Sequential Art Workflow: With iClone and Stable Diffusion. Great for Comics and Storyboards

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
all right so here we are inside of iclone and we are going to be attacking uh the creation of a storyboard panel so but just because we're doing something specific uh it doesn't mean that this is the only thing you can do all right so I'm gonna show you what we're going to be building we're going to be building a similar panel to one of these and this is something that I built inside of uh or with the assistance of stable diffusion uh so I'm gonna take you through the steps into how to create a panel like this there's a lot of challenges that that have to be overcome but I'm going to show you how to streamline all of that nice and easy all right um the first thing that you can see is that there's a very interesting angles on the panel itself which are not so easy to achieve with uh the basic uh open pose model uh but you can do that by way of prompting a little bit and the other thing is the full control of two or more characters and um and that's one of the things that I'm going to be showing you a shortcut for uh as we create a finished uh product so here we are inside of iclone this is the open poser model and uh it's fully rigged so it'll accept any kind of motion capture data any data that already exists in your iclone collection any data that you can bring in from miximo and it's very easy to use uh you can grab this character and just start posing them around like you do with any other 3D software it's a lot easier and more convenient than working with it in blender you know you have to mess around with all sorts of like uh uh rigging stuff I don't even know what they call that stuff it's a little bit above my head to be honest I tried working with a model that somebody else made for blender and I damn near pull my hair out so I decided to make my own so yeah this is a fully customizable character and you can pose them just like any other 3D model like uh you know that's made for manipulating inside of open posts um so okay so the other thing is that it does accept uh uh Motions like pre-made Motions like I said you can work with um you can work with any motion that's inside of your iclone by just dragging and dropping it onto the character so now you have a character that pretty much moves around and uh yeah if you just want to grab just random pose or whatever of a thing that's already been made you can already just you know all right so I'm gonna work with that so that is the you know you just ex you export that frame and you use that as your reference so that's kind of what I'm going to be doing and I'm going to be doing a panel similar to the ones that I showed earlier uh and this is just a storyboard panel uh you can use this technique for building comic books uh even full-on animation because you can export the frame sequences and use the batch processing and stable diffusion to uh control the motion of a stable diffusion character so that's another beast in and of itself but I'm just gonna focus on the basic uh task at hand and you can evolve that this procedure this workflow into your own thing so yeah you know what I actually kind of like this pose here it's like a very Dynamic looking thing I'm gonna go ahead and keep it so I'm going to select this and I'm just going to select remove motion object animation so now the entire thing got stuck at that frame basically well I spoke too early let's do it again so we select the character remove object animation and that should have removed the the motion yeah that moved I removed the motion it didn't remove it from the camera and that's one of the reasons I'm doing this video without any editing because I want to also see the little mistakes that I make so if I select this camera here I can remove the object animation of the camera itself so now I'm stuck with that one frame that I liked so let's grab this character here and that will select them and uh we'll with while holding down the Ctrl key I'm going to grab this green arrow and that's going to allow me to quickly duplicate so now I can just post this guy around and we'll find another animation to work with him again you can you can pretty much work with whatever animation you want you can make your own you can pose them manually you can do whatever the heck you want see if I can okay here's a good one very interesting looking so the thing is we want to create a dynamic pose or a frame that involves the two characters interacting that's uh that's not that easy to do in stable diffusion uh to control that and I'm going to show you the trick in a minute so basically these two characters are kind of fighting I don't know what the hell they're doing they're this guy's getting ready to punch but he's blocked he's defending against a previously thrown punch that should be good enough right and uh you might have decent random results if you export this as is but what I recommend is uh once you have the the action that you want you frame it and uh you click on this export thing and you export it as your dimensions of what you're going to be doing my panel is two to one so I'm going to frame it like this like a cinematic film and I'm just gonna frame it like that uh a lot of what people are doing are in portrait format two to three or they're doing the the square thing you know 768 by 768 or five total by 512 I mean how you export this that's all up to you at the end of the day so um so I'm just gonna set up my my frame the way I want it Maybe push this out a little bit more just like