--- title: 'Kling AI Motion Control: Ultimate Guide from Beginner to Pro' source: 'https://youtube.com/watch?v=P_hhQyDd5E8' video_id: 'P_hhQyDd5E8' date: 2026-07-24 duration_sec: 1285 channel: 'Arca Artificial by Lordwind Enrique ' --- # Kling AI Motion Control: Ultimate Guide from Beginner to Pro > Source: [Kling AI Motion Control: Ultimate Guide from Beginner to Pro](https://youtube.com/watch?v=P_hhQyDd5E8) ## Summary This video provides a comprehensive guide to Kling AI's Motion Control feature, explaining how to transfer movements from a base video to an animated character or superhero. The creator covers the interface, key concepts like image match vs. video match, five rules for success, and advanced tips for maintaining consistency. ### Key Points - **Kling AI Pricing** [01:48] — Kling AI is a paid tool with a basic plan at $8.80 per month, providing 660 credits. - **Image Match vs. Video Match** [02:05] — Video match prioritizes the base video for smooth movement, while image match prioritizes the reference image for artistic fidelity. Choose based on movement type: video match for fluidity, image match for consistency. - **Rule 1: One or Two Characters per Shot** [05:40] — Using too many characters confuses the AI, leading to poor results. Stick to one or two characters per shot. - **Rule 2: No Editing Cuts** [06:33] — Abrupt shot changes disorient the AI. Use continuous video for best results. - **Rule 3: Good Lighting and Clear Movements** [07:04] — Record with good lighting and ensure movements and gestures are well-marked for the AI to transfer accurately. - **Rule 4: Avoid Weird Positions** [07:49] — Positions like lying down or face down will not work and waste credits. - **Rule 5: Video Length and Framing** [08:03] — Videos can be 3-30 seconds (max 10 seconds for image match). Ensure the reference image has ample space for movement to avoid cropping. - **Formula for Creating Reference Images** [09:55] — Use a prompt like 'keep the same pose and composition' to ensure the character stays in the same frame as the base video. - **When to Add Prompts** [12:14] — Add prompts to create objects (e.g., a sword) or environmental reactions (e.g., fire). The AI transfers movement but can generate reactions if specified. - **Linking Facial Elements for Consistency** [18:23] — Use the 'link facial elements' feature to upload multiple angles of a character, improving consistency across movements. ### Conclusion Mastering Kling Motion Control requires understanding the balance between video and image matching, following key rules for recording, and using prompts strategically. With practice, you can create professional-looking animations for various use cases. ## Transcript you can imagine, without chroma key, without costumes, and without a budget of thousands of dollars. Cleanmion control allows you to transfer your movements, your face, your expression. It an animated character, or even a superhero. The truth can be anything you imagine, like this, no, no, no, not this, not this, this is on the channel, I'm Loren Enrique, captain of the artificial ark, and today I'm going to give you the ultimate guide to mastering clean motion control. You're going to learn all this to go from zero to advanced mode. You already know that. Let's begin. Control, if you've never used it or don't know what it is, here's what you need to know. When encounter this interface. Simply clicking on this symbol where it says " generate" will take you directly to either video generation or motion control. Up here you'll also have four options where you can go through all the generations that Clean currently has. In the first square here is where you will upload the video that has the movement, the one you want to be the basis of your final product. In the second one, where it says adder image, that's where you're going to upload the image that will replace the person in the base video. In this case, he's a magician. And down here you can write the promise, you can leave it as is. I'll explain this to you later . For those who don't know how Clean works, it's a paid tool, but it's well worth it paid tool, but it's well worth it . In fact, their most basic plan is $8.80 per month and they give you 660 credits. To start working with Clean Motion Control, you need to understand what image match and video match are. Because? Because this question is not very well explained, I'm here to explain it to you in the easiest way. If we go back to Clint's interface , as I just showed you, where you upload the reference video and the reference image, at the bottom you have something that says character orientation match that says character orientation match video or character orientation matches image. What does this mean colloquially? In the match, the most important thing for the AI ​​is the video, where the image will have consistency, but there is a smoothing because what the intelligence will do is reconstruct this will do is reconstruct this image adapted to the video. However, the video will be super smooth because its main reference is the base video. In contrast, in Match Image it's the opposite. The image is the main reference. The artistic aspect of that image is the main reference. Are you going to take the video? It may lose a little of its naturalness, but it will be taken as if the main thing is the image. I might have some examples for you to see. I looked for this video that many people have seen, which is a very brutal scene because it is a single take where the person walks and many things happen. What I wanted