MANY PEOPLE HAVE THE SAME DOUBTS WHEN THEY FIRST GENERATE AN AI IMAGE:
Why does my picture look natural? The subject is not focused, the space is flat and the whole body lacks a sense of truth。
That's because you have a problem with your design
MANY PEOPLE HAVE THE SAME DOUBTS WHEN THEY FIRST GENERATE AN AI IMAGE:
Why does my picture look natural
The subject is not focused, the space is flat and the whole body lacks a sense of truth。
Too often the problem is not that the hints are too few, nor are the models too strong, but at the most basic point — the structure。
The image determines the visual logic of the image:
The first sight of the audience, the way the sight moves, the whole picture is stable。
AI WILL TRY "AUTONOMOUS DRAWINGS" WHEN CREATING IMAGES, BUT IF YOU DO NOT GIVE CLEAR INSTRUCTIONS IN THE HINT, AI TENDS TO CHOOSE THE SAFEST APPROACH - TO PUT THE SUBJECT IN THE MIDDLE, TO GIVE LIGHT EVEN, TO BALANCE THE SCENE WITHOUT FOCUS。
The result is that the image is "ordered but weak" and it looks like it's not a picture。
AT THE HEART OF THE IMAGE CONTROL, IT'S ACTUALLY TO MAKE AI UNDERSTAND, "WHAT KIND OF VISUAL STORY DO YOU WANT IT TO TELL?"。
This lesson gives you three basic skills that help you directly control the structure logic in the hint。
Skills one: Control of the main position - determines the visual focus
1. Rationale
The “subject location” of the image determines the first sight of the audience。
AI DEFAULTS TO USE A CENTRAL IMAGE, I.E. TO PUT A PERSON OR OBJECT IN THE PICTURE。
It's the safest way, but it's also the easiest way to look stupid。
The more common way of photography and painting is "Rule of Thirds"
The picture is divided horizontally and vertically into three equals, forming the Nine Palaces。
The four intersections are where human eyes are most naturally focused。
When the subject falls in one of these areas, the picture is balanced and respiratory。
2. Operational applications
An example:
a girl standing in a field, finematic lighting
THIS TYPE OF HINT AI WILL BE CREATED BY DEFAULT, WITH THE RESULT THAT THE PERSON IS IN THE MIDDLE, WITH A BALANCED BACKGROUND AND NO SENSE OF DIRECTION。
We'll read:
a girl standing in a field, posted at left one-third of the conflict, soft light from right
The subject is thus left, the light comes from the right, and the person and environment are connected. The image immediately had a sense of space and direction。
3. Practical formulation
You can add a position description directly to the hint, for example:
Scene Needs Example
subject identified at left one-third of the frame
subject at right one-third of the frame
subject placed at top one-third of the frame
subject identified near
WHEN YOU CLEARLY TELL THE AI SUBJECT'S IMAGE LOCATION IN A HINT, IT'S LIKE A PHOTOGRAPHER'S "IMAGE FOCUS."。
Space breathing can occur only when the main subject is deviant and light corresponds。
Remember: the main thing is deviating, so the picture is deep。
Skills two: Control the visual motion line -- direct the sight
1. Rationale
The motion line determines how the viewer moves in the image
In short, it's where the audience is "seeing" and being directed to
AI DOES NOT UNDERSTAND THE ABSTRACT CONCEPT OF "MOTION LINE", BUT IT UNDERSTANDS SPECIFIC ELEMENTS SUCH AS "ROAD, LIGHT, SIGHT, RAILING, SHADOW"
WHEN YOU CLEARLY WRITE THESE GUIDANCE STRUCTURES IN A HINT, AI AUTOMATICALLY CREATES A DIRECTIONAL, DEEP IMAGE
2. Operational applications
Example of error:
a man walking on the street, cinematic lighting
AI WOULD PUT PEOPLE IN THE MIDDLE, LAY THE BACKGROUND, NO DIRECTION。
Improvements:
a man walking on the street, raised at right one-third of the conflict, leaving lines from bottom-loft toward the subject, sunlight diagnon from top-right
At that point, the viewers will follow the road towards the characters and be drawn to the subject by light. The images naturally form a visual flow, with a hierarchy of information。
3. Common aerodynamics
Type of hint description Effect
line lines from foreground towards the subject leading to the image
diagonal light from top-right firing viewer's eye
perceived motion lines
subject looking toward distant light
The action line is not a decorative tool, but a tool to guide the sight。
WHEN YOU WRITE THE LIGHT DIRECTION, THE ROAD EXTENSION, OR THE LOOK OF A PERSON, AI CAN CREATE A "SPACE PATH."。
The image is thus more solid, and the eyes of the audience are "walking in the painting."。
Skills III: Controlling the image balance - stabilizing the overall vision
1. Rationale
The balance of the image is not symmetry, but the distribution of the visual weight
If one side of the picture is too bright or too heavy, the other side of space will “slash”。
A "asymmetric balance" is used in professional photography to offset each other by color, brightness, volume or detