This is once again a perfect mix between comedy and information. Perfect video. Perfect overview and testing. Perfect sense of humour and timing. Thank you for your hard work.
Great analysis! Impressive breakdown of each AI art generator, highlighting their pros and cons. Your dedication is commendable. Keep up the excellent work! - Samson
I appreciated very much your evaluation using artistic terminology and perspectives. So many people who review image generators just say whether they think it looks cool, so I appreciate this idea of learning about art or illustration through image generators and the inaccuracies they have right now.
mid - if its all about aesthetics dall e - if its about prompt following/conceps ideogram - if its both after looking at results i personally tested i will say ideogram, mid is paid and is just aesthetic money grabber, its main motive is give visually appealing images and take money with no complains. dalle lacks aesthetics too much, ideogram is inbetween. and if other models are included hands down its always stable diffusion for me.
Great video. I've been with midjouney since v2 and it seems the best at rendering realistic anatomy. I just started with ideogram and I have just been rerunning my midjourney prompts on it. It still needs work on the face and hands but at the free 100 generations a day, I'll stay with it awhile.
Despite the rapid pace at which technology evolves, it cannot surpass the talent of a reviewer like yourself in delivering a comparison video such as this. I appreciate AI for its ever-changing capabilities, but I value it even more when it is presented in a clear, understandable human language, just as you have done in this video. What makes a video stand out is the effort of individuals like you who make everything seem more entertaining to watch.
Midjourney rocks! Definitely pleasing to the eyes. 🎨 For certain works I do also love Leonardo, which struggles with text to result, but for the simple requests it's great!
For some reason I'm very happy my guess in the poll for the dog images was right hahaha. Your video is great, thank you for the great comparison! Also I enjoy your humor, the rule them all reference was fun! About the comparison itself, I was very curious about ideogram, but honestly I'll continue with Midjourney due to its better aesthetic and overall image balance and compositions.
From my rather limited experience with them and watching videos like this one: You cannot prompt Dall-E the same way you prompt Midjourney. People are not ultra-realistic or hyper-detailed, if you think about it. Any terms with "realistic" in the name don't work as you would expect them to, if you are used to Midjourney. Especially "photo-realistic" and ultra-realistic" will make it look less...realistic. In the sense that apparently most people expect. I noticed that from just a couple of these videos. I think that adding "hyper-detailed" will make it even worse. That's because Dall-E adheres close to the prompt and these terms are exaggerations of what is just plain realistic. Midjourney sometimes seems to rather ignore your prompt and force its own style. Try to use "live-action", instead.
Thank You for doing this. I learned a lot. Like the fact I have no idea what I;m doing LOL! I was thinking about your last video and you made it seem easy, now II'm not so sure I can do it. You point out flaws I would not have noticed. Thanks again! Jazz it up!
I love the new Ideogram upgrade. I have been following you since you began pretty much and bought your course on UDEMY.. so you got at least a coffee or proper tea out of me. You are always entertaining and this video is over the top helpful. I have shared in some of my AI groups.. so spreading the word bro.. keep it silly...
Midjourney's aesthetic is unmatched. I think anyone with an artistic eye appreciates it. I think it's easy for an AI art generator to get 80 percent of the way, there's then really hard to get that last 20% right, and that's what Midjourney does.
i appreciate your insights from your art background. i hadn't picked that up in earlier videos, maybe because they were less image oriented. felt a bit like a class in art composition. well done, i'm subscribed to the channel and now your email. many thanks for the work that you put in.
MJ was the easiest to use. Dall-e has the most comprehensive prompt understanding in my opinion. But the move is to let MJ describe Dall-e outputs or use it as the style reference. Try that out peeps. Enjoy. (i use stable diffusion in comfyui mostly, but i use dall-e pretty regularly for dataset fodder)
Great comparison Samson! When it comes to pricing another feature that is important to creatives BUT not pointed out enough is privacy of creations. If you use these for work that last you want is to have client concepts in the public gallery. MidJourney’s $60 plan offers this. I don’t think Ideogram nor Dalle offer privacy.
Great video, I was looking for a good comparaison, thx! I love Dalle3 for its ability to tell a story within an image. Give it a poem and you'll get a beautiful picture that make sense (with extra limbs). I find Midjourney to be more factual, which is good if you have a very clear idea in mind but if you are looking to explore the depth of the model knowledge about concepts, it's less interesting. For me, once they fix hands, faces and texts, Dalle will be the best one for what I like to do. Defnitely will try Ideogram though, could be a fair compromise.
Didn't even know that Ideogram was up! And bro, I absolutely love your energy and the light heartedness you bring :) I mainly use Dalle-3 in Bing Image Creator (BIC) as it allows for more words. Haha your joke about the Ai personas had me rolling haha. Haven't used Midjourney in a while so am probably gonna get it soon. Would love to see some pointers of how you got started with youtube and any advice to grow audience?
Ideogram is cheaper for private and has a new version out 28 Feb, it is less restrictive than Midjourney and less banned words. However like he said the interface and features are behind midjourney. Although, they are only on their 3rd version I believe.
Tüyo: videolarınıza manuel olarak ingilizce altyazı eklediğiniz takdirde otomatik çeviri ile başka dillerde ingilizce bilmeyenler tarafından kendi dillerinde anlayarak videolarınızı izleyebilirler.Hem böylece daha geniş kitlelere ulaşabilirsiniz
with the bunny, mj got very little out of that prompt. dalle and ideo both had the meal, the menu, holographics, cyberpunk, glasses, coat, and reflective table
I find it bizarre and a bit frustrating that stable diffusion would do so poorly in these polls when it's the only one you have actual full control over, can train yourself and run locally for free. The mind boggles.
The thing with Midjourney is that you can really manipulate what it is doing. So with the right flags and parameters many things can be ironed out in no time.
I've just tested it when seeing your video title and the tool is really promising. Especially the way skin and eyes are rendered for faces that can sometimes be so impressive. Many thanks again man !
I find with the rapid improvement of existing AI's and new ones popping up it is hard to keep up. But it really depends on use case. Canva seems fine and user friendly, for social media but if I was to design luxury wallpaper for POD - midjourney would better.
Regarding a beautiful girl in a cafe, I would like to see what would have been rendered with half the information in the prompt? as a photographer I would not have used any of the camera settings that were suggested in the prompt. I’m enjoying the videos, lots to learn 😊
Great content! I know midjourney allows you to upload a pic and then it uses that as a reference, can you do a segment where you upload the same file to all 3 AI’s and see how they handle the same prompt? Unless Dal-E and Idiogram don’t allow for upload files?
I would like if you could test out the natural style for Dalle, using the API. The default one is vivid and I think maybe chatgpt uses that style. Great vids btw
First of all I love MJ, however your prompts are optimized for MJ. You once did the opposite prompting Dall-e prompts into Mj and the results where also biased to Dall-e. Altough I love MJ I do believe you need to prompt the software in the format it understands the best. So to compare the output please prompt according to the software. Altough in the end MJ will still kick ass 😊
W Leonardo AI przeszkadza mi nie aktywna wyszukiwarka. Jest to funkcja udostępniająca różne obrazy na inny temat. Z tego powodu nie wystąpił Leonardo. Obecnie Playground AI muszę powiedzieć, że pracujemy w nim łatwo. wirtualne przeglądanie swojego archiwum . MJ jakoś mi nie pasuje. Czy możesz utworzyć skład tych programów z Playground?
I greatly enjoyed your comparison of the leading AI image generators. However my experience is that given the same prompt, the same image generator will create a variety of images, differing in both style and content. Showing only a single image from each could be misleading. This leads me to wonder how the images were chosen. I don't recall this being mentioned. Usually DALL-E and Ideogram produce 4 images at a time, for example, so presumably some curation was involved.
