ฝัง
- เผยแพร่เมื่อ 8 พ.ค. 2024
- I put two of the best AI art generators to the test, to see which one is better for you.
We look at features, prompts comparison, speed test and much, much more.
Here are my images:
www.figma.com/file/6lrMz5ONAw...
📧Join my newsletter
delightfuldesign.eo.page/w7tf5
---
👨🏫Check out my AI courses:
www.udemy.com/user/samson-vow...
---
💼Business inquiries
samsonvowles@gmail.com
---
--- My top resources:
📖 Grab My AI Secrets! Dive into my handpicked prompts e-book.
delightfuldesign.eo.page/6sw6r
🎨 Unleash Your Creativity! Experience Adobe’s cutting-edge AI tools.
tinyurl.com/2sj2kwk6
🤖 Boost Your Efficiency! Automate with ClickUp.
tinyurl.com/4p9yxew5
🎤 Find Your AI Voice! Clone your voice with precision.
elevenlabs.io/?via=b38559
🎥 Revamp Your Content! Edit shorts effortlessly with AI.
www.opus.pro/?via=7fcbf8
✍️ Supercharge Your Writing! Generate content the smart way.
www.jasper.ai/?fpr=samson
📹 AI-Powered Videos! Elevate your content with Kaiber.
kaiber.ai/?via=samson
📢 Ads Made Easy! Create compelling ads in a flash.
free-trial.adcreative.ai/2yui...
👤 Stand Out Online! Craft your unique AI avatar.
www.synthesia.io/?via=delight...
🎬 Level Up Your Channel! Get coached by AI on TH-cam growth.
vidiq.com/samsonvowles
🖥️ Present Like a Pro! AI-driven slides for impactful presentations.
slidesai.io/?via=samuel-vowles - วิทยาศาสตร์และเทคโนโลยี
🎯 Key Takeaways for quick navigation:
00:56 🔄 *Mid Journey excels in changing parameters, fine-tuning algorithms, and managing works in its gallery. DALL•E operates mainly in chat GPT form with a standalone website version.*
02:57 🤖 *Mid Journey lacks an API, while DALL•E has an available API for users.*
03:11 🎨 *In a prompt battle, Mid Journey tends to offer more refined details and nuanced outputs, while DALL•E produces more saturated and broader-colored images.*
05:58 🧩 *Mid Journey outperforms DALL•E in creating seamless geometric patterns for elegant wallpapers.*
06:11 🖌️ *When designing logos, Mid Journey wins for better incorporation of diverse visual elements, although improvements are suggested for both platforms.*
07:20 🤖 *In the Pixar animation style, Mid Journey provides a more characterful and emotive robot, while DALL•E's output is more saturated and feels slightly childish.*
08:48 🎨 *Mid Journey's AI influencers appear more realistic and natural compared to DALL•E, which tends to produce oversaturated and less realistic images.*
13:12 🐉 *In an illustrated fantasy scene with a dragon, DALL•E adheres more closely to prompt specifics, but Mid Journey's composition is considered more interesting.*
14:45 🎌 *Mid Journey's tattoo design of a Phoenix in Japanese style feels more authentic, while DALL•E's version has a more graphic and symmetrical style.*
16:49 🦊 *In anthropomorphic animal illustrations, DALL•E tends to produce whimsical and childish images, while Mid Journey creates more dramatic and emotive scenes.*
18:25 🌌 *Mid Journey's surreal landscape with floating islands is preferred for its balanced composition and visual integrity over DALL•E's more scattered approach.*
20:58 🕍 *In detailed Gothic Cathedral illustrations, both Mid Journey and DALL•E showcase strengths, with Mid Journey offering a more atmospheric feel.*
23:36 🏠 *Mid Journey's approach to futuristic interior design prompts is refined and demonstrates thoughtful composition, while DALL•E tends to be more saturated and less considerate of overall composition.*
24:16 🤖 *Darly's unique strength lies in collaboration with ChatGPT, allowing interactive design discussions and iterative improvements.*
26:23 🕒 *Mid Journey's processing speed is comparable to Darly, and recent updates make it 2.7 times faster than the original V6 version.*
27:17 🚫 *Both platforms have content limitations, with Darly being more censored, restricting images of famous people or using artists as inspiration.*
28:38 💬 *Mid Journey boasts a more active and supportive community on Discord compared to Darly's evolving community.*
29:32 💲 *Pricing varies; Darly comes with ChatGPT Plus for $20/month with unlimited images, while Mid Journey starts at $10/month with GPU time restrictions and offers higher tiers for more features and GPU time.*
30:52 🎨 *Mid Journey leans towards Ultra realism and artist control, while Darly excels in specific object relationships, Illustrated, or painted styles and benefits from collaboration with ChatGPT inside ChatGPT Plus.*
Made with HARPA AI
Excellent observations! Truly insightful. I completely concur with the comparison.
Very professional video, really fleshed out the details on this comparison. Great!
Excellent analysis of these tools. I understand them better than ever.
Very good job with comparing the 2 tools. The commentary was top notch and subjectivity very fair.
no.
love the runway green screen removal background. Also just image adding the motion brush AI tool features to theses pics. Amazing
One of the best videos, THANK you
Thanks for the comparisons. I apricate your content.
Thanks for watching!
I like them both. they both serve a purpose and have a different angle which benefits one idea vs another which opens my mind to even more ideas and creativity
this is really helpful contents for making a decision. Great Job! Thank you very much!!!
Best comparison video ever!
Thanks for creating this great video! I’ve been using MJ5, just started using 6, the coherence of MJ is frustrating but I’m hopeful it’ll improve, the quality of the imagery is just incredible! And it’s improved with v6!
I like your prompt style, very descriptive and detailed, makes me imagine it in my mind’s eye! Thanks
Glad it was helpful!
I liked this type of video. Very interesting. And great commentary
Thanks for this, I appreciate it
A useful comparison, efforts appreciated!
Glad it was helpful!
Thanks for taking the time to make this video! I'm deciding between Midjourney and ChatGPT Plus. Might just get both! 😆
Yes! I have both
get both. dalle is amazing for illustrations and for refiining what you want over many generations with the help of gpt4, but it cant hold a candle to mdjourneys photorealism and pure splendor. its worth the combined 50 or 30 a month!
To sum up:
MJ makes photos
Dall-e makes paitings
Great comparison. I've been using MidJourney for about a year now. I've generated close to 40,000 images. As someone who is a perfectionist and expects depth and granularity from an AI art generator, I have to say that I have a strong preference for MJ. DALL-E is like Corel Draw where as MJ is like Photoshop. These image generators are both mind blowing and deserve due credit for the monumental accomplishments they're making.
I started doing this last April using Bing at first, then Night Cafe. I avoided Midjourney for several months just because of discord. I like doing more artsy stuff but loved Midjourney 5.2 for everything else. Dalle-3 refuses to do anything that could be called artsy, and Midjourney 5.2 had some limitations in that area. But Midjourney 6 is amazing. The only area where Dalle-3 beats Midjourney 6 is with complex subject orientation, but you have to work with ChatGPT to really get the benefit from that. I do love using ChatGPT for writing prompts, and I have built some custom GPTs, but Dalle-3 is not reason enough to justify having a plus account. I am using Midjourney 6 for over 95% of my artwork now.
MJ is nothing like photoshop? DALL-E is blindly getting an art commision and MJ is supervising it, you have no control in the end
Both Mj and DALL-E have distinct 'styles' IMO, a lot of the time you can tell by looking at the images which tool they were generated in (though not always). I suppose that is just the way with AI tools at the minute.
Great! I like both, and I often ask GPT for ideas and I compare prompts from GPT in both DALL-E and Midjourney. Furthermore I'm starting to use Stable Diffusion through Pinokio
Nice content man, very precise. Thank you
Glad you enjoyed it!
I think your videos are really good, I really enjoy them. Can I maybe ask you some questions?@@aisamsonreal
Hi, I want to know how you made the three-dimensional effect on the picture at 0:03 at the beginning of your video. It looks cool.
I don't know what did he use but I know this AI tool that is free for 3D effect on images: LeiaPix.
