However, once SORA comes out, it feels like everything will just come to an end. I wonder if there will be many drawbacks, though?😔🙂
Sora is game changing. However u don't think Pika or Runway won't have an answer before or just after do u?
@@WayOfTheZombie. I also primarily used Pika and Runway, but seeing how Sora is far superior, I hadn't considered whether they could keep up with its capabilities. Anyway, it would be great if everyone could show their own distinctive advancements
Remember, SORA may take a few months and at that rate many of the competing tools will not only have advancements but also new features.
@@curiousrefugehmm yes. I think you are right. It looks like there will be a big revolution in artificial intelligence. soon.
If a Sora character turns around or disappears temporarily it comes back with a lot of changes at the moment.
"Explore an idea" sounds like the key to making AI content.
Yes! It's all about experimentation and exploration :)
1. Artlist
2. Pika
3. Eleven
Love this video. Great work Curious Refuge
My competition entry 1. Pika 2. Eleven Labs 3. Artlist
Thanks so much for watching!
What I don't like about Pika is the fact that it's a toss of the dice. If the render does not match the prompt, or if the render is just terribly warped out of shape, too bad... so sad you just tossed credits out the window. But with others like Runway, at least you get the free option to choose an appropriate image to use.
Pika is made by 20 year old kids. They just copy and paste code from GitHub.
Couldn't agree more!! Pika is mentioned everywhere but it's quite disappointing when you see the results
Hkw would that even work long term? This doesn't make sense economically
We appreciate the feedback - we've found these apps do great at different things. We appreciate you engaging with us!
I think a combination of all the tools out there is a way to go. Runway, Pika, Haiper, whatever gets the job done! : )
TIP. You can select two or three icons at the same time. Just hold down SHIFT while selecting. Then the image is at the same time a character, a style and a hint to the image.
That’s a really great tip! I’ll be sure to mention that in a future video.
Please also indicate how to do this on a Windows PC if you do the video.
1:00 that diamond cat tho...
Consistent characters? 😂 Lol The woman in the example if just different in every generation!
Really?! We thought it did a good job!
Thank you!!!
1. Pika 2. Eleven labs 3. artlist
🔥🔥🔥🔥🔥🔥
🔥
Pika is looking like an excellent solution, especially since it's cheaper than runway.
Pika is a great solution!
1. Artlist, 2. Pika, 3. Eleven labs
great like always
Thank you! Cheers!
Character Consistency = Santa read enough letters.
1. Artlist 2. Pika 3. Elevenlabs :)
Damn, sat down to do some work and then CR has a new upload
Everything stops, when a new CR drops
1.Artist
2.Pike
3.11 Labs
pika #1: artlist #2 and elevenlabs #3
i’m impressed you pronounced gaussian correctly… everyone always says “goss-iam” and it makes me cringe 😂
Thank you!
1. Eleven Labs 2. Pika 3. Artist
Hey Caleb, is there a reason I don't have the option to copy link? I only seem to have the option to copy message link, which doesn't allow for me to use this cref command.
To answer my own question: this option is only available if you're using the Discord desktop app, otherwise you have to enlarge the image and select "copy image address".
Correct! In Discord, you need to make sure the image is already inserted into discord, then click on it to expand, then right click to get the copy link option .
Thanks!
the gaze and head angle of character is almost everywhere same though
True! A bit more testing and we can get more expressive generations :)
Very Nice
Thanks!
1. pika labs, 2. Eleven Labs, 3. Artlist
I wish controlling the characters facial expressions was easier too
Try some inpainting :)
are these films, or scenes ?
Short films!
1. Eleven labs 2. Pika labs 3. Artlist
1. pika 2. artlist 3. Eleven Labs
why is pika adding new features (that are terrible anyways) before getting the lipsync feature to a decent level? it is already a very unique feature still in the field. if they get it right, it can sky rocket them ahead of competition.
We agree - with a bit more focus the feature will be a huge selling point. We think they can pull it off :)
1. artlist, 2. pika, 3. 11 labs
why i don't have option to create images via website????
You may have too few generated images to be able to do that. I blieve you need over 1K images now.
@@curiousrefugethanks
@@curiousrefuge is there any way I can add two characters into one photo? For instance, I have an image of a girl and a photo of a cat. How can I use Midjourney to merge these two into one image while ensuring they fit the script? For example, if I want the girl and the cat eating rice.
Sora should atv least let us make Will smith eating spaguetti
Haha we are sure that will come out very quickly once it's generally available :)
Like a knife, a weapon available to everyone is not a weapon anymore it's just a mere tool... #hail_to_the_open_source_and_public_research
We appreciate that perspective!
@@curiousrefuge Thank you 🙏
Pika eleven labs artlist
Very Nice 👍💯🤩🤫🤔🥰😉
We appreciate you watching!
1. eleven labs 2. pika 3. artlist
I’d really like to know what you think about the future of filmmaking… if AI takes over, won’t the original works dry up leaving AI constantly remixing and regurgitating the same source material? I honestly can’t see how this can end well, seeing as studios and financiers are always going to opt for the cheapest solutions. Snake eating its tail comes to mind. It’s going to decimate the film industry unless the paying public choose real people and real footage with their wallets. Personally I can’t think of anything more boring and pointless than watching a film made by AI.
Well, there is room for both (not a zerosum game) right?. For example, like animation, some people believe generative AI will be a subsection. Or, it could be a mix like LiveGen (live action + live generation). We believe ART in all forms will continue to live on and evolve.