Hey, I got an idea what you were to poll about asking people if they would like if you actually put links to the things you actually talk about in your videos.
Honnêtement je ne comprends ces retours si positifs et tout cet engouement sur le nouveau IDEOGRAM 1.0 , car après des centaines de générations sur Ideogram, et des milliers générations sur Dalle via Chat-Gpt4, ben il reste indétrônable. Ideogram ne lui arrive pas à cheville, sans exagération aucune. En tout cas, c'est indéniable en parlant de génération d'illustrations et de contenus comics books enfantins, ou albums illustrés ou style BD franco-belge. Pour les autres genres je ne sais pas me prononcer… (Un gros avantage cependant sur la gen de phrases ou de titres dnas les images pour Idéogramme.)
Man if you are able to understand how MJ thinks with the prompts and control the parameters with the URLs there is no comparison the others are not joking
Useful analysis of these three apps, maybe you could've included Leonardo and Stable Diffusion too. Was kind of baffled when you labeled Dall-E as 'woke'. Also, speaking in a 'funny voice' is probably not the cutting edge of comedy these days. Anyways, I've more or less enjoyed your videos so keep it up
is anyone else finding child corn on ideogram???? there is a user online.. collecting likes..and creating images of little girls .. .in distress.. with some more description which i don't dare add here or i will be flagged.
True. His use of the word “woke” is incorrect. In fact, there was no need to use it all. But our opinion won’t matter. He’ll keep using it the way he’s heard whomever he follows online use it. Though incorrect, it’s his right. Just as it’s my right to not subscribe nor watch another of his videos.
- [00:00] 🖼️ AI Art Gen Comparison:
- Comparison of idiogram, Dar, and mid Journey.
- Idiogram excels in text rendering.
- Public prefers mid Journey.
- [03:37] 🧠 AI Art Gen Features:
- Dar excels in chat interaction.
- Mid Journey balances simplicity with advanced options.
- Idiogram lacks full-screen prompting.
- [06:24] 🎨 Prompt Battle:
- Mid Journey wins public poll.
- Detailed critique provided.
- [12:09] 📷 Human Photorealism:
- Evaluation of idiogram, mid Journey, and Dar.
- Strengths and weaknesses identified.
- [19:18] 🐉 Fantasy Illustrations:
- Examination of AI-generated fantasy.
- Detailed analysis provided.
- [20:13] 🎨 Artistic Styles:
- Comparison of idiogram, Mid Journey, and DI.
- Each generator's style discussed.
- [23:28] 🚗 Dynamic Action Shot:
- Evaluation of Mercedes renders.
- Strengths and weaknesses highlighted.
- [26:08] 👧 Childish Illustration:
- Review of student illustrations.
- Characteristics and drawbacks discussed.
- [28:38] 🖋️ Font Design Challenge:
- Assessment of font rendering.
- Quality compared across generators.
- [31:35] 🐰 Surreal Illustrations:
- Examination of cyberpunk bunnies.
- Realism and style discussed.
- [32:58] 👠 Product Renders:
- Assessment of women's shoe renders.
- Strengths and weaknesses identified.
- [34:23] 🔮 Tarot Card Review:
- Review of tarot card illustrations.
- Adherence to prompts analyzed.
- [39:14] ⏱️ Speed Test:
- Comparison of idiogram, Mid Journey, and Darly.
- Performance in completing prompts.
- [40:10] 🎨 AI Gen Comparison:
- Comparison of idiogram, Mid Journey, and Darly.
- Restrictions and leniency discussed.
- [41:04] 💲 Pricing Comparison:
- Comparison of pricing plans.
- Monthly pricing and inclusions provided.
- [43:12] 😄 Ideogram Features:
- Ideogram's meme creation showcased.
- Humorous meme samples discussed.
- [44:33] 📝 Gen Personalities:
- Comparison of Ideogram, Mid Journey, and Darly.
- Characteristics described.
This is once again a perfect mix between comedy and information. Perfect video. Perfect overview and testing. Perfect sense of humour and timing. Thank you for your hard work.