that and uh yeah so if I were to click this render button that's what my image would look like right and uh but I don't want that because it may or may not work like that uh but the fact that these things are interfering and overlapping each other I'll almost always creates issues so what we do is we go to the scene uh panel and we'll turn off one of the characters first uh we'll turn off oh that's the camera we'll turn off that guy first and uh I'll go to the render option and I'll render out the frame uh just gonna go to the poses thing I'm gonna call this guy oh I'm going to create a clean folder here it's gonna call it tutorial I'ma call this guy uh player one all right so it looks like this and then I'm gonna go to my scene again I'm gonna turn this one on and I'm gonna turn this one off I'm gonna call this one player two now unfortunately there's no keyboard shortcut that I know of that allows you to easily just export the image boom boom to speed up your workflow so that kind of sucks um you could probably do something else with screen captures and stuff having saved to a folder so you have to go through all that it'll save you a few seconds of your life but it's still not that bad player two okay so that's saved so now I got two independent pictures so the next thing that we do is we're going to fire up our stable diffusion let me see if I already have it running okay yeah it looks like I already have it running and uh so what I need is the user interface so here we are and I'm gonna I'm not gonna close my iclone chest yet because um if for some reason something doesn't work I can always go back in there or if I want to push the pose a little bit more I um I can come back to iclone repost the character just a little bit tweak them out and then export the new image to use so in order to do this particular workflow you're gonna need control net and you're going to need an extension and another extension uh which is [Music] the remove background extensions okay and these are already on your list you just pretty much have to enable them uh but basically if you don't have them just Google remove background extension for stable diffusion and of course control net and there's a million tutorials on YouTube that'll show you how to do that so try to install it uh which just basically involves putting the link on here and here clicking this install button but um but yeah that's uh that's outside the scope of this tutorial this one I'm gonna assume that either you're going to do this or you have already done it so those are the only two requirements for this technique so in the prompt I'm going to write um uh uh a shot of a boxer defending Advanced a punch I'm gonna run it with nothing so you can see what happens just with that and we'll kind of get an idea uh for my model uh you know you can choose whatever model you want I'm using this storyboarding model which has kind of been trained to do uh uh somewhat painterly techniques you still have to finesse it sometimes to get it to make be more sketchy but uh I'm gonna use that one because I am uh demonstrating how to create a storyboard panel that's kind of the main idea of this tutorial so I'm running it now and I left it at 512 by 512 and this is the image that it generates right and as I generate more and more images it's always going to be random unless I lock down my seed um but that's that's good enough I'm gonna go ahead and lock down this seat because I kind of like that sketch uh look and feel so I'm going to keep that I'm going to click on the this uh button uh reuse seed from last generation that's going to keep it right there so that's good uh but of course we have no control over the over it um I I could in fact run it again and it'll give me a completely different image all right so but I'm gonna keep that image and this is the this is basically what I'm telling stable diffusion to do a shot of a boxer defending against a punch now we're gonna make this into a um what was my Dimensions 768 by something three something uh in iclone let's see 768 by 384. that's a two to a two to one ratio if I can find my way back one more time confer 384 okay did it again okay three eight four by 768 I'm keeping the same seed I'm just gonna roll it again so you can see the aspect ratio change and you'll see that I have no control over what stable diffusion really does um but I'm going to now influence this pose by bringing in the first image player one and using it as a as a guide for what I want so I'm going to enable it I'm not going to select anything for preprocessor because we're not pre-processing it uh we did it inside of iclone so we kind of pre-processed it manually uh and the model that we'll be using is the open pose model all right so now that that's set up we're gonna run it again so you can see what happens so now you can see the pose that we created is following a bit more accurate the uh the post that we we made in an iclone uh there's a you know there's there's a few uh hiccups here because this arm is supposed to be in the back and that arm is supposed to be in the front uh that's just a product of the randomness and you might have to generate a couple before you get the right one but you know you'll get it if uh if you need any uh you can use your prompts for example we don't want the back and we want the character facing the camera so that's how we'll modify it will influence it let's see what happens now so now table if you should notice that the character is facing the camera let's see if it obeys they're still not obeying let's try one more thing we're gonna give it more weight facing the camera we're going to give it a weight of 1.5 and in the negative prompt we're going to add uh multiple characters because we don't want that second guy being generated and that might help it so it got rid of the second guy okay so I see that it's trying to work now I can see this arm is being an arm that's facing forward while that one is also a forward-facing arm but it's in the front so it's a little bit wacky um I'm gonna go ahead and roll this a little bit more I'm gonna let uh the CG I'm gonna put move the CG CFG scale a little higher to let uh um let's stable effusion try try its hand at being a little bit more creative there we go I think we got it so you have to finesse it a little bit you know use your prompts uh so my prompt in this case was a shot of a boxer defending against a punch and then in parentheses facing the camera uh with a weight of 1.5 and the way that you add weights like if it's something very important that you want to emphasize is you select a part of the prompt and use control up and keep pressing up until you get the desired weight number or press down to go down and wait so that's how you do that's a quick uh Nifty little trick that a lot of people don't don't know uh let me get rid of that all right so I have an image that I like I'm going to keep that seed uh I'll also in the negative prompts I I told it basically don't put multiple characters characters on my scene and uh don't show the back so that should help the the system get a better idea of what I want because it's not always going to be perfect but you can see now that it pretty man it pretty much matched the pose that I wanted almost exactly and uh here's the other trick if you have and this also this is a beyond the scope of this demonstration but if you can take the time to train your your uh stable diffusion uh you can actually maintain more consistent characters so in this case I have a character that's kind of based on my look uh I'm gonna bring him in now and this is done by using the the Laura that you create when you're training it okay and up and and if you guys are interested I can uh make a different tutorial that kind of goes a little bit into more detail as to how to create uh these lauras this training models so I'm gonna I'm gonna bring in the Ibis Fernandez at a 0.6 it's going to give them that little chin hair and spiky hair uh more of a Hispanic look to him so if I generate him just like that um I don't I don't think I need to put the keyword in but if I do I guess we'll we'll just add that in but see how it comes out so I do get the semi-spiky hair I don't like this particular art style so I started influencing a little bit too much so I'm gonna add the words here my name a shot of Ibis Fernandez a boxer defending against the punch and uh thank you I'll put a gray scale pencil sketch I wanted to find the art style a little bit more a grayscale pencil sketch of a shot of etc etc etc yeah and I think this should give me the final image of what I want I don't like where it's going with the foot but it fixed it okay so it ended up giving me this particular look and feel that I do kind of like but if you're not if you're not happy with it you can always do a multiple roll situation where you're gonna put like maybe like draw me four panels at a time and uh we'll roll it and then we'll pick one of those just just to see how many options it'll give us because sometimes you'll have to roll a bunch of times and uh and I mean and this it is what it is sometimes it'll take you 30 seconds you'll get it right away and sometimes you'll just have to uh finesse it until you get the exact image that you want and um in my case I think I was almost there so it's generating all the images at once and we'll be able to review them all and when I go for the second character I'm not going to take that much time okay I'm just gonna show you how the second character is generated and how it's composited nice and quick so it looks like it's almost done okay so the arc style is pretty consistent uh so that's always good we're gonna keep that seat so that the next thing that we do draw is going to keep that that consistency okay for for my taste I mean you can go full on photorealistic you can go on anime style whatever it is that you're going to be doing that thing is all up to you and your creativity and the models that you're using um sticking with storyboards because that's kind of what I'm working on right now uh and hopefully this is going to give you guys a better idea on how to better create your own storyboards or comic books graphic novels and all sorts of uh cool little uh ideas right so I kind of like this one over here uh I'm gonna stick with this one uh let's see here yeah I'm gonna stick with this one I'm going to open up the folder where it's being saved to and there is my image so here I'm going to copy it and let's see if I have my Photoshop running I do somewhere um so I'll go to photoshop because I was working on this storyboard before so I'll go to file new and it already knows that I have something on my clipboard so it already knows the the size that I want it in just Ctrl V to paste and now it's on there right so I'll move that out of the way let's create the second image so I'm gonna keep everything exactly the same all the parameters are exactly the same and I'm gonna clear out this image here for the pose and I'm going to bring in player two image and that guy is like this and uh so this guy I'm gonna change his name even though I don't have a trained model of him uh I'm gonna use a a a a character name I'm gonna call him give me give me Dean a boxer and what she's doing I'm gonna say that he's throwing a punch because it's you know you still want to help even though we have the motion there you still want to help guide stable diffusion as to what it is that it's seeing and what it is that you want from that so it's a boxer ducking and throwing a punch and and he is also facing the camera I believe so I'm gonna leave all that there I'm gonna take away this Laura uh you can add other things like uh so now that you established that you have a name that Jimmy Dean is a boxer you can also say you mean thing where's a mohawk and a tutu I mean Mohawk oh yeah that would be fun a mohawk a tutu and ballerina shoes all right that should be fun so let's generate it perfect so here's the thing once this image is generated oh I forgot I had left the the setting there for four Images so I'm gonna I'm just gonna have to sit through it and see what it comes up with in those four images and I'll just pick one of those that I like the best um [Music] so you know if you're working really fast you can easily get away with just doing one image at a time you'll once you get your rhythm going you have the seed that you're going to work with your desired uh art style and everything you can continue using that over and over and and nine times out of ten you will get lucky and just get it right on the first image so it's just a matter of you know just playing around with it at first just to warm up the system uh so here's my images there's one there's two there's three and there's four so I kind of like this one this one's a little bit more accurate based on my my pose here so I'm going to keep that one he's wearing his tutu and ballerina shoes and the mohawk that's beautiful so now here's the thing normally if you're gonna composite these things you got to go through Photoshop and remove the background manually and do all sorts of uh of fun stuff right that we all hate so uh what I'm gonna do is I'm gonna because we had that extension that removed background extension actually is going to do it from here it's going to use its own depth mats to isolate the character and cut them out of the scene so I'm going to send them to extras and at the bottom of the extras area we don't really have to do anything here we're just going to select uh uh from the remove background area I'm just going to use the first one U2 net I have no idea what any of these are all do but I know the first one works uh most of those work pretty much the same I I haven't noticed the difference yet uh so we'll click generate and we have our image with the background completely removed we can see we still have that extra limb there and I could have gotten rid of that limb if I wanted to by putting it in the negative prompt you know add in the words extra limbs you know stuff like that uh but yeah it's no big deal you know we're working in Photoshop so I'm going to open up my export folder and let's see here here's my image that I'm working on and from the export folder which is right here I believe no that's the wrong one uh this is the export folder it has its own little thing under the extras which is drag and drop it there and that's pretty much it you know we can start finessing it our degree of actual manipulation that we have to do to the file doesn't really it's very minimal so I'm just going to make this guy a little bit larger maybe that's too big all right and then we'll just use some of photoshop's built-in tools to remove certain things that we don't want but this one's going to use the edit [Music] or the image you know it's under where is it it's in the content aware field you got to make sure the correct layer is selected though that's what it is content aware feel I would have had the wrong layer selected and it didn't have any content to to work with uh so we're going to be removing this image here and really all you have to do is select the surrounding areas just draw a little brush stroke there and it's going to use that data to fill it in and uh yeah we're pretty much done we hit okay and that thing is gone and we'll do the same thing for this third leg here actually I I don't I don't think it's even that big of a deal we'll just put it out because it's against uh transparent stuff make sure you got the right sometimes you gotta rasterize these layers and before you can cut parts of them on there it is okay so there we go uh we have a storyboard panel that we created it didn't really take that long uh you know uh for some traditional artists uh you know you could probably draw this faster than it took me to uh to generate it but that's not the point the point is I can I'm not a traditional storyboard panel artist and I'm not able to do this kind of level drawings um that fast and I hate storyboarding so for me being able to do this is you know it's kind of a blessing uh so now I'm able to create really good storyboards without with very minimal effort uh so there it is uh we just made a storyboard panel using two characters that are interacting with each other uh you know you could have I could have made uh like let's say this guy's face was over here would you know cut this piece off here and have the the head pop up to the foreground so you can make them interact even more but uh yeah that's an easy way to have multiple characters doing exactly what you need them to do uh using stable diffusion and uh it all starts with this open post character which uh uh yeah you can use to uh you know you can do it with uh any kind of motion capture uh predefined uh poses inside of iclone uh it's a lot easier to work with because you it's it's like a fully rigged puppet you know let's see here modify animation we'll go to edit motion layer you're gonna raise his arm like this and we'll we'll just kind of play around with the camera so it's going to make a character that's post like this uh in my render tab which is going to change it to 512 by 512 which is what a lot of people are working with by default uh and I'll just kind of frame my character the way I want them to be you know I'm cutting off I can cut off parts of the body I don't have to do the full body I can just frame it the way I want like if it's a portrait and I can export that and use that as my guide that's my portrait I'm going to export a second one a full body one so I can demonstrate that I'm going to turn them around this way all right so uh we'll do one more quick demonstration and uh we'll uh call this video complete so I'm going to bring in this uh portrait shot change it to 512 by 512. we're going to use a different model see uh a portrait of a man be in thought cyber Punk Style night rain fog I will do in 80 millimeter lens Boca it's bokeh is that blurry that happens when you put things out of focus and what else can we throw in there we'll just leave it at that for now we have this influence going on and uh okay so here's gonna be my story uh my my storyboard sketch of that scene using this pose that I brought in looks like it's gonna give me multiple versions because I keep forgetting to put that number four back to a one I really only wanted one so for demonstration purposes but that's fine you know we'll see the multiple renderings and uh you can see that sometimes it doesn't get it right sometimes it uh but the post works you know um so here he is this uh aside from the hand that's not going up the the pose is pretty is fairly consistent that one's a little bit more consistent we got that bokeh effects in the background okay so you can see that because we cut off parts of the arms it doesn't have that elbow joint there but it has the other one bringing in it it does kind of struggle sometimes that does kind of tend to break it a little bit but uh you you get the overall thing and then you can of course you know tweak it further uh because one of the things is you know this character is facing the camera the man faces the camera so that should help it readjust because sometimes it'll take the pose but it'll put the camera in the wrong place so that's another thing to look for and give it a weight of 1.5 I'm only going to generate one I don't want to generate four again so you can see here that it's starting to give me the the angle of the shoulders it got them wrong got the arm right it built something here because it did see something there but I don't think it translated it into an arm but that you know that's still kind of fine you know I could go back into iclone and readjust my position let me see here oh tab so whenever that happens we can always go to iclone and just kind of readjust it you know we can always go back in here and uh maybe just zoom in a little bit more so it's not so broken and we'll render it out like that color portrait again so I want I want you guys to see this video as an unedited uh thing because you get to see what really happens you know the kind of mistakes that you're likely to make or the kind of little things that the glitches that you're likely to run into and some of the solutions that I come up with in order to get that that resolved so now it should be a little bit better because you can see more of the image there you go it's almost exactly the way that we have it in the pose so that that's really not too shabby you know we can if we want to turn them around a little bit more we can go back into iclone rotate the camera around a little bit more to push the angle uh we can adjust the head and then we come back here make him look up down uh raise his arm a little bit more so yeah that's actually not bad uh so it still gave me random um color styles but that's kind of the setting that I have on it right now uh let's try it with a different model and we'll get a different result but we're still we'll still use the same seed so now I don't have the anything model this should give me a more comic bookish animish type of style and uh so but this is not really about you know uh models and styles and stuff like that this is all about the open post model that you can use within um within uh Cyclone to basically control what you do in stable diffusion a little bit better so in this case it it thinks that it's using cameras the man faces the camera uh or maybe you got it from there but I'm going to hit put here the word camera we don't want we don't want any cameras in the picture so we put the word camera in the negative prompt so that should remove it from his hands so now you have a guy with no cameras and he's doing the post pretty much like like it was intended inside of the iclone thing um yeah you can clean up the picture more using Laura's and finessing your finessing your your um your your prompt and in your settings a little bit more um so let's let's try the other one the full body one let me see here let me try to make a really pretty picture um I'll use this high-risk fix with a r ESR again times four upscaler I'm gonna I'm not really upscaling it I'm just having a double processed so I'm just going to switch it to one I'm gonna leave everything here the same well high res I'm gonna do four high-risk steps and uh let's bring in the other one the full body image we'll render this one just as is and uh