Speaker 1: We're about to combine the incredible power of Chat GPT and Runway to create lifelike cinematic videos just like this. I'm going to break down exactly what it can achieve and show you how you can do it too. So how do we write effective prompts for video generators? Well, the key is that we have to break down different parts of the shot and explain them with precise detail using language that is specific to the situation. So we're going to look at a few different variations of this. The first is going to be text to video, which is where we'll define an entire prompt to create the video. For example, if we're using a text prompt, we're going to make sure we're defining the medium, which will be either a cinematic shot or an animation. We're going to define the subject or the protagonist. What is the character that we are looking for in the shot? We're then going to define what type of motion that they are engaging in. How are they moving? Are they dancing? Are they walking? Are they talking? We'll add in the scene, which is the surroundings, the environment that our subject is in, and define any scene motion. Now, it's particularly important here to add some dynamic cinematic elements, and that's why we're asking ChatGPT to help us come up with ideas that are going to improve the fluidity, the motion, and the cinemagraphic qualities of our work. So this could include having rain or waterfalls coming down in the background. Perhaps there's some smoke or fog, or even some lights flickering on the distant hilltop. Now, importantly, is we then define what this has been shot on. The type of camera that is being used has a huge effect on the quality of footage that you get out. And we can also define the camera settings, which allows us to play with things like the depth of field, and also the exposure. Now, the next element is to consider the camera motion, which is how is the camera moving in relation to the subject and the environment. Is it static? Is it zooming in? Is it zooming out? Is it coming and tracking an individual moving at the same speed as them? This gives us a whole host of different cinematic techniques. And finally, we add our aesthetic intent. This allows us to define a brand, a feel, and an atmosphere to our work. And again, ChatGPT can come up with a whole host of specific, articulate, nuanced approaches that will really take your footage from generic to stunning. So to collapse that even further, you really have to think about three things, which is the subject, the scene, and the camera. Now, I've put together a whole prompting guide for this process, which you can get for free. And that's available in the link in the description below. Inside that, you'll have all of the prompts that I've used in this video laid out and organized effectively. You'll be able to go in here and simply change any of the variables that you want to get exactly the same effects. So let me walk you through the prompts that we're going to be using in ChatGPT. And we're going to be saying, you'll act as a prompt engineer for generating AI video prompts. I will provide a short content idea, and your job is to enhance the prompt by making it detailed, specific, and coherent, focusing on the key elements needed to build a strong video prompt. Use precise language related to cinematography, photography, and aesthetics. Your goal is to create clear, cinematic, and realistic prompts. I will give you a content idea, and you will transform it into two detailed video prompts. Remember to keep it concise, remove brackets, and be specific. Next up, I go into the guidelines for the video prompt, which summarizes the formula that we have just discussed. And finally, I add in an example input, which will be the general concept that I'd like to have, and an example output about how we've taken a simple input of a man walking through a rainy street at night. And I can show you what this comes out as in Runway. And then it has turned it into this much longer, more specific prompt. If you're watching this, you're probably just as excited about implementing AI into your business and workflows as I am. That's why I've teamed up with HubSpot to bring you their latest AI guide completely for free. It's called the ChatGBT and Work Bundle, and it's packed full of a whole host of useful tools to help you implement AI into your business. You'll learn things like when and where ChatGBT fits into your workday, how to seriously level up your productivity with AI, and how to smoothly integrate AI across different areas of your business. My personal favorite is the ChatGBT Decision Flowchart, which helps you decide exactly when and where you should be using ChatGBT. Now, this is incredibly useful because I believe that AI thinking or AI prompt thinking is the way that we can understand when and where to apply AI into our workflows. There's also 100-plus example prompts, which are real gems of the bundle. Now, I like the Refinement Checklist, which helps that your AI-generated content does not sound like AI-generated content, which is incredibly important because if you're like me, sometimes you're going to receive an email and you're like, my God, that is absolutely not you talking at all. This has simply been generated by ChatGBT, and I certainly feel a sense of disconnect with the people who share those emails with me. I know you're going to find this super useful, and the link is available at the top of the description. And a big thanks to HubSpot for sponsoring this video. Now, back to Runway plus ChatGBT. And then it has turned it into this much longer, more specific prompt, which is giving us a lovely sense of drama. And you can see this is the output, and it is a completely different world of footage. Let me show you some prompts that were run before using this method and what they look like after applying this technique. So here is another example where we have a woman walking in a street, and