Ai I set the table, pick up the spoon and scoop, what are you doing with SUNO.AI? Are you making a lot of songs with SUNO.AI? Music is a great weapon for us. It is possible to make direct profits, but you cannot deny that owning the music you create has a value that cannot be measured in money. If you are concerned about copyright, please once again check the SUNO.AI Terms and Conditions. Please take a look. According to the terms and conditions for paid plan subscribers, all commercial use rights and copyrights belong to the subscriber who created the song. Now, let's start producing a music video that will give wings to our song. This time, the 3D animation musical Raincoated Girl produced by Kashiinai. Did you see it well? It's my first time trying it, so it's a bit clumsy, but if I practice and make it a little more, I think you and I will be able to make it well . Girl in Raincoat was submitted to the Runway International AI Film Festival with a first prize of 15,000 dollars and a first prize of about 20 million won in Korean money. There is still about a month left in the submission period, so let's try for 20 million won. It's not thumbnail aggro, right? I will cover the video I am introducing today in more detail . It is a bit longer than the usual videos, so please subscribe in advance and set all alarms. When you have time, watch slowly and follow along. Brothers and sisters, to make good food, the ingredients must be good. The same goes for music videos. Music videos The ingredients that go into the food must be good. Stories, images, videos, music, special effects, sound effects, etc. become the ingredients. We will trim them with a knife called editing and cook them. Now, let's go make the ingredients. Step 1 Story and image through CHAT GPT and GRAVITY WRITE. Today, I will write lyrics for a 3D animation musical. Musicals should be more interesting and have a story with twists and moving emotions, so I will use GRAVITY WRITE and CHAT GPT . Well, first, let's go to GRAVIA, GRAVITY WRITE . GRAVITY WRITE is an AI tool specialized in content writing and is good for various writing such as blogs, YouTube scripts, game scenarios, etc. Now, first, sign up with your Google ID. Among the various types of writing, we use film. Select the making tool. Enter the genre and style of the movie and the overall time . Enter the basic idea you thought of in advance. If you don't have an idea , you can get a basic story idea through CHAT GPT. It comes out as soon as you click on it. It comes up as soon as you click on it . It even comes with a fingerprint. That's great. I wrote this in English. English is more natural. Of course, Korean is also supported. Now, let's go to CHAT GPT. The reason we go to CHAT GPT is because it allows more modification than Gravity. To CHAT GPT I go and add my thoughts through conversation and adapt the scenario. You can use the scenario created by GRAVITY as is, but if you use more human imagination, a better scenario can come out. Now, a more refined scenario has come out . It's great. Now go back to GRAVITY because GRAVITY creates an image prompt to create a picture that matches the scene . Translate it back to English and go to Gravity's movie production tool . If you click, the AI image creation prompt tool will appear in the middle . Click Genre and Please rewrite the style and paste the scenario refined through CHAT GPT into the video script. Now, an image prompt has been created on the right . This prompt is actually a prompt for Midjourny, an image creation program. Midjourney has a free plan. There isn't one, and it's a bit expensive because it's paid, so I'll change the prompt a little and change it to the Leonardo AI type we'll use today. Please go back to CHAT GPT. I tried to view the image using Dali 3 on CHAT GPT in advance, but it may be because it's popular these days, so the server It's very slow , so I give a command to CHAT GPT to change the Midjourney prompt created by Gravity to a Leonardo prompt. Leo understands commands that are not complicated and intuitive like Midjourney, so he uses the following unnecessary phrases and commands. And the most important thing is the consistency of the characters. In AI stories or movies, the consistency of the characters that appear is the most important. Currently , this part is difficult for Dali 3 or Leonardo AI , and in mid-transition, there is no way to maintain the consistency of the characters. It's better. Since we're writing content instead of expensive mid-game transitions , we have to repeat the phrases that describe the character in the prompt for maximum character consistency. In my case, it's a girl in a red raincoat with short black hair and green eyes. I repeated this part, and when I finished making it, there were descriptions of the types and colors of the high-heeled shoes, shirts , etc., and the characteristics of the fox . So, if you watched the finished video closely, you could see that the styles of the high- heeled shoes and high- heeled shoes were changing. When maintaining image consistency, you must make the character's characteristics specific, clear, and unique, and repeat the prompts so that you do not make the same mistake. Now, the prompts for each scene according to the scenario are complete. Now, here are the materials. What is it? There are song lyrics and songs. So, through CHAT GPT, I ask you to organize the above scenario into song