Some time ago I disassembled ancient characters AI narrated monologues, mainly for womenHistorical FiguresCharacters. Click back to see it if you didn't see it.
Using AI to make historical legend story narration monologue video, 10 minutes to create explosive video
Historical FiguresnarrativeWe will not go over the story of the explosion of the class of videos. Today we will share another not equally realistic sense of historical figures AI narration, through the inner narration of the characters, with epic and grandiose images, so that people are very interested in the life of the characters.
The whole practical work is only 4 steps: 1. copy generation, 2. text generation, 3. audio generation, 4. video generation
step by step
Step 1: Historical Character Monologue Copy Generation
First open deepseek and enter the following command:
https://chat.deepseek.com/
Instruction prompt word template:
I'm a short video blogger who explains historical figures in the form of monologues, and I'm asking you to help me write xx's first-person monologue text
①Tells the key experiences of his life, digs deeper, and shows the inner world that fits the character's personality.
② The point of view and tone must be consistent with that character's position and personality.
③ The beginning needs to complete the identity contrast within three sentences, actively deconstructing the character's sacred aura and conveying the character's core traits
There should be a summary or sublimation at the end that is in line with the character's temperament, stating his or her creed in life, his or her historical position, or his or her thoughts for future generations.
⑤ Spoken language, about 1,500 words.
Example: Xiang Yu
According to the output, we organize the text for you to see if it is very telling in place.
"They say I'm the 'no two in a thousand' God of War? Heh, but a reckless man who refuses to read, a fool who treats swords as reason, and a defeated general who ...... can't even protect his beloved woman!"
"When my uncle Xiang Liang forced me to study, I threw down the bamboo slips; when he asked me to learn the sword, I broke the bronze. When he scolded me for not being able to write or fight, I yelled back, 'What's the use of memorizing names and surnames? What's the use of chopping up a man? If you want to learn, learn to be the enemy of ten thousand people! 14 But I only learned the art of war in a superficial way -- at that time, I thought that everything under heaven was only broken with one fist and split with one sword!"
I saw Emperor Qin Shi Huang parading around, with his golden carriage and gorgeous cover pressing the people to prostrate themselves. I pointed at him and laughed at Xiangliang: "He can be replaced!" My uncle was so scared that he covered my mouth, but I was burning in my heart: in this world, we should let the power of carrying the tripod turn the world upside down!
Zhang Han's 200,000-strong Qin army surrounded Julu like a barrel, and the lords cowered in their camps as quails. Song Yi drank to make peace. I cut off his head! Sinking the boats in a cauldron? It's not strategy, it's madness! Sinking the ship and smashing the pots, three days of grain and grass, I led the Chu people to attack the Qin army - the sword cut and rolled for the hand, the horse trampled and crushed the mountain of corpses! Nine victories in nine battles, the sound of killing cracked the clouds. The lords kneeled outside the Regulus Gate, not daring to raise their heads.
After this battle, the word "Xiang Yu" has become a living hell. But they do not understand: I smashed the yoke of the tyrannical Qin, but also put a heavier iron shackle on myself - the power can crush the city, but can not pinch the hearts.
That old fool Liu Bang entered Guanzhong first, but he didn't do anything to the people? That's a joke! I massacred Xianyang, killed Ziying, and burned Afang Palace for three months! My strategist advised me to set up a capital in Guanzhong, but I scoffed: "Not returning to my hometown is like walking in the night with a brocade on your back!" How good is Pengcheng? Warm wine from Jiangdong, sweet songs from Wuzhong ......
Fan Zeng pointed at me and scolded: "The vertical son is not enough to conspire with!" At the Hongmen Banquet, I saw Liu Bang cowering and toasting like a dog, and I actually softened my heart....... When Yafu shattered the jade bucket, I realized that: in this battle for the world, those who don't want to live are afraid of those who don't want to face!
On the night when the song of Chu was drilled into my ears, Yu Ji's sword was colder than the song. When she fell into my arms, her blood stained the moon red. I sang the Gaixia Song as the dead souls of 8,000 Jiangdong sons cried out in the wind: "The force that lifts the mountains covers the world, but can it lift the hearts of men? Can it cover the destiny of heaven?"
