Great Tutorial, can you show us how you made the final video with all its effects (zoom in, move from one panel to the next, page flip effect etc)? What software did you use for that? Thanks!
How you make on minute 5:08 to 5:10? In 5:08 he's very thin and has bad hands instead in minute 5:10 the right size and right hands how did you do it? Thank you very much
Leonardo let's you control the poses of the characters much better than Midjourney, although the quality isn't has high. Midjourney unfortunately has no real way of controlling character motions outside of the prompt. Image references don't seem to work that well.
@@taoprompts thank you for that tip. I haven’t used or noticed Leonardo for a long time because it just seemed hard to understand but I wasn’t giving it a try which if I did I would’ve researched how. I just tried to use the free and it was of course crazy with no guidance. You probably just got Leonardo a big customer. What plan do you suggest to do that level of control and how much could I generate like a real world estimate of just about how many images because that’s be the next thing I would ignore until I get a error message saying I’ve used all my monthly credits right in the middle of a project
@@m.j.mcintear793 The image guidance features definitely makes a huge difference in what it can be used for. For the finished comic panels I made in this video, it took around 3,500 credits. Although I did a lot of experimentation and generated images with the highest possible quality. You can always try the cheapest $12 a month plan and then upgrade to the $30 plan if you like it or need more credits.
Dear Tao, your work is really great. It's very interesting that you do zero to 100 things and don't limit yourself to just teaching theory. this is fabulous. Please continue this way. I have a question. In this story, June looks young and about 20 years old. Now if you want to generate an image of June in her 40s, what method do you use to preserve Jane's original face and make her look older?
Thanks man, that's a good question, I had some trouble with that also. One thing that kind of worked as using June's character reference image, and then attaching another photo of an older looking woman as a content reference. Then using a prompt like "an old woman with red hair and green eyes and wrinkles on her face."
I am really curious, you are showcasing alot of different AI. But which one do you prefer for anime? Really interested in your opinion. And ofcourse I understand this changes by time since new AI appear and others get new updates ect.
Midjourney has best looking images. Leonardo gives more control. The absolute best to use is Stable Diffusion Automatic1111 repo, but that will take a lot of extra setup on your end
I would be specific in the prompts about exactly what pose you want "standing, sitting, arms pointing outwards". But using pose reference will give you more fine grain control
sometimes i spent hours looking for what I need .Today after 3 hours of watching videos and 4 hours trying by my own I found what I was looking for here.tnx
Question.. I do this in MJ Alpha.. which do you like MOST or do you feel they both are now compatible , work flow wise.. I find MJ a bit easier to use, because of how Leonardo makes you access the Style Ref, Image Ref is clunky compared to just toggling on and off a small icon with MJ.. so curious which one you like to use in your real day to day.
Midjourney creates high quality & better looking images in general, but it doesn't give you as much control over the pose of characters like Leonardo Ai has with pose reference and content reference. So if I want a specific pose for characters I would use Leonardo, for better aesthetics I would use Midjourney
I am working on my comic right now and it will take me years to finish it. My only concern is that in near future everyone can create comics with the click of a button and all my work will be for nothing.
Hey man, Ai is advancing very quickly. Based on my experience as a researcher/scientist in Ai image & video: 1) For high production quality tasks, Ai still needs human labor. For example, consistently generating good images of hands doing complex tasks is extremely difficult, if not impossible. 2) Human motion/poses is also a hard task. I've spent over a year trying to generated realistic complex human motions and it is still a struggle to generate realistic motions of people, especially with different body types.
After you're finished and you feel like all your work is for nothing. Then that's a you problem, not AI. Plus, all because you work on something for years. Doesn't mean it's going to sell, A.i created or not.
Quick question my man. Does Stable Diffusion or Mid Journey has an option to take reference like this and generate consistent image based on a predetermine model? And how consistent it is in a scale of 1 to 10?
how long exactly do you need to generate the comic strips as shown in the video? In a commercial environment, is it possible to generate a story of 40 ~50 pages in 2 weeks?