I like the depth of field in Midjourney presentation. The characters are more realistic and true to life and Dall E is more illustrative. I am a older person and seem to like what I see in Midjourney a little bit more but I do see that Dall E seems to follow instructions better than Midjourney
Well done comparison and critique! Looking forward to checking out your other videos. If you do more of these side by sides in the future (please?!), I would just say maybe you could have the prompt text below the side by side images as you are comparing them, not just before you show them.
Awesome, thank you!
Great video thanks. I’m doing your children’s picture book course too, where you use Midjourney (when it was free) for that. Seeing as now I’ll have to pay to use AI art, would you say to pay for Midjourney or Dall-e? It seems as if Dall-e might be better for children’s art anyway in this case? Especially if I need certain characters to interact.
Superb reaction - your words had me laughing with their sensitivity and spot on commentary
Thank you so much 😁 I appreciate that
I would say that MJ is much better when it comes to photorealism, but I would give Dall-e the edge in illustration. One big difference between the two is that Dall-e takes the user's prompts very literally, and MJ tries to interpret the user's intent. A good example was the Pixar style robot image. While he preferred the MJ image, it really wasn't what he asked for in the prompt. He asked for the image to be reminiscent of Pixar animation style, and the Dall-e image was far more Pixar-like. It looks as though it could be a still from a pixar movie. The character is cute and marketable. The background, in particular, is much more like what you'd see as a background scene in one of their movies. He also asked for the robot to be shown in a playful manner, and the setting to be bright and cheerful. Dall-e followed those prompts, whereas the MJ image had a more serious tone with more muted colors. Merely comparing image to image, I like the style of the MJ creation over the Dall-e alternative, but when comparing the images to what the prompts asked for, MJ strayed while Dall-e stuck to them. As I said, when it comes to photorealism, MJ is the hands down winner, but when it comes to illustrations, Dall-e gives you what you ask for. He seemed impressed by the rust and weathering of the MJ robot, but that was an MJ decision, it wasn't in the prompts.
The whole Midjourney community and the company itself, are easily the best. They have excellent support, seem to be very transparent, spend lots of time on what matters, the users input etc....
Anything dependent on Discord will never be the best
When you spoke about pricing, you said Dall e was $20 per month with no restrictions on the amount of images, but do you not also have to pay for each image you generate, as well as the $20?
Not only was your presentation awesome, it was super well done! My only complaint was that you favored MJ more than I did, BUT, in your conclusion, you made several valid points that I totally agreed with. Bottom line is that it is mostly subjective as it should be. I'm very, very glad that I stumbled onto you and happy to be a new subscriber. I look forward to viewing more of your videos going forward!
Thanks for your video comparison. However, at this stage it's a little unfair to even put them in the same building. MJ is The Louvre. Dall-E is a porta-loo in The Louvre car park. Unfortunately someone has pushed it over and the contents are spilling out onto the street.
That's a superb review of the apps and a thorough comparison. I share your thoughts, on both these apps. I use them at work daily. I use SD to upscale and refine the images from MJ for use in my advertising projects. A colleague is urging me to give a go at Comfy UI, which looks smashing for some examples he produced. I find the process laboured, and the output misses what I get in MJ, everytime I try to use it. What are your thoughts on that?
Thanks for the tips!
Videos like this are very important going forward.. it proves what human is.. how humans analyze, thank you for stepping up and taking on the challenge, you really did a good job, so midjourney wins this but dall e 3 is somewhat free and that means latent spaces take more time to interpolate In dall e 3
Glad you enjoyed it!
Was going to run these through Firefly, but didn't see the link for the prompts. Would love to see how Firefly stacks up against MJ and
Dall-E. Of the two, definitely prefer MJ...by far! Edit: typed first prompt into Firefly and it more closely resembles MJ; it's actually quite good and stacks right up with MJ.
Updated with the link
www.figma.com/file/6lrMz5ONAwyoNHlD5CUvnV/Untitled?type=design&node-id=0%3A1&mode=design&t=SorWbVb36j8xhMgB-1
Great breakdown. I halted my MJ monthly for GPT Pro (Dalle). I totally see how MJ produces more interesting, realistic works of art, but with Dalle I can collaborate with a (censored) super supportive, always positive art genius. Plus, I get the worlds most powerful LLM for everything else that provides. Each has it's merit. I have to admit, I do love the output of MJ better. But the user experience (and added benefits) of Dalle take the cake.
