Seedance 2.0 + Nano Banana 2: Master AI Video Creation

1. Introduction: So as a kid, I really wanted to make Hollywood Level films, but I didn't have any budget. I didn't have any resources, except my toys. So that was my entire production studio. But now it's actually possible with Seedance 2.0. Whether you want to make a commercial action pack sequence, showcase your work or create stunning effects. It's all in one place. Most AI tool gleich. Reactor morphs into something else, video falls apart midway. You spend more time fixing than creating. CNN two point oh is built different. In this course, I'll teach you prompt engineering that actually works how to use CNN step by step and how to create visuals so eye catching people will stop scrolling just to ask you how you made them. You don't need a film degree. You don't need expensive equipment. You just need a write system, and I'll give you exactly that. So enroll now, your Hollywood era starts today. 2. How to Access to Seedance 2.0: Welcome back, everyone. Now, let me show you how we can access to Seedance 2.0. We have three ways. First, go to dreaminiaab.com. Once you need to go to create or you can click here and you can go to video. Right now, as you can see, it's coming soon. So in this case, you need to install a pan. So I have this free VPN extension. If you turn this on, once you have your location, you can refresh your page, which is dreaminia.cabcad.com. If we click here, as you can see, now we have this video. Now if you click here, you can access to the previous model of C dance, and you can also access to Seedance 2.0. If you are using Seedance 2.0 for the first time, you will have your first free generation. So my higher recommendation is instead of generating five second video, click here and you can generate 15 second. So this way, you can value your first free generation through this remna dot CAP. Second method is go to Capcutofvisual website cbcud.com. Once here, as you can see, we have Seedance 2.0, you can click here and you can access to their previous model. You can also access to Sora two and VO CDFs. You can also go to Video studio. From here, you can go to your Canvas. You can create new project. You can click here, go to video, and you can select Seedance 2.0. You can also change your aspect ratio and your video quality. And right here, we have the duration. Now the dreamnia version, you can also access this in the CapCut app. Now open CapCut app and you can go to create with AI. Once you click here, as you can see, now we have the Seedream version where we can access this video generation. But as you can see, we can't really use the VPN right here, so the best way is to use this in your browser. In the last method, we need to create a new project. We need to go to AI Media and go to AI video. Right here, as you can see, we can access to Seedance 2.0, Seedance 2.0 fest. We have the previous version, and again, we have Sora two and VO three. And additional, we have Runway Gen four as well. So these are the three official method where you can access to Seedance 2.0 and their previous model. 3. The 3-Part Starter Formula: Welcome back, everyone. Now, if you have no idea about prompt engineering and you want to generate videos, then there is no point of generating videos because without the prompt engineering knowledge, you will lose your credit and you will generate random videos. The basic prompt contain three things. First, the subject, what the subject and second the action. What is the subject is doing and the setting, where the subject is doing that action. I have the first example. As you can see, we have the subject, we have the location, and we have the action wind on an open hilltop at sunset. So as you can see, we are describing our scene. So if you don't have any idea, you basically just type forest or a lawn tree. But if you really want to generate impressive video, then you will have to put some details of your shot. Now we have the bad example where I just type forest, and as you can see, we got this random video. So this is what I was talking. Next, we have this example, an old lighthouse, which is the subject standing on rocky cliff. This is the location against a store my gray sky. So this is the setting. So as you can see, we are describing our scene. Next, we got this example, a rusted iron gate slowly creaking open in an overground garden. And we have the final example, dark storm clouds moving fast over a flat open desert at sunset. Now with the basic prompt structure, as you can see, I have managed to generate these impressive videos that I have just shown to you because I have contained three things again in my prompts. The subject, what is the subject is doing and where the setting of our scene. So if you include these things, then you will have your basic prompt formula. 4. Adding Mood and Color: Welcome back, everyone. So to take our basic prompt to the next level, we need to add mood and the feeling of our scene. So in this prompt, we have a narrow stone street shining after rain at night. So this is the basic prompt. After that, we have the mood and the color, sad and quiet feeling, dark blue light in the puddles, which is the warm yellow glown from old street lamp, which are these And if I play my clip, as you can see, the basic prompt, we managed to generate this amazing scene. Next, we got this example ocean wave hitting Black Rock at sunrise. This is the basic prompt, and now we have the feeling calm and peaceful feeling, soft pink and orange color in the sky. So these are the color, and this is the feeling and have a look. Now I am using the basic Sea dance 1.0 mini to generate this video. And even with the basic model, we have the impressive result. Next, we got this video an pine forest just after sunset, dark and quite deep green and blue shadows between the trees, soft silver light from the moon, and have a look. And again, this is the basic model that we are using. That is why we are getting this kind of result. Next we got this example, a field of yellow flowers moving in the wind and golden. This is the basic prompt. Now we have the feeling happy and warm feeling. Now we have the colors, bright yellow and orange colors, soft blood in the background. And check this out. Now if you really want your generation to the next level, then you can change your model. So as you can see, I have generated this clip using Sora too. So this is the model. Now we have this final example, broken factory walls covered in green plants. This is the basic prompt, lonely and empty feeling. This is the feeling. Now we have the color, dull gray and dark green colors, one thin line of warm light coming through a crack. And if I play my clip, have a look. So by adding modes and the colors in your prompt, we'll help you to take your prompts and your generations to the next level. 5. Common Beginner Mistakes: Oh. Welcome back, everyone. So the common beginner mistakes is that they always try to make their prompt like they add these kind of words like ultra lilistic super, perfect, amazing cinematic masterpiece. These kind of words does not generate any video. The AI will guess and they will create something that you don't have any idea. A dramatic video with a surprise ending, the AI have no idea what you are trying to generate. Make it look like a movie, so you have to describe your scene because with this kind of prompt, the AI won't generate anything useful. Nice landscape with mountain and tree and stuff, this kind of mistake is acceptable. Show me something scary and intense. This is also a beginner mistake. So again, the AI have no idea what you are trying to generate. You have to convince AI your visuals and you have to type your visual detail in your prompt. So let me show you some fixed version. So first, we got this clip, Sharp mountain peaks at sunset, warm orange light on the rocks, white short falling the full sky. You can see, I am describing my shot, not just trying to make the AI think what I'm trying to generate. Now we have this from a single street lamp flickering on an empty wet road as thick fog slowly rolls in from the left and have a look. So if I have to type generate something scarier, the AI won't be able to generate this video because this is the visual that I have in my mind, and if I want the AI, generate this video, then I have to describe it. Now, instead of typing, make a movie scene or something like that, you can type the director style, like I have this style of Stanley Kubric a long empty hotel hallway, where is still cool white light from above, perfectly center shot, and check this out. Next, we have this example, snow covered mountain totop above a valley, pine tree in the middle, low clouds between the peaks white shot, and have a look. Now, if you generate your videos with the same prompt, you will get some different result than the previous version. Have a look. With the same prompt, I got my visuals, but I have different settings seen. And final, we have dark empty hallway in an old broken building at night, walls peeling, windows broken, one flicking light above, very still. So instead of typing, make something cool or something like that, describe your scene, put some details in your scene in order to make the AI make the visuals that you have in your mind. 6. Shot Types and Angles: And Welcome back, everyone. Now to take your prom even to the next level, you can add types of shots and angle in your prom. So if you search different camera shots, as you can see, we have different kinds of camera shots, and we have different kinds of angle based camera shots right here. So I have some few examples. This is very close shot of hot coffee. So as you can see, I have this short eye which is very close slowly dripping into a white cup, steam coming up, dark background, soft light from one side, and have a look. You can also describe where the light is coming from, from the left side or the right side. Now, next we got the top down shot. This is my shot type of a busy city road at night, car lights moving blow, white frame showing the full street grid. Now we have this example, very wide shot. This is my shot type of tall snow mountains against a deep blue sky, a tiny rocky valley far below, big and silent. Check this out. Next, we have the low angle shot from the ground looking up at the big ocean wave, sky high in the frame, water spraying above the camera. Now, next and final, we have this looking up from the floor of an old house camera pointing up at the high ceiling, colored light coming through stained glass windows above. So by typing short types and angles in your prompt, we'll take your prom to the next level. Not just that you can type the camera shots and the angle shot. You can also describe something like if you are recording the video through your phone. Like in this example, I have type water spraying above the camera. So you can describe some tiny details to make your videos look better and professional. 7. Camera Movement: Welcome back, everyone. Now, if you really want to take your prompts to the next level, you need to add some camera moments. So if you go to Google search for different types of camera moments, you will get these basic camera moments like dolly, pen, tilt, et cetera. You can include these keywords in your prompts. You can describe your camera moments. So I have this example, a drone slowly flying down in a white circle above an autumn forest, red and orange leaves blow soft white fog in the low areas between the trees and have a look. So we are describing the camera shots camera moments in this scene. Next, we have camera slowly moving forward inside a candle light old house, getting closer to the altar at the far end warm gold light growing brighter ahead and have a look. Not only you can describe your camera moment, you can also describe the focus the focal length of your camera. Like in the scene, I have camera slowly shift focus from a single water drop on a leaf in the front of a bloody mountain range far behind it, soft golden light and have a look. The camera is shifting the focus from here to here. Next, we have camera slowly turning right across a flat white salt desert at noon, wide open horizon, hot air rising from the white ground, no endpoint insight, and have a look. Now we have the final example camera completely still on an empty train platform at three in the morning, one light flickering above, departure, board clicking through with no trains, tense and quiet. So we are using the basic model of dance. That is why we are getting these kind of glitches. 8. Lighting in Simple Words: Oh. Welcome back, everyone. Now to take your videos even to the next level, you can describe lighting like hard light, soft light, lighting, beam, et cetera. So we have the first example, a tall glass building at sunset, warm orange light covering all the windows, a long soft shadow stretching across the ground below. Not only in the light, we are also describing the shadows as well. Now next we got this example, sunlight coming through tall trees in strong straight beams, rays of light you can see moving through the mountains, mist, deep dark shadows between the trunks and have a look. So we are describing the light moment, et cetera in the scene to make our prompts and generations better and professional. Next, we got this scene, a deep cave lit by a thin crack of blue glowing light from far inside. So as you can see, I am describing my light and the light condition, the light bouncing of still water on the cave floor. So the light is bouncing on this cave floor and have a look. You can basically explain the light beams light rays in your prom to make your prom better. So these are some extreme examples for your better understanding. Next, we got this example an old bridge just after the sunset, cool blue sky above, warm, yellow Latin on each side of the bridge, soft reflections in the river below, and have a look. Now we have the final example, an broken rocket on a red rocky planet, the sun far away from behind it, creating a bright glow around its edge light beams through floating dust, long shadow in front. So this way, you can take your generations to the next level by describing your lights. 