Hey great video! Just a quick question - which of the paid plans allow me to add images for character references? Is it all of them, or only the most expensive plan? I wasn’t able to figure that out based on the description of the paid plans they gave on the Leonardo AI website
@@taoprompts THAT WAS A PERFECT VIDEO.. I JUST HAVE A QUICK QUESTION ABOUT HOW YOU DID THE ''FLIPPING BOOK'' EFFECT ON THE ANIMATED VIDEO IN THE BEGGINING OF THE VIDEO ?
First of all, great video! I noticed that many of your manga videos were created using MidJourney. Is there a specific reason why you switched to Leonardo?
Great Job @taoprompts . Please Please Please i will like you to do a video on the motion comics .. software and how u get to do the motion part of the video for comic ai
Hey, the video editing is done in Capcut. The main feature you need to know how to use is keyframes, which is what does all the zooming in and panning. Other than that I just added in premade transitions between the frames. I may do a tutorial in video editing in the future though.
Thank you! Can we see a step by step besides what you showed, to make sure the character looks a lot like a person one may know as a character reference as well as a certain "Superhero Outfit" for image to image consistency?
Hey man, for character consistency (with clothing), the best way right now is to finetune a flux model like I show here: th-cam.com/video/5Z8fwEeWfRg/w-d-xo.html Leonardo is good but won't give you that same level of consistency.
Free Ai image generators tend to be much more limited. You could probably get reasonable looking characters, but probably not be able to control the poses with the dynamic action shots.
OpenPose for Stable Diffusion will let you control the pose, and it's all free. In fact that's probably what LeonardoAI uses. Setting up your own local AI generator will have more options and capabilities, but it will also be more difficult to learn. Plus you need a graphics card capable of generating images at a decent rate. LeonardoAI and other websites might be using the same base software, but it's their settings, models, and user interface that make them far easier to work with. TaoPrompt has come up with a great workflow. If he did this with free software, the process would be much different.
This is a fantastic tutorial. I'm going to use this. Thank you brother! Do you have an affiliate account to Leonardo? I'm about to get the pro version and if you have a link I'll use it to support you.
You got a good result, but I believe is more cost efficient to run stable diffusion locally or with google colabs, in the case of people that doesn't have a high end GPU you can use your google drive to store all your settings in colab and add features like controlnet, checkpoints and LORAs, you can also train your own characters LORAs, because this comic is very small and you can also use 3d models to get open poses for controlnet; unfortunately putting aside the great developments and improvements of AI, I will say in her current state can not be even be considered a tool in comics creation and animation at less not in her current state, even multibillion companies as Disney are not getting it work right yet after ingestion MILLIONS, and the whole western entertainment industry is being roasted for using cheap CGI and VFX generated by AI; at pro level is still contra-productive to use AI, because is very hardware demanding, the process of creating train and fine tune models is also time consuming and as you mentioned you need to generate thousands of images to get one right if you consider that you also need to make the editing, post-production and tie all together you end spending the triple of time than if you work with a traditional pipeline and honestly after giving a try I wills ay the production cost is not significant if you take in account the final result will be of inferior quality that the one produced with the existent pipeline, then for now I useless in my workfield and I will say even with the fast pace in development we will still need to wait at less a decade until AI efficient tools for this file are accessible to the majority of us, right now the best bet are 3d models, Blender is free and cheap to customize, with only a couple of thousands you can get a specific version of blender that fit with your needs andBlender is awesome for NRP, right now you can see how is used already in many netflix series like DOTA, Castelvania, Blood of Zeus and blue eyes samurai, the combination of grease pencil, hand painted textures and toon shaders make blender perfect for comic creation and animation. The best of working with 3d models is that you have full control over the production process and you are not gambling like with AI, I always check videos like yours to see if is any new improvements but we are in a stalemate, no new advances, just think about it, how long and how much you could spend in leonardo to generate a grapich novel of standard size (164 pages) and then you realize is out of question to even try to do it and your method can work to create a 2 to 5 comic pages but for a graphic novel or a comic saga you will definitely need to train LORAs and just one, summary is not even worth to give it a try yet.
For the highest production quality, you still need a lot of manual labor. But if you take a look at channels like this one: www.youtube.com/@BibleNutshells/videos , it uses Ai for animation and has been very successful and probably fairly profitable. You don't need super high quality animation/art to tell a good story and entertain people, which I think ultimately is what matters the most. Of course it also depends on what type of content you are creating. For a film like Lord of the Rings, it will be a long time before Ai can fully create that.
What other guides do you want to see?
Maybe talking AI Avatars?
Would love more of this!!
MidJourney version of this.
Maybe anime video/animation using pika, runway or any others
Is it possible to do this on Midjourney, using image reference for the pose/action guide and cref and sref to keep the characters consistent?
Really nice!! Thanks dude!
Thanks David!
Great Tutorial, can you show us how you made the final video with all its effects (zoom in, move from one panel to the next, page flip effect etc)? What software did you use for that? Thanks!
Thanks! Edit was done with premade effects in capcut
Huiii...bangWowbangWowbangWoooow...question please: Pose to image: Is this only in a payed version or in free too? Cheers, Tom
I have paid version
@@taoprompts jo´
How you make on minute 5:08 to 5:10?
In 5:08 he's very thin and has bad hands instead in minute 5:10 the right size and right hands how did you do it? Thank you very much
Great work! Very creative and interesting use of the different AIs.
Thanks for sharing!
Thank you. It's good to experiment with different Ai's out there, they all have their strong points.
When you said generic characters, meaning it will struggle with mythical creatures like a minotaur?
I’ve been running away from Leonardo but I should give it a try. Been waiting for more of this style of video for Midjourney. Is Leonardo easier
Leonardo let's you control the poses of the characters much better than Midjourney, although the quality isn't has high. Midjourney unfortunately has no real way of controlling character motions outside of the prompt. Image references don't seem to work that well.
@@taoprompts thank you for that tip. I haven’t used or noticed Leonardo for a long time because it just seemed hard to understand but I wasn’t giving it a try which if I did I would’ve researched how. I just tried to use the free and it was of course crazy with no guidance. You probably just got Leonardo a big customer. What plan do you suggest to do that level of control and how much could I generate like a real world estimate of just about how many images because that’s be the next thing I would ignore until I get a error message saying I’ve used all my monthly credits right in the middle of a project
@@m.j.mcintear793 The image guidance features definitely makes a huge difference in what it can be used for.
For the finished comic panels I made in this video, it took around 3,500 credits. Although I did a lot of experimentation and generated images with the highest possible quality.
You can always try the cheapest $12 a month plan and then upgrade to the $30 plan if you like it or need more credits.
@@taoprompts I can't thank you enough
@@m.j.mcintear793 I'm more than happy to help 👍👍
Can you do a video on animating images or backgrounds using Leonardo?
I will test Leonardo's animation tool out, it does seem fairly limited based on what I've seen though.
Dear Tao, your work is really great. It's very interesting that you do zero to 100 things and don't limit yourself to just teaching theory. this is fabulous. Please continue this way.
I have a question. In this story, June looks young and about 20 years old. Now if you want to generate an image of June in her 40s, what method do you use to preserve Jane's original face and make her look older?
Thanks man, that's a good question, I had some trouble with that also.
One thing that kind of worked as using June's character reference image, and then attaching another photo of an older looking woman as a content reference.
Then using a prompt like "an old woman with red hair and green eyes and wrinkles on her face."
@@taoprompts Thank you
Once its saved and completed, where do u sell it?
Hey man, great video! But do you create the mini comic movie with voices?
I made the voices with elevenlabs and the video editing was done in capcut.
Very interesting demonstration, woooow!!! 5* But, how did you do the sound and the voice over? Thanks once more! Dan 👍👍👍
Thanks Dan!
The voiceover was done in elevenlabs, sound is music from Artlist
@@taoprompts OK. Thanks a million once more Tao 👍👍👍
I am really curious, you are showcasing alot of different AI. But which one do you prefer for anime? Really interested in your opinion. And ofcourse I understand this changes by time since new AI appear and others get new updates ect.
I haven't seen any that are good at animating Anime. Although most of them do pretty well with 3D pixar style cartoons.
how do you create the flipping book effects?
I use capcut built in effects
Do you think it's possible to do this In mid journey?
It would be harder to control the exact poses, but you could replicate characters and create backgrounds for them.
@@taoprompts sir do you think you can make tutorial for comics using mid journey if can only thank you 😊
@@jaysontuonan I will plan on it!
Just subscribed from S.Korea. Learned a lot from your detailed explanations which showcase your talent. Greatly appreciate it.
Welcome to the channel 👍. And Hello to South Korea!
It leaves too many art direction decisions in the hands of A.I. and ends up being a visual mess.
what ai platform is the best to use?
Midjourney has best looking images. Leonardo gives more control.
The absolute best to use is Stable Diffusion Automatic1111 repo, but that will take a lot of extra setup on your end
How to control anime character pose with Prompts?
I would be specific in the prompts about exactly what pose you want "standing, sitting, arms pointing outwards". But using pose reference will give you more fine grain control
how can you keep the same clothes everytime?
Here's an example of finetuning a flux model to get consistent clothing & characters: th-cam.com/video/5Z8fwEeWfRg/w-d-xo.html
sometimes i spent hours looking for what I need .Today after 3 hours of watching videos and 4 hours trying by my own I found what I was looking for here.tnx
Glad to hear this was helpful 👍
Finally, what is the best site to create comics with a consistent character?
Agree, sometimes better to generate characters and background separately and paste them together in Photoshop.
That seems to work best if you want consistent backgrounds, it hard to get backgrounds and characters right in a single Ai image
Can this be done in midjourney? If so how?
You could do something similar in Midjourney, but you won't be able to get the fine-grain pose control over the characters.
You can remove background with Leonardo ai, it’s their built-in feature ❤
I did try that. I found the Leonardo background removal wasn't as consistent as Photoshop's, but it can definitely be used
Question.. I do this in MJ Alpha.. which do you like MOST or do you feel they both are now compatible , work flow wise.. I find MJ a bit easier to use, because of how Leonardo makes you access the Style Ref, Image Ref is clunky compared to just toggling on and off a small icon with MJ.. so curious which one you like to use in your real day to day.
Midjourney creates high quality & better looking images in general, but it doesn't give you as much control over the pose of characters like Leonardo Ai has with pose reference and content reference. So if I want a specific pose for characters I would use Leonardo, for better aesthetics I would use Midjourney
Finally, what is the best site to create comics with a consistent character?
You can use Midjourney or Leonardo, although you'll probably need to do you work in Photoshop to get the exact panels you want
I am working on my comic right now and it will take me years to finish it. My only concern is that in near future everyone can create comics with the click of a button and all my work will be for nothing.
Hey man, Ai is advancing very quickly.
Based on my experience as a researcher/scientist in Ai image & video:
1) For high production quality tasks, Ai still needs human labor. For example, consistently generating good images of hands doing complex tasks is extremely difficult, if not impossible.
2) Human motion/poses is also a hard task. I've spent over a year trying to generated realistic complex human motions and it is still a struggle to generate realistic motions of people, especially with different body types.
A good idea is a good idea
Focus on the story and basic ideas of the scenario. In 1 year from now you will able to do it
After you're finished and you feel like all your work is for nothing. Then that's a you problem, not AI. Plus, all because you work on something for years. Doesn't mean it's going to sell, A.i created or not.
perhaps skills just won’t be as valuable in the future, ideas and the sense of what rhymes with the pubic is.
can you turn an old comic book page to AI
Yeah, you could use it as an image reference for the Ai
@@taoprompts could you do just one page of a comic book and edit it with voice and SFX
@@taoprompts just bring page live please
Quick question my man. Does Stable Diffusion or Mid Journey has an option to take reference like this and generate consistent image based on a predetermine model? And how consistent it is in a scale of 1 to 10?
sure, it's called controlnet in stable diffusion. It works about the same as this video
how long exactly do you need to generate the comic strips as shown in the video? In a commercial environment, is it possible to generate a story of 40 ~50 pages in 2 weeks?
It's possible if you have multiple skilled people working on it
bro do you have any other channels?
This is the only one I'm currently running.
Awesome tips! I'll see how it works from creating book trailers. Thanks for making these tutorials. They're the best ones on
Great to hear you enjoyed! Book trailers would be a cool application I haven't thought of.
Hey great video! Just a quick question - which of the paid plans allow me to add images for character references? Is it all of them, or only the most expensive plan? I wasn’t able to figure that out based on the description of the paid plans they gave on the Leonardo AI website
I have the cheapest $12 / month plan it all the features work for me
@@taoprompts cheers man, thanks a lot!
yoooo soooo good! appreciate you man!
Thanks! I'll make more guides like this one in the future.
Thank you!! Now i get equpment for ❤
Glad this helped! Leonardo ai can do some pretty cool stuff
What an amazing tutorial!!!
Thanks man, I appreciate that
Thanks a lot!
This was great, btw!
I love this. Thank you
Thank you for the tutorial! Did you do the final animation of your comic strip in Canva?
I did the animations in Capcut
@@taoprompts THAT WAS A PERFECT VIDEO.. I JUST HAVE A QUICK QUESTION ABOUT HOW YOU DID THE ''FLIPPING BOOK'' EFFECT ON THE ANIMATED VIDEO IN THE BEGGINING OF THE VIDEO ?
That was a interesting twist on the comic ..
Thank you very much, I was looking for a video like yours!
Great to hear this was what you're looking for! Leonardo Ai is a pretty cool tool, it gives you a lot of different options for creating images
What did you use in Canva to do that? Thanks!
speech bubbles
Leonardo ai is not for free so anyone please tell me a tool to make character reference
Why make a sad comic dude that got me to tears
Cheer up, it's not over for them yet 🙏
this is im waiting for im have see like this but i dont kwn how to start thank you man
I'm glad this helped 👍, I'll post other guides like this in the future.
First of all, great video! I noticed that many of your manga videos were created using MidJourney. Is there a specific reason why you switched to Leonardo?
Thanks! Leonardo has pose control which I showed here
Thank you 👍😊
Interesting but you left out a huge step of how you did the AI voices, which was the only part I was interested in ,...doh!!
elevenlabs for voices
@@taoprompts cheers
can you suggest an AI that has the feature of the Leonardo AI with free features?
If you have a good GPU you can run comfy UI locally for free and use the same features (ControlNet)
awesome thank you !!!!
It seems like the process will take a lot of generating just to produce a single panel
Yeah, it does take a some work. But you can get some nice results if you're willing to put in the effort
Let's gooo🔥
Appreciate you man!
This could not be done with MidJourney?
You can make comics in Midjourney, but it will be much harder to control the poses.
Great Job @taoprompts . Please Please Please i will like you to do a video on the motion comics .. software and how u get to do the motion part of the video for comic ai
Hey, the video editing is done in Capcut. The main feature you need to know how to use is keyframes, which is what does all the zooming in and panning. Other than that I just added in premade transitions between the frames.
I may do a tutorial in video editing in the future though.
Nice guide. Will take some tips from this for sure
Thank you 🙏. I make more practical guides like this in the future.
Thank you! Can we see a step by step besides what you showed, to make sure the character looks a lot like a person one may know as a character reference as well as a certain "Superhero Outfit" for image to image consistency?
Hey man, for character consistency (with clothing), the best way right now is to finetune a flux model like I show here: th-cam.com/video/5Z8fwEeWfRg/w-d-xo.html
Leonardo is good but won't give you that same level of consistency.
great instructions
But this feature are paid
Will you find something like this in free
Free Ai image generators tend to be much more limited. You could probably get reasonable looking characters, but probably not be able to control the poses with the dynamic action shots.
OpenPose for Stable Diffusion will let you control the pose, and it's all free. In fact that's probably what LeonardoAI uses. Setting up your own local AI generator will have more options and capabilities, but it will also be more difficult to learn. Plus you need a graphics card capable of generating images at a decent rate.
LeonardoAI and other websites might be using the same base software, but it's their settings, models, and user interface that make them far easier to work with. TaoPrompt has come up with a great workflow. If he did this with free software, the process would be much different.
This is a fantastic tutorial. I'm going to use this. Thank you brother! Do you have an affiliate account to Leonardo? I'm about to get the pro version and if you have a link I'll use it to support you.
Thanks, I'm glad you found this helpful! I do have a link here, thanks for supporting: app.leonardo.ai/?via=yutao
Can someone please say what is used for creating the movement, "animations" and visual FX for this? After effects, canva?
capcut
@@taoprompts thanks!
Awesome!!!