The conversational aspect of Dalle in GPT4 is excellent but the over censorship and hallucinations as the conversation continues makes it less and less coherent and useful. I've cancelled my GPT sub and am going back to MJ simply for the quality of images as they are truly outstanding and beat Dalle the vast majority of the time.
and the fact you cant even try it without paying
Ya for sure, it's not perfect. And sometimes it just won't do what you want, no matter how simple. It's odd. But as they say in AI, 'this is the worst it will ever be.' So, I know in a few years this will all be ironed out for everyone. It'll just be about which package you prefer--which subscription. I wonder if these AI generators (Dalle, MJ) will in a few years time both boast text to video as well. I assume they will.
@@Bhodisatvas
Hi great video. Are you aware that you can invite the MJ bot to a private chatroom? That has easily eliminated the annoyance of doing work in a public chatroom.
So glad I went down the MJ route instead of Dalle. Not missing a thing!
Hmm...
Pic 1 (photorealism): Win for MJ, easy. Only one of those looked anything like a picture.
Pic 2 (geometric pattern): MJ made the nicer wallpaper, but failed to include circles as per the prompt. I'd still use the MJ generation every time.
Pic 3 (Logos): I'd give this one to Dall-E because MJ got overly ambitious on small details. But frankly, I'd reroll both attempts.
Pic 4 (Pixar): How did MJ get 'postapocalyptic' from friendly, cheerful and colorful? Dall-E win by process of elimination. Samson, did you completely forget what you prompted for while evaluating this one?
Pic 5 (Cyborg): I don't like the helmet. If you're that augmented, what's the helmet for? Meanwhile, the Dall-E has no humanity remaining. That's a robot, IMO. I dislike them both.
Pic 6 (Influencer): Only one of those isn't straight out of the uncanny valley. MJ win for sure. Though even MJ still needs to work on skin texture. Her ears are wonky and she looks like she buys makeup by the truckload.
Pic 7 (Fantasy): I'll give this one to Dall-E for having more dragon in the image and actually following the prompt regarding forests, rivers and architecture. Even if the MJ picture is less cluttered.
Pic 8 (Tattoo): Hard one. I have no idea about japanese designs, but I don't think they had scales? Also, the Dall-E one seems like it'd be a better tattoo.
Pic 9 (Anthros): Neither really fit the description, I think. The MJ has the mood right, but Dall-E got the prompt right and doesn't look like a hobo.
Pic 10 (surreal landscape): I prefer the MJ pic, though it doubled the double moons.
Pic 11 (Gothic Cathedral): MJ Win, both in ternms of architecture and in terms of atmosphere.
And I kind of quit at this point. That's 5 wins for MJ, 3 wins for Dall-E and 3 fails where I actively disliked both images.
Thanks for your views, interesting to hear another perspective
Blending Midjourney's artistic flair with DALL-E 3's realism enhances content creation. Midjourney excels in abstract, creative imagery, adding a unique creative touch, while DALL-E 3 specialises in lifelike detail, enriching the visual narrative. This combination offers diverse visual appeal, appealing to enthusiasts of AI, digital art, and practical content creation. It's a valuable technique for creators aiming to elevate their digital artwork and storytelling, merging the strengths of both AI tools for more compelling and varied visual content in TH-cam videos and blogs.
They don't look unique tbh. They look like Midjourney
@@Kombo-Chapfika While Midjourney might not scream "unique" to everyone, it's got its own flair. It's like comparing apples and oranges with Darly - both have their pros and cons. Midjourney's knack for detail and mood sets it apart, but hey, Darly's not far behind with its vivid colors and patterns. It's all about what catches your eye!
the dragon by Dalle looks so much like a John Garland painting that I have a feeling that it could contain huge parts of this works...