9. Using Style: Oh. Welcome back, everyone. Now you can also add styles in your prompt. Like you can add studio style like Pixar dream work, et cetera. If you want to add director style, you can add that by typing their name like Wes Anderson, Christopher Nolan, et cetera. So in my first example, I have small town train station at Sunset, camera perfectly centered, soft pink and cream colors, flat and neat look, and at the end, I have Wes Anderson style, and check this out. We have Wes Anderson style generation. Now next we have a research building in the snow during a storm, very white shirt, call blue and gray colors, heavy and serious feeling, Christopher Nolan's style. Now you can change your model and you will get better results. This is just the basic model. That is why we are getting this type of mediocre results. Next, we have a wet rooftop at midnight, neon sign reflected in puddle on the ground, rain falling in slow motion, deep red and green colors inspired by Wonka i films. I'm sorry if I'm not saying the name correct. You can type any director's name. It doesn't matter. You can add their style in your prompt. Not only the person name, director name or the studio name, you can also type all film noise style. So I have a deep underground train station with no people, lights flickering on and off. Steam from the vents, strong dark shadows and hard white light beams, all film style, and have a look. Next and the final, we have the famous Gibil style, a small forest clearing at early morning, soft mist between the trees, bright green leaf, warm golden light coming through the branches, painted and dreamy look, studio gible style and have a look. Now we have this second variation of the same prompt with a different model. 10. Color Grades in Simple Words: Welcome back, everyone. Now you can also add color grades in your prompt like warm cool slit, warm all fame look, deep dark shadow, film grain, high contrast, et cetera. You can add these type of color grading keywords in your prompt and make your video cinematic. Now we have the first example, top down drawn shot of Tokyo Road at night. This is the basic prompt, warm orange in the lit area. You can see, we have warm air on in this light area, cool blue in the dark shadow areas. So we have cool dark color in the shadows area. So we are describing the color grading through the words. Very deep dark blacks and neon lights reflected on the wet ground. So as you can see, we are making the video cinematic. Now without the color grading keywords, you will get basically this type of result. Now we have this example, an old costal fort at sunset, faded and washed out colors. So as you can see, this is faded and washed out color. Greystone, dull silver sky high contrast between light and dark area. So we have this high contrast. Check this out. Isn't this look like a movie scene? Next, we have an wooden dock over calm water and sunrise. This is the basic prompt warm brown and amber stones all over, soft film grain texture, quiet and nostalgic feeling. Next, we have this example, steam coming out of a drain on a wet city street. At night colors look like all film with a warm middle tone and slightly lifted dark area, soft grain all over the image. So by adding the color grade in your prompt, we'll make your prom look like a movie. Next and the final example, we have melted rock flowing in the ocean at night. So this is the basic prompt. Very bright orange in the hot area. So we have bright orange color in the hot area, very deep black all around it. So this is a deep black, blue, green steam rising up, strong light beam from the glow. Check this out. So by adding the color grading keywords in your prompt, we'll make your prompt look like a movie scene or make it cinematic. 11. Adding Film Texture: Welcome back, everyone. So you can make your videos look more cinematic by adding some simple texture words like film grain, heavy grain, dark edges, dust and scratches, et cetera. These types of texture keywords make your videos look more cinematic. So let me show you some few examples so you can better understand what I am talking. First one, we have a field of halt, melted rock at night. A long thin light trek stretching sideways from the bright and glowing point. Orange and black colors, rough grain on the whole image. So as you can see, we added this keyword, rough grain on the whole orange, and as you can see, the image that we are getting look crispy clear, even with the basic model. Next, we have a cinema sign with half the lights burnt out, rain at night, fine dust and small scratch mark on the image, warm film colors, flickering neon, dark corner on the frame, and have a look. Next, we have an all broken greenhouse, red glass roof, green plants growing over rusted metal, slow camera move forward, faded and washed out colors, soft film grain all over, dark edges on the frame and have a look. We have dark edges around our image. If you notice right had, we have dark cadges. Now next, we have a narrow alley at night, one old gas lamp flickering, wet stone ground below, thick film grain, a soft warm glow bleeding around the lamplight, deep dark shadow area, and have a look. Isn't this look cinematic? Even with the basic model, we are getting this type of result. Imagine what we will get if we use the latest model. Now next we have an old steam train in the mountain station, documentary film fiel, tiny dust specks floating in the beams of light on the platform, warm brown tones, rough grain all over, and check this out. If you're thinking why the train is moving backward is because we are using the basic model, so this is not accurate. If you want to get the accurate result, then you have to use the latest model. Now we have the second variation of the same prompt and with the same model. 12. Adding Technical Details: Welcome back, everyone. Now you can achieve professional results by adding some technical details like frame rate of your video, lens type, your depth of field, slow motion, time laps, et cetera. These type of keywords will help you to achieve professional results. So in the first example, we have very close shot of a single raindrop hitting a still puddle surface in slow motion at 1:20. So as you can see, I'm describing my slow motion. Rings spreading out soft gray sky above. Now I have typed my aspect ratio of this video. You don't have to type that because nowadays, we can easily use the interface to select the aspect ratio of our video. Now, this is the result of this prompt now we have another variation with the same prompt and with the same model. Now next we have Strom clouds building over mountains peak 60 times faster than normal speed, camera not moving, white shot, dark blue gray sky, very white cinema screen, expect ratio, plays back at 24 FPS. So as you can see, these are tiny additional technical details that I have added in my prompt and check out the result. Now, next, we have drawn shot slowly pulling back from a single lighthouse to show a white S around it. Smooth 30 FPS. This is my aspect ratio. You can select the expect ratio from here. You don't have to type that. Warm afternoon light, soft faded coastal color. Have a look. Next, we have vertical phone screen. So again, you don't have to type the aspect ratio. You can click here and you can change your aspect ratio. Very close shot of a steam coming off a black coffee cup on white marble, 30 FPS warm yellow light from the left side, soft bloody background. Check this out. Next and the final example, a brown leather bag on a glass table, city skyline at sunset through big windows behind it. Camera focus shifts slowly from the back in front to the bloody skyline behind. Warm and cool colors 24 FPS very wide cinema screen. You can skip that and you can select your expect ratio from here. And now, if we look at the video, as you can see, the camera focus is shifting from the back to this skyline behind it. So you can also play around with your focus shift. 13. Think Like a Director: Oh. The biggest dum to Pavel is a change in mindset. Stop just describing what is in the scene. Start deciding what the viewer should feel and use every detail to make that happen. You need to ask yourself, what should the viewer feel in the first to second of the video? Where should the viewer's eye go first? What is the main point of the focus? What does the camera movement say? Is it warm and close or cold and pulling away? What color style best serve this movement and feeling? And finally, what can I remove to make the scene stronger and cleaner? So think like a director. As you can see in this example, I have a single chair facing a big window in a huge empty office at night. Camera slowly pulling back to show how alone and small the chair is in the big dark room. Cold blue city lights, soft and still 824 movie style or film style 24 FPS. Check this out. So you have to put the details of your scene in order to make your scene look professional. Next, we have a lighthouse light slowly sweeping through thick fog, low camera angle from the rocks below, faded gray colours, fog hiding all the details, white shirt feels heavy and unsetting 24 frames. Have a look. So you decide what the viewers should feel and see. Now, next, we have a long broken hallway with one dim light far at the end, camera slowly creeping forward toward it, getting closer to the dark film grain dark corner on the frame. Complete tally still tense and check this out. Now, next, we have hot melted rock pouring into the ocean at night, very close shot at the point where they meet. Slow down to 120 FPS, long light streak from the Brighton hot point, orange and deep black, powerful and beautiful white cinema shot. Check this out. And final, we have an old clock tower at blue R clouds rushing fast above it in the time lapse while the clock and bailey move. White shot. The fast sky makes the frozen time feel strange and heavy muted colors, white cinema screen 24 FPS. Check this out. If you notice in all these example, I am pulling your focus, I am taking your eyes to focus on one thing. In this example, we have the clock. In this example, your eyes follow this lava, and in this example, your eyes focus on this light and on these areas as well. But the main focus is this light. And in this first example, I am pulling your focus to this tear. And finally, your eyes follow this lighthouse light. So as you can see, you decide what the viewer should see and feel. Think like a director. 14. Generate Stunning Images With Seedream: Wow. Hm. Welcome back, everyone. Now I'm going to show you how we can generate stunning images using CDeam. So first, you need to go to dremnia.cbc.com. Once you, you need to go to create. Once you go to create, you can click here. You can select AI Image and from here, you can select your model. As you can see, we have from image 3.0 to image 5.0 it, which is C dream 5.0 it, but we don't have 5.0. So to access to the latest model of CDream you need to go to CapcAT. Once you hey you need to create a new project. Once you create a new project, you need to go to AI Media under this media page. Once you can select AI image, and from here, as you can see, we can select 5.0. Now, you need to have a plan in order to access to this model because as you can see, we have this diamond shape which means you need to buy a plan. We also have nanobanana Pro and we also have the previous model of CDream. To generate image, you can basically type your prompt. You can click here to generate your image. You can also select your aspect ratio from here, you can select your resolution from here. Default, we have two k, and this is the only option that we got for 5.0. If you switch your model to maybe 4.0, as you can see, now we can switch to four K, and you need to also have a plan in order to use the four K resolution. Now we can also upload an image from here. We can add an image here to generate consistent correct. Dreamnia.cpa.com, we have the similar interface. You can type your prompt generate image. But as you can see, it will took two credit per image and it will by default generate four variation of your images. We can select the resolution from here, we can select the aspect ratio from here. Now let me show you some examples so you can better understand what is the difference between the previous model and the latest model? Which one is the best for you to generate images. Now before we take a look at the images, first, how to write prompts for image model. You don't have to come up with your own prompts. You can simply go to aistudio.google.com. Once you can click here to expand this menu. You can go to playground. You can type your prompt here. Now I have typed this prompt. Give me Optimized prompts, 54 test image models. Give different variety in prompts like action, beauty, stunning, et cetera. Don't include any animals in person. Now it gives you different variety of prompts that you can test for your image model. You can also go to cloud.ai. You can also use this. It's very useful for your prompt. As you can see, I have several images, and these are all the images that we got from aistudio.google.com. These are all the prompt tests. So first of all, we can select an image. We can go to AI Media and we can see the prompt of the current image. As you can see, I have used 5.0 model. We can also select image. As you can see, we have four variations, so we can select any variation. We click here, as you can see, we have 5.0 model. With this prompt, we got these results, and this one is 4.0. Now, I believe 4.0 is the previous model, but it gives you the cinematic result better than 5.0. The 5.0 gives you artificial results that looks AI, like it will give you a look that it is generated with AI. With 4.0, it gives you a cinematic look. As you can see, we have different variation, but if you look at the 5.0, it will give you different variation, but with the same style of camera and the scene. Whereas in the 4.0, we have different composition. We have different camera style as well. Now with this prompt, as you can see, we have 5.0, and we got these results. And with the same prompt, we have 4.0. Now, tell me which one is better, 4.0 or 5.0. Let's select this image, and this is 5.0. Check this out and we have 4.0. Now again, we have 5.0. With the same prompt, we have 4.0. Now, tell me which one is the best 4.0 or 5.0. We have this image with 5.0. We got this result, and with the same prompt, we have 4.0. I have this image. We have this prompt. With 5.0, we got these images which look stunning. But with 4.0, as you can see, check out these results. Now we have this example where I have this ring. With 5.0, we got this result which not look bad. But if we look at the 4.0 result, as you can see, we only have one variation. If you want to generate multiple variation, then you can select your image. You can click here in AI media and you can also click Regenerate. Now this time it will give you one or four variation of the same. As you can see, we have now different variation of the same prompt. Now we have this example with this prompt, as you can see with 5.0, we got these results. And now if we check out the 4.0 result, as you can see, we have the stunning result. Now, if you are wondering why we have this crop, why we have this aspect ratio is because I have by default selected the fit. If you want to generate a different aspect ratio, then you need to select your aspect ratio from here. Now we have this example with 5.0. With this prompt, we have this and with a 4.0, check this out. With 5.0, we have this simple prom design, and with 