ADOBE AI Firefly Explained

today we have our major announcement in the AI world as Adobe is bringing its own AI Suite of tools onto the market and they have a number of exciting and important features we're going to run through them all so you understand the latest state of play and the AI art World Adobe sits in a unique position for the ability to exploit the emerging AI art tools it already is the biggest player in the design Suite industry and it can automatically add some of the most powerful features that are coming out straight into its own mature software this will give it a lot of Leverage and power over the existing players in the AI Art Space so mid-yony Dali and stable diffusion better watch out because Adobe has come to the table first up we have Firefly which is a generative AI made for creators with the beta version of the Firefly model you can use everyday language to generate extraordinary new content one of the key and most interesting aspects about Firefly is that it's been trained on a data set for which it was explicitly given permission to use much of the criticism for earlier models of AI generation come from the fact that it was using art that it did not get the permission to use it was crawling the web and you could easily go onto a website to see if your art or your face had been used to train these models many artists were in uproar that their own unique style could be easily imitated in seconds but now Firefly has changed the game and actually released a model that is only using open source Works however it will remain to be seen whether this will impact the quality of the images that come out as well as also massively impacting the ability to interpret existing artists Styles it's very good for artists and perhaps not so good for AI artists the cultural approach to how we deal with this new technology is going to happen massive ramifications for all areas of society and is going to take time for regulation and attitude to form around how we want to use these tools looking forward Firefly has the potential to do much much more Firefly will move Beyond static imagery and allow you to use vectors and even edit text and video they're planning to bring in generative graphic design tools which is something that many AI art generators are still struggling with the ability to combine text imagery and layouts in a meaningful way and this is going to change very rapidly so hang on to your hats ladies and gentlemen as the world of AI takes another step into the future and with that we'll also see 3D models being excellently outputted in AR one of the features I'm most looking forward to is the generative text effects that they're previewing on their site let's take a look at their advertisement video with Adobe Firefly you're able to add generative imagery on top of type forms you can simply enter in a text prompt like melting chocolate and suddenly you will get the word yum with melting chocolate applied on top creating some mesmerizing beautiful typography imagine the opportunity for typographic designers to now rapidly create unique and excellent fonts for many situations this is a great opportunity to actually leverage AI straight away and create your own business proposition so you can create a number of fonts and sell them very easily you can see how simple it is to adjust some of the effects on the text you start off by a choosing a base shape of the font you have the options of acumen alphlam Cooper black Poplar Postino and sanvito and then on top of that the effects are rendered let's take a look through some of the other features that adobe has revealed so with their text to image generator first of all you have a visual style interface rather than it being in a chat application like mid-journey which attunes it a little bit more to a program like Leonardo you can see that they have a beautiful visual approach that allows you to change some of the aspects this gives us the tools to rapidly generate and prototype our images faster and faster they've segmented the process into a number of component parts first you choose the aspect ratio followed by the content or the medium then you apply an aesthetic style and after that you have the ability to adjust the color composition lighting and other features beyond the text image effects they have you also get some in-painting features which allow you to edit inside of the images so you can mask this gentleman's shirt put on the nice Red Jacket damn I'm not so sure that is a good fashion choice but I'm not here to judge we are here to explain the features there is a smart portrait feature which allows you to instantly adjust the smile the age the jaw and the eyes within the face of your subject this gives us great control for changing the feel and content with the smart portrait feature it's excellent to use it inside of comic books where you can easily adjust the Expressions on your subject's face there's options for eye wrinkles so you can give them more natural smiles also eye-openess and you can even adjust the length of the bang depth to image is the ability for the AI to understand the relationship of 3D objects in a space this allows us to adjust the contents of objects in a space so for example you can see how different pieces of furniture relate in a room and it can identify say that there is a sofa and change the fabric the Furnishing or the cushions on top of this sofa so this is great for interior design so imagine you have a dilapidated room or building and you want to imagine how it would look with different styles applied on top perhaps an art deco style perhaps something slightly more like a urban Farmhouse with depth to image you can easily see and mock up differences in seconds this is great if you're planning to refurbish your home or if you're flipping a house and you want to see what it could look like with a little bit of tender love and care extra template is an AI generative tool that allows you to create layouts for invitations and other graphic Microsoft did release not so long ago a beta version of a similar tool called Microsoft designer I reviewed that and found it to be a little bit underwhelming