lyrics . It is organized into two types: English and Korean lyrics. For more information about creating lyrics, click the link in the upper right or do it again next time. I will introduce it briefly . INTOR, INSTRUMENTAL, VERSE1 VERSE2, CHORUS, OUTRO, END I asked you to write the lyrics with this structure. Now the lyrics are complete. Now, I will go to SUNO.AI and create a song. This step : Creating a song through SUNO.AI SUNO Let's go to .AI. The song style we will make today is a musical style. First , let us copy and paste the lyrics that CHAT GPT created and modified. If there are any parts that were cut, remember where they were cut. If the lyrics are too long, they can be cut . You can attach it later with a clip attachment. The style is a Disney animation musical song that Kashinai is trying for the first time, and the vocals are written in FEMALE female. The reason for writing them twice is to make the commands more clear because SUNO sometimes stutters . I'm writing this for you. Please write the title. You can't change this style and title each time you add it until the song is completed. Then the program will get confused. Now click on create. Wow, you really make a song in a musical style. SUNO, you're so cool. Now I compare the song with the lyrics and check to see if I'm singing it correctly and if there's anything that sounds too AI- like, and I play a game called "What's my score?" It's really fun. It's like being a producer judge on Wanna One. This candidate advances to the next stage. This guy is eliminated. Try playing producing while doing it . If the singer has difficulty pronouncing something or the lyrics are a bit long, they can be corrected. For the SUNO.AI input prompt, click the link in the upper right again and check it out later. Now, the first two candidates are two without any finalists. I was eliminated. Now, the third and fourth contestants are coming out. While I'm preparing for the contestants, I'm going to modify the lyrics a little more to repeat the ending part a little more fluently . Now, let 's listen to the contestants' songs. Haha. The contestants' skills are not that great these days . It's really difficult to risk judging. Haha, now contestant number 4 is advancing to the next stage . Check where participant number 3's song lyrics were cut off. Sing until Open the World and say, " The lyrics were cut off at the With Usheril part, so rewrite the song lyrics from the With Us Seril part to the end. You can copy it. In the lyrics field, delete all existing lyrics and paste them from the lyrics of the continued words to the end . Oh, and if you click the three dots on the right side of the song of participant number 4 who advanced to the finals, Continue from this clip, the second from the top, will appear . Click this. If you press Continue, the second part of the finals will be created with a chain-shaped icon. Now, we will enter the second stage of the finals screening. The core of this stage of screening is the I CAN HERE part. This is the part where the emotions are the most heightened in the musical story . Advancement to the finals is determined depending on how the singer handles this part. In this way, the song continues up to the third part of the third stage. If you are satisfied with the ending smoothly, click the three dots to the right of the singer who advanced and click Import All Songs. Now for the finals. Well, there are some shortcomings, such as the sound quality, but since SUNO.AI has not yet been updated, this participant is selected as today's grand prize. Clap clap clap. This grand prize winner will be given the honor of singing the theme song of the musical animation. Now, the song is complete. And now let's go to step 3 image creation: Now, Step 3 : Create an image through Leonardo AI Today, not through Midjourney, but through Leonardo AI. I'm going to create an image. My brother Leo from the art school is very generous and gives out 150 credits for free every day. And even though it's free, you can use it commercially as you like. However , images that are public and published on the website cannot be used commercially. The website is public. You can view the image only for reference purposes. Leo can only use images that he has created for free or for a fee. Furthermore, every time you change your Google ID and log in , 150 credits of coins are generated, so is this really a piece of cake ? Before creating an image in earnest, refer to the images created by existing Leo users and note the prompt commands of the style you like. Although Leo is intuitive, there are many different combinations, so it is best to look at the created images and refer to the image style you want . Now, press image generation and set the default values. There are various combinations such as photoreal Alchemy style. This photoreal and Alchemy style are not available in the free version, but only the paid version. It is more like a realistic photo like Mid Journey. It is a function that allows you to create images . Number of images refers to how many images are shown each time you create one. If you do more than two from the beginning, the coins will be consumed too quickly, so start with one image and wait until you get the picture style you want . We'll do a test and set the screen ratio to 16:9, which matches the movie size. And if you click the style in the upper left corner, various styles will