The real test of any ai image generator should be consistent characters. While single image renderings are good test but ultimate test is only character consistency and that should define the value of any ainimage generator because that leads to the ultimate use of any visual medium as it's ability to be used in storytelling. So please help us know which one is best for creating consistent characters.
Yeah its flux and considering flux can make videos out of still pictures too its more flexible than lets say midjourney even tho lot of people seem to like the style mj puts out for its pictures. MJ used to be the best but nowdays people want to have more so id say flux is the new king
Completely agree. So far Flux is the only one doing this. I have have had some success with custom GPTs that I have trained on character sheets and detailed instructions but there are still big continuity errors and DALL-E’s art styles are way too limited. I would like to see more of these platforms focusing on consistent characters
I think you have to run each prompt 3 times and then look at the avg for each platform. You never know when one platform will give you a stinker of an image. Doesn’t mean the platform is more or less preferred. I think you’d have a more accurate assessment.
"Wow, Recraft Version 3 really seems like a game-changer in the AI art space! The ability to handle detailed prompts, realistic styles, and even create SVG designs makes it a powerful tool for artists and designers alike. I love how you tested it against other AI generators, especially with famous figures and dynamic action scenes. Great breakdown and insights, can't wait to see how this tool evolves!"
Thank you so much for another very useful video. It’s wonderful to hear your opinion too; it’s very valuable. I loved how you developed this lesson. I’m so grateful for everything! I’m already part of your academy and newsletter. Loving this more elegant, black setup with your new hair!
Indeed. I tried photorealism last night and wasn't impressed at all. Being able to generate and save logos, or whatever, in SVG without the need to trace an image is a big time saver. That's what I'll use it for.
Great comparison test. I don’t think it’s about which one is better but more about the image style of the output from the same prompt. I like them all but are different. Thanks
I tested out Recraft V3 and found that it was impossible to get a character to look directly at the camera. I do like the the idea of it not generating that standard look by default, but it's disappointing that I can't specify it to do so. Even if I upload a reference image, it turns the head slightly smh. Maybe you can test things out and find a resolution to this issue.
Funny how Scarlet Johansson in Wonder Woman suit makes the suit looks like a Marvel one instead of a DC comics one for Midjourney (looking more like the suit of Captain Marvel than the one of Wonder Woman), the dataset of Midjourney influences that to the point it's no longer Wonder Woman at all (but got Scarlet Johansson almost right though...) guess Black Window (Johansson) is really tied to Marvel in their dataset. Other than that, feel like models generally struggle real hard with freckles and imperfections
I'd argue that Recraft did what you asked because you said the Sea Horse should be on the left of her, not the left of the screen. All 3 of the images show the character basically facing the sea horse, however, in Mid journey her right shoulder is more forward, but in the Recraft and Flux images her left shoulder is more forward. I'd give it to Flux, Recraft and then Midjourney, especially since Midjourney didn't really draw a sea horse.
Is there already a Ai app where i can import/feed the app with a collection of only one and the same full 3D figure/model, so after that I can use it in any form and poses?!
What these comparison leaderboards forget to incorporate is how much control you can have over the outcome in the sense of creating consistent characters and various in-painting features. Especially for storytellers and as a pre-production for video generation, that is very important. I have a particular character created in MJ, but it's hard (or impossible) to re-create or adjust that character on other platforms.
Thank you for your video, it was, as always, interesting! :) I looked at your figma and it seems that you put the same prompt for the fluffy animal as in the car prompt... ;)
I've tried every one of the image generators and for photorealistic faces, nothing compared to CgDream. Now if I could just get Midjourney to accept these incredible images as a character reference I'd be on cloud 9.
I tried ideogram with your prompt and was blown away. I used "Ornate signage in the spirit of a pirate ship that says "I am the master of my fate and the captain of my soul." It was perfect -- flawless!
@@goliathonscave9834 ya i did say imo heh but that said a lot of us in the ai community are 2nd guessing the rankings..youll see some vids coming out discussing that. but either way, dig the breakdown of these 3
It's difficult to compare different models with a single image or a single prompt, because the optimum wording for one model is not always the best for another. Plus some models have parameters the others don't have that can improve a specific result.