You got a good result, but I believe is more cost efficient to run stable diffusion locally or with google colabs, in the case of people that doesn't have a high end GPU you can use your google drive to store all your settings in colab and add features like controlnet, checkpoints and LORAs, you can also train your own characters LORAs, because this comic is very small and you can also use 3d models to get open poses for controlnet; unfortunately putting aside the great developments and improvements of AI, I will say in her current state can not be even be considered a tool in comics creation and animation at less not in her current state, even multibillion companies as Disney are not getting it work right yet after ingestion MILLIONS, and the whole western entertainment industry is being roasted for using cheap CGI and VFX generated by AI; at pro level is still contra-productive to use AI, because is very hardware demanding, the process of creating train and fine tune models is also time consuming and as you mentioned you need to generate thousands of images to get one right if you consider that you also need to make the editing, post-production and tie all together you end spending the triple of time than if you work with a traditional pipeline and honestly after giving a try I wills ay the production cost is not significant if you take in account the final result will be of inferior quality that the one produced with the existent pipeline, then for now I useless in my workfield and I will say even with the fast pace in development we will still need to wait at less a decade until AI efficient tools for this file are accessible to the majority of us, right now the best bet are 3d models, Blender is free and cheap to customize, with only a couple of thousands you can get a specific version of blender that fit with your needs andBlender is awesome for NRP, right now you can see how is used already in many netflix series like DOTA, Castelvania, Blood of Zeus and blue eyes samurai, the combination of grease pencil, hand painted textures and toon shaders make blender perfect for comic creation and animation. The best of working with 3d models is that you have full control over the production process and you are not gambling like with AI, I always check videos like yours to see if is any new improvements but we are in a stalemate, no new advances, just think about it, how long and how much you could spend in leonardo to generate a grapich novel of standard size (164 pages) and then you realize is out of question to even try to do it and your method can work to create a 2 to 5 comic pages but for a graphic novel or a comic saga you will definitely need to train LORAs and just one, summary is not even worth to give it a try yet.
For the highest production quality, you still need a lot of manual labor. But if you take a look at channels like this one: www.youtube.com/@BibleNutshells/videos , it uses Ai for animation and has been very successful and probably fairly profitable.
You don't need super high quality animation/art to tell a good story and entertain people, which I think ultimately is what matters the most. Of course it also depends on what type of content you are creating. For a film like Lord of the Rings, it will be a long time before Ai can fully create that.
Thank you brother for this. really appreciate it
That's really cool.
Excellent tutorial! Thank you!
Great to hear you like this! I wasn't sure if people would be interested in something like this, so I'm happy to hear you found this helpful.
@@taoprompts Oh yea I feel blessed to watch this. Please keep it up!
You have the best channel.. good stuff
Thanks, I really appreciate that 👍
Please do how to make anime with ai tools.
I will look into it. Ai still has a hard time animating anime style images though,
This is what I'm looking for ♥️
Subbed, Best Workflow guideline in the whole internet🙏
Thanks for subbing! And great to hear this helped
Bro what's the AI WEBSITE NAME
app.leonardo.ai/
With SD you can get character, style and clothing consistency, but it takes a lot of practice and the right computer to get there. Nice video 💪🏻
SD is great to use with a ton of customization options, although it does take some extra setup.
@@taopromptscould you make a tutorial for setting up SD and creating comics with it?
Thank you very much for the tips and knowledge
Glad this was helpful, Good luck with your comics!
Thanks alot, this video is exactly what I was looking for.
Great to hear that, Leonardo Ai is a very underrated tool!
Thanks
Wow, cool
PS.. Comic Life 3 is old school, but CHEAP and makes it easy to assemble comics especially add speech bubbles..
Thank u lm not a good artist but l can tell a good story
That's great! Ai can help with your illustrations for sure
There should be a great tutorial for creating NSFW comics.
You'll probably need to create locally on your own desktop & gpu for NSFW work
Comipo comics.
but character referance is not free it's a paid feature show us something and it's free
Rice gum?
God...I hate Leonardo AI everything looks cartoonish or airbrush art.
There's pros and cons to every Ai image generator. Leonardo gives you way more control over character poses and image composition.
Where is.sliding animated step?
that's done in capcut
@@taoprompts where is that part video bro?