appear. Also, since each style has its own picture characteristics, you can change these and press them to find your own style. It may be a bit complicated, but the biggest advantage of styling like this is that you can maintain the character's tone consistency. And since it's a NEGATIVE prompt, it tells you to exclude things like morphing effects and six fingers as much as possible. And now, 3D. IMAGE STYLE I created an image by combining two 3D LANDER STYLE, but I don't like it a little. So, you have to keep practicing with other free Google IDs until you get the style you want . Now I finally got the style I want . Albado Base Style and Ray Traced Style. I created an image by combining the two. In order to ensure consistency of the most important character , the character's characteristics, that is, a girl with short black hair and green eyes wearing a red raincoat, are repeatedly entered in the prompt for each command . Consistency issues are a key issue for all image creation AI tools. Midjourny is ahead in maintaining the most consistency, and Dali Oppa and Leo Oppa, which we use today , are also following suit. Nevertheless, the reason why Leonardo AI is used today is This means that it is possible to create an image-to- video video. I think Leo Art University's older brother is much better than Pica Labs, the leader in existing video creation tools, in terms of image quality and moving expressions. That's why I recommend Leonardo AI. Now , the image created in advance in step 1. As a prompt, images suitable for each scenario scene are completed. If you click Image to Motion at the bottom of the created image, a video is created. If the image creation intensity is too strong, a morphing effect or a flaming face may occur, so set the number to 1 or 2 up to 3. I recommend choosing only the intensity. Now, here's the video . The face is distorted in the back , but Leo provides a total of 4 seconds of video, so it's enough to cut the back part in editing and only increase the front part. To save coins, in the case of character depiction, it's enough. Be sure to use only the intensity level of 1 or up to 3. In this way, the video for all images is completed. It is better to add insufficient close-ups or insert cuts while editing . Record narration through 4-step type cast. Search for Typecast and sign up with your Google ID. Please be sure to review the copyright issues before using all AI tools. Typecast can be used for free, but for commercial use, you are required to copy the sentence and indicate the source like this . So, I copied and pasted it into the description box of the final musical movie I made. Type Cast even makes a video, but since we only use audio narration, we choose a voice actor that fits our musical image among several voice actors. Once opener time, which starts the narration, is the most important. Because of this, we select a voice actor that feels familiar to us . I like the voice actor named Icarus the most. Select this voice actor and enter the story in the dialogue field. And when you listen to it in advance, you really like the story. Now, in the middle of the story, the main character's voice is heard. Add a character to change it. Select the gender and age range. Create a separate input window in the part where the voice changes. I chose the last name Millie. Once all the last names have been selected in this way , I make small changes to the overall feel of the voice. The nuance varies depending on things like commas, periods , and exclamation points, so you can change it to find the feeling you want. When everything is complete, download and save it. Lastly, don't forget to indicate the source . Now, all the materials to make a musical are almost ready. Story Lyrics , song, image, video, narration. Now, we will go to Capcut and start cooking in earnest. Cooking requires seasoning. We will make seasonings for effects such as screen transitions, special effects, and sound effects through Capcut Chef. Now go to Step 5 Capcut. Before I take a moment , no, I'm not talking about Capcut . Go to the Capcut download address linked in the 5th step Capcut video description and download Capcut. There are some people who say they can't see this Auto Lyrics, so I'll have to call Capcut's head office in Zhongguo. Now, we will put all the necessary materials on the cutting board called import . We will import all the videos we have made. We will also import the musical song we created. We will also import the narration. Now, all the materials for basic and editing have been placed on the cutting board. Now, music. Drag and place the and narration on the music layer at the bottom . We will arrange the video appropriately to match the music and narration and adjust the length for temporary editing. The detailed explanation on how to use Capcut can be a YouTube channel in itself, so I will omit the detailed explanation. Now, continue watching the video while listening to music. Don't add the effect or seasoning for the screen change yet. Just cut the ingredients step by step and add them to the music called the recipe. For the knife for cutting and inserting, press the shortcut key "B" to create a knife, and the chopsticks for moving ingredients are the shortcut key "A", so just use these two. You can use it to make food. Shortcuts can be set to suit your needs through the shortcut menu at the top right. Usually , there is the alphabet qwe at