I think you will agree, MidJourney is clearly the best for pure artistic style, if you want to create art, MidJourney is the leader. Like I have said before I really hope MidJourney never tries to be like the others. I have no connection to MidJourney but nothing even comes close to the aesthetics and artist stylized outputs you can get. Now MidJourney is not even close when it comes to real life normal photos, or cartoons, text, etc... MidJourney is like an artist, more creative, more ways to be abstract, with unlimited diverse styles you can generate over many models. But I will certainly add recraft to my growing toolkit. Art and style MidJourney is the leader. but poor on text and realism or normal photos. Text and realism 1. Ideogram, 2. Flux 1.1 pro, 3. FreePik, now looking at Recraft. others I use much less lately Dell-E, and Leonardo
The woman with the colourful seahorse, from her point of view because of the seahorse being slightly further away (as for being out of focus a little), the seahorse is actually and literally ¨to the left of her¨. If the seahorse should have been on the left side of the picture, next to her, that would need other wording. Otherwise I like the video very much!
I believe Scarlett Johansson might not be a good example, after OpenAI included her voice, they got sued. So I can imagine certain projects being wary of her in particular.
Recraft has many models from which you can create different styles. You need to do a "real" comparison though with these different models in order to make a true comparison!
Id give a big minus points for MJ on that scarlet johansson because shes clearly not wearing the prompt you asked for. Id say flux wins that round just for costume accuracy and its the most cinematic anyway.
It is quite unfair to compare these products using them as is, with default out of box settings. Midjourney has lot of finetuning keys like sref, personalization, chaos, stylize and what not. Same applies to Flux Dev model, it is quite versatile, though also limited in many ways (camera controls, artistic styles, stereotype characters, etc.), but you can greatly expand it's potential with custom workflows and Loras.
Recraft is much more realistic. Midjourny creates images that are very much AI. The image of the Asian person in a vest was a great example. It was also the only one that actually had a vest on.
tell me one, that accepts image references, upscale with variations, variations in specific areas, character ref, using an image as a complete reference, and more.@detective2221
On the contrary (sorry) but the interface isn't very user friendly, and it produces some of the strangest results. Such as when prompting a simple male or female portrait they're always staring up into the sky like they're looking at a passing helicopter, or posed strangely like they're peering around a corner. Weird. Midjourney still seems like the best, now that Playground no longer has canvas and "fix faces" :(
Great video but there's one thing in relation to the female and the seahorse. Technically RECRAFT is the only correct image, the seahorse is to her left, as she standas the other 2 are to the left of the frame which is in fact a fail :) IMO that is!!!
Recraft does the best celebrity faces ever seen at a close up to medium close-up range however it completely throws out the resemblance beyond that which is extremely annoying unfortunately
I find the midjourney images of people very unrealistic. The skin texture in particular always looks like plastic, nothing like a real person, even when within the parameters of "cinematic". They scream to me "AI generated". Nice on style, very weak on realism.
@@TPCDAZ Yeah, I get that. But I came here for AI and Tech info, not politics. But creators just can't do anything without making it political. It's the same reason no one watches MSM & he is bringing it to the channel for no reason.
@@mikefiatx19 Actually you made it political. He just chose a famous face to include, having watched the video he never once mentioned politics. Would you have been up in arms about this is it was Kamella Harris? I doubt it. SO you made it political, not everyone cares about politics enough to lose your head over. Who cares who wins, all changes again in four years anyway and the next liar can come in. Just enjoy the video and leave the politics to the politicians.
tried it - no matter what I did it repeatedly put glasses on my character and apparently 30 is the new 60 or vice versa - whatever, my 30 year old character looked like a retiree - POS
Tested this, useless for any serious creative work. Both enhancing and compositing, even in the setting "almost identical" creates absurd results, ignoring prompts, texture and composition. How can this mess possibly be ranked ahead of Flux Pro etc?
As young girls navigate the vibrant and often intricate landscape of their formative years, they find themselves exploring a world filled with fascination and allure. This period is marked by a blossoming curiosity about relationships and the dynamics of attraction, where they begin to discover what captivates their interest and ignites their imagination. In this enchanting journey, discussions about physical attributes and personal preferences can take on a playful yet meaningful tone. Young girls may find themselves drawn to the idea of attraction in a way that sparks excitement and wonder, as they learn to express their feelings and desires. This exploration is not merely superficial; it lays the groundwork for understanding deeper connections and the importance of mutual respect. Embracing these topics with a sense of fun and openness allows for a rich dialogue that can empower young girls. By creating a safe space for these conversations, we encourage them to explore their identities and boundaries, helping them navigate the complexities of relationships with confidence and curiosity. This thoughtful approach nurtures their self-awareness and appreciation for both themselves and others, leading to fulfilling interactions in their journey ahead.