Great video! Why does it seem Dall-e is not good for any input you do? Mid journey seems like the clear winner
I will say this. I write stories and need the illustrations to contain the right elements, expressions and posses. So ... although MJ images are always better looking. I choose DALL-E because accuracy of scene is more important than technically fantastic art. Very often MJ just straight up ignores my requests in favor of a beautiful face.
I'm also a novelist. I never tried DALL-E, but I like MJ for it's fine tuning features. Yes, as you say it tends to ignore my prompts at times, but there are prompting tools that can be used to fine tune whatever you want to create. For example, there are times where I get a result I really like, but with minor things that need to be fixed. Well, I can use a editing tool to fix a region that needs regenerated and take it from there.
I definitely prefer to use all of these combinations.
We would like to get more functional with AI Image generators, but we are struggling to create a consistent character within. How to have the same character for different scenes? which platform can do this job the best and with the least complications?
So in Dall e and mj you can get consistent characters in mj you can use the previous images and use vary region. In dall e if your prompt is refined well it will always create the same image and you can use the gen-id to create that image in multiple scenes ect…
I will try my best, thank you @@user-cg8th4gj3u
However you can get either image - and both types of image with the proper prompt - you would simply edit one prompt to make it more comparable to the other - we are just looking at an interpretation, I would follow up with , worn and rust but full of heart, highly textured surroundings and so on..
12:04 , damn !!!!!! look at the hair details on MJ
i believe this is all about prompt adherence
Great comparison. It very much feels like MJ is more moody overall, Dalle is more playful. Thanks for the hard work!
Thanks for watching!
thanks for this detailed comparison.
To me, your comparison revealed that DallE was mainly trained with Disney-content, and I mean the stuff for kids. While Midjourney was trained with all kinds of stuff from photos to art paintings to (maybe) Japanese anime. Disney content isn't a bad thing, they really have high quality standards, but the kids stuff is too polished, too composed and it always carries this forced happy mood where you expect someone starts to sing and dance every second. Emotionally, it's one dimensional. Midjourney seems to be better in carrying all kinds of moods, not only the happy ones.
But now to the downside of BOTH. They are censored. And not only in a way to prevent illegal stuff, but to be child-friendly. PG10 I would guess. The problem is, most paying customers are adults. For software that is created and maintained in the free part of the world, it's hard to believe how strong they enforce censorship in their products. How hard would it be for them to offer two different models, one child-friendly and one that's almost uncensored for people who are old enough to decide for themselves what they want to see and create?
Does mid journey provide any apis? Like dalle3?
are you using DALL.E3 standard or HD?
Whats weird is that in the DALL-E Bing interface it specifically suggests using the prompt "In the style of Van Gogh" but I have encountered it generally refusing to include celebrities and sometimes artists.
I think that's because Van Gogh is dead, and his work is in the public domain, while living artists are not allowed due to copyright concerns.
I wouldnt call this video bias at all. He did the prompts and he called out what the results were. MJ results came out alot more realistic whereas the other is more for fun.
Did seemed to me that Midjourney was going to get the palm leaf whatever it did, I am left questioning how neutral this appraisal was. In the robot example, for example, Midjourney completely ignored most of the brief - something it is very prone to doing. Just like a real graphic artist then ...
Dall-E was probably the first to make it mainstream, but it has remained very much at the level of 2022 in terms of creativity and style. While MidJourney and LeonardoAI have come very far from those early iterations. Firefly and Meta still have some catching up to do (although Firefly is exceptional at realistic / not-over-the-top beautiful portraits)
The difference is Midjourney is doing a lot more hand holding. Dalle's 3D capabilities for are exceptional if you know how to use it beyond a simple prompt.
Gooood day, sir!
Hello there!
I would say that MJ did a lot better copying the best artworks of real humans, whereas dall-e tries harder to be more original, but thus creating less beautiful / balanced images. MJ simply takes bit and pieces of your instructions and grab some images from its database, while dall-e actually listens to your detailed instructions and tries to create an original image. In the long run, I think dall-e would be the better Ai art generator.
I wonder why you haven't show the generation process
The issue with the fantasy dragon image is that there are things that don't make sense. The castle in the foreground makes it seem that the dragon behind it is larger than a building. Also the valley is lit by sunshine but there is a night sky. MJ wins that one hands down IMHO.