4.0, have a look. This one is also generated with 5.0, and this one is with 4.0. And this is my prompt. Now if I select this image, as you can see, with 4.0 with the reference image, as you can see, we have this result and with 5.0 have some cases, 5.0 perform better than 4.0 and any other model. Like if I show you this example, 5.0 perform better than 4.0. In this image, 5.0 perform a little bit better than 4.0, but the 4.0 will give you a real photography look like if you are shooting this product, this jewelry with your camera. Now with this prompt with 4.0, we have these amazing results with a 5.0, check this out. Tell me which one is better this one or this one. Now we have this example with this prompt, we have 5.0, and with 4.0, have a look. We even have different kind of variation, whereas if you look at the 5.0, it will give you the same image slightly tweak some settings here and there. Now with the 5.0, we have this image and with 4.0, check this out. Now with this prompt, we have 5.0, and with 4.0, we have this image. We can also regenerate to get more variation of this. With 5.0, we have this image. We also have some error in the image. As you can see, we don't have any rider on top of this bike, where if you look at the 4.0 with the same prompt, check this out. Isn't this look amazing, cinematic and real? Now we have animation test. With 5.0 with this prompt, have a look. We have this variation, and with the same prompt, I generated some more variation with 5.0, and this is from 4.0. Now, in this case, I believe 5.0 perform far more better than 4.0. Now we have more variation of the previous prompt and have a look, and this is the 5.0. Now, in this case, the 4.0 perform better. Now in most cases, 4.0 perform better than 5.0, but in some exceptional cases, 5.0 perform far more better than 4.0. Now we have this bullet time test. This is my prom with 5.0. We have these images and with a 4.0, check this out. This case, I believe 5.0 perform better because it will give you a photograpl look, whereas the 4.0 will give you this look like it's happened in real life. Now we have this test with 5.0, we have this one, and with 4.0, check this out. Have this example with 4.0, we have this variation and with the 5.0, check this out. I really like the 5.0 result like A, but the 4.0, it's something else. Now we have another test, store motion animation test. With 4.0, this is my prompt. If you want to generate this type of store motion animation, then you can include hyperrealistic lay, this keyword in your prompt to generate these kind of images. With 4.0, as you can see, we have these strunning images, whereas if you look at the 5.0, this is for tattle looking images. I mean, this looks like the previous model, and this one look like the latest model. Whereas, this is the previous model, and this is the latest model of CDR. Now we have another test with this prompt. W 4.0, we have these images. Where if you look at the 5.0 result, I really like the 5.0 because it will give you this cinematic look. Now we have this test with 4.0, we have these images, and with 5.0, check out. I really like this composition of 5.0. As I told you earlier, the 5.0 in some exceptional cases perform far more better than 4.0. Now we have this prompt with 4.0, we have these images and with 5.0, check out. With 4.0, we have these images. With 5.0 with the same prompt, we got the Now final with 4.0 with this have these images and with 5.0, check this. This is the same prom that I used previous. If you look at this one right here, as you can see, in 5.0, we don't have any hands controlling the bike, whereas if you look at the 4.0, we have the correct composition and cinematic look and depth of field of the scene. Now the pro tip is that don't use one single model, use multiple model. Even use nano Banana P or nano Bana two, then you can compare all your images. You can select, you can finalize the best image across all the images that you have generated. 15. How to Generate Consistent Character: Welcome back, everyone. Now let me show you how we can generate consistent character. Now, to create our character, go to labs Google slash fxlasTol and slash Flow. As you can see, I have my image right here, which is this one, and I have typed this prompt. Person wear black sunglasses and a metal mask of silver color with red wave line on it. We have our character. Now we can place this character anywhere that we want. Once we have the character, we can go to CapCut. We can go to AI image. We can add the image of our character and we can add the prom. Now, to get the prom to write the prompt, we can go to simply cloud.ai, we can upload the image of our character, and we can type this prom. Optimize enhanced prom for cinematic look. We have this column, and under the brackets we have gravity shifting VFX sequence, give ten ideas including the person in the picture. You can see, we have ten ideas. If you don't have any idea, if you run out of ideas, then you can ask the AI for the idea. If you ever want more ten ideas, then you can type, give me ten more, and it will give you more ideas with the person included in the image. And as you can see, it is now generating ten more ideas. After you get the prompt, you can add your arrector, write the prompt, select your model, select the aspect ratio and the resolution, and boom, we have this cinematic sequence with the correct. Now with the 5.0, as you can see, we have different results. We have this sequence with the 4.0, as you can see, and with a 5.0, check this out. With 4.0, we have this result and with a 5.0 with the same prompt, we have this one. With 4.0, check this out. We have this sequence, and with the 5.0, have a look. With 4.0, we have this look and with 5.0, check this. Now we have this prompt and with 4.0, check this out. I mean, check out these results. And with 4.0, we have this one. Most of the cases, 4.0 perform better than 5.0 because it will give you more variety, more variation of your images than 5.0. Now we have these images as well. We have this one. With 4.0, we have these images. With five, we have this one. With 4.0, check this out. With 4.0, we have these cinematic looks. And final with 50, we have these. Now as you can see, the Cloud give me more ten ideas. We can simply copy the prompt from here, go to capcod.com. We can add the character right here. Now we have the character. We can simply paste the prompt right here, select the aspect ratio, select the model, and boom. Now, once you click on the generate, if you switch the model, as you can see, we don't have the generate button. So we can simply click here select this layer. We can simply paste the prompt, add the character, select the model. And boom, we can generate these images with a new model. Now, once your image is generated, you can simply switch the model and you can click here to generate your images, and it will generate your images, and it will include these images with this model in this folder right here. So as you can see if you notice in one single layer, we have multiple images. So basically, the layers in the image generation and the video generation models are work like a folder. As you can see in one single layer, we have multiple images, which means we have these tiny folders of images. If I select my images, as you can see, we can switch the model to four point maybe four or we can switch the model to 4.0. And if I click here to generate, it will include these images with new model in this layer inside this layer. So basically, inside this layer, we have multiple images. So the layers work like a folder. Now the images are generated, have a look. We have this result and with the 5.0, check this out. And the 4.0 is right here. Now we have more variation and this is the final variation. Check this out. Isn't this crazy. And inside this layer, we have the 4.0 images. If you select the 5.0 image result, the image model will update right here. It will tell you that this image is generated by your selected model. So as you can see, this image is generated by 5.0, so we have 5.0. If I select this image, which is generated by 4.0, it will also update the model right here. Now instead of asking AI for some VFX shot, you can also at the AI person in the image is doing a chocolate commercial. Now as you can see, we have these prompts, we can copy this prompt. You can go to CapCut, add your character, ph your prompt. And as you can see with 4.0, we have this terrible result. Even with 5.0, we have this bad results as well. So instead of the CapCut C Dream, I have used the flow platform, which is by the Google. I have my character. You can upload your arrector from here. And once you have your character, you can click here, add your arrector to the prompt, you can paste the as you can see, this is the result that we got with the same prompt, this is Seedream result. Now, tell me which is better seed dream or the Nano Banana two. Now I have these images as well. As you can see, these are all stunning images. Not just chocolate. You can make your character promote or do anything. As you can see, the character is promoting the soda. We have this version. Check this out. We have this image. We even have this image, and we have this. Now I just simply change person is promoting soda can and we have this prom, and these are the prom that I have used in my generation. After this, I have make my person promoting a car brand, and these are the prompts. And as you can see, these are the results of my person, my character is promoting the car brand. Have a look. Isn't this stunning and amazing? So you can make your character do anything that you want. And this is how we can create a consistent character. 16. How to Steal Any Prompt: Welcome back, everyone. Now, let me show you how we can steal the prompt of any image that you want. What if you saw something on the Internet and you want to replicate that? Well, we can do that by simply going to atodogogle.com. You can create a new chat and you can type this prompt act as prompt generator from image to generate similar image. Now, after that, you can upload your images that you want to replicate. And once you press Inter, it will give you the prompt that you can use to generate the similar images. Once you have your prompt, you can go to flow and you can simply page the prompts right here. As you can see, we have this image, and this is a prompt that we got to generate the similar image. As you can see, now we have the similar image. Now, this is my second example. This is not same as this image, which I also generated by flow, but as you can see, it pretty close to my original. Now we have this example. This one is also generated by flow, and we have this prompt that we can generate the similar images. Now you can replace your chocolate bar by selecting your chocolate bar from here, you can upload your chocolate bar right here and you can type replace the chocolate bar in the image. And it will replace your chocolate bar. Now we have another way to replicate the images to steal any image prompt. To do that, you need to click here to expand these options once you can go to build. Now, once you go to build, as you can see, we can build any app that we want. So I use this prompt to build an app that input images and optimize image and get prompt to generate similar image, prompt generator. This is my app. I have already built this app. After my first prompt, what I did, I make some changes. So you can make some changes by typing a prompt right here. You can update your app. With this prompt, I have a single upload option. We can only upload one single image. So I just change that by typing add multiple images. After that, I add add multiple image input and generate prompt for each image. Now as you can see, now we can click here, we can upload multiple images and we can click here to generate the prompts of these images. As you can see, this is our first prompt of this image. Now we can simply copy this and we can simply paste prompt right here and click here to generate this image. Now let's copy this prompt, paste my prom right here, and final, we can copy and we can paste this prom right here. Now we got the images. As you can see, guys, these are our original images, and the results that we got are similar to the original images. And this is how we can steal prompts of any image that we want. 17. Model Comparison: Oh. Welcome back, everyone. Now, before we even generate any AI video, which model is the best for you? Which model can benefit you and can save you a lot of credit? Well, first of all, go to media and under AI Media, you can go to AI video. Right here, you can see we have image to video, we have Omni Reference, text to video, we have MultiFrame as well. You want to generate videos with only your prompt, without any image, you can select text to video. You can even select image to video, and without any image, you can generate text to video. You can click here to expand your prompt area, and as you can see, this is my prompt. And with the same prompt, I have different model results right here. So the first one is, let me just restore this. And as you can see, I am using CDance 1.0 fast. It takes ten credits per generation, and as you can see, this is the result that we got. Can see with clear and precise short prom, we can even generate this kind of stunning result with the oldest model of the Seedance. Now right here, we have 1.5 pro and it takes 75 credit. And if I show you the result, as you can see, even with the sound effects, we have this result. Now with the same exact prompt, as you can see, I did not change anything at all. We have Seedance 2.0 fest. We have 480 P quality, five second generation, it takes 35 credit. Now with this model, if you increase your resolution to 720 P, it will take 70 credit per generation. Now, even with Seedance 2.0 fest, 480 quality, we have this result, which looks stunning. It looks like a Hollywood film. Check this out. Now with C dance 2.0, 480 quality, five second duration, it will take 40 credit. The first will take 35 and the simple will take 40 credit. We have five credit difference between these two. Now here is the result of C dance 2.0, and check this one. Both are amazing. Now right here, we have another example with C dance 2.0 with 480 quality, five second duration. This is my prompt. As you can see, you can read it right here and have all Now, this really surprised me. With Seedance 2.0 fast with 48 equality with the same exact prompt, check this out. The results are more efficient and stunning. Now if I replay again, as you can see, we say five credit, and we have even more stunning video than the Seedance 2.0. Now final with the same exact prompt, we have Seedance 2.0 fast, 20 quality, five second duration. It costs 70 credit, and check this out. Now you don't have to always select the higher quality. You can select the lower quality