Recraft is over censored garbage. It will not let me draw a simple political cartoon. When it does draw something it is the complete opposite of what I asked it to do.
As older women navigate this vibrant phase of life, their connection to the natural world often deepens, particularly with the fascinating array of birds that inhabit their surroundings. Among these, the larger bird species can evoke a sense of wonder and admiration, embodying strength and beauty that resonates with many. Big birds, such as majestic swans, regal peacocks, and striking roosters, play significant roles in both the ecosystem and human culture. For instance, swans glide gracefully across lakes, their elegant movements and striking white feathers captivating onlookers. These birds symbolize grace and tranquility, often inspiring those who encounter them to reflect on the beauty of nature and the serenity it brings. Peacocks, with their dazzling plumage and intricate feather patterns, represent the splendor of self-expression. Their vibrant colors and elaborate displays during courtship rituals can captivate the imagination, reminding us of the importance of embracing one's uniqueness and confidence. Older women may find themselves drawn to these birds not just for their beauty, but for the messages they convey about individuality and self-acceptance. In more pastoral settings, the presence of large cockerels can add a lively dynamic to the landscape. These birds are known for their vibrant colors and robust personalities, often serving as the spirited guardians of the flock. Their morning calls herald the dawn, marking the start of a new day and infusing the atmosphere with energy and vitality. Engaging with these big birds can foster a sense of connection to the rhythms of rural life and the cycles of nature, encouraging a deeper appreciation for the simple joys found in everyday experiences. As older women continue to explore their interests and passions, the world of big birds offers a rich tapestry of beauty, inspiration, and connection to nature. Whether through gardening, birdwatching, or simply enjoying the sights and sounds of their feathered friends, these experiences can enhance their sense of well-being and contribute to a fulfilling, enriched life.
Thanks for this analysis, but it is irritating to have your face on top of the images that you compare. Next time make yourself smaller and leave the screen to what's important.
The real test of any ai image generator should be consistent characters. While single image renderings are good test but ultimate test is only character consistency and that should define the value of any ainimage generator because that leads to the ultimate use of any visual medium as it's ability to be used in storytelling. So please help us know which one is best for creating consistent characters.
simple it's flux since you can create loras.
@@thibaudherbert3144even if you use lora it's not 💯% so yeha we need an ai which can produce consistent character , poses and different expression
Yeah its flux and considering flux can make videos out of still pictures too its more flexible than lets say midjourney even tho lot of people seem to like the style mj puts out for its pictures. MJ used to be the best but nowdays people want to have more so id say flux is the new king
Agreed.
Completely agree. So far Flux is the only one doing this. I have have had some success with custom GPTs that I have trained on character sheets and detailed instructions but there are still big continuity errors and DALL-E’s art styles are way too limited. I would like to see more of these platforms focusing on consistent characters
I think you have to run each prompt 3 times and then look at the avg for each platform. You never know when one platform will give you a stinker of an image. Doesn’t mean the platform is more or less preferred. I think you’d have a more accurate assessment.
"Wow, Recraft Version 3 really seems like a game-changer in the AI art space! The ability to handle detailed prompts, realistic styles, and even create SVG designs makes it a powerful tool for artists and designers alike. I love how you tested it against other AI generators, especially with famous figures and dynamic action scenes. Great breakdown and insights, can't wait to see how this tool evolves!"
Impressed by Redraft, a lot. Every shoot feels more realistic. In Midourney best model, the skin ever looks like plastic to me.
Thank you so much for another very useful video. It’s wonderful to hear your opinion too; it’s very valuable. I loved how you developed this lesson. I’m so grateful for everything! I’m already part of your academy and newsletter.
Loving this more elegant, black setup with your new hair!
I think the most important feature of Recraft is that it can generate vector images or vectorize loaded or created images and download them as SVG....