we've applied this process, and now we have this beautiful cinematic and stunning piece of footage. So to use this, all you have to do is go into our dear friend ChatGBT. We're going to start a brand new conversation, and we're going to paste in the prompt. Now, immediately, we can go ahead and choose an input. So I might go for something like, man riding a tiger through Indian rainforest. Now you can see immediately, we've got two much more detailed prompts coming out. So let's put all of these into Runway and see what we get. One of the keys to writing good prompts is simply having an understanding and a broad range of vocabulary to be able to define an idea in your mind and then articulate that into language that can be interpreted by the AI. So here we have the first input with just the simple basic prompt. And you can see it's not so bad, but it doesn't have a real cinematic feel. Now, this is the first prompt that we got out of the process, and you can see that the camera movement and the relationship between the two is much more engaging. Now, what we can do is we can actually ask ChatGBT to enhance these prompts with our own intent. And this is where things start to get really interesting. So I might say, I want it to be a tracking shot with a heroic, attractive warrior prince. It should be a period piece with green and purple tones. And here we have our heroic prince. So let's try one more of these. So this time I've asked for a beautiful mermaid swimming with thousands of fish and sea creatures. So here we have the initial video that comes out of Runway, and this is the step up. And you can see we're getting a better use of light here, and the refined color palette gives us a real sense of immersion. Now, in this variation, we get a more creative host of fishes. Now, we've had a look at how we can work with text to video, but what I think is the best process is using image to video, which is where we're taking an image from an AI art generator and putting it into the video generator. And I like to use, most of the time, Mid Journey for my image generation. So what we'll focus on now is actually the differences when using an image inside of your video generator. So I'm going to be working with this image of a female gladiator, and I'm also going to be doing it in vertical this time. So what you can do is come back into the Runway session, and we're going to upload an image. I'm also going to be doing this in Minimax, which is a alternative AI video program, just to show you that it can work in both programs or any AI video generator. And the advantage of Minimax is it has free credits every day. So you can even try out this entire process and follow along for free. So once we've uploaded our images, we're going to come back into the Prompt Guide, and we're going to go to the Image to Vimt Prompt Engineering. From here, we're going to copy this entire prompt, and we're going to come back into ChatGBT. We're going to paste this prompt in. Now, the difference with this prompt is that here, we already have a lot of information about the starting point of the scene that is given to the AI with the image. It already understands a lot about the scene. It has the protagonist defined, and it also has the environment in the shot. Now, this means that we don't have to spend so much time defining exactly what is the environment and what is the character. But what it doesn't know is three things, which is one, how is the character going to move? Two, how is the environment going to change? And three, how is the camera going to move? And that is what we have tweaked inside of the prompt. We told it that these things are now much more important, and we need you to bring this information to the front. So to help ChatGBT, what we can do is upload the image and then give the general input that we are asking for. So my short concept is an orbital shot of a female gladiator for battle. We can then add this is the image that will start the shot. Now, it's useful to add the image in because it helps ChatGBT have additional context for working with our prompt design. They can really see what is inside of the image. So now I can take this prompt in to Runway. Now, sometimes ChatGBT might go AWOL, and you can simply bring it back on track by giving it the prompt again. Now, we'll put these into both Runway and Minimax. And in Minimax, there is an option to check Enhance Your Prompt, and we're going to uncheck this. This is because we have the art of prompt writing mastered, and we do not want Minimax applying its own spin on our prompt. So here we have the first edition coming out of Runway, and you can see we don't get so much motion. It's somewhat in slow motion, and it's not a particularly cinematic or engaging shot. However, in this version where we use our prompt, you can see that we get much more dynamic camera movement, and we get some more engaging, environmental pieces of atmospheric animation, such as this dust being cast across her shoulders. Now we can look at the Minimax edition, and you can see this was the initial version where we simply used an image prompt. And here is the version where we used our fine-crafted ChatGBT prompt. And as you can see, the camera movement and the atmospheric movement is highly enhanced. Now, I want to show you a more advanced technique, and this is about using ChatGBT to create unified aesthetic prompts. Now, this is perfect for creating a series of shots that are going to have a consistent look and feel. So if we want to maintain an aesthetic or stylistic intent so that all of our shots look like they form a cohesive piece of art, then this is the perfect process. Now, there is another prompt for this, and what happens with this prompt is you simply enter in a concept and it applies a number of specific effects to your prompt so that whatever subjects you put in, you're going to get out a more enhanced prompt with consistent stylistic