1 artist, 2 11labs, 3 pika
1. pika 2. Eleven 3. Artist
Fascinated by youtube channels that makes videos that probably are irrelevant in 3-6 months. Thanks for covering development of AI technology.
We constantly make videos each week because we love and are passionate about this space!
🔥
Right?! This is crazy stuff :)
Is your sync off at 7 minutes?
Thanks for the catch!
to Midjourney Consistence Character - My question ? What actually is a character ? A human being, an animal or just a simple object ?? What I want to say is, that in an AI based project even non-living things like a ship or a building need a certain consistency. Any experience with Objects or something else in conjunction with --Cref ?
Good question in terms of the scope. This will require more testing - but I believe the purpose right now of cref is "character" reference for humans.
@@curiousrefugeThank you. I have tested it in relation to objects and of course it doesn't work. It probably can't work at all yet. But in the end this has to work to let a character move in an environment and interact with things.
Do you know if an AI photo/video generator that can take consumer product imagery as inputs and reimagine them like MJ does with characters? It seems like what's out there doesn't exactly replicate the product/logo.
Good question! For us, we prefer photoshop and inpainting. We believe object reference will be next!
Yea I think an object reference/consistency tool would be very useful.
Can I use this feature in the Discord server Mid journey? and where is Midjourney Alpha?
This is really a game changer.
Yes it's available in both Discord and the website (which is the 'alpha')
1 pika
2 eleven labs
3 artist
this is a real ground breaker, can you tell us if Midjourney is able to create consistent objects in a set?
Object reference isn't available yet but we believe it will be soon!
Where do I get the app?
What is the tool you mention at 18:08?
LensGo!
❤❤❤❤❤
chapter 12 says "Same Altman" in case that was an accident.
Same ‘ole Sam haha
Model agencies and models will be replaced…..
They CAN be replaced but that doesn’t mean they will be. I’m really hoping the public will always want authentic real people in their media space. Human to human connection is undervalued.
It will be interesting to see how it plays out. There will certainly be some models but not all.
When I use the consistent character option using a selfie of me, it doesn't seem to work as well as it did for Caleb because it generates images of characters that don't look like me. I mean, am I really that ugly jeez haha😂
not sure if anyone had the same experience?
It’s so weird because some peoples are more easily replicated. My gut tells me that Midjourney is creating its own image of you, then it creates the clones.
Try a reference photo with flatter lighting. That might help :)
Eleven labs 2. artist io 3. pika
That hipster looks a lot like you. How many tries did it take to get you in there? My face isn't coming out too well. blend or the new Alpha site?
hahah! Well I appreciate you watching. What prompt did you use?
@@curiousrefuge With V6 I've used sref, weights "::3" and ce 0-3 with multiple variations of blend. Ive lost count. Ive tried different things on and off for a month. Maybe I'm too picky.
Could it be that Artlist has hired some fiverr promo to comment an obviously fake top 3 here?
No but that's a funny idea :)
"Consistent"
huh?
@@curiousrefuge It is a bit tiring that many apps talk about consistency, when the morphology changes from a little to a lot in each generation. Let's not talk about the clothes. Semi-consistent would be an acceptable expression, at least it wouldn't be a blatant lie.
now everyone can create a fake Instagram influencer...
So true…
@@curiousrefuge My brother does too.🥴😞😓
Oh the horror...
Keep scrolling like you do today.
What changes?
@@jd2161 reality changes. or rather stops.
sepia is pronounced "see-pee-uh"
Woops! We appreciate the correction :)
Er...herm. None of the MJ character outfits you showcased were the same as the source image. Fine for concept art, maybe, but let's not get carried away with hype just yet.
With a bit more testing we are able to get very similar outfits (so long as they aren't too complex)
@@curiousrefuge OK but don't p!$$ on our leg and tell us it's raining.
OLD MAN YELLS AT CLOUD: I do not like your Game of the Week. And you never give the results. You should just give the results at the end of the video (forget the free swag gimmick) or forget the game and just tell us which is which so we can make an informed comparison. Yes, people love free crap, but no one is bringing their friends here because of the GotW.
If you really want a give-away (in exchange for comments), just tell people to make a specific comment and you'll pick a random winner, instead of creating stressful secrets for all the rest of your fans.
I do otherwise love your channel and Film of the Week and appreciate all your reporting!
We appreciate your engagement and comment. We'll definitely take your feedback and reconsider how we do things. We certainly don't want this to be 'stressful'. Thanks so much for the feedback and wathcing!
the idots at hollywood signed a contract to not use ai. But independents will smash that and make good ai films that aren't constrained by the hollywood strike contract. All actors need to be completely replaced by ai. period..
Have fun in your artificial cartoon world. I’ll always prefer real people over cartoons (which is what AI generated images are, let’s be honest)
holly wood is bad. that is all you have to know. And if you pretend yo udon't know, then you are an apologist for hollywood, which means you are summarily dismissed from all forms of credibility. take your pick. either way, you are exposed.@@jamesr6562
Hell no
We tend to agree that Hollywood not taking advantage of this is a likely not a smart and long term move. If they embrace it they can create some amazing things!
pika is one of the worse tools I highly don't recommend to use it junk
They all have their pros and cons. We appreciate you watching!
renaissance? thats barock rokoko style outfit. its centuries after renaissance. why is midjourney not capable to differ styles anymore?
It's capable of any style! What do you mean?
1 - Pika
2 - Eleven Labs
3 - Artlist