Appreciate Ideogram, I think it's way ahead in terms of text and making logos especially
Great analysis! Impressive breakdown of each AI art generator, highlighting their pros and cons. Your dedication is commendable. Keep up the excellent work! - Samson
Ideogram is really good at understanding prompt but the quality is not that good
seen this in the first set of images. not bad though.
Midjourney always had difficulty adding details to complex prompts and doesn't take all the words into consideration.
Ideogram wins because free
I appreciated very much your evaluation using artistic terminology and perspectives. So many people who review image generators just say whether they think it looks cool, so I appreciate this idea of learning about art or illustration through image generators and the inaccuracies they have right now.
This is really awesome. I watched quite a few of these kinds of model comparisons and this one was, by far, the most artistically comprehensive. Cool
Which do you prefer?
Whichever is free
mid - if its all about aesthetics
dall e - if its about prompt following/conceps
ideogram - if its both
after looking at results i personally tested i will say ideogram, mid is paid and is just aesthetic money grabber, its main motive is give visually appealing images and take money with no complains. dalle lacks aesthetics too much, ideogram is inbetween. and if other models are included hands down its always stable diffusion for me.
Great video. I've been with midjouney since v2 and it seems the best at rendering realistic anatomy. I just started with ideogram and I have just been rerunning my midjourney prompts on it. It still needs work on the face and hands but at the free 100 generations a day, I'll stay with it awhile.
Despite the rapid pace at which technology evolves, it cannot surpass the talent of a reviewer like yourself in delivering a comparison video such as this.
I appreciate AI for its ever-changing capabilities, but I value it even more when it is presented in a clear, understandable human language, just as you have done in this video.
What makes a video stand out is the effort of individuals like you who make everything seem more entertaining to watch.
Midjourney rocks! Definitely pleasing to the eyes. 🎨 For certain works I do also love Leonardo, which struggles with text to result, but for the simple requests it's great!
For some reason I'm very happy my guess in the poll for the dog images was right hahaha.
Your video is great, thank you for the great comparison! Also I enjoy your humor, the rule them all reference was fun!
About the comparison itself, I was very curious about ideogram, but honestly I'll continue with Midjourney due to its better aesthetic and overall image balance and compositions.
From my rather limited experience with them and watching videos like this one: You cannot prompt Dall-E the same way you prompt Midjourney. People are not ultra-realistic or hyper-detailed, if you think about it.
Any terms with "realistic" in the name don't work as you would expect them to, if you are used to Midjourney. Especially "photo-realistic" and ultra-realistic" will make it look less...realistic. In the sense that apparently most people expect. I noticed that from just a couple of these videos. I think that adding "hyper-detailed" will make it even worse. That's because Dall-E adheres close to the prompt and these terms are exaggerations of what is just plain realistic. Midjourney sometimes seems to rather ignore your prompt and force its own style.
Try to use "live-action", instead.
Thank You for doing this. I learned a lot. Like the fact I have no idea what I;m doing LOL! I was thinking about your last video and you made it seem easy, now II'm not so sure I can do it. You point out flaws I would not have noticed. Thanks again! Jazz it up!
Glad it was helpful!
I love the new Ideogram upgrade. I have been following you since you began pretty much and bought your course on UDEMY.. so you got at least a coffee or proper tea out of me. You are always entertaining and this video is over the top helpful. I have shared in some of my AI groups.. so spreading the word bro.. keep it silly...
I appreciate that sooo much. Thanks for being here and your support
Great video, entertaining and full of substance.
That Ideogram mafia boss rabbit looks like the mafia boss I met in a dream last week but he was no rabbit. Thanks for this inspiering video.