however this does look like an improvement I hold back my full response until I see it working I think that being able to put text images and layouts together in a meaningful way is actually one of the hardest creative skills because it takes complete understanding of the meaning and the importance of the piece it's not just replicating the real world it's actually having an understanding of visual communication and the notion of hierarchy and how people will navigate through not only the images and text but actually the process of understanding a piece of information is much more complicated but I've been pretty astounded in the last eight months so I'm prepared to eat my hat you can see that these example templates that they have outputted are fairly basic there is not a lot of information on them you would expect to also have an address some contact details perhaps more of an explanation of the itinerary of the day at the moment all we know is that it's Mary's sixth birthday we've got no idea what time we need to be there whether she wants a present or will there be drinks that's the most important thing tell me will there be drinks Adobe Beyond this they are demonstrating conversational editing and this is a feature that allows you to iterate in a chat-like way it's almost like using chat GPT and integrating a AI art generator inside of that which is quite an interesting way to approach the problem and I do see us moving more and more to chats like applications for all of our means having the flexibility to work in a conversational chat-like way and also move into a more specialized editor or give us greater freedom to work more rapidly and have more control when needed I'm interested to see how text to Vector comes out you can see here that their example is very low Fidelity and it's almost like taking early versions of mid-journey and having them output it as vectors but I think that even if you specify in mid-journey right now that you want a vector that it will output something that's very easy to translate into a vector using Auto Vector generation tools they also demonstrate some features we've become accustomed to 2 such as combining photos or blending photos as they are called inside of mid-journey where you can take different images and combine them to create unique aesthetic combinations where it takes the content and the style and merges them together in interesting ways they demonstrate AI upscaling inside of their tour which is a great feature to have and especially if you have complete control over the resolution and the size of the upscale that you want Adobe is looking to create AI generative tools that it can apply to all of its existing Suite of design software one of the interesting features it's displayed here is to use AI to generate brushes to use inside of its painting softwares here you can see that it's asked it to generate a bra at the moment you need to sign up to join the beta and have Early Access to this and to do that all you have to do is come to the link in the description below click on join the beta and from here you'll be taken through a short form you have to request access and put in your name email address and your region and tick all of the types of Creations that you have done in the last 12 months make sure not to tick none of the above little sneaky one at the bottom there at the end it asks you what do you currently use or want to use a generative AI for and here I will simply put a make a nice stuff go ahead and tap on next and you all you have to do now is sit back and wait a few moments later people are understandably excited by the new announcement the text of vector is excellent and something we wanted for a while and the upscale at the end was an expected bonus call to see you incorporating the more or less smile somebody else is less impressing we can do all of that into stable diffusion already but it is very good to see it in a simpler and directed interface for creators and this is the key is that adobe has a large experienced team of product designers and developers who are able to turn the experience of using these AI art generators into something that is seamless and reliable Paul says it's weird that I'm most excited about the magical upscaling of images well pull each to their own it's interesting that a lot a lot of people are mentioning the upscaling features one that they are most interested in I am most excited about the text effects text to text effects well does that make sense the text to text image effects I'm playing out with those as well as exploring the text to template design which I don't have high hopes for but I am excited to see someone has mentioned that we're going to see hundreds of advertising Industries go underwater or at least massive layoffs I would counteract that this gives people better tools to work faster and raise the quality of their work so if you're able to incorporate these tools quickly I don't see you going underwater of course it's going to allow people to create work more easily and they might not necessarily need very entry-level designers to help them do that but that's great for everyone because we will be able to work faster do more and be more entrepreneurial with our own lives this is an exciting time and it has taken a little while for Adobe to actually bring these features out out which is to be seen in a large multi-billion dollar organization however the sleeping giant has awakened and it won't be long until we're going to have a major battle on our hands and what do you think about Adobe Firefly how will it change the landscape of air generators thanks for watching and make sure to watch this video next for more on the latest AI art news

Channel: Samson Vowles - Delightful Design

Views: 16,627

Keywords: adobe firefly, adobe firefly beta, adobe firefly ai, ai art, generative ai, adobe io, firefly tutorial, generative ai future, adobe, ai, ai news, ai art news, adobe firefly app, adobe firefly images, adobe firefly text, adobe firefly announcement, adobe firefly generative ai, adobe firefly beta price

Id: 8lrYfNgyW9s

Length: 13min 4sec (784 seconds)

Published: Thu Mar 23 2023