Consistent Characters - Using New GPT's

hey everyone in this quick AI animation related video I wanted to share some new character consistency methods I stumbled across via a new GPT available on the new chat GPT store here we go first of all we hop into chat GPT and we can go to the new explore gpts Tab and we can search for character and you'll see there's one called consistent character GPT you can then click to start designing your character and then go through answering the various questions so we're going to have a male character please provide a name we're going to

call him Bob it then asks for characteristics 12-year-old white American messy dark brown hair yellow hooded top with black skull graphic wearing skinny black jeans oversized black frame glasses blue eyes oversized black Converse trainers it then prompts for a theme for the image and you can choose from photography Pixar Animation 2D flat illustration or anime pastel and I'm going to go for Pixar Animation it then asks what to to vary in each image and I'm going to go for expression action and time of day okay and then we're going to give it our prompt for

the first image happy walking forward daytime and there we have our first image and it's got the messy hair Pixar style oversized black frame glasses the yellow hoodie with a black skull those black skinny jeans and the black Converse we can now prompt for our next image and I'm putting sad sitting at desk facing camera dusk you can then keep answering questions and generate different consistent character images rear views side views close-ups um different environments to create the images you want to use in your work potentially taking them into Photoshop I really am very impressed

by this consistent character GPT it does what it says on the tin it's actually consistently making very good characters I'm not cherry-picking the images I'm sharing here each time it's producing something aligned to our character there are some small variations in the design of the skull on the jumper between shots you could stop here and use these character images in your project but I'm going to go one step further and now use these images to kind of train up a prompt for Mid journey to allow us to insert this character into some new MID Journey

V6 Generations okay I'm hopping over to Mid journey and I'm going to quickly show you how to use prefer option set and then do a second one where we're going to use those images to create a pre-trained prompt so if you press forward slash and type P and choose prefer option set and we can give it a name so I'm going to call this set to and press plus one more and choose value you can then write a text prompt that you might use regularly in your normal mid Journey generations for example cinematic Dynamic lighting

bocker D- a69 and press enter and that will create cre a new custom option called set two you can then press for/ choose imagine and write out a new text prompt man escaping from Alcatraz then type-- set two and press enter and it will replace that set two with our cinematic Dynamic lighting Boker AR 16x9 and now going to do it again but this time making use of those character images so in mid Journey via the plus button I've selected the various images choosing ones I think give the most variation ones with eyes closed big

steps sulking with a hood up side profile and then various ones where we're seeing the hands as well I'm just going to press enter so they upload to Mid journey and then I'm also adding a cropped in image of that skull by itself which I just quickly did in Photoshop just to help give mid Journey a bit of extra guidance for that graphic I then press SLP choose prefer option set give our character a name let's call him Bob 2 press plus one more press value and then drag down each of these images as you

drag them down it will bring in the URL press space after each one and drag in the next one and then at the end you can add a text prompt as well which will be reused each time you use this pre-made prompt I've added the text 3D male Pixar character dark brown messy hair blue eyes oversized black frame glasses yellow hooded jumper with black cartoon skull graphic black skinny jeans black oversized Converse trainers and then press enter at this point I was going to say it's time to go for a run have a cup of

tea sit back and have a beer as it can take quite a while to go through the process but I did find that it actually just gets stuck on Mid Journey bot is thinking but it turns out you can actually use the prompt relatively quickly I'm not sure how long it takes maybe less than 30 seconds so you can press dismiss message and then write out your prompt for/ imag d-bob 2 and then text to describe our scenes dancing with hand in the air and then Das Dash set two which is the other Shand prompt

we created earlier on and press enter and it will write out the prompt with all of those image references that text description plus the shorthand cinematic Dynamic lighting Boer AR 16x9 okay and now we have these images of our character and admittedly it's not anywhere near as consistent as the ones you were seeing in darly 3 in the GPT but it does show how you can potentially utilize those character images to direct your mid Journey output okay and there we go hopefully this video looking at character consistency using these new gpts on chat GPT and

looking at ways of being able to use those images within mid Journey has been useful and opened up some new possibilities for your projects all right so as always please press like subscribe comment I'm John Draper thanks very much [Music] cheers