Sorry if already been asked/answered but midjourney has an alpha site where you can generate. Been live for a couple weeks.
only for users with 10.000+ images generated.
Only if you generated 100 images for 100 days, lol.
You correctly mentioned the word "audience" (Fox comparison), but you never mentioned the audience target was never prompted in all examples, leading IA generating pictures based on the audience of their choice / default (same with the robot / Pixar).When comparing pictures all of us naturally tend to prefer the picture that is closer to the audience we belong. Always be specific about the audience (adult, child, US citizens, Chinese citizens, professionals , etc) to make sure you are comparing the same thing and not leaving away this field that matters a lot in image generation through IA interpretation. That is my humble opinion, I really appreciate your great work.
Thanks !
Welcome!
Which do you prefer?
Sd
mj
Midjourney so far
Midjourney mostly, but Dalle is better at following complex instructions. So if you absolutely want what you want, sometimes Dalle is better. I’m doing repeating patterns though much of the time, and Midjourney is the king of that, and almost the one and only. V6 seems to have upped their tiling game as well, which I’m thrilled about.
please review your Alphonse Mucha pronunciation, its Mu - ch -a : ch as "Koch" the ch without k
As a Dall•e 3 user it’s all about you prompt, comparing the two is not really practical when they both use different ways of interpreting a prompt. You can indeed get a just as good as Mid Journey if you “know” how to prompt, using bokeh, distance etc etc I can indeed recreate just as good, and did with the cyborg image by prompting correctly. It’s just maybe Mid Journey is easier. And yes you can use certain images in Dall•e 3, like Star Wars, Marvel, and painters ‘in the style of…’
The MJ content looks, as you mentioned, more alive, more realistic, but it rly often just ignores the prompts. Cant one just use dall e and get the same effects as MJ has through further instructions, just with the quality of prompt following of dall e?
MJs basic plan is useless for anything serious with the monthly limit of 200 prompts, so its like 30$+ a month, while I can get dall e for 20$ on top of chat gpt thats super useful for my IT job. But the MJ images rly look more pleasing...
I thought it was possible to create images on the website already for people who made more than 10000 images...
Guess not. Because I still can't.
Yeah. I had the same problem. Even though I’ve made more than 10k
is chatgpt4 or midjourney better?
MJ looks more creative to me in terms of regenerating the new ideas, dallE is more illustrative and artistic looking shots for example painting and illustrations, even comic book
for creating a attractive images go for Midjourney whereas if you want a detailed image go for Dall E.
If you enjoyed this, you may enjoy this prompt battle with Stable Difussion th-cam.com/video/AVV9jm2E9lI/w-d-xo.html
If you have a complicated image with multiple interacting elements -- a very specific image -- it's not even close. There are images you can make with Dall-E that are _absolutely impossible_ with MJ.
Dall-E-3 is _fun_ -- MJ is more of a chore. But if you _really_ want to see those pores in the face of the busty Asian model with elf ears, or whatever, then MJ it is.
It seems like everybody is sleeping on Leonardo which has more features than all the other generators combined
Bing image creator gives you 4 photos too and is dalle-3
Plus its free
While Dall-E3 has garnered many kudos, we still think Midjourney is the one to beat. What do you think?
Please add pixel art to the comparison
Its not fair to use the same prompt for each platform. Each platform requires specific prompt language to get what you want so using the same will not get you the best of each.
I canceled my MJ for gpt pro, MJ is expensive and it has so many things you can't do, I'm not saying it's bad but Dalle is reasonably priced and it can put texts well in images as well as the flags, I always suffered with flags with MJ, it can only get first world countries flags correctly if you ask it for an African country flag it will never get it.
Dally 3 me recuerda a midjourney v4.
Aun asi me quedo con dalle3 y todo su paquete de componentes.
Dalle3 new blur mode inside copilot.
Why are both so terrible with creating accurate historical weapons especially muskets. They are also not too accurate on equipment, uniforms and accoutrements. Which seems strange since there are so many references on the internet, I do find that Dall-E does better with reference images than Midjourney.