Indeed. I tried photorealism last night and wasn't impressed at all. Being able to generate and save logos, or whatever, in SVG without the need to trace an image is a big time saver. That's what I'll use it for.
Firstable where do you access the Recraft?
aitutorialmaker AI fixes this (AI driven Tutorials). AI art model beats Midjourney.
Great comparison test. I don’t think it’s about which one is better but more about the image style of the output from the same prompt. I like them all but are different. Thanks
Great tutorial! as always ;-)
Question: How can we make consistent characters for a children's picture book?
I tested out Recraft V3 and found that it was impossible to get a character to look directly at the camera. I do like the the idea of it not generating that standard look by default, but it's disappointing that I can't specify it to do so. Even if I upload a reference image, it turns the head slightly smh. Maybe you can test things out and find a resolution to this issue.
is there option how to create consistence figure or character based on one image
Why is ideogram not among the comparison models? It's up there among the very best.
He checked the list and used the top 3. Ideogram was 4th.
Funny how Scarlet Johansson in Wonder Woman suit makes the suit looks like a Marvel one instead of a DC comics one for Midjourney (looking more like the suit of Captain Marvel than the one of Wonder Woman), the dataset of Midjourney influences that to the point it's no longer Wonder Woman at all (but got Scarlet Johansson almost right though...) guess Black Window (Johansson) is really tied to Marvel in their dataset.
Other than that, feel like models generally struggle real hard with freckles and imperfections
Thank you, I am going to check out Recraft!
I'd argue that Recraft did what you asked because you said the Sea Horse should be on the left of her, not the left of the screen. All 3 of the images show the character basically facing the sea horse, however, in Mid journey her right shoulder is more forward, but in the Recraft and Flux images her left shoulder is more forward. I'd give it to Flux, Recraft and then Midjourney, especially since Midjourney didn't really draw a sea horse.
A couple of months ago you did a video with a title that said Midjourney was "way too censored". Have you run into any censorship issues with Recraft?
this is the best your hair has looked. thanks for the informative vid!
Is there already a Ai app where i can import/feed the app with a collection of only one and the same full 3D figure/model, so after that I can use it in any form and poses?!
Did you use the new version of Flux?
A great and helpful analysis
And in comparaison with Leonardo ?
What these comparison leaderboards forget to incorporate is how much control you can have over the outcome in the sense of creating consistent characters and various in-painting features. Especially for storytellers and as a pre-production for video generation, that is very important. I have a particular character created in MJ, but it's hard (or impossible) to re-create or adjust that character on other platforms.
Why can't I clip your latest video to share it?
Thank you for your video, it was, as always, interesting! :) I looked at your figma and it seems that you put the same prompt for the fluffy animal as in the car prompt... ;)
I've tried every one of the image generators and for photorealistic faces, nothing compared to CgDream. Now if I could just get Midjourney to accept these incredible images as a character reference I'd be on cloud 9.
@detective2221 ask Ai :p
@detective2221very happy
@detective2221learn English
ideogram missing from the testing? easily the best rn (imo)
maybe recraft vs ideogram video is in order
I tried ideogram with your prompt and was blown away. I used "Ornate signage in the spirit of a pirate ship that says "I am the master of my fate and the captain of my soul." It was perfect -- flawless!
For real.
He discussed the top 3 rated. Ideogram is 4th so it is only "easily the best" by your own bias.
@@goliathonscave9834 ya i did say imo heh but that said a lot of us in the ai community are 2nd guessing the rankings..youll see some vids coming out discussing that. but either way, dig the breakdown of these 3
Loving the Fash-cut. But the moustache needs to be a little narrower :D
The best image generator is flux 1.1 for real photorealism and midjourney is the best for consistent character
Bruh this AI technology is hitting different, fr fr it's giving main character energy and making other models look straight mid, no cap!🔥🔥🔥🔥🔥
It's difficult to compare different models with a single image or a single prompt, because the optimum wording for one model is not always the best for another. Plus some models have parameters the others don't have that can improve a specific result.
I think you will agree, MidJourney is clearly the best for pure artistic style, if you want to create art, MidJourney is the leader. Like I have said before I really hope MidJourney never tries to be like the others. I have no connection to MidJourney but nothing even comes close to the aesthetics and artist stylized outputs you can get. Now MidJourney is not even close when it comes to real life normal photos, or cartoons, text, etc... MidJourney is like an artist, more creative, more ways to be abstract, with unlimited diverse styles you can generate over many models. But I will certainly add recraft to my growing toolkit.