from this video was the person's walk. I uploaded a reference image of a grandmother in the supermarket, and the first thing I'm going to show you is how video matching works. As you can see, the grandmother walks and the entire video captures the full camera movement that is in the base video. Go up, look, the grandma goes up, it gets distorted a little. The boxes also distort a little, but it's all based on movement. The same movement that exists here. Notice that in the image, the camera goes up a little, it gives the person up here some air and she runs away. It's the same thing that happens here. The camera goes up, gives it a little air and all that. What happens if I select a match image? Things are different. image. It's no longer blurry. Grandma looks perfect. He runs away. Yes. It captures the person's movement, especially because they're moving in a straight line; the camera is moving backward while they're moving forward, and it captures it perfectly. In this case, Match Image will give me a better result and better generation because the movement is in slow motion and it will be matched without sacrificing the image. But if the movement is the opposite, a little more abrupt, you have to take the video match. I also have another example for you so you can continue understanding, and things you can do. I have this shot which is a dolly shot to the right. The person is walking, he has his lying down and my image, as you can see, is the complete opposite of how he is. Notice that the base is the same, it's like the same building, the same area, it's animated, yes, but the person is looking straight ahead. According to the theory, in Match Image, he will take the reference of the movement, which is the person moving while lying down, but he will do it from the front. I'll show you the result. The character is walking in the same place, lying down in the same position, but facing forward. It maintains the [clearing] overall consistency or takes the main consistency of the image, but adds movement. Now that you know how video and image matching work, you can create whatever you want. However, I'm going to give you five rules to help you avoid making as many mistakes as possible and wasting money mistakes as possible and wasting money on videos that might not work. The first rule is this: one or two characters per shot. Look what happened when I tried to do something with a lot of characters. I made the frame of the grandmother where she's going to be fighting with the people. fighting with the people. Notice that the AI ​​doesn't perceive the movement as such; that is, it tries to perceive it, but it makes a mess because with so many people, it doesn't know where to focus, what the main focus is. Even if you call it grandma, there are so many people that the AI goes a little crazy. Two, without editing cuts. Another thing that affected this Bruce Lee video is that it has many shot changes, and they are abrupt and fast. That's why the AI ​​goes crazy, because it gets disoriented not knowing what the main focus is. Even if you put it on her, she gets disoriented because there are many abrupt changes in the feeding. That's why you should always try to make it a continuous video to get the best results . Third, when you're going to record a video to use with clean motion control, try to have good lighting so that if you're going to be in full body view, you look well lit. It was going to be medium, like the videos I've been watching of myself, medium but well lit, where the movements and the mouth, hey, if I'm going to speak, the gestures should be quite speak, the gestures should be quite well marked so that the AI ​​can pick it up. Remember that the AI ​​in clean motion control transfers your movement, it doesn't create it, it transfers it. So, if I do it this way, the video will be like this. Don't invent anything you haven't done yourself. But this video you record has to be good. If what you record is wrong, the result will be wrong. Four. No weird positions like lying down, face down, no, none of that. This AI won't work and you'll lose your credits. Five, the time of the videos. In Clean Motion Control, videos can be from 3 to 30 seconds long. Note that there is one detail you need to know: in Match Image the maximum can be 10 seconds. And there's an important point that many people don't mention, which is whether the movement you're making in the video is super wide, okay? The image you are going to upload, which is the one you are going to change, also needs to have ample space to allow for the movement. Otherwise, the guide will cut it. I might have an example to show you this. When I was testing the video match, the image match, I had this video, I had made this first image, but notice that the person is framed here at a three- that the person is framed here at a three- quarter view, okay? And here it is in full mode, whether it's in match video or match image. And the result may be the following. The person does walk forward, but tries to go walk forward, but tries to go sideways and it's super weird. Because? Because the person's position is not the same, it doesn't have the same feel as here, and it's super weird. That's why when I created this image, when it maintains a similar depth, so that the person appears standing, then it has already changed and it has indeed taken on what is known as the mash image. So, keep this in mind. Very good. And as always, we have a formula for everything in the ark so that you can solve this problem. But before I give it to you, I want