choices. Now, these stylistic choices include lighting, colour palette, the type of camera that is being used, the subject and scene motion, and compositional choices. What you can do with this prompt is simply take this and then choose what elements you would like to keep the same every time you create a video. So let me show you how this one works too. You can simply take the prompt, come back into ChatGBT. Now, you may want to start a new chat if you're working on an entirely new project. You can paste this in. Now, you can give it a few different ideas and say, give me prompts for each of these different shots. So I've asked for three ideas. Idea one, a man walking on a high ledge, Prince of Persia style. Number two, Prince of Persia sword fighting thugs. Number three, Prince of Persia kissing beautiful women. Now, we have three whole prompts. So now we have our three videos that have come out from our method of creating unified text-to-video prompts. And as you can see, we start to get a consistent aesthetic which allows these clips to be put together to create more of a cohesive piece of film. But let's take this one step further and use the most advanced technique that I have developed. And that is using this prompt, which allows us to create both an AI art prompt and an AI video prompt from one single concept. So you can take this whole prompt from the prompt guide, paste it into ChatGBT, and then simply give it an input. Then we can take the first prompt and go into your favourite AI art generator. This is specifically crafted for Midjourney. And you can see we're getting this lovely shot. Oh my God, which one do I want to take? I think I'll take this one. I like this one because it shows there is some room for taking our character from the start of the scene to the end of the shot. So we're going to save that image, come straight back into Runway, and we're going to go and upload it. Now the final part is to take the video prompt, paste it in, and go ahead and press Generate. Now this is the perfect process for one, maintaining consistency across the visual elements of your shots and also adding in the most stylistic and cinematic visual motion elements that we possibly can using this prompt building method. The outputs that you get are much more controlled and directed as well as being closer to what it is that we are looking for. This gives us a whole host of possibilities moving forward into the future of AI video making. And some of the regions and opportunities this opens up are in marketing, advertising, and filmmaking. Now let me show you a couple of examples of some of the most advanced uses of this so far. Here we have an advertisement of this beautiful product that was entirely generated with AI. Now here we have a cinematic short film that was created that really starts to bridge the gap between a piece of AI video generated content and that of real cinema. Now I'm going to take you through some of my favourite additions that I've created using this method. Now this first one plays on the use of beautiful contrasting colours with the reds highlighting her face and the blues in the background giving us a wonderful cinematic colour palette. She also flickers her eyes so beautifully here. Just look at the way she casts them at the camera. Now in this edition we have a lovely depth of field where we have these blurred lights in the background and it gives us a wonderful sense of depth. So I've got these to play forwards and then to play backwards so we can have a look at them a couple of times. Now this is one of my absolute favourites where we're playing with the light cast with some foliage and this brings these wonderful contrasting shadows across her face and the way they move swiftly across her forehead and cheeks and then show the undulations of her bone structure is quite mesmerising. Here I was playing with the use of rain to give this melancholic and moody feel and again we're using a strong contrasting colour palette dividing the scene in two where we have these lovely warm yellows conflicting against the cold blues. Now in this one we're keeping a similar colour palette yet what we play with here is the light coming from behind the protagonist and you can see that as it becomes into full view as she arcs the nape of her neck we get this wonderful radial blur coming out. Now in this edition I wanted to play around with a few more botanical elements so we have these reeds moving gently in the breeze. Now this example plays with the wonderful actions of this actress, of this lady and the way she plays with her hair is extremely natural and flirtatious. In this sequence of shots I've attempted to maintain a consistent aesthetic using colour palettes as well as camera movement to keep a continuity between the shots. We're using a more dynamic sense of camera movement following our protagonists and these quite washed out blues and yellows. Here I was playing with the smoke to give us this atmospheric feel and this is one of my absolute favourites. This man standing, looking wistfully at this train passing by. There's lovely lights being cast from the train onto the platform here as well. Just look at how the light changes across the tarmac, quite stunning. And in this next edition we're playing with glitter and sparkles because there's nothing like a little bit of glitter and sparkles to make the day go a little bit faster. Now you can apply this methodology to any AI art model and to any AI video model. So you can be using Mid Journey or Flux and then you can be putting these through your favourite AI video generator whether it is Kling, Minimax, Luma, Pixverse or Runway. Check out the free guide in the description below and if you're interested in seeing the entire process for doing this specifically for AI art generators, check out this video. And thank you very much for watching, thank you for being here and most of all, I wish you a delightful day.