Art and style MidJourney is the leader. but poor on text and realism or normal photos.
Text and realism 1. Ideogram, 2. Flux 1.1 pro, 3. FreePik, now looking at Recraft.
others I use much less lately Dell-E, and Leonardo
Good video. Thx 🎉
The woman with the colourful seahorse, from her point of view because of the seahorse being slightly further away (as for being out of focus a little), the seahorse is actually and literally ¨to the left of her¨. If the seahorse should have been on the left side of the picture, next to her, that would need other wording. Otherwise I like the video very much!
where is the consistent character feature? how it beats midjourney ??
I believe Scarlett Johansson might not be a good example, after OpenAI included her voice, they got sued. So I can imagine certain projects being wary of her in particular.
great video, thanks
Recraft has many models from which you can create different styles. You need to do a "real" comparison though with these different models in order to make a true comparison!
Recraft has a new text mode for long text witch is not just a normal prompt...
You said that the sea horse should be to '...her left.' It was to our right but to her left.
Id give a big minus points for MJ on that scarlet johansson because shes clearly not wearing the prompt you asked for. Id say flux wins that round just for costume accuracy and its the most cinematic anyway.
It is quite unfair to compare these products using them as is, with default out of box settings. Midjourney has lot of finetuning keys like sref, personalization, chaos, stylize and what not. Same applies to Flux Dev model, it is quite versatile, though also limited in many ways (camera controls, artistic styles, stereotype characters, etc.), but you can greatly expand it's potential with custom workflows and Loras.
Recraft is much more realistic. Midjourny creates images that are very much AI. The image of the Asian person in a vest was a great example. It was also the only one that actually had a vest on.
I think some of these models avoid exact replicas of celebrities to avoid lawsuits.
nice new hair style
TBH, I liked the Midjourny images in most comparisons.
tell me one, that accepts image references, upscale with variations, variations in specific areas, character ref, using an image as a complete reference, and more.@detective2221
@detective2221like what
cartoonish more like it
its model hidden behind a paywall, trial period available
at least sd models are freeware
On the contrary (sorry) but the interface isn't very user friendly, and it produces some of the strangest results. Such as when prompting a simple male or female portrait they're always staring up into the sky like they're looking at a passing helicopter, or posed strangely like they're peering around a corner. Weird. Midjourney still seems like the best, now that Playground no longer has canvas and "fix faces" :(
Recraft doesn't let you use image references (that I know of) so it automatically loses.
Midjourney's censorship is unhinged.
Самсону привет!
hiya
0:25 the hand holding the baby looks real bad, not talking about it and scrolling fast by makes me to leave
I have no desire to have a service for my generations. So flux is still king for me.
Great video but there's one thing in relation to the female and the seahorse. Technically RECRAFT is the only correct image, the seahorse is to her left, as she standas the other 2 are to the left of the frame which is in fact a fail :) IMO that is!!!
Recraft is pretty bad at following my prompts. Ideogeam considerably better. I like the menu and canvas approach though
Who actually even uses midjourney? Stable Diffusion XL and pony are so good
i love these types of videos! keep it up
Clearly a MJ boi 😅
OMG...you just start messing around with it and trigger a moderation alert. Recraft can go to hell. I'm really tired of the censorship.
Recraft does the best celebrity faces ever seen at a close up to medium close-up range however it completely throws out the resemblance beyond that which is extremely annoying unfortunately
Its not really free, now is it, its limited to daily trails, not free endless.. Stop doing click bait.
Mid... what.... Never heard of her.. Does she have a great free tier and plans..?????
I find the midjourney images of people very unrealistic. The skin texture in particular always looks like plastic, nothing like a real person, even when within the parameters of "cinematic". They scream to me "AI generated". Nice on style, very weak on realism.
Hiii❤ watching👀.. Now
So Midjourney wins most but the title says it beats MJ? that's what we call clickbait.
Recraft gladiator wins by default for no morphing
I found it to be garbage, for now no one will beat Midjourney in realism, and no one will beat Flux in accuracy
Title said "free" for a closed-source paid model. Clickbait? Isn't that a straight up lie? 🧐
Video could have been made without being political but you had to throw trump into it. Well done pissing off half your followers for no reason.