to ask if you're finding this video helpful and enjoying it. Please subscribe and hit the bell icon so you know when I upload a new video or when I'm adding new information to the community tab. free for you. All I ever ask is that you subscribe, like, and comment so I can reach more people and those more people can learn with me, just like you. Like it. We continue. Once you have your reference video or base video, recorded it , and followed the five steps I gave you above to avoid making any mistakes, then you need to create the image to create that fictional character or the character you want to include or that will replace you in your video. I always that will replace you in your video. I always use the same formula, and it is as follows. And with this prom, what you're telling the guide is, "Hey, keep the same proportions and make sure the person or whatever is going to be in the same frame as what's in this image." I'll show you the same way I did it. For the intro. What I did was record these videos saying certain phrases with pauses to know where I was going to cut. What I did was take a main frame where I could see myself framed in the shot. As you can see, I'm putting the promo I 'm giving you, and that's how I dried up all the variations where I was. In this case I did want to change the background to make it a little different, but notice that the character is always in the square where I am and that's what you have to look for. I always included a reference point to ensure the consistency here was as realistic as possible. But based on this, your character will always be in the same frame as you, with the same air as you. And when it comes to creating the video, all you have to do is upload the video. Okay, this is a piece of the video that I cut because I made it 6 seconds long. I uploaded the picture of Superman, as you can see, it's here. The video generates my movements and the image of Superman. And as long as you follow this formula, no matter what you create, it will always give you a good result. And something important that I haven't told you, in none of my videos, as you can see, I 've put a promo on it. Because? Because the model is so perfect, especially in these cases, I perform an action and I just want my image to perform the same action as me. In these cases where there is nothing to add, don't put minutes ago, clean motion control transfers the movement and there's no mistake about it. If the framing is the same, the movement is clear, mistakes. mistakes. In what cases should you add prompts? I'll Spartan video, notice that in the base video I pretend to have the sword in my hands, and when I made the video, notice that he did the hand gesture, but he didn't do anything. In other words, he didn't put or create anything. That confirms what I'm telling you, that it only transfers. What did I do here? And yes, I put it, he says, hey, I do here? And yes, I put it, he says, hey, I put it, the warrior has a sword in his hands and when I generated the video, well, look, I already have the sword and he puts it and holds it in his hands. In this case, this can happen, or sometimes it might vanish, like, I have the sword and at some point in the fight it vanishes, but normally it will always keep the object you tell it it has in its hands, so it will work. Another example of when to use video where I was supposedly a broomstick to act as a staff. That was my intention, to show him riding. But here's where I confirm that this function doesn't create movement, but it can create reactions or actions that can happen in the environment. I'll keep showing you in this pron. I'm going to teach you two things, and I 'm sure nobody tells you this outright: the first thing I'm going to do is show the magician walking in the cave, and the intention is that he makes a movement with the staff and the staff bursts into flames, okay? To illuminate going to teach you. My mistake was that I made this move, but I wasn't moving. If I had stepped back and started doing this and stepped forward, okay? It's like there's a forward movement and he transfers it and he was going to achieve it. That was my mistake. And the second thing is that in these cases, if something happens at some point in the video, you can tell your daughter. I'll show you. Look at the video, I tell you 1 2 3 seconds I take out the movement of the vaculus. I'm putting myself on the prom. In the first few seconds he sees, that is, the magician sees around, but exactly in the third second, it tells you here, in the third second he moves the staff and it catches fire. And besides, and this is another thing I'm going to show you here, when the light illuminates the face, I'm making it look seated, okay? Let it illuminate the cave, let there be a reaction to when, hey, it's dark and the light turns on. I put all of this the light turns on. I put all of this in the prom and the result is the following. walks, boom, I light the staff, the fire. But the important thing here is this. Look how the generation reacts, lighting up my whole body, my whole face, my face like , oh yeah. And the fire that's advancing here, how it goes. I think this is amazing because it gives you a lot of room to create scenes and add-ons scenes and add-ons for any videos you want to make. This can be used in real-world professional use cases. Yes, I could give you, look, more than 10 professional use cases where you can take advantage of a tool like this, but the video would get super long. And I'm going to mention this in the blog post, which you'll