Artificial Intelligence is here to stay and as I see it we only have two options: start debating whether it is ethical and whether architects are going to be replaced by machines in the not too distant future or take advantage of the potential that this software offers us to use it to our advantage, that's why in today's video I come to show you a tremendously useful tool for your project deliveries, a program capable of transforming a hand-drawn sketch into a photorealistic render and the best of all is that it does it in just a few a few seconds okay, I'm going to be honest with you, I've been thinking about making this video for a long, long time, in fact I've been about to record it explaining how to obtain the same result with stable diffusion, which is an open source artificial intelligence engine designed To generate images from a text that we enter, that is, all we have to do is write some sentences in English with the main parameters that we want and the program will return four images with proposals that are more or less close to what we requested. With this Software, which by the way is free, you can achieve very, very crazy images like these interior views of cathedrals that will never exist or these utopian cities built on oil platforms above sea level. You have all the images on Instagram in case you want to check the As a result, the truly interesting thing about all this is that not only do they serve to create architecture, we can also train Artificial Intelligence with photographs so that it learns our face and generates results as absurd as a pencil version of ourselves, a self-portrait of Van Gogh, a kind of Batman from the future or what we would have looked like if Darth Vader had taken us to the dark side with his laser sword and everything. In today's video I am going to show you another tool that is not this one but that is much easier to use. What I'm going to talk to you about today is called defy and not only is it easier to use than this one, everything is done online from a browser, any browser you have. And the good thing is that anyone can obtain very good results in a short time, you don't need a computer. not much power nor a quality graphics card [Music] come on, let's start, the first thing you have to do is go directly to their website, I'm going to leave you the link in the description of the video so that you can access it easily and simply here in the box where it says enter the email because we are going to enter our email. It took me a couple of days but it is true that I requested it on a Friday and the access arrived on Monday but in general it is usually quite fast, I tell you that I know people who swim in One day they have managed to have access, once you have access, this is the graphical interface that the page has, as you can see, I am going to show you a little bit about what the basic tools are like, we start with a description of what the platform is, as you can see, here is a tab that, well, you can choose between architecture or generic because this is an Artificial Intelligence tool that is designed specifically for architecture but if you are not going to do architecture, you can click here on generic since we are interested in architecture, we are going to leave this marked Here too You can choose between a façade style, let's say, and Soon you will be able to create plans, which is quite interesting. You can even make a video later explaining this process, since right now it cannot be marked, let's leave the one that comes by default . Let's explain the different tools. First we have text creation, which is what you have here. Okay, here what could be done is in English, as I told you before, write what you want to achieve. No, for example, we can simply give a very basic description. Okay. I'm going to give it generate so you can see how it works. You can see how little by little the machine is working and generating four images that are in principle different from the description and the text that we have given it and you can see how quickly what it has achieved. This would be the description of a contemporary cabin in the forest with white walls, you see that there are four images that are very different from each other can be opened and compared, okay, you see that the result is quite good simply to obtain from a text another of the functions that it offers, we are going to see is that if you If you click this button, this same image will be generated in HD quality because if you want to use it for plans or for presentations, the image that is coming out of the computer right now is low resolution. So if there is one in particular that you like, you simply have to click on the button. HD and it generates it in a higher resolution, the image in HD, as you can see, has a higher resolution and the details are much better, which is interesting to use it in projects directly. Not here you see the difference between the default and the HD, you can also see that Even in the field there are more details not only that they gain in size but another of the buttons that is interesting is this one called image variations that we can give you and what the program is going to do is generate four other images from this one that We liked the ones that it has already generated for us and they are going to give us four other options based on creating something similar to what already exists and these would be the four variations that the program has given us regarding the image that we had asked for, you see. How good There are certain concepts that it preserves And that idea of a contemporary cabin in the woods Well, it can help us to create variations of the same image. Another thing that is very good, especially when you have been with Artificial Intelligence for a while, is that many times the words or Perhaps you are not clear about the concepts, you are not sure, of course, put no then, what this program offers is here, let's say that a list of words with suggestions allows you to create no then so that you can see a little idea of the things that can be done. of architecture we are going to try, for example, to select a bird's eye view, we are going to create a super skyscraper with a modern futuristic design designed by Taja Hadid and Rem and we are going to press generate to see what we get here in the text box you can see the words or The concepts that we have selected to generate the following images I say this in case there are any that do not convince you, you can simply delete them by deleting