of the Seedance 2.0 fest, and you can generate stunning video 18. Fix Failed Generation: Oh. Welcome back, everyone. Now, first of all, let me show you how we can avoid failed generation, not even the bad generation, how we can avoid failed generation. Most of the cases Seedance 2.0 or failed the generation, it won't generate your videos. It gives you some option to change in your prompt, but don't select and don't agree on the changes that it want in your prompt, because it will generate a completely different video. So let me explain what I mean. So this is my prompt, as you can see, and I have failed three times. After that, it give me some option, and I have selected that option, and it just generate this video. Where most it just smiling, staring at the camera. Now this part of the video was for the three second of my whole video, but it generated this for 15 seconds. So I lost 255 credit. So tip number one, you can copy your exact same prompt like this. Now go to the Google Translate. You can paste your prompt and you can select Chinese simplified. You can copy this prompt, and then you can paste your prompt. And as you can see, for five second clip, I am succeed at generating my clip. Before we fail because we are using English language, but now it generates my clip because we are using the Chinese language. Just one clip. As you can see, I have several examples where I switch my prompt from simple English to Chinese, even in some cases, the simple English work, and in some cases, the Chinese version also failed. Let me just translate to English. So if you fail to generate, even if you are using the Chinese language, then you can agree on some condition that suggest. Example, let's say, remove or simplify the description of discomfort. You can go to AI studio. You can paste your prompt, and after that, you can say apply these changes to given prompt, and you can paste the following changes that you get from this Seedance 2.0 or chat box. Now, before you even do that, you need to look at one thing. Are these changes are completely changing your original prompt or they are just tweaking some words here and there. If these are changing your prompt completely, then don't make this. If these make your words, some few words here and there, then make these changes. This is the second tip to fix your failed prompt. Now, after I agree on the changes, here we have the result. As you can see, now again, I'm using the Chinese language. Now, here's what happened if you blindly follow the changes that it gives you and you make them in your prompt. As you can see, guys, we have this prompt, and I am failed at my first try. So in my second try, it gave me an option, and I have selected that option, and this time, it just gave me this video. Originally, I want a fight between the person and the robot, but this time it give me this version. So originally, I want to make this video. But after I selected the changes that I get from the Seedance, it completely changed my video. So as you can see, we don't have anything about the fight. We just got a person looking at the robot and seeing the robot at the closer. You are using human faces, then it won't generate your video because right now it is not accepting the human face. So as you can see, I got this error, modify the prompt to describe the character is a more stylized or abstract way. Then I say simplify the prompt, and as you can see, it asks me to make these changes in order to make the AI create my video. So in my next generation, I created a new project with brand new images with no visible faces, and as you can see, this time I'm succeed at making my video. Check this out. And I did not get any error at all. Now, sometimes you have a bad generation because you are using a bad image and you are confusing the AI with your prom. Let me show you what I mean. So I have this image and I have this prompt. As you can see, I'm using the 2.0 fast model with four ATP quality five second generation. Check this out. Have this weird person walking on the ceiling. Now in my next generation, what I did, I changed my prompt a little bit, and I also use a different image where we have the ceiling visible in the picture, and I also use Seedance 2.0 four ATP quality five second generation. Handheld shot person jump and rotate and start walking like a vampire on the ceiling towards the camera. As you can see, this is a simple prompt, and it conveys the message clearly instead of this weird prompt where we are confused, even I'm confused while I'm reading this again in this prompt, we are clear what we are trying to achieve. And if I play my clip, as you can see, we fix the prompt because as you can see it fit in the whole frame. That is why the person is walking in this style. Now we have another example. Now, originally, if I change this to Chinese, I use the Chinese prom, and as you can see, we got this result. The person is not walking it just stays there. So why I'm showing you my bad generation instead of the last one which is good, the person was walking. So if you got bad generation, don't lose H. Regenerate, it will help you to get your desired result. Now, bad selection of the image is also the reason why you are getting bad generation. As you can see in this example, if you see we have this hue commercial, we have this weird move part of the commercial is good. The reason why we are getting this kind of error is because we are using the wrong image. We don't have any space for AI to work. We have a still shot of the shoes of these shoes. They all aligned at the center of the screen. So the AI can't really make some motion movement in your scene. So that is why it stuck there and it's just make this morphing. If you are getting bad result, also try to change your images. 19. Basic Text to Video: Two Welcome back, everyone. So let me show you how we can generate stunning and amazing text to video that grabs attention. First, go to astudo.google.com, create a new chat and type act as prompt optimizer and enhance given prompt. Now, what it will do, it will enhance your given prompt. So we have this prompt a samurai swing his sword and Shockwave hit and destroy a giant mountain in a distant. As you can see, this is a pretty basic prompt. Now, this is the prompt and enhanced prompt that we got from AI Studio. As you can see, we have three options. You don't get the options, you can give me three variation of this enhanced prom, and it will give you the three variation or you can type more variation if you want to. Now we have these variation, and again, I just change my prom to a samurai with metal face mask, swing his sword and shockwave hit and destroy a giant mountain in a distant. Now we got this prompt. So I just use this one, the first one. Now we are in CapCut I use Seedance 2.0 fast, 480 quality, five second duration. And as you can see, you can check out the prompt. This is the exact same prompt that we got from AI studio. Now have a look. Check this out. Isn't this amazing? Now, this is the result of Seedance 2.0 fast 480 quality, we have 720 quality, and with the ten second duration, now I was not expecting this, but hey we have this one. I cost us more credits, and the results, as you can see, are not very stunning than the 480 quality. Now we have the final generation. We have Seedance 2.0, 720 quality, five second duration. It cost me 85 credit per generation. And check this out. What I notice, if you really want to generate stunning videos, you can select the lower quality as well. You don't have to select the higher quality because the higher quality does not mean you will get perfect result. We have this example, as you can see with the higher quality and we have the bad generation across all these generation. Now in the same chat, what I did, I type another prompt. Camera logged onto robot transforming into modern combat robot fighting at enemy in battlefield. After he run out of bullets, he fights them with kung fu. Write detailed prompt, write a decent fight choreography, and check this out. We have this prompt. So I use the first one. As you can see, we have ten second generation. If I expand this, you can see we have the same exact prompt. Now we can restore this and have a look. Check this out. Isn't this amazing? Isn't this look like a Hollywood level film scene? Even the sound design are so crispy here. Oh my. Isn't this beautiful? Now we have another example. As you can see, we have this enhanced prompt. With the same method, I got this enhanced prompt. With the Seedance 2.0 480 quality five second generation, we have this is. Check this. Isn't this amazing and cinematic? Now we have this example. Even without any enhanced prom with your basic prompt, tanks stormed into the city, crushing part vehicles and crashing through the building walls. We have Seedance 2.0 48 equality, five second and Hao. With the basic prom, you can even generate these type of cinematic videos. We have another example reload under fire, extreme close up hand, plumbing a magazine wild bullardpark nearby. With 2.04 80 equality, check this out. Now we have another example in infinity Corridor hallway strates endlessly as a figure with yellow ski mask walks never reaching the end. With 2.0 for eight equality, this is what we got. Now we have this example, a gunfight in the lobby, an action hero in a yellow ski mask faces off against 30 security guards in a red ski mask in a marble lobby where pillars explode. With 2.0 fast for eight quality, this is what we have. Check this out. Let's wheel over this. Now, how to write these types of prompt. Basically, you need to go to astodo.google.com, create a new chat, and you can ask the AI, give me 50 ideas for action Back sequence. If you want the action Bag sequence with the robot, you can type, give me 50 action Bag sequence ideas including robot. You can say, give me 50 ideas for battlefield, give me 50 ideas for Kung fu fight scene, et cetera, et cetera. This way, you can get multiple ideas and you can try them self. Now you can also go to capcad.com, go to Video studio, select Canvas Project, create a new project from here. Now, as you can see, I have several projects, so I'm going to open this one. Now, this is the prom that I got from AI Studio. Now, with the basic prompt, the Seedance 2.0 will automatically switch to different camera angles, even if you don't mention them. So if I play my clip, as you can see, we have different kinds of camera angle sort, we have closer, we have white sort. So by default, the C dance will automatically apply multi-camera angles into your Dance. Here, as you can see, I have 720 P quality with Seedance 2.0 with five second generation. Now you can also include dialogues in your video. You can merge multiple scenes. I have this example, four character having conversation. First anime style boy ask this if I go to translate, how do you say hello in your own language? So this is the first boy saying this second boy American style, Disney style boy says in English, hello. Third Japanese Ink watch boy says in Japanese, this fourth three D style Arab boy say in Arab Mava. Then they all smile. If I play my clip, on yen, Sama Neha. Hello. Cichva Mahbum. Check this out. Isn't this amazing Yama Neha. Hello. Cichva. Mahabum. So you can also add dialogues in your scene if you want to. Now we have another example. This is the prom that I got from AI Studio, and this is the result that we got from it. Check this out. Now we have another example, Pacific Rim style Kaiju Inter in the city camera focus on Kaiju, and you can read the prompt. It's a very long prompt. And as you can see, without the camera mentioned in my prompt, if you look at the result, check this out. I use Seedance 2.0 720 B equality. The 480 will cost you 40 credits and 720 will cost you 85 credit. And with the basic prompt, a field of bright yellow sunflowers under a blue sky. Now, as you can see in my prompt, I did not mention any camera angles or shot. So by default, the Seedance 2.0 will automatically add multi-camera angles in your video. 20. Multi Shot: And Welcome back, everyone. If you want to have multi-camera angle shot in your video, then you need to type this keyword multi shot. You can add a column. After that, you can type your prompt. I got this prompt from astudio.google.com, and this is the final output. As you can see, we are switching from different camera angles. You can also mention your camera angles at the end, and it will switch between those camera angles. Now, the tank is firing gun. This generation is not accurate, but this is just a simple demonstration on how to use multi shot. Now another pro tip if you really want to save your credits. If you don't have any idea, go to astog.com. Give me ten ideas, includes. Now in includes, you can type your keywords. You can type what you want to include in your prompts idea. So I type robot, Giant Monster, and alien don't include animals and humans. Now after that, I type switch between different camera angles during sequence. Now these are my angles, close up, wire shot, handheld, depth of field. Write five second prompt. You can also type ten second, 12 second and 15 seconds. As you can see, we have ten prompts idea. Now how to check are these ideas are good or not. You can simply go to abstGoogle FXTolFO. Create a new project, type all these prompts, and as you can see, you can get the idea what the prom is about. What's included in the prompt. Now, I use this prompt, so I run of credits as you can see, so I just use CDs 1.5 pro to demonstrate on how to use multi-camera shot. And this is my final output. As you can see, guys, we are switching between different camera angles. Even with the previous model, we have this amazing generation. Now we have another example. I have used Seedance 2.0 480 quality five second duration, include multi-camera shot keyword in my prompt, and this is my prompt, as you can see, now this is my final output. We have second angle, we have third angle. This way, you can use multi-camera shot in your prompt. Now we have final example. This is my prompt and after that, instead of using multi shot in the start of my prompt, after my prompt, I have typed change camera angle every 2 seconds, alternative between close up, white shot, Dutch angle shot, but I view et cetera. If I play my clip, as you can see, guys, we have different camera angles. Check this out. This is how you can use multi-camera shots in your video. 