find in the description so you can read it when I upload it, and you'll know, look, if you can do crazy things with this . I'll give you three examples that I think are the most suitable for the channel's theme. The first case is cinematography. I The first case is cinematography. I recorded this at home. a rather strange movement in my room, and 10 seconds of video. And room, and 10 seconds of video. And this is the result I obtained. prom thing I told you about regarding time, and I'll show it to you. Look, notice that I'm telling him the guy is in a parking lot, okay? Look to the sides and in the third second he starts shooting and as you could see in the third second he shoots and this is the control you can have with this. If you know the exact second you perform the action and tell them, then it will work. And as you can see, it's a normal video. No, I'm not making the best films in the world either, but it's an example to show you how it works. Second case, imagine you sell Second case, imagine you sell products on Amazon or TikTok or Instagram and you want to make videos to reach different niches or different types of audiences and well, you don't have time, you don't have much of a budget and you can record a video like this. When I want to train, I always take this creatine because it's 100% pure creatine, and my fanciful intention was, look, I want to reach people who exercise in a gym as well as people who are in their office. They are two different audiences, so to speak, and this is the result. And for the third case, imagine you are a teacher and you want to do something dynamic for your students. I recorded this what the bones of the face are. For what the bones of the face are. For example, the nasal bone, the mandible, and the jawbone. First, I'm doing some super improvised scripts and it's something simple to give you an idea of ​​how you can act. But what happens if I take my bone class or my product and use another tool where I give it a female voice, a girl voice, change the voice, or use a different language? Hey ! It's not in Spanish, but in English, it's in Arabic, it's in Italian, it's in Portuguese. And I'm starting to reach other niches with the same product. I created a video, changed the image, changed the voice, starting from a single video, the same with the classes, because look, I am an English teacher and I want to teach my students in a dynamic way, my students who are children of such and such ages, so I want to show them what they are like , but in a dynamic way. This works, and there are millions of other things you can do with Clean Motion Control. And here's one last tip to help you maintain the best possible consistency when creating your videos. And this option is the one that shows you below, which is to link facial elements to improve consistency or coherence. How does this work? If you click here, you'll see that you can create elements with the function . You can create databases about a person. For example, if I click "create," you can upload a video or an image here. If I click " attach image," okay? I can find this girl myself. Boom! And here it will tell you to upload some different angles. So, look, I'm going to upload this, this, this and this. I think there are only three. Correct? There are only three of them. I have all four different angles. I can add a voice, okay? I can add Laura's name to it and give it a name. Since I already did it, I have it here. What did I do? I took one of the videos I had already recorded and the result is as follows. An animated character result is as follows. An animated character And what this does is that if there are many movements, okay?, to the sides and such, it doesn't invent, but maintains consistency based on what you uploaded. This is your case, I'll leave it to you. This feature is there and many people overlook it. So if you've liked this so far , boom, subscribe. Before we important points. One, your reference video is the ceiling of quality. If the quality is bad, the generation quality is bad. Second, if you're going to be standing and the movement is complete, record your whole body. If you're going to be in the middle, it's not like your hands are moving this way, because this movement here isn't going to work. It must always be visible. If you're going to speak, always speak clearly and clearly so that the AI ​​can understand you, especially if it's in Spanish. Three, the prom controls the background, the aspects that happen, not the movement. Four, video match for fluidity, image match for artistic fidelity. image match for artistic fidelity. Five. Always use the phrase "keep the Five. Always use the phrase "keep the same post and composition" in your prom. It will save you a lot of headaches. Now, with all this information, you Now, with all this information, you can go and create whatever you want. can go and create whatever you want. You set the limit. Remember that you have a free Telegram community for the artificial arc and the link to access it is in the description. Also, if you have a project that you don't quite know how to give it momentum, how to turn it around because you don't know how to approach it, I don't know, I don't know how to create it, visit the Arca page because I have 75-minute one-on-one consultations to help you solve your problem. If you want to learn how to create videos with artificial intelligence using the latest technology on the market, I'll leave this video here and here's my face so you can subscribe directly. This is the artificial ark. See you. Yeah.