them with the keyboard. You see that in general it is quite fast. The images here are amazing, not really amazing, just with a text, we have created an architecture like this, not that it may have its defects or its inaccuracies, but I believe that as a basis for us to have ideas for projects as references, it is not brutal, it is truly brutal, I believe that in the end we must take into account Take into account the limitations that Artificial Intelligence itself has , but I already said that to obtain at that moment in which we are trying to guess what our project is going to be like, we need many references or to see things that give us ideas because Artificial Intelligence is incredible before moving on. To the next section, which is this one here, I am going to show you this box called exclusion, which simply opens a dialog box here so that we can write what we want, not simply with this description that we have given you, if there is something that we do not want to appear. well we would simply have to write it here in this text box It is simply like the negative not what instead if here we are writing the positive what we want to appear in the images here we have to write what we do not want the images to appear another of the The options that I am going to show you are this one here, the Lora model. They simply add a series of models that may interest us. Well, due to different parameters, not for example. You see that in terms of plans, there are already designed ones that are called architectures, which is not what it is . directly because this type of style that we can use for when we can create plans as at the moment it is not possible because it is not very interesting but there are other options for example to create projects in the style of Zahaja no Although the name of Zahaja is already in the in the names In the labels that we can add in text, you see that here there is a series of let's say that a model that is what generates you is therefore some images very similar to the renders that are used in, for example, with this description that we had marked down here you see that there are more styles. I don't really like these two down here that they have been adding that it does generate a very different visual style, it's not like that a little bit like that of a pretty cool comic and let's say that this result would be applied directly to the text that we have already entered before. So we are going to use it so that you can see what the example is like and what results it gives and you see the difference with the images from before, this would be the skyscraper that we have asked for style Hahaha Only with this visual style that we have marked No, you see that it is much more suggestive that it can even be used for the plates, that is, a drawing of this type is beautiful to see, okay, let's move on to the next section What I'm going to do is Delete all these labels that we had. Here I'm going to remove the image of the Lora model and I'm going to show you this section here, which for me is one of the best it has and that's a bit of what's interesting about the video. We are going to create a project, we are going to create a render from a hand-drawn sketch to make the video faster. What I have done is download some images of sketches directly from the Internet. You see, here is an image that is apparently no pencil and it is what we are going to use as a base to create so what we do is here where it says let's upload this image and here where it says diagram Style this is a bit of the reference that we want to use for the materials for example of the render so well As the image is of a house, so a little modern, what we are going to do is we are going to go to Google and we are going to look for an image that will serve as a reference that we want to use for the materials and we are going to select one that, for example, draw attention a little to the style okay this one you see that it has wood a little bit of stone it looks like glass let's go to the image we return here to the page and upload the image that we have downloaded for reference here we have the guide for this one it is not in 0705 it What we are going to do is raise this one a little more because I am interested in making it look more like the sketch. We are going to raise it to a 0.8 and this one, which are the materials, we are going to lower it a little more so that in some way we can say that the machine can play a little with the image we are going to generate it to see what it gets us and it has already started generating with the two images that we have given it and you are going to see how the result is truly crazy and in just a few seconds you can see how the program gives us has generated from a sketch made by hand 4 photorealistic renders that are quite similar to the original, you see here in the compare with the original button, you see above all what the volumes are like, not that it preserves them quite well, there are issues that The best thing is not clear, for example, this part here does not fully understand if it is a pavement if it is a swimming pool, but that is what is interesting that there are no images, since we cannot find the one that most closely resembles what we see. Interesting, not from there the same thing, if one of them catches our attention is except for the theme of the pool that has been considered that instead of a path in a pool it is quite suggestive, we can create the HD version or we can create four variations of this image, which is what has caught our attention, one thing we are going to do so that you can see the tremendous potential that this tool has is we are going to eliminate here the two images that we had as guides and what we are going to do is load For example, a sketch of frangiri, the sketch that he made specifically for the Google engend museum in Bilbao and what I am going to do is upload an image, a photo of the real Google home and I will use it a little as a guide, we are going to lower that also to 0.4 The interesting thing about this is that you see that the tools, let's say the options that the program has, can be combined with each other, so here I go to text and what I am going to mark. There are some tabs here and a very interesting thing that I also wanted you to see. is that what this program also has is here in output Style we can choose If we want it to be a sketch, that is, a sketch or a model, then I am going to select model so that you can see the result that it is going to give. to give to