On the banks of the Wu River, the pavilion chief rocked his boat and called out sharply, "Jiangdong is still king!" I suddenly laughed out loud. Is heaven going to kill me? No! It is I who refused to cross the river to bury my 8,000 sons and daughters in the Central Plains. If I live alone, how can I face my father and mother in Jiangdong? How can I face this body of pride!
"They say I lost to Liu Bang? Bah! I merely refused to win on my knees!
Emperors take turns to sit, but on this earth, there must be someone who stands and dies, to show the world: the spine of a hero can hold up a collapsing sky!"
"Tell future generations: the sword can be broken, blood can be shed, and the flag of the Overlord will never be stained with mud and dust!"
Step 2: Generate a diagram
After generating, we have to continue the dialog to get the AI to generate a cue word for us into the mirror screen image. Enter a description of the cue word here, and the command cue word template is below:
Help me to generate this monologue text into a mirror screen prompt words. Requirements: ① visual tone: epic, realism, both film and television level rendering (light and shadow, heavy, realistic details); solemn, majestic, historical sense of gravity (through color, smoke, light and shadow to strengthen the) ② scene: the ancient city (Qin and Han Dynasty architectural style, eaves, walls), large-scale army (neatly arrayed red flags), the smoke of war / distant mountains, to create a "grand narrative "(iii) The clothing of the characters is in line with the (iii) The clothing of the characters is in line with the characteristics of the dynasty in which they live (iv) The picture is mostly scene-based, showing only the back of the characters.
Here when testing, there is a point where you can give requirements in the birth of the number of subscopes, for example, 30 subscopes, depending on our copywriting, 1500 words in 3 minutes or so, the lens in 20-30 appropriate. Otherwise, the text will appear too long video screen is not enough.
With the cue word, let's go straight to the chart.
Here because the number of pictures is more, we still choose to use the beanbag batch out of the picture, you can also use that the dream of a separate sub-division out of the picture, the core is the pursuit of the effect of the convenience of modification.
Beanbag batch out of the picture / namely dream single out of the picture
Open Beanbag, https://www.doubao.com/, click on Image Generation, copy all the cue words for the subplot, and copy the image sent to Beanbag for generation.
Imagine AI out of the picture
Open Instant Dream AI: https://jimeng.jianying.com/
Click on the image to generate
Click to enter the prompt above
Click on the model to select 3.0
Click size 16:9
Below is the instant dream draw card out of the chart.
All generated after picking the appropriate watermark to download to the local, not suitable must be generated several times.
Step 3: Conch AI Generated Voiceover
Here why say use conch, after using you will know, directly find the counterpart video, download the audio through the cut screen, and then use the conch clone to generate their own timbre, and then put the text into the generation of the whole audio.
Conch International: https://www.minimax.io/audio
In the international version, the first time you log in with your email address, you'll get 1,000 points, which is roughly 10,000 kanji for recording
Click on video, click create yout voice clone, upload the recorded audio, select the voice of the voice, click on clone
Wait a moment we can get after the cloning of the voice, in the click text to speech, for text to speech, enter our text for output on the line.
Currently the domestic version does not support cloning but you can set the character tones: https://www.minimaxi.com/audio
You can also refer to the setting character General Tone
Once cloned, just enter our copy to generate it.
Step 4: Graph Generation Video
Then we took the previous images and generated dynamic videos for all of them. The descriptors are directly using the visual words written by DK.
Open Instant Dream AI: https://jimeng.jianying.com/
Click on the video to generate
Click on input to generate a descriptor
Click on the model to select 3.0
Duration 5~10S
The following is an instant dream demo
Other videos in turn follow this method to generate the completion of the watermark-free download to the local.
This way the images, the audio, all the material is organized and the final edit is done.
All it takes is for us to import all the local images into the cutout. Line up the images and copy in turn, and then add the copy voiceover.
Then align the audio with the picture these, and then do some keyframing of the image
Finally, just add music effects card points and all that as you see fit.
Test points: the text is easy to come up with, just the text of the split-screen map of this piece should be a major watershed, need to consume a lot of time to modify and back and forth pumping the map, more troublesome, the other dubbing and simple comparison is simple.
Practical difficulty: ★★★★★✰✰✰
Tools and methods have the last is also a test of perseverance and creativity, the lack of one can not~
That's what I'm sharing today.Everyone can get hands-on, easy to learn, hurry up and try it~!