here we go So based on the blurriness here I can see that it is getting the the poses uh pretty much the way I want uh it's gonna draw the hands on the sides so let's see ours up let's add a little hint of all her art style 0.7 so I can see the arm up there but at some point during the high-rise fix it puts the arm down okay it looks like it's going to keep it up there this time so there you have it there's our image let's see I'm gonna do one more render and see what we get you should have told it to do something with like much brighter colors he wants to put a camera on his hand he thinks it's taking a picture oh it thinks it's on the phone yeah I can see that based on the image there it looks like he might be on the phone you're at the work phone [Music] let's see here and uh yeah there's a lot of things that you can do too as long as you don't overlap the poses you can also get away with doing character turnarounds let me I'm gonna wait for this to finish and I'm gonna demonstrate that yeah it's kind of cute I mean it gets the general pose because of the way that the angle is and everything so the this arm here should be at the bottom but it doesn't read it like that um so you'll have to like finesse the uh prompt a little bit better all right so now I'm going to go into my clone I'm actually going to do a another pose on this character content I'm just gonna make him standing like that and we'll go to this guy here which is going to delete him we'll go to the front and we'll just post him we got one two three okay it'll go to modify I'm going to make a bigger image this time around and it does you don't have to do all this you render your stuff out any way you like it's just ideally you want to render your things in a manner that's similar to what you're going to be using inside of stable diffusion so in this case it's a two by two to one aspect ratio 10 24 by 512 she's going to render that out we call it model and we'll switch over a model feet of a character let's say a man standing in a relaxed pose we'll just leave it at that uh nice and basic uh we'll bring in the reference image and we'll increase the size of this guy M24 and which is going to hit the render button uh I'm not sure if I need this high risk fix for something that's already this big but whatever this is the last demonstration so we'll take a we'll take a shot so I could have done like four different variations of this so that it'll allow the system to try multiple versions and one of those is going to be fine uh in this case oh it's going in a completely different direction it's giving me additional poses that I didn't ask for but uh okay sometimes that happens when you do the fire race fix it goes in one direction and then it changes Direction and draw something completely different during the upscale procedure but you can see that I have the first pose there and I have this pose and uh I believe this one is coming from the models but that's really not what we want we want to be able to like really control it so we're going to get rid of this high-risk fix for now and we're gonna roll the dice again oh we're good we'll get rid of that seed we'll roll the dice again see there you go because it's giving me the exactory that I wanted but during the high the high-res fix I have it set to a certain way that it kind of starts over I don't know why it does that but it's something to watch out for it does it sometimes and that's uh that's unstable diffusion side so it did it again do I did I leave it on I got restore faces why did I turn that on let's interrupt it oh it might also have to do with the fact that I've selected character sheet so that's my bad let's just say a man standing on a relaxed post yeah when you put character sheet it's going to try to fill in the whole thing with different poses uh yeah I should probably should have remembered that um so here we go because you can see when it starts working you can see the diffuse model there the actual poses that I want it's just that at some point it changes course on me and it starts drawing something else but these are things that you're likely to run into so by seeing me run into them uh hopefully you'll have a better understanding as to how to deal with them so I mean technically it's giving me the exact poses that I want uh ideally this third post should be looking away from the camera but that also has an easy fix so you can keep rolling the dice you can keep uh playing with it have it generate images and you know you might have to do it three four five times and it will give you the the actual uh poses that you want in terms of the direction to the camera but uh there's also other things you can do like you can send it to in paint and uh you can say you can just kind of mask this guy out and we'll we'll try to regenerate him uh looking away from the camera so we still have a man standing in a relaxed post with his back through the camera that might be enough we'll we'll find out I'm going to select the masked content original and uh on we only want to generate what's been what's being Mass because if I leave it here it's going to try to regenerate all of this uh sampling method uh I think I was using what was I using DPM plus sde Caris so let's see here anything else that I need to keep track of I'm going to bring in the control knit model so he knows the pose that I want it's just going to sample because the images match the same size it's just going to grab it from this region right here you got to make sure you enable it and select the model open pose right there okay let's uh let's see what happens