People have different political opinions. Get over it
Take care my sensitive little man
IDK if he pissed off half of his followers, nobody else seems upset about this lol
@@TPCDAZ Yeah, I get that. But I came here for AI and Tech info, not politics. But creators just can't do anything without making it political. It's the same reason no one watches MSM & he is bringing it to the channel for no reason.
@@mikefiatx19 Actually you made it political. He just chose a famous face to include, having watched the video he never once mentioned politics. Would you have been up in arms about this is it was Kamella Harris? I doubt it. SO you made it political, not everyone cares about politics enough to lose your head over. Who cares who wins, all changes again in four years anyway and the next liar can come in. Just enjoy the video and leave the politics to the politicians.
tried it - no matter what I did it repeatedly put glasses on my character and apparently 30 is the new 60 or vice versa - whatever, my 30 year old character looked like a retiree - POS
Tested this, useless for any serious creative work. Both enhancing and compositing, even in the setting "almost identical" creates absurd results, ignoring prompts, texture and composition. How can this mess possibly be ranked ahead of Flux Pro etc?
Meh.. Cloud services😒
Why even bother doing these tests if you keep pickign Midjourney even when its giving far worse result.
To be honest I did not like it - after half a dozen tries I did not like anything it produced
As young girls navigate the vibrant and often intricate landscape of their formative years, they find themselves exploring a world filled with fascination and allure. This period is marked by a blossoming curiosity about relationships and the dynamics of attraction, where they begin to discover what captivates their interest and ignites their imagination.
In this enchanting journey, discussions about physical attributes and personal preferences can take on a playful yet meaningful tone. Young girls may find themselves drawn to the idea of attraction in a way that sparks excitement and wonder, as they learn to express their feelings and desires. This exploration is not merely superficial; it lays the groundwork for understanding deeper connections and the importance of mutual respect.
Embracing these topics with a sense of fun and openness allows for a rich dialogue that can empower young girls. By creating a safe space for these conversations, we encourage them to explore their identities and boundaries, helping them navigate the complexities of relationships with confidence and curiosity. This thoughtful approach nurtures their self-awareness and appreciation for both themselves and others, leading to fulfilling interactions in their journey ahead.
Recraft is over censored garbage. It will not let me draw a simple political cartoon. When it does draw something it is the complete opposite of what I asked it to do.
As older women navigate this vibrant phase of life, their connection to the natural world often deepens, particularly with the fascinating array of birds that inhabit their surroundings. Among these, the larger bird species can evoke a sense of wonder and admiration, embodying strength and beauty that resonates with many.
Big birds, such as majestic swans, regal peacocks, and striking roosters, play significant roles in both the ecosystem and human culture. For instance, swans glide gracefully across lakes, their elegant movements and striking white feathers captivating onlookers. These birds symbolize grace and tranquility, often inspiring those who encounter them to reflect on the beauty of nature and the serenity it brings.
Peacocks, with their dazzling plumage and intricate feather patterns, represent the splendor of self-expression. Their vibrant colors and elaborate displays during courtship rituals can captivate the imagination, reminding us of the importance of embracing one's uniqueness and confidence. Older women may find themselves drawn to these birds not just for their beauty, but for the messages they convey about individuality and self-acceptance.
In more pastoral settings, the presence of large cockerels can add a lively dynamic to the landscape. These birds are known for their vibrant colors and robust personalities, often serving as the spirited guardians of the flock. Their morning calls herald the dawn, marking the start of a new day and infusing the atmosphere with energy and vitality. Engaging with these big birds can foster a sense of connection to the rhythms of rural life and the cycles of nature, encouraging a deeper appreciation for the simple joys found in everyday experiences.
As older women continue to explore their interests and passions, the world of big birds offers a rich tapestry of beauty, inspiration, and connection to nature. Whether through gardening, birdwatching, or simply enjoying the sights and sounds of their feathered friends, these experiences can enhance their sense of well-being and contribute to a fulfilling, enriched life.
Thanks for this analysis, but it is irritating to have your face on top of the images that you compare. Next time make yourself smaller and leave the screen to what's important.
Midjourney is a woke joke
STOP USING AI ART LAMES
they all look the same. it's boring.