21. Different Shot: Welcome back, everyone. Now let me show you how we can generate multiple videos in one video. Now you heard it right. In one video, we can generate multiple video. So in AI studio, I have typed this prompt, World War Robots give 50 ideas seen set in war. So we have World War ideas include robots. I have also typed World War robots classics styled, and these are my prompts. Now you can also type World War I instead of human. It's robot fighting. So we got these ideas. Now we can generate multiple videos in one video. Let me show you what I mean. So as you can see in my prompt, I have typed shot one, shot two, and shot three. We have even shot four. Now, this is a 15 second video with Seedance 2.0, and we have 720 equality. As you can see, I have generated four videos in one video. Now here's how it works. You can type shot one. This is my first shot. So I have type shot one, five as the S means seconds. Colon. Now, if you want the AI to make this video for 2 seconds, three, four or five, you can mention that second right here. You can even type ten second, et cetera. We have the first video. After that, I have type shot, two, three second, then we have this video. Then we have third one, which is short three, three, a, we have the third one, and we have the fourth one. Four, four second. My video is 15 seconds. You have to keep this in mind. You have to divide your seconds and you have to combine them and make the combination equals to this second. So as you can see, we have four, three, this is seven, three is ten and 105 is 15. So 15 is equal 15 second. And if I play my clip, as you can see, guys, this is my final output. So I combine these in such a way that it looks like a one final video. Now you can also use your control and use your scroll to zoom in and zoom out your interface. You can also click here and you can select your Zoom using this slider as well. Now we have another example. As you can see, this is my short one for five second. Show two, five second and short three 5 seconds. So 53 time is 15, and we have 15 second video, and here we have the final output. Now we have another example. As you can see, I have shot one for three second, we have shot two for three second and shot three, four, four second. If you combine these together, it will make ten, and we have ten second video. Now, as you can see, we have this prom, you can use Google AI Studio to get these types of prom. You can get ideas from Google AI Studio. Check this out. Isn't this amazing? Now we have this video, so I have five second video, so I have type, shot one has two, two second and shot two, four, three second. If you combine these together, we have five second, and we are generating five second clip and here it is. Check this out. Now we have this example, shot one, two second. We have underwater explosion, shot two, three second, we have NepamFest and have a look. Now for these types of ideas, you can go to AI Studio and you can give me 15, ten, or maybe 20 ideas, the world ending situation, and it will give you the ideas according to your prompt. Now we have this example. We have shot one, 3 seconds, show, two, 2 seconds. We have first mostran emerge from the sea. The second shot a giant wading a yellow, ski mask, batted in the city, et cetera, and check this out. We are combining two videos together in one generation. Now next we have this one. We have short one for 2 seconds and shot three for three second. And check this out. This way, you can take your generation to the next level. Now next we have ten second clip. So as you can see, we have two short for three second and the third one is 44 second. So if you combine these together, it will make ten, and we have ten second generation. If you want to look at my prom, these are my prompts. And in all my generation, I have used Seedance 2.0 four eightB quality. Check this out. Even 480 P quality, you can generate Hollywood level with you. Now, if you really want to make cinematic shorts, you need to switch from AI studio to Cloud AI. The AI studio is for the basic or intermediate level. The Cloud AI is for the pro. Now, as you can see, I have typed this prompt. I want to optimize this prompt. So I have typed optimize and enhanced prompt for cinematic movie style. And by the way, this is my prompt. Shot one person with yellow ski mask running away from giant Alien Army. Ada comma, then I define my camera shot. Camera tracking shot of person from front. Show two, person stop, turn fear into anger punch alien behind him. Ada cooma then a type shot three person standing in center while all Alien mid air bullet time shot shot four, person with Aura power fights alien, punching and kicking them. As you can see, I got four shots. Now using CapCut, Seedance 2.0 480 quality, 15 second video. Just tie my prompt and I divide it duration. So as you can see, the cloud AI does not define any duration because I did not describe my duration. So I just divide my duration. So short one, three second, shot two, three second. We have third shot for 5 seconds because this is a bullet time, so I want to make this longer than the other shot. So the final shot is for 4 seconds. And now if we combine these together, we have three second, three second, this is six and six plus five is 11. And at the end 11 plus four is 15, and we have 15 shot. And check this out. This is my final output. Isn't this amazing? I mean, we can now just basically make Hollywood level films in our garage, backyard, in our living room, et cetera. After that, I want to make bullet time effects video. So I have type optimize and enhanced prompt give ten more bullet time effects video. So as you can see, it asks me what I want. So I just type, give ten ideas. After I got ten ideas, I have type give ten more, so it gives me ten more ideas. I have used one of these prompts in my CDance two point fast video. Now accidentally, I have type short one, short, one, short one, but the C dance understand it anyway, and here we have the final result. Check this out. We have this bullet time effect. A. Now we have this one last shot. So in AI studio, I have typed this prom, optimize and enhance prompt. A decoma after that, I have type switch between different camera shot. Addecoa each shot must be two second long. Create five shots, intense card drifting rays with action, don't include human. So we got this prompt, and these prompts are two second long. Now in CapCut, I make 15 second video, so I divided them in 3 seconds. So five time is 15. So we have 15 second video. I have used Seedance 2.0 720 quality, and this is my expect ratio, and check this out. Now, by using the same prompt, I have used four ATP quality, Seedance 2.0 with the same duration and same aspect ratio. Check this out. Now at the end with the same exact prompt with the same duration, I use C dance 2.0 fast with four ATP quality and check this out. Now if you compare all these three shots, so I believe the short one, which is Seedance 2.0 720 equality is the best amongst these two shots. Now, if you are getting these types of error, you can fix that in post. So as you can see, if I use my arrow key right, we can go to the next frame. We can add a cut right here and we can add a cut right here. Now you can select this clip. You can hold your mouse right here and you can make it reverse and now check this out. Oh. Now if you are wondering the sound design behind this video is also in reverse, we can easily fix that. So what we can do, we can press Alt key. We can rack this video up. So we have the duplicate clip right here. We can hover the mouse right here, we can reverse this. So we have the forward clip. Then we can right click Extract Audio, select this clip, press backspace to delete. Now we can select the original clip, go to the audio, and we can lower the volume to minus infinity TV. Now if I play the clip, now we have the forward audio. And this is how you can combine your multiple generation in one generation. 22. Single Continious Shot: Welcome back, everyone. Now, instead of making multi shot shot one, show two, we can create one single continuous shot. So I have 15 second generation with 2.0 fast 720 P quality. We have this prompt, single shot action sequence, no cuts, no scene transitions. We have the scene war intense shot camera handheld, motion baler, low behind a group of robots moving cautiously along a wall. And we can also define the audio. So we have machine gear audible. And if I play my clip, as you can see, guys, we have this intense one single continuous sot with no cut and no transition at all. Now next we have single shot action sequence, no cuts, no scene transitions. We have the scene war intense shot camera handheld, motion blod tanks, fire cannon at each other, camera violently shaking with explosion. A single green tank destroy all four tank in empty New York, also destroying building while toging tank cannon. If I scroll a bit down, you can see we have raw tank audio. So I want the raw tank audio behind this video. So I want the audio effects of this row tank mechanic. And I have also used Seedance 2.0 felt for ATV quality, and we have 15 second generation and check this out. Have a look. Even with the lowest quality and with the Seedance lowest radiation, we are generating this Hollywood stuff. Let me play this one more time. I mean, we are creating with a few words Hollywood level clips right here. Imagine what we could do in five years. 23. Image to Video: Welcome back, everyone. Now let me show you how we can generate stunning image to video. As you can see, I have this image. We are here in capcad.com. You can simply go to Widostudio and create your new project. Once you here, you can click here and you can upload your media from here, upload from device, you can select this, and you can also click here and you can upload your media. Right here. Now, once you have your picture, you can type your prompt. So I have typed this basic prompt. Person running towards camera, a giant robot is chasing him. So I want a chase scene. And if I play the clip, as you can see, we have this stunning, beautiful cinematic video. And as you can see, with a few words, we can turn our basic image into cinematic masterpiece. Now in my second example, I have this different image, and we have this enhanced prompt. Now, first of all, to create your character, you can upload your image and you can type any prompt that you want. You can create your own character. I have this character. Person wear black sunglasses and a metal mask of silver color with red waveline have our character. Now, once you have your character, you can go to Cloud AI. This is the best platform to write enhanced prompt. And as you can see, I have the basic prompt. Enhanced prompt, person in a space sus alien monster, Ada Coma person running Aicha, both land in ocean, add a coma cracking a beer and eat alien, then again, Aaka Now the reason why I have added this coma because if you want to add something new, then you have to add a comma. So it is very useful for AI to understand that you are making these different changes in your I say person watching it in a distance. Now my prompt is end, so you can add these brackets, add the column. Inside, you can write your prompt and after that, you can write some technical detail. So as you can see, after these columns, I have typed this technical detail, write a 15 second sequence, also add seamless transition between scenes with handheld intensne single continuous shot, depth of field, shot on cinema camera lens 35 MMT 1.2. Now, if you're familiar with the camera lenses, normal lenses contain the focal length, F 1.24 or something like that. The cinema camera contain Instead of F, we have T in the camera focal. After that, I say cinematic lighting. Now we got these prompts, and at the end, we have the master prompt. Now, if you don't get the master prom, you can type write a complete video prompt in one single paragraph. As you can see, we got this prompt, you can copy this prompt from here. Now, once you have your prom, you can upload your image and you can paste the prompt. As you can see, this is the same exact prom that we got from Cloud AI and check out the sequence. Check this out. Even the sound design is seamless. Now with our character, I have a different scene. Enhanced prompt, and we have this colon. Inside the colon, I have a person trying to escape a prison full of robot. He is only human there. He's flee with a helicopter with intense chase. Now, after that, we have the technical detail. Write a 15 seconds sequence, also add seamless transition between scenes with handheld intensn, single continuous shot, depth of wheel, shot on cinema camera, lens 16 MM 1.2 cinematic lighting write it in one single full video prompt, and as you can see, we get the prompt. Now, after that, I say write three version and we have different kinds of version. Now, before you page this directly into CDNs, you might lose some credit. So we have to be careful. We have to go to flow, which is labs dot google dot sshxTolssF and you can test your prompt. As you can see, I have tested this prompt, and I have selected this. Now the reason why I have selected this one because it look better than the other version. Now you can also use this image. Now maybe you might be thinking why I have to use this image because it contains all the reference in one single image. If you look at the other images, we have single reference. In this, we have multiple reference. It is better for AI to understand what we are trying to achieve with our prompt plus image. Now, once we have the image and we have the prompt, you can go to CapCut or you can go to CapCut software, which is this one, go to AI Media. AI video. You can upload your image. As you can see, we have the image. We have the prompt. I have selected Seedance 2.0 for ATB quality, 15 second video. We have aspect ratio right here. And as you can see, we have the cinematic masterpiece. Check this out. Now I make my character wear this suit using the flow platform, and I have typed this prompt. Now enhance a person flying in helicopter. In Vietnam, a joint monster can be seen in a distance. Person fire machine gun, at joint, giant hit helicopter with his hand, helicopter crash, person survive, person get up, get a Bazooka, and fire at joint. Giant get knocked and fall on ground. Simple prompt, as you can see, we got the enhanced version of this prompt. Now I'm using C dance 1.5 pro with 720 P equality with 12 second, we have our prompt, and this is the result that we got without any image reference. Now with the image reference with C dance 1.0 with 720 P equality, ten second duration, this is what we got. This is just the test footage. I want to test my prompt, so that is why I use the previous version of C dance because I don't want to lose my credits. Now, after this, I have used my picture. We have the prompt, and I'm using Seedance 2.0 fast with 480 p equality and with 15 second duration. And I have selected this aspect ratio. And now have a look. Check this out. Now, this is the only error that I got in this video. If you ignore this, everything look fantastic. Now we have this image, and this is the prom data I used to generate this image. After that, I use the same prom data I used to generate this image, and by using the Seedance 2.0 with four ATP quality five second duration, this