the top left of the keyboard, so you can set it as a shortcut to work quickly. Once all the ingredients are in, check again for the first time to make sure nothing is missing. Editing is done according to the music. If the BGM musical song and the voice actor's narration overlap, the volume is adjusted by setting keyframes for the audio volume. This part will also be explained later in the Capcut special lecture. Now, the editing window is becoming more and more complicated. Editing is indeed. It is an ancient work , but if you do it step by step and see the finished product, the hardship will disappear. Give it a try. The video work we are learning today is a work that can be applied to all content. The video can be applied to making videos for meditation music channels, sleep music channels, etc. It can be applied to the creation of various contents such as children's animation channels, so please try it slowly several times to create your own. If you want to make it really easy , there are many AI tools that automatically create videos, so you can use those. However, it does not feel like instant food . It cannot be erased, and it is questionable whether it will be selected by the YouTube algorithm. Now that the basic editing is done, we will now add seasoning such as screen transition techniques such as dissolves and fade-outs, audio adjustments, etc. The biggest key to seasoning is the keyframe. If you press the dot on the right side of the edit point, the color changes to white and a keyframe is created. For example, if you want to set the starting point of the audio to minus infinity, take a keyframe point, and then upload the audio again at the desired point. A key frame point is placed at that point. Then, the audio volume gradually increases or decreases according to this key frame command. This key frame is the core of video editing, and not only the audio, but also all changing movements, such as fade in and out animations that change the screen size. It is the core of , so make sure you understand it. And the advantage of capcut is that it is very effective. It is best to try the mask function , chroma key function, tracking, etc. The editing I am showing you today is a bit of a difficult editing, so it may seem more complicated, but Next time, I'll let you know with an easier video . Ha ha ha ha. There are various audio effects as well as video effects, so audio effects can be added in between. For audio, you can add speed control or echo effects separately. It can be applied in a variety of ways Now that the temporary editing is over , let's go to the next step to create insert cuts and effects that are missing in the video. Step 6 Runway ML's Motion Brush Now let's go to Runway ML. Please sign up with Google and come in. Some coins will be generated for free. You can use it for free, but the free version has a logo mark at the bottom right and the image quality is poor, so we recommend that you practice for free and then sign up for a paid service only if you need it. The biggest advantage of Runway is the motion brush. You can give movement. Recently, the runway has been upgraded, so you can use up to five motion brushes. In other words, it is possible to provide each of five types of animation. It is certainly a very groundbreaking feature. However, there is still a problem with the morphing effect, which is a limitation of the video creation tool, that is, the face is distorted more than necessary , so the character expressions that Kashiinai is working on today are limited. If you increase the intensity, it will be distorted a lot, so only apply effects such as flames and hair and water with a motion brush to create movement. The motion brush function is best used on abstract photos, not people, because distortion or morphing effects are allowed because they are abstract. This is because when you use it on a character, a strange shape appears. I think this will also develop further. I think it would be good to create a fix brush and not move that part at all . In this way , the additional insert cut video for the comprehensive version was completed. For reference, as briefly explained earlier, Runway ML will be holding the 2024 AI Film Festival for a long time, following last year. The first prize is 20 million won, and various prizes and benefits are offered up to fifth place. I'm preparing, so I hope you all watch my lecture today and follow along to take on the challenge. Looking at last year's 2023 winners, only about 20 percent used AI. Today , we are 100 percent AI, so I think there's a good chance of winning if you put in a little effort. If you challenge yourself with an abstract image, it will be easier to create a video and go well with the film festival image. Give it a try . You will know that the 20 million won thumbnail is not an aggro . Now that additional inserts are ready, go to Capcut and make comprehensive edits. I'll end today's lecture with this. My neck hurts from talking for too long. AI is not omnipotent, but it is definitely a tool that makes your life very convenient. 2024 will be the year when AI becomes more active. Any questions you may have while watching the video? If you have any questions, please leave a comment and I will answer as long as I'm awake. Hey, brother, did you click subscribe? Subscription is loyalty and popularity, so see you again in the next video. Turn on all notification settings.
