IT'S NOT THE VIDEO GENERATION CHAIN THAT OFTEN AFFECTS EFFICIENCY AND QUALITY IN THE DAILY USE OF AI VIDEOS, IT'S NOT EVEN THE TOOLS WE USE, IT'S WHAT WE WRITEPrompt word.
Whether Midjourney, SD, or Dream, DALLE... can't stop the hint。
FOR THE CREATORS OF AIGC, THE IMPORTANCE OF THE NARRATIVE IS SELF-EVIDENT。
I will present them in two separate articles:
- OneAN ALL-EMBRACING FORMULA FOR THE MAIN AI MAPPING TOOL
- oneA field-based training method for fast-tracking tips
In this article, we focus first on the All-Assisting Word formula。
Phrasing - Summary of formulae
There's a lot about itAI PaintingThe formulas, but most of them are too rigid, have to remember their backs and, in practice, have to be applied in a rigid manner and cannot form their own understanding. So I put together a more comprehensible formula。
Summarizing: narrative elements + lens language + color set, i.e
1. What is the image: Subject + Environment (background) + Details (subject details + environment details)
2. How do you do it
3. How to tune the image: Hue (colour + light) + parameters (size + mass)
With that understanding, I feel that it is easier for us to create the picture we want one step at a time without a formula. Here we share the details of the hint。
Phrasing — narrative elements
What's the picture, the subject and the background. This constitutes the main element of a picture, and by describing it in detail we can make a more controlled and rich picture。
I'll show this by example。
For example: a lady, in the living room. This constitutes the underlying narrative element of a picture。

enter the hint above into the midjourney where you get the above four pictures。
And we found that with these two elements, we got a different picture than we wanted。
AND THESE ARE THE DIFFERENT THINGS THAT WE DON'T HAVE IN THE FORMULA, SUCH AS PAINTING, CHARACTER DETAILS, ANGLES, ETC. WHAT WE DON'T DESCRIBE, AI, WILL BE A RANDOM MATCH FOR US, FREE TO PLAY。
BUT THIS FREE PLAY IS NOT USELESS, AND SOMETIMES WE CAN LET AI BE FREE TO PLAY, AND THEN WE CAN ADD FROM IT THE ELEMENTS THAT WE LIKE TO SELECT TO THE NEXT HINT。
FOR EXAMPLE, I FEEL LIKE I WANT THE LIGHT OF THE FOURTH PICTURE AND THE PERSPECTIVE OF THE SECOND, SO I CAN ADD A CORRESPONDING DESCRIPTION TO THE NEXT HINT, AND BY USING THIS KIND OF THOUGHT IN AI TO FILL THE GAP IN OUR IMAGINATION。
I now fill out the details for this image:
A woman in her twenties with a light and pink make-up, with long hair on her shoulder with a microfilm, in a light-coloured dress, sitting on the sofa, with two legs wrinkled, with a book in her hand. The living room is bright and the sun is pouring through the window. The tea table contains a bunch of white flowers and a cup of coffee。
Here's what we're looking at

It can be seen that by describing the person and the environment in detail, we basically control the person and the scene. We've identified the subject, and we're going to ask the second question how to shoot。
Phrasing — lens language
The image style is the overall picture, and the effect of the different styles is very different。
The most common are written, sketches, Disneypix, and other artist styles such as Miyazaki, Wang Wei, Van Gogh, Monay, etc。
Let's show you how different styles work
3.1 Factual style

3.2 Synopsis

3.3 Disnipix style

Wang Jia Wai style

3.5 Miyazaki style

It's a picture, a view and a perspective。
MORE OFTEN THAN NOT, THE SPECS AND PERSPECTIVES NEED TO BE EMPHASIZED IN THE HINTS THAT AI ITSELF OFTEN PRODUCES A SUITABLE GRAPH APPROACH, UNLESS I PARTICULARLY WANT A GRAPH FORM, WHICH WAS SPECIFICALLY MENTIONED IN THE LAST TIME。
View: Vision, Panorama, Midscape, Periphery, close-up
The perspectives are generally: front, side, back, angle, over。
Through their combination, different images and expressions。
Vision: Usually used to show the environment, with no person or person in a small position

Panorama: display of complete character and environment

Midview: Half body of the character

Proximity: Highlighting emotions, strengthening details, appropriate preservation of the environment

Special feature: Turnover, acoustic, emotional, strengthening details

It is not the narrative of the subject and the background, but the part of the video (how to film) from Part II that tends to leave the gap。
THE FLEXIBLE APPLICATION OF THESE COMBINATIONS HAS ALSO MADE VISUAL THINKING A MANDATORY SUBJECT FOR AI DRAWINGS。
Finally came to the third part of the introduction。
Phrasing - how to be transposed
ONE OF THE ADVANTAGES OF AI PAINTING IS THAT IT CAN BE ADJUSTED LESS LATER THAN IT IS DONE. REGARDLESS OF THE SIZE OF THE IMAGE, THE COLOUR LIGHT CAN BE PUT IN PLACE DIRECTLY AT THE LINE. IT'S NOT FREE FROM SCENES AND DEVICES。
Here's the main thing about color and light。
Colours are primarily the same colours that need to be consistent with the same scene when we create successive animations or lenses. And the light is mainly cold and warm. And it's softer, softer, and it's hard-lighted。
Soft light, hard light, contrast


Cold, warm, contrast


I LEARNED ABOUT THE AI PAINTING TIPS, FOLLOWED BY TRAINING IN THE ABILITY TO WRITE THEM THROUGH CONTINUOUS PRACTICE。