is what we got. Now, with this prompt, and this is the same exact prompt that I used to generate this using the flow platform and with the same Seedance 2.0 for it to be quality five second duration, this is what we got. Now, by using the flow platform, I have generated this image three d style and running away from a man's foot, trying to crush it. Camera handheld shot, camera focus on and camera flow and cinematic shot. Now we got this image. I have downloaded this image, and with the same exact prompt, we can go to Seedance 2.0 with four ATP quality, five second duration. This is what we got. Now I again generate the second variation, and this time, I say ant running away from a man foot camera handheld shot camera follow ant because in this version, the man just crushed the ant, so I want the ant to escape the man's foot. And for this version, I also use Seedance 2.0 fest four ATP quality and check this out. And this time, the ant survive. Now we have this image, and by using this prompt, I have generated this image. Now with the same exact prompt and we got the image. I use Seedance 2.0 with four ATP quality, five second duration. This is what we got. Now we have this example. As you can see, guys, the character is now astronaut running away from mutant mouse. They both land in ocean. Then a shark, eat the mouse and person is left alone. Now, this is my prompt, as you can see guys, and I use shot one and show two type style, and I have used Seedance 2.0 fast for ATB quality 15 second duration. Now to write these kind of prompts, you can go to Cloud AI and you can upload your character and you can optimize your prompt. Optimize and enhance prompt for movie look person in space running from a mutant mouse in space station. Mouse almost cut person, person get out from space station, mouse follow him, both going down onto Earth as they enter hit Earth zone, a shock wave appear, both land on ocean, then both came out from ocean. Mouse coming close to person, but a shar cut mouse, person then take a deep breath of relief. And these are the prompts that we get from the Cloud AI, and these are the exact prompt that I used in this video. Now in my second generation, as you can see, I have used 2.07 20 P quality and 15 second duration, and I have used this image. Now you can simply go to flow platform and you can paste the exact same prompts one by one that you get from the cloud in order to generate images like this. Now, next, what I did, I replace my character image in space station, and I write this prompt. And as you can see, we get a different result this time. So I use these prompts, and these are my results. As you can see, guys, if you get higher quality, you will get higher definition cinematic result. Check this out. I mean, it looked like a complete movie scene. Now next we have this example and to generate these kinds of videos, what you can do simply go to Cloud AI. You can upload your character, and you can type optimize and enhance prom for cinematic look. And under this quotation, you can say gravity shifting VFX sequence, give ten ideas including the person in the picture. So we got the ten ideas, including the person in the picture. So I say give ten more, we got ten more idea. Now, next you can also say give me ten ideas for AI video generation one man versus Army Huge battle, cinematic depth of field, and we got these prom now to test this prom, you can go to simply flow platform and you can test this prompt. You can paste your prompt one by one to generate the images, and you can use those images. So as you can see, we have these images, and we can use the same prompt that we use to generate the image. And as you can see, we have this cinematic video. We have this one. And we have this 11 man versus Army. Check this out. And I have Seedance 2.0 four ATP equality and 15 second duration. Next, we have this one, and by the way, these are the same prompts that we get from Cloud AI. And final, we have this one. Take this out. 24. Start and End Frame: Welcome back, everyone. Now let me show you how we can generate stunning start and end frame videos and how we can avoid bad generation. So first of all, go to flow and you can generate your image. So I have this prom mechanical giant robot standing full body shot. This is the first variation, and this is my second variation. And this is the one that I use in my video. Then we have this image, rusty car full shot. We have this image, and I use this image as well. Now simply go to Cloud AI and you can use your images. So I have my character robot and rusty car, and I have this prompt. Optimize enhanced prompt for cinematic movie style multi scene prompt. We have the prompt inside this quotation person driving car in a distant giant mechanical robot destroying city. Man crash car into robot, person climb onto the arm of the robot and running towards his head and punch the robot. Robot knocked down, sparkling person standing next to it. Camera focus on robot then focus shift to person. As you can see, guys, we got this enhanced prompt. Now go to CapCut, go to Image udo, select your first frame, and select your second frame. And as you can see, we got these prompts. We have shot one as four, and as you can see, I divided my shots in this paragraph. Now, I have also used Seedance 2.0 for ATP quality, and this is my final output. Check this out. Now to this point, the video look fantastic except the first frame which we use this one and except the last frame. As you can see, if I play my video, at the end, we have this clip. So instead of selecting random start and end frame, you can go to flow and you can generate your start frame and you can generate your last frame. Let me show you another example how to fix these kinds of bad start and end point header. Now to get enhanced wrong first, I go to Cloud AI, select my character and optimize my prompt. Camera single continuous sequence focus on tire full acceleration, tire spinning in full speed. Then move front focus on subject, person drive car in high speed, as car move forward, dolly effect happen. Camera goes left sideway, show drag race, person win race. And as you can see, we got this optimized prompt. And I use this whole prompt in the flow platform, and I got these kinds of images. Now to get the last frame, which is this one, what I did, I used the last line which is this one to generate the last frame of my video. And as you can see, we have this now to have the character drive the car, what you can do, if I click, as you can see, you can upload your character, and then you can use the prompt. And I also used, let me show you, if I hover my mouse and click and as you can see, I have used Nano Banana to generate these kinds of images. Now next what I did, I select my start frame, which is this one, and I have my last frame, which I showed you earlier, which is this one. Now, this way, you can generate consistency and you can avoid bad generation. So as you can see, I have used Seedance 2.0 for ATP equality and 15 second generation, and we have aspect ratio, which is this one. Now, if I play my clip, as you can see, we have this masterpiece. Check out the video. Isn't this crazy? And now have a look. We have the last frame. You can also use dream mania for stunning start and end frame videos. So we have the person and I have type person standing in empty Street. We have this picture. Then I use the picture that we got from flow platform, and I have type person with full body metal, Bonic armored gandem style with a real metal look, liter rusty camera close up. And we got this armored picture. Then to optimize my prompt, I go to AI studio and I have type act as prompt optimizer from single prompt. Now we can optimize any prompt. We can give it any prom and it will optimize it. So I have typed this person as hand and Gandom armor assembled piece by piece with his body. While camera do 360 degree camera movement, when peacefully assembled camera Zoom into person phase, we have the optimized prompt. I use this prompt, go to dymnia.capca.com. And what you can do, go to Create. You can click here and you can select your star frame and frame, first of all, you need to select first and last. Then you can select your first frame and your last frame. And now have a look. Now, if you regenerate your clip, you will get better result. Now in AS Studio, as you can see, we have these different options. So I use one of these option. We have this 360 degree camera orbit. And for this one, I have dynamic transformation sequence. And if I play my clip, this is what it looks like. Check this out. Isn't this amazing? And by the way, I did not use Seedance 2.0 items. To use the previous version, which is C dance 1.5 pro. Now again, you can go to Cloud AI. You can upload your character picture and you can optimize this prompt. Gravity shifting VFX sequence give ten ideas. If you get ten ideas, you can ask for ten more. After that, you can go to flow and you can generate these VFX ideas. And as you can see, we have several ideas. So what I did, I selected this as my first frame, and I selected this as my last frame. And if I play my clip, as you can see, guys, we have this clip. Check this out. 25. Omni Reference: Welcome back, everyone. How can we generate Omni reference video? First of all, go to Cloud AI and you can upload your reference images, and you can optimize your prompt. But before even we optimize our prompt, how can we even generate these kinds of reference? Well, simply go to the flow, and as you can see, I have typed Army of robot, Army of Viking. We have rusty time travel motorbike, and we have modern empty New York. And last, we have Army of Samurai with mass. Once you have your reference, now we can optimize the prompt. We got the prompt. Now now let me show you how we can avoid bad generation. And this is example of bad generation if you do that. Now, as you can see, we have the person and we have the first prompt. After that, we have the bike because we are referenced this as a bike. Here how Omri reference work. You need to reference your character and your object, and then you can define the relationship between those two, the subject and the object. Now, in this example, this is the bad example of Omni reference because I referenced my character, the bike. But after that, I did not reference my subject writing pike. And first, the person writing Pike. And at the end, we have only the Pi because we did not reference the person is writing Pi. So how to fix these kinds of bad generation. So first of all, go to flow and create your reference. So as you can see, I have type time travel car. We have the first variation, the second variation, and after that, I say time travel car, all four angles. We have all four angles. Now we can go to Cloud AI and we can use this as a reference. And we can use the character as well. Now I have typed this person time travel to different places in a futuristic car, write a 15 second sequence with single continuous shot and have a. We have this long prompt because I say write it in one single video prompt. Now, after that, I also mentioned in bold text person and car and places. And now check this out. We have the car name and we have the person name which is time traveler. Now in this paragraph in some places we have time traveler only, so we can reference our character with the time travel only. Now, here's how Omni reference work. You need to first of all, switch to Omni reference. You can select your model. Now as you can see, we have two reference. Now, before we copy this prompt, I'm going to copy this, go to Drea Mania, and I'm going to say this is and we can shift plus and we can add the rate. We can select the image. So I'm going to select myself and I'm going to paste Time traveler because I am the time travel. I'm going to write this is at the car. And now I'm going to copy the Delance DMC 12. Now after that, what you can do, you can add E Coma. You can copy the same prompt and we can paste the prompt right here. And now you can also, if you want to increase the duration, you can also increase the duration and you can generate your clip. So I have already generated my clip, so I'm not going to do that. And here we have the final output. As you can see, guys, we have consistency in the arrector and the card. Have a look. Check this out. Now we have one final example of Omni reference. So first of all, I have generated this image futuristic robot boxing all four angle. As you can see, we have all four angles. This is my second variation. I use this one, go to Cloud AI, and you can upload your person. And I'm say person fighting with a boxing robot in a ring, slow more depth of field, 120 MM cinema lens T 1.2. Write a 15 seconds sequence with a single continuous handheld shot. We have the sequence I say, write it in one single prompt. Also mention person and robot in bold text because I want to reference the robot and the person. Now we have the first reference, which is the mask fighter, which is our character, and we have the robot, the second reference. In this example, what I did, this is the person and Ade Coma, this is the robot, and then we have the prompt. And check this out. This is the final output. So as you can see, with the Omi reference, you can generate consistency and cinematic shots. Check this out. And I'm using the basic C dance 2.0 fast. I'm not even using the C dance 2.0. 