Is this magic prompt?😁
definately appreciate your effort of going such deep man
thanks a lot CalmScapeTV
Midjourney's aesthetic is unmatched. I think anyone with an artistic eye appreciates it. I think it's easy for an AI art generator to get 80 percent of the way, there's then really hard to get that last 20% right, and that's what Midjourney does.
i appreciate your insights from your art background. i hadn't picked that up in earlier videos, maybe because they were less image oriented. felt a bit like a class in art composition. well done, i'm subscribed to the channel and now your email. many thanks for the work that you put in.
MJ was the easiest to use. Dall-e has the most comprehensive prompt understanding in my opinion. But the move is to let MJ describe Dall-e outputs or use it as the style reference. Try that out peeps. Enjoy.
(i use stable diffusion in comfyui mostly, but i use dall-e pretty regularly for dataset fodder)
Great comparison Samson! When it comes to pricing another feature that is important to creatives BUT not pointed out enough is privacy of creations. If you use these for work that last you want is to have client concepts in the public gallery. MidJourney’s $60 plan offers this. I don’t think Ideogram nor Dalle offer privacy.
Great video, I was looking for a good comparaison, thx!
I love Dalle3 for its ability to tell a story within an image. Give it a poem and you'll get a beautiful picture that make sense (with extra limbs). I find Midjourney to be more factual, which is good if you have a very clear idea in mind but if you are looking to explore the depth of the model knowledge about concepts, it's less interesting.
For me, once they fix hands, faces and texts, Dalle will be the best one for what I like to do. Defnitely will try Ideogram though, could be a fair compromise.
Ideogram is behind on the artistry of the image but it has a lot of positives for sure.
Love your channel. Welshboy here living in Bali. I have an Irish Pub in Kuta called 'Gracie Kelly's'. If you're in Bali pop in for a Guinness!
Ha! Great. Greetings from Ubud. I'll drop in next time I'm down
nice. can you please also do Ideogram vs Leonardo ai. I'm in a dilemma which one to choose for premium
Midjourney is unmatched in all spheres !!!
Dall.e 3 is great if you take it's output (that often has better composition) and upscale it with a good SDXL model.
Didn't even know that Ideogram was up!
And bro, I absolutely love your energy and the light heartedness you bring :)
I mainly use Dalle-3 in Bing Image Creator (BIC) as it allows for more words.
Haha your joke about the Ai personas had me rolling haha.
Haven't used Midjourney in a while so am probably gonna get it soon.
Would love to see some pointers of how you got started with youtube and any advice to grow audience?
Amazing video! And thank you for sharing your authentic view and knowledge about this crazyyy AI world.
Ideogram is cheaper for private and has a new version out 28 Feb, it is less restrictive than Midjourney and less banned words. However like he said the interface and features are behind midjourney. Although, they are only on their 3rd version I believe.
Tüyo: videolarınıza manuel olarak ingilizce altyazı eklediğiniz takdirde otomatik çeviri ile başka dillerde ingilizce bilmeyenler tarafından kendi dillerinde anlayarak videolarınızı izleyebilirler.Hem böylece daha geniş kitlelere ulaşabilirsiniz
For me, ideogram won most of the tests, the image layouts are dynamic, colors are balanced and easy to prompt
Didn’t know about this one , thanks
with the bunny, mj got very little out of that prompt. dalle and ideo both had the meal, the menu, holographics, cyberpunk, glasses, coat, and reflective table
I find it bizarre and a bit frustrating that stable diffusion would do so poorly in these polls when it's the only one you have actual full control over, can train yourself and run locally for free. The mind boggles.
Yeah, I think UX plays a huge role in this
The thing with Midjourney is that you can really manipulate what it is doing. So with the right flags and parameters many things can be ironed out in no time.
I've just tested it when seeing your video title and the tool is really promising. Especially the way skin and eyes are rendered for faces that can sometimes be so impressive. Many thanks again man !
I find with the rapid improvement of existing
AI's and new ones popping up it is hard to keep up. But it really depends on use case.
Canva seems fine and user friendly, for social media but if I was to design luxury wallpaper
for POD - midjourney would better.