generate And this is crazy because you see the results, that is, they give us four images of something that could be a model of a sketch and the result is amazing. I have to tell you that it was a legacy with the models in the race. and having had a tool that generated images for me, well, you see, not from models, it would have been a lifesaver for me, the career but brutal. Here is an option that is like the history, here we can see everything we have generated and I am going to show you because there are some images that I am testing before preparing the video, you see, for example, the towers that I was creating from one that caught my attention simply by inserting text and giving it a real madness that the program can create. What seems brutal to me is the issue of the models, not with Frangi's image I asked him to create a model for me and you see the details, that is, it seems that it has been built with wood or I don't know if cardboard very good No, but it even uses several types of materials but the result is certainly very crazy, it would be very worrying if the teacher told us to bring that model to class when in reality it has been created by Artificial Intelligence But well, I'll leave it here in case you It can be used for the race or for a client to give you a feeling like this of how you can see something created, it is quite interesting, you see, here I was also testing zahajadiz-style complexes with an aerial view before moving on to create interior images Because I also believe that It's a very interesting thing I'm going to show you this here is the part like Advance here you can choose if you want the program to be more average or if you want it to be more creative OR you want it to be more precise from the text or the images that we are giving No right now I have been generating images with this in the middle and here we can choose to also choose the images that it is going to give us from four to nine from one to nine Better said I have it in four and the size of the images , this is also a little how long it takes to return the result to you. I usually leave it by default. If in any case there is an image that catches our attention, what we can do is click here on the HD option to that we create this image with high definition quality what we are going to do now so that you can see that it also works very well it is good I am going to delete the texts that we have entered I am going to delete the guide images and we are going to create an image of a interior we give it here in straxer we are going to upload an image I have downloaded interior sketches so that the video would also be faster and you will see how well it works we are going to upload this sketch it is an interior sketch and for the style we are going to choose for example this modern style here what I have discovered is that many times when we try to make an interior the program does not quite understand it, maybe it thinks it is on the outside so what works for me It is simply here in the text to put interior view in English and with We are going to generate this data that we have given and we are going to see the result. These are the images. Well, there we have everything, but in general, if we compare it with the original, you can see that the result is quite similar. Here I tried to simply leave it a a little more creativity to the program The size of the image I kept And this is what we have obtained, you see that it works Also very good for creating interiors from hand-drawn sketches I am going to show you what happened when it was not put the interior part You see that in these images here it is left like the sky, you can't see any trees, it looks good, it gives the feeling that this is open, that it is an exterior and that the house begins here and we didn't have that part, let's say interior of the view of the house no So if what you want to create is an interior, it is important that you put the interior part in the text so that the program clearly understands that what we want is an interior no, you already see it marked here and well, images as crazy as These ones that I am showing you can be very absurd or they can really serve as inspiration when we are facing a problem. interesting project that we can change the view, we can tell it to be a model, we can tell it to be a hand drawing, we can tell it a thousand things, we can create something from some sketches as we have shown you, but the possibilities are endless and well, you see, you don't know. that can be created simply with images and giving it that wooden tone. If you compare it with the original, this was another test that I did before making the video. This is the original, the original image and these are the renders. You can see that it perfectly preserves what It is the chimney the vol the overhang And from there it is to vary a little what interests us in terms of materials in terms of light in terms of the background the clouds you see, not everything is quite customizable these images that you see here I believe also from simply Frangi's sketch and you see what it generates for us in models, it is not a real madness, it is something that can be passed off as a model, the problem is only that the teacher tells us to take it to class , there we go to be Dead but if we can't plant it on our sheets without any kind of problem, for me the most interesting thing about the program is understanding that the artificial intelligences are starting to be very green. You see that right now the only thing they generate are images from a text but this will continue to advance in an exponential way surely Soon we will be able to create plans Soon we will even be able to adapt plants as we move the mouse but the result, you see, is super interesting enough to be very attentive to everything that is coming in this world of Artificial Intelligence and that we will be able to transfer in some way to architecture and that's it for today's video. If you found this topic of Artificial Intelligence applied in architecture interesting, leave a comment below so that we can make more videos similar to this one. And if you want to discover everything that is hidden in other AI tools like stable, I tell you the same thing, you have a box below where you can write your suggestions at the end. For me too, it is important to know if this content that I am creating for you is useful to you during the race. or during professional practice and nothing more than we will see each other very soon in the next video [Music]