26. MultiFrame: Welcome back, everyone. Now let me show you how we can generate multi frame video. So first of all, go to Cloud AI and I have this idea. Optimize and enhance prompt and under this quotation I have right is subject running away from a giant enemy camera front shot, gift ideas in kitchen like mouse versus cheese. As you can see, we have ten ideas around the kitchen. Then I type, give me transition between scenes and it gives me this transition between the scenes and we can use transition, these prompts, and it will run in a loop. Let me show you. First, what I did, I used these prompts to generate my images using flow platform. As you can see, we are in flow and these are all the prompts images. So I have used four images. First, I use this one, then I use this one or maybe this one. Then for third one, I use this one, and for the last, what I did, I loop my video and this frame. Now, as you can see, first of all, you need to go to dreamina.capca.com. You can click here and you can select Multi frame. Once you select Multiframe, you have only two options. Either you can select C dance 1.0 fast or C dance 1.0 mini. If you use this one, as you can see if I select my frames. So as you can see, I have selected these frames. Now, this is the first prompt. You can paste your first from right here. This is to be your second prom, your third, and your fourth, which is the last. Now, if you switch to 1.0 mini, as you can see, it will skip the first prom and it will start your video from this frame. Now to seamlessly transition from one scene to another, we can use the same prompts and we can paste it. But what if you want to transition from the first scene to the fourth scene? Now in this case, what we can do, we can copy the first line exactly the same. But instead of the last line, we can copy the line of the third transition, the last line of the third transition, and paste it right here. We can replace it. Now, let me show you the final output. So as you can see the mouse run, we have the second scene and we have the third scene, and it loops back to the first scene. Now to give this video a loop effect, what we can do, we can press Alt key. We can select the video, and we can hold the lt key and drag this to the right side. It will make a duplicate copy. At the end, as you can see, we have this mouse running. So we can just select this and we can select this. Now we can go to the transitions. We can go to Classic. We can select any of the transition. We can increase the duration of the transition. And now, if I play my clip, as you can see, guys, now, the video is running in a loop. Now we have the second example. You need to go to flow platform, and I have generated my character using this from three D Pizza style, a person with a yellow ski mask ready for boxing in a locker room. Then I selected this image, and I have typed person punching back, and I have used this image, and I have typed person standing in front of his competitor, which is a robot. So they both appear in rocker room. So what I did, I changed the background. So use this image, which is the locker room image, and I say person standing in front of his competitor, which is a robot after a comma, I say in a boxing ring. So as you can see in some few words here and there can change your entire image. Now I use this image, and I have eye person at corner sweating, camera, clows of shot. And then I use the same image which is this one and I have typed robot at corner sweating oil. Then I use this image again, and I have typed person punch robot. Then I use this image, and I have typed close off robot face getting punched in slow motion. And I have used this image again, which is this one. And I say robot knock down closer. And at the end, I use my same image which is this one, and I have typed person breathing heavily standing next to robot knocked down on ground. Camera low angle. Now I have used these same exact prompts. Now in Capcot I have used these same prompts. So in order to generate multiframe videos, what we can do? So let me just cancel this, click here and now go to AI video, go to Multiframe. As you can see in multiframe we can't really select the model like we can select in imania. You can select the quality if you want to. Now, as you can see, I have also selected my resolution, which is 720 P. Now to generate multi frame videos, you can select your images. Now I'm going to select the locker room person, and we can click here. We can select the second image. Now, as you can see, as we select the second image, we have this text appear, and we can also control the duration of this prompt. Now if we go back to this clip, if I click here, after my first image, I have this prompt. After my second image, I have this prompt. After my third, we have this prompt. And if I play my clip, as you can see, guys, we have this video. Now the reason why we are getting these kinds of error because we are using the oldest version of dance. Now we are here in dremna.capcad.com. I have selected C dance 1.0 fest. We have multi frame, and I have selected all my frames at once. Now you can also rearrange them if you want to. So as you can see, this is my first frame. So I'm going to drag this at the start. After that, we have both boxing, so I'm going to select this one. After that, we have the close up, and then we have the close up of the robot. Then, sorry, I'm going to rearrange this right here. And I'm going to rearrange this right here. He punch, robot, get punched, knocked down, and at the end, we have this. Now, if we total this, as you can see, guys, it's make 35 second clip. So we can lower the duration. So I'm going to select the first frame and I'm going to paste my first prom right here. And for the second prompt, I'm going to lower my duration through three second. I'm going to copy my prompt, which is this one, and I'm going to paste this right here. Now I'm going to also lower my duration through three second. And for the third shot, I'm going to select this one. Paste this right here, lower the duration, copy this and paste. And I'm going to also lower this to four second maybe. Let's change this to paste the prompt, confirm. And for the slow motion, let's make the five second. So I'm going to paste this right here, and for these two, we can lower the frame to three second. And for this one, I'm going to also lower the frame to 4 seconds. We are using 120 credits, which we are getting for the daily basis for completely free right now at the time of this recording. Robot knockdown. For the last one, we have this. I'm going to confirm it. Now let's click to generate this video. The video is ready. Let's see which one is the best. Either cap cut multi frame C dance 1.0 or the CDance 1.0 fast. So right now, I believe it's performed better than the normal CDNs 1.0. In most of the places, we have less error, but we also have this bad generation. So I believe the cap cut version performed better than the fast version. Now let's look at another example. So go to flow, and I have used old childhood cartoon shows. As you can see, we have several shows, and I turn them into three D. So if I click here, as you can see, I use this picture and I say, turn it into Dreamwork three D style. And this is my second variation using nano Banana two. Now we have Adventure Time. We use the same exact prom. We have DragonbllZ add add and Add, we have Scooby do. As you can see, I'm using the same exact prom to turn them into three D. And last, we have some Rye Jack. Now in Capcad, I just upload all my images, and if I click here, as you can see, I say animate all these images, and I turn this on. And it will apply your single prompt to all of these proms. So let me show you how it works. Now, as you can see, we have so many images. If you want to rearrange them, you can also rearrange them if you want to, but I don't want to rearrange them. Let me just track this and drop I'm going to click here and I'm going to apply all. If you change the duration and done, it will change the duration of all these clips. As you can see, now all the clips are now in 2 seconds. You can select any text prompt. It doesn't matter. We can select apply to all, increase the duration, click Done, and as you can see, all the duration is now 3 seconds long. Now we can click here to generate this video. So the video is ready. This one is 480 P quality, and this one is 720 P quality. And if we take a look it's animating the images very accurately, but we have so much morphing effect going on. We have so much error because we are using the oldest version of C dense. Now let's take a look at 7:20 B equality with all the same images and exactly the same prompt. Check this out. I mean, it does perform a little bit better than the four eightP quality, but we also, again, have so much morphing effect going on in the video. We also have some character appearance. I don't know why this character appear. In the next film, it does not have this character. So after watching these two, I believe you can use the 480 P quality and you will be good to go because the 720 P quality will cost you 2000 plus credits, and the result does not surprise me at all. Like we just watching YouTube video and we change the quality from 480 P to 720 P, and the rest is the same. Now we have one last example. Go to Cloud AI, upload your character and optimize your prompt, like we did in the last lecture. As you can see, we got the optimized version of our basic prompt. Now we can go to flow platform and we can use the prompt to generate our image. Remember, you need to upload the reference of your character, and then you have to paste your prompt in order to get these kinds of images. As you can see, we are building this storyboard. We have this story right here. If you got the bad generation, don't worry, you can regenerate your prompts in order to get better results. As you can see, this is my first variation, so I did not like it. So again, regenerated, and this time we got this one. Now go to mina.cbc.com. You are using C dance 1.0 Mini, what you can do, you can click here. You can say animate these Imgs. You can select your second and you can select apply to all. If you get the high credits, if you don't have much credits, what you can do, you can then individually select your particular scene and you can lower the duration of that scene. After that, you can simply just click here to generate your video. So the video is ready, and we have both versions, C das 1.0 fest and 1.0. First, we have C das 1.0 fest, and you have to keep this in mind. We also use the same prompts that we used to generate these images. We have so much morphing errors in the video. Now let's look at the result of C dance 1.0. I mean, we have less morphing effect than C dance 1.0 fest. We have much more cleaner result than the 1.0 fest. I mean, check out this scene. I believe the C dance 1.0 performed better than the 1.0 fest. I know this one is not a masterpiece, but it does perform better than C dance 1.0. So this is how you can generate multi frame videos using C dance. 27. Jewellery Commercial: Yeah. He. Welcome back, everyone. Now I will show you how we can make a jewelry commercial using C dance two point oh. So go to aistudo.google.com and you can type this prompt act as prompt optimizer from single prompt to multi scene prompt. Now we can get multi scene prompt from a single prompt. Now, I have typed this symbol prompt. Jewelry Commercial, focus on only jewelry, showcase jewelry without any model. We got the prompt. Now, if you have a reference image, you can also use that for consistency. So as you can see, we also have the reference image, and I have typed this four sequence prompt for each item shown in image, each shot must be two second long. Switch between different camera angles during showcase jewelry don't include human or animal focus on items. And as you can see, we have these proms. Now, I have used these proms. If we go to the flow platform, as you can see, these are all my results. We also have these two as well. Now next, go to dreama.capqad.com, and we can also access to dream dance two point oh fast from here. We can also access to CDance two point oh as well. I have used Omni reference, and with my reference, as you can see, I reference all my images like this style. This with my first image, I type my first prompt. For the second shot, I said, I have typed this with my second image, and I have typed my second prompt. And I did all this with the other images as well. And this is my final output. Now, as you can see, at the start, we have this error where the image is not even moving. We have a steel image. Now to fix that problem, what I did, I use image to video using the previous model, which is C dance 1.5 pro base. Even with image to video, we have some error. So next, I regenerate my clip using the same prompt and the same image and even with the same model. And this is my final result. Now we can download this, and we can mix this with our final commercial. So let's download this from here. Let's download the commercial as well. Now create a new project and you can rag and rob your videos right here. Now I'm going to place this on top of my clip, and as you can see, we can also go to this speed and we can also make it faster. So I'm going to say, so let's hide this first. And so the second shot is starting from here. So what we can do, we can enable this, go to the speed, and we can click here and we can make this faster and check this out. And now we have the final commercial. And this is how you can create a jewelry commercial using CDance 2.0. 28. Multi Styled Animations With Seedance 2.0: Welcome back, everyone. Now I will show you how we can make multi style animation using C dance two point oh. So first, we have this three D animation where I use a Chinese prompt because the single English was not working, so I convert my prom to Chinese. I also use C dance two point oh fast with 720 P equality and with 15 second generation. Now you can go to Google Translate. You can translate your prompt. You can also convert your prom to simplified Chinese. So I have this prompt, three D pizza style animation. Short one close up of a person wearing a yellow ski mask practicing boxing in a locker room preparing to hit the punching bag. Second, we have human and robot face to face in a boxing ring. Third, we have close up of a person sweating profusely in a corner. And fourth, we have closeup of a robot covered in oil and sweating profusely in a corner. We have two men punch each other and sixth, we have slow motion close up of a robot being hit in the face. Final, we have closeup of the robot being knocked down, followed by a person painting beside the felon robot shot from a low angle. And with this from, this is my final output. Check this out. Now next we have this anime style video, and for this, I also convert my prompt to Chinese. I also use the C Dance two point oh fest. This time, I use the lower quality for ATP, and we have the same duration which is 15 seconds. Now, as you can see, I have these prompts, and to get the prompt, what you can do, go to astudioggle.com and you can type this pro ideas, POV anime style, cooking, washing clothe, repair bike, typing on keyboard, et cetera, in detail. And as you can see, we got these prompts. So what I did next, I just copy each prompt one by one and I divided them in different kind of shot. Like we have shot one for three second, et cetera, and these are my final outputs. So you can make multiple videos in one video. So as you can see, I have multiple shots in one shot. Now, we also have this version where I use 90s anime style, short one, four second blue robots say in Chinese, this, and as he speak, his eyes blink with words. Second show closeup of the green robots say this in Chinese. As he speak, his eyes blink with words. And third, we have white shot ship attack Earth fire giant cannon. So you can also make dialogues in your scene. You speak. And for this video, I also use S dance two point oh fast with 480 B quality. Now we also have this version with the same prompt I used the previous model, which is S dance 1.5 pro. Was Coco deme Jimi or Kenchustema. So haka Bekideska. We also have the second variation with 720 equality and with the same model. Wata or Kenchema So Hake Bekideska. And finally, we have store motion animation. Now to get the prom, you can go to Cloud AI and you