Regarding a beautiful girl in a cafe, I would like to see what would have been rendered with half the information in the prompt? as a photographer I would not have used any of the camera settings that were suggested in the prompt.
I’m enjoying the videos, lots to learn 😊
Great content! I know midjourney allows you to upload a pic and then it uses that as a reference, can you do a segment where you upload the same file to all 3 AI’s and see how they handle the same prompt? Unless Dal-E and Idiogram don’t allow for upload files?
❤❤ love the videos
I would like if you could test out the natural style for Dalle, using the API. The default one is vivid and I think maybe chatgpt uses that style. Great vids btw
No prompt e-book?
Thanks!
No problem!
First of all I love MJ, however your prompts are optimized for MJ. You once did the opposite prompting Dall-e prompts into Mj and the results where also biased to Dall-e. Altough I love MJ I do believe you need to prompt the software in the format it understands the best. So to compare the output please prompt according to the software. Altough in the end MJ will still kick ass 😊
W Leonardo AI przeszkadza mi nie aktywna wyszukiwarka. Jest to funkcja udostępniająca różne obrazy na inny temat. Z tego powodu nie wystąpił Leonardo. Obecnie Playground AI muszę powiedzieć, że pracujemy w nim łatwo. wirtualne przeglądanie swojego archiwum . MJ jakoś mi nie pasuje. Czy możesz utworzyć skład tych programów z Playground?
Nice video thank you
Thank you too
I greatly enjoyed your comparison of the leading AI image generators. However my experience is that given the same prompt, the same image generator will create a variety of images, differing in both style and content. Showing only a single image from each could be misleading. This leads me to wonder how the images were chosen. I don't recall this being mentioned. Usually DALL-E and Ideogram produce 4 images at a time, for example, so presumably some curation was involved.
Самсону привет! Hi Samson!
Hey, I got an idea what you were to poll about asking people if they would like if you actually put links to the things you actually talk about in your videos.
What are you looking for ? Happy to offer any missing links, just let me know
The skeletons ribs cage is rotated :)
accents are great, everything is great, you've come a long way, you shall rule the tube whohahahaarrrrghbrrrrrrrrrr
I’m using stable diffusion Forge 1111 and literally making real life images.
Honnêtement je ne comprends ces retours si positifs et tout cet engouement sur le nouveau IDEOGRAM 1.0 , car après des centaines de générations sur Ideogram, et des milliers générations sur Dalle via Chat-Gpt4, ben il reste indétrônable.
Ideogram ne lui arrive pas à cheville, sans exagération aucune.
En tout cas, c'est indéniable en parlant de génération d'illustrations et de contenus comics books enfantins, ou albums illustrés ou style BD franco-belge.
Pour les autres genres je ne sais pas me prononcer…
(Un gros avantage cependant sur la gen de phrases ou de titres dnas les images pour Idéogramme.)
Man if you are able to understand how MJ thinks with the prompts and control the parameters with the URLs there is no comparison the others are not joking
Useful analysis of these three apps, maybe you could've included Leonardo and Stable Diffusion too. Was kind of baffled when you labeled Dall-E as 'woke'. Also, speaking in a 'funny voice' is probably not the cutting edge of comedy these days. Anyways, I've more or less enjoyed your videos so keep it up
Thanks for the feedback! Highly appreciated.
Ideogram and Dall-E are for amateur teenagers who enjoy creating cartoonish clipart.
is anyone else finding child corn on ideogram???? there is a user online.. collecting likes..and creating images of little girls .. .in distress.. with some more description which i don't dare add here or i will be flagged.
yall gotta stop the use the word woke to meancensorhip its anti racist word not a far right meme...
Far right lol
True. His use of the word “woke” is incorrect. In fact, there was no need to use it all. But our opinion won’t matter. He’ll keep using it the way he’s heard whomever he follows online use it. Though incorrect, it’s his right. Just as it’s my right to not subscribe nor watch another of his videos.
ruclips.net/video/H6wph16fbZM/видео.html The First 3D Animation Serie 100% A.I !!