crossing our fingers but yeah so don't think uh just because the software or stable diffusion doesn't do things exactly the way you want them to the first time around that it doesn't work or that you're going to have a hard time doing it it's uh it's really not that difficult once you get a feel for how to do things uh almost everything has a solution if stable diffusion gives you a hand with 16 fingers you can go in there and fix the Hand by doing pretty much the same thing I just did right now um so yeah there is our pose and there is our you know generated model model sheet and of course you know once you have your own characters trained or you get the right seat for the right characters that you want uh you know sexy females whatever animals uh uh you know posing them around it's not not really a problem the only other tips that I can give you is uh it works best for stuff like this with multiple characters under one single render pass when the skeletons or the the the posing models don't actually overlap each other it just it works a hundred times better you might get decent results every now and then but you won't have as much control over it and um and yeah you know make use of uh make use of the same seed and make use of the feature that uh the remove background extension because that's going to allow you to just drag and drop the two images that are pretty much matching in terms of camera angle and uh rendering style and you'll just drag and drop it onto your Photoshop and you know just put it on the layer and you're pretty much done um so for me that is what I am doing now is I hate storyboards man I hate storyboarding and uh but this is something that at least it's bringing up bringing back that fun aspect of storyboarding I guess that I personally never found [Music] um okay and uh yeah I mean I've never been able to create storyboards of this quality ever in my life and um I could never imagine actually paying for somebody else to have something like this uh made for me uh because it would be very expensive I'm sure and uh um I just I just can't see myself paying for that you know I I don't have that kind of a budget but having this superpower uh I'm able to have this type of content for my production my pre-production my pre-visualization stage um but yeah hopefully this gives you guys uh some ideas on how to work with the open pose model uh there are some freely available uh stuff that you can work with within stable diffusion but it's all manual it's kind of weird and uh you only like this one for example you can only work with one character at a time and yeah you really do have to pretty much um pose everything it doesn't allow you to bring in motion capture it doesn't allow you to bring in pre-made motions so for me I just I don't care too much for that then there's the open pose editor which is like a little drag and drop thing where you can make all sorts of weird here and uh it's not I mean it's it's it's it's pretty terrible in terms of the workflow you can still get some decent results uh sometimes but um but yeah it's not it's not it's not it's not very very user friendly uh so aside from that um there there's a there's one that somebody made for blender which is you know just as bad as the as the open post editor requires you to mess around blender I mean if you if you're a newbie with 3D stuff and I'm mostly a newbie myself so um doing all those tweaks it's just I can I can't imagine myself going through all this clicky Dicky Behavior just to end up with one post that in the end you know that could just draw myself and get it done fast so that's not the point right uh the point is just to be able to work fast and efficiently and get stuff done and get a pretty decent quality out of it so so yeah uh doing this inside of iclone is probably the easiest and most user-friendly approach that I've uh implanted and um hopefully you guys get some good use out of it all right so if you have any questions uh yeah it looks like we've reached almost a one hour mark so if you guys have any questions feel free to leave it in the comments or reach out to me by email ivis toontitan.com uh this model will be available uh probably by the time this video is up on the re-illusion marketplace and uh you'll be able you can download it from toontitan.com as well I usually put my stuff a little bit cheaper at toontime.com because I don't have to pay commissions uh to anybody but uh yeah if you want to get it from toontime.com I mean from re-illusion and uh uh you know go through their licensing stuff uh go ahead and download it from there and uh because they have that really easy installer that you could just you know once you buy it it becomes available in your in your character Library whereas if you buy it from toontang you kind of have to just drag and drop it manually into the right folder which is still not that big of a deal but um yeah so by the time this video is over um this uh By the time this video is up uh this model should be up there if you want to help support the channel if you want to help yourself uh if you find this useful go ahead and get it you know it's uh it's gonna be super cheap and uh hopefully you get some good use out of it so until next time if you have any questions let me know um Peace foreign
Info
Channel: Ibis Fernandez
Views: 6,999
Rating: undefined out of 5
Keywords: ai, ai art, iclone, open pose, openpose, openposer, sequencial art, stable diffusion
Id: 1Ee-bLdaEAc
Channel Id: undefined
Length: 59min 1sec (3541 seconds)
Published: Wed Apr 19 2023
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.