can type this prom, optimize enhanced prompt for cinematic movie style. Clay mission, a mouse, and a cachaseGive three ideas. We got the ideas. Now you can simply go to flow, and as you can see, this is my image that I got from flow. We also have this one and we have this one. So I use this image and with the same prom dad we use to generate this image with CDs 2.04 ATB quality with image to video, as you can see, I have selected my image and we have the prompt. This is my result. We have some error in the video, but as you can see, we have this storm motion animation. Now, next, we have this one with this image and with the same prompt that we use to generate this image with five second duration, C dance two point oh, this is my final output. Check this out. Isn't this crazy? And final, we also have this example where I use this image, and this is the prom that I used to generate this image. And with C dance 2.4 ATP quality, five second ation, this is my final output. Check this out. So this is how you can make animation using C dance two point oh. 29. Pizza Commercial: Welcome back, everyone. Now let me show you how we can generate a stunning Pizza commercial. First of all, as you can see, if I click here, we have our character, and we have this prompt. So for the prompt, we need to go to Cloud AI. Now in Cloud AI, we need to upload our person, and we need to type the prompt. So I have typed optimized and enhanced prompt. Guy in a picture, order pizza from three different hotel, three pizza delivery guy battle to deliver pizza fast in a race face. First rider with red face, schemas, second rider with yellow face schemas, third rider with green face schemas. Yellow guy won by delivering the pizza to the guy in the picture, hand him over the pizza. So we have this detailed prompt. So what I did, I copy this line, as you can see, and with the person reference, I paste it right here in the flow platform, and I have generated these images. So this is the most accurate image that I found across all my images. So what I did, I use this image. Then as you can see, to test out my prompt, which is this one, I paste it in the flow to get the idea what the prom so as you can see, we got these. Now, next what I did, I copy this same exact prompt. Go to Capcut, go to AI Media, and you can go to AI Video. You can upload the first frame and the last frame that you got from flow. Now for the last frame, what you can do, you can copy the last line of your prompt, go to the flow, and as you can see, you can use your reference, which is this one and you can type a prom like person standing at the doorstep with steaming pizza box into the hands. So we have the last frame, which is this. Next, what I did, I convert my prom to Chinese because the English was not working for this commercial. And I have also used 720 P 15 second generation with C Dance 2.0 Fest, and this is my final output. Now to improve my commercial even more, go to Cloud AI, upload your reference, and you can type your prompt. After your prompt, you can type technical detail. Like write a 15 second sequence, also add seamless transition between scenes with handheld intense scene. Single continuous shot, depth of field shot on cinema camera. Lens 16 MMT 1.2 cinematic lighting, write it in one single full video prompt. Now as you can see, we have the prompt. Now I also write write prompt in short one, show two, et cetera style. So as you can see, we have short one, show two, short three, et cetera. I say, write it in one single video prompt. We have the prompt. Now I say write three more variation. So we have three more variation. Now next what I did, I copy all these prompts one by one, and I have tested each prompt. So if we go to the flow, as you can see, I have tested all these prompts. Now in all these prompts, what I did, I have selected this one because I really like this one. Now I have also again used CDance 2.0 first version with 15 second generation, and as you can see, this is my prompt. With the reference in the image in one single image, we can generate more accurate video. So if I play my clip, check this out. Mm. Isn't this amazing? And this is how we can generate a pizza commercial. 30. Shoe Commercial: Welcome back, everyone. Now I will show you how we can generate a Shu commercial using CDance two point oh. So first for the prompt, we need to go to astodgole.com, act as prompt optimizer from single prompt to full multi scene prompt. Now, after that, you can type any prompt and you will get a multi scene prompt. So I have typed running shoe commercial, focus on shoe, first show, assembling shoe floating in air, focus on shoe, man wear shoe, transition to monster wear shoe, transition to robot wear shoe, and then alien. Shoe brand name turbo tagline, no matter who you are, we believe in equality, where the comfort. So as you can see, after that, we have these. Now if you don't get multiple prompts, you can also write a prom like give me more variation of the same prompt or give me three or four or five variation of this prompt. Now as you can see, we got multiple prompts. We have first the assembly, then we have the human, then we have the beast, et cetera. After that, I have right now make a master prompt combining all six scene in this style. Short one, show two, and then at the end, we have short six in one single prompt. And now as you can see, we have this master prompt. So I use this prompt. Now if we go to the cabcut.com, you can access C dance from here. And as you can see, we have the same exact prompt. I just divided the duration. As you can see, we have three second, we have three second, we have two second, and I'm generating a 15 second commercial using the 720 B quality and the C dance two point oh. And here we have the final result. Check this out. Now with the same exact prompt, if you make a five second commercial, you will get something like this. So as you can see, we have the same prompt. I accidentally forgot to remove the second, but it understand it anyway. And as you can see, we have five second generation, 720 B quality. Check this out. So by using the same prompt, if you regenerate and lower your duration, you will get something like this, a fast paced commercial. Now, after we get this master prompt, what I did, I copy this prompt, paste it right here, and then I have typed this prompt. Optimize and improve prompt, make it cinematic, and we got this prompt. Now next what I did, I use this prompt, and here we have the result. As you can see, we have some error in the video, and I'm also using the four ATP quality for this prompt. Now after this prompt, I then copy the same prompt, which is this one, and I also type at the and optimize and improve prompt, make it cinematic. And this time, as you can see, we have this prompt. We got this one. Now with the same exact prompt, what I did, I used C dance two point fest with 720 B equality and 15 second generation, and here we have the final result. Now, still, as you can see, we have some flaws, some error in the video. So next what I did, I lower my resolution with the same exact prom and with the same model. And here we have the result. Now, as you can see, guys, we have this error. The shoe is facing this way while the person is tapping and his hue is facing this way. So to fix this problem, I just split my clip like this, and then I select my clip, and if we go to this one, we can mirror this, and as you can see, now the shoe is facing this way. So this is my final version after I fix this problem. And here we have the final commercial. Now, if you regenerate your same exact prom, you will get some amazing results. So if you don't get the result that you want, try to regenerate, it will help you to improve your result. And this is how we can make a shoe commercial using C dance 2.0. 31. Dentist Commercial: Welcome back, everyone. Now, I will show you how we can make a Dentist commercial using Seedance 2.0. First of all, go to astudio.google.com and type this prompt. Act as prompt optimizer from single prompt to full multi scene prompt. Now we will write a single prompt and it will give us a multi scene prompt. So we have this basic prompt. Three D Illumination studio style Dentist commercial with voiceover. Monster feel pain in tooth, then dent is commercial, then monster doing brush, moster clean tooth with wire at the end, monster smile and happy. As you can see, we got the prompt. Now sometimes it happened with Seedance 2.0 that it will give you an error. The video is failed because it will give you suggestion that we can't really make this video because you need to make these changes because it violate their guideline. So what you can do, you can read these suggestion. You can copy the suggestion, and then you can type your prompt and at the end, what you can do, you can apply these changes to given pro as you can see, these are the changes that I copied from the CapCut from this paragraph. And after, as you can see, I got my prompt. To test if the prompt is working or not, I copy my prompt, and as you can see, guys, we have this video. Now, sometimes you also need to translate your prom to Chinese if your prompt is also not working, even if you make the suggestion changes, then you can convert your prom to Chinese. Simply go to Google Translate, paste your prompt and select the Chinese simplified. It will translate your prom to Chinese and check this out, we have the final commercial. Isn't this simple as that? Now in this prompt still, we don't have any voiceover. So what you can do in post, you can go to CapCut and you can copy the voiceover. For example, like this, we can copy this. Now we are here in CapCut. We can go to the text. We can click here to add this default text. We can paste our text right here and we can go to text to speech. We can select any voiceover. Now I have selected this voice story teller and I'm going to generate my speech. You can select AI to preview your voiceover actor, and now we can delete this and check. Student, keep your sparkle legendary. And now we can simply just download our video from here. We can import our video, and at the end, from here, we can add this voice over. We can also go to the audio. We can lower the volume and check this out. Student, keep your sparkle legendary. You can also select your audio, go to Voice hanger, and you can also apply some effect. And now we can select any voice filter. I have selected music enhancer and check this out. We have the final results. Mod student, keep your Sparkle legendary. And this is how you can make your dentist commercial using Seedance 2.0. 32. Suit+Watch Commercial: Welcome back, everyone. Now I will show you how we can make a suit commercial using C dance two point oh. So first of all, I have generated this suit. So first one is hanging black suit. We have hanging black and wet suit, and we have hanging golden suit. Now, next what I did, I use my character, and I use the suit as a reference and have type person wear this suit. Now, this image was not looking right, so next I change my reference and I say person wear this suit. Next, I change the whole prompt. And as you can see, this time, the black suit is looking good. So I have tired person wear this suit in front of a black hark or studio background, and I use the same prompt to generate all the other images with different. This up. Now next for the prompts, we need to go to Cloud AI. We can upload our reference images, and I also typed this prompt. Ten second sequence commercial cinematic, focus on four suits, showcase them in fast pace sequence. Add seamless transition between shorts, optimize and enhance prompts, and we got this master prompt. Now next what I did, I convert my prompt to Chinese, and I have also uploaded all my images. So if I click here. So now, as you can see, we can upload this reference like this. So I have uploaded this reference, and then I paste my prompt that I converted into Chinese using Google Translate. And this is my final result. And by the way, I also use four ATP quality C Dance 2.0. Isn't this amazing. Now for the bonus, I use the same method to generate these watches. As you can see, we have wrist watch silver, and we have wristwatch golden, and we have wrist watch black and red waveform pattern. And finally, we have futuristic wrist watch closer. Then again, I use Cloud AI to generate my prompt. As you can see, we have these reference images, and we have this prompt. 10 seconds sequence commercial cinematic focus on what is, showcase them in fast pace sequence, optimize and enhanced prompt. We got this prompt, so I use all my four reference and with my prompt, check this out. This is my final output. Now, this is my first generation. That is why we got some error. I believe if you can wt your prompt into Chinese, you will get better result. And this is how we can make a suit and watch commercial using C dance two point oh. 33. Chocolate Commercial: Welcome back, everyone. Now, I will show you how we can make a chocolate commercial, even if you use the oldest model of Sedance which is C dance 1.0 fast. First, to get the prawn, we need to go to Cloud AI and we can type this optimize and enhance prompt for cinematic commercial for chocolate. Focus on chocolate, CDI shots, spiller shot, swilling shot, vortex, et cetera, right prompt for 10 seconds sequence. And as you can see, we have these prompts. Now we can go to flow and we can generate images using the same prompts that we got from Cloud I. So first, I generate my chocolate bar, simple design chocolate bar name get Choco. This is my chocolate bar that I created. And with my reference image, as you can see, guys, I have my reference image, and then we have the prompt. So this way, you can make a consistent commercial. So as you can see, this is my first shot, check this out. We have the second over the head swing vortex full back. And if I click here, check this out, we have the same pro. So I use the same prompts to generate these images each one by one. Now if you don't get the result that you want, you can also regenerate your prompt. And as you can see, I have regenerated my prompt. Now I have also convert my prompt to Chinese. Now I also don't understand the Chinese language here how I convert it and use in multiframe video. So first of all, select your first prompt. You can copy your first prompt. Go to Google Translate. You can paste your prompt, and you can select Chinese simplified. You can copy this prompt. You can go to multi frame video, select your first frame, and before the first frame, you can paste your prompt, which is this one. Then you can repeat this process to all the other frame as well. As you can see, this is my final output. Now, check this out. Isn't this amazing? Even we are using the oldest model of Sedance, we generated this cinematic commercial. This is how you can simply create a cinematic chocolate commercial using see dance.