AI image generation now is more of a mistake or a interpretation of many things, or more abstract like you're referring too, but make no mistake, IMO, when these technologies like image generation get so good and perfect, is when it will get boring. For me, the best images I have seen in my life have been through Midjourney, most of the elements that are in those images... are not correct, but you know what they are. It's not 3d rendered perfect intentions but those nuances make the best mistakes I've seen. NO human would ever create most of what Midjourney creates, too many influeneces, no human is that influenced. That being said, you're right, it does seem like it should be more dialed in, perhaps it's not meant too? Who knows.
You need to tell DALL-E to make a hyper-realistic image, the same goes for images with darker more disturbing content.
The quintessential truth being said right here: 17:05
The Dall-e3 influencer also has 6 fingers on her right hand...
Midjourney is the clear winner in terms of the images themselves. The style and the composition win for me in almost every case. The image with the fox detective is the one that Dali would win because here, a kid-friendly image seems to be more fitting in my opinion. Also, the Pixar image felt more "Pixar" for me with Dali, regardless of which image in general appealed to me. In general, the Dali style feels outdated, highly airbrushed and unappealing, on top of that most of the images Dali produced were very busy and cluttered with no real considered framing and composition. It is very cliche-looking AI art. Where Dali shines is in its understanding, it can be very frustrating working with MJ when it ignores important parts of your prompts or misunderstands them as you demonstrated in the image with the cats and the dog. If MJ can improve its understanding of what is actually wanted in a prompt I think there is no beating it.
you are using over complciated polluted prompts, gpt has contaminated your promtps, that is why most of the dalle images you made were not good. Tip*Tell gpt to not alter your prompt and dont go off discussing itwith gpt, Otherwise you will end up with "GPT" looking images, insert a properly engineered prompt instead (filename is the prompt so you can check if your prompt was altered).
Facts!
Artistically DallE can never keep up to MJ but you need to be trained in order to understand this
Stable Diffusion gives much better results than both if your willing to tweek a little
@@TheCcamera You have just exposed the fact you dont understand what I'm talking about.
@@D_VAULTZ if you or anybody else is able to get artistically equal or better images out of DallE with a different promting technique, please let me know I would be very interested in this
Midjourney the king
Midjourney for me.
The midjourney robot has NO GD LEGS! It's TOO doom-laden.
Midjourney has looked way better since the start
More brithsh I say! XD Nice video thanks for the review...
Thanks for watching!
Brittania all the way
Hey Samson! would love if you came on the MJ server and played an actual game of Prompt battles with the gang, We've been doing it since the start of Midjourney! My name on there is KinkyKong, and i'm one of the Green names in the guide section. I host a lot of the events on Midjourney. would love to colab on something!
Prompt battles is a great event thanks for hosting every week! Hopefully he'll see your comment
@@jaya7760 AWH thanks so much! I love doing it!
4:15 No way LOTR scenes are much better than... Oh wait "On Amazon Prime"?? ... Yeah, sounds about right.
Yeah I stay off of discord,too many negatives for me ,too many stories of people getting hacked and or stalked 😊
It is simple: Midjourney is vastly better at images and composition while Dalle is better in understanding the task and accounting for every detail specified. Pity that Dalle generates such poor images compared to (brainless) Midjourney.
I agree I don’t like DALL-E
... so I'm still ont he side-line wwanting to hop on an play BUT too many hoops and workaround in the 'motion' arena, so stills is the go for now and then I discover just how stringent CENSORSHIP is with blanket capped to all extremes of expression to a middle ground 'vanilla' of accetable boundaries MJ and others seems to be. I find that frustrating and annoying as an artist who enjoys more avantegarde and wild art forms. even basics such as any hint of nudity seems to be shot down (rediculous given the historical context of art and life drawing etc) ... open source options do not currently have these extreme censorship restrictions but not up there to the same degree of QC and nuanced detailing capacity as MJ.
Dalle just doesn't understand framing or composition while midjorney does.
👋
I love you man, but your videos are not like before. Still love you though man... Cheers!
In what way?
@@aisamsonreal you seemed more alive, more enthusiastic. Your still the man!!!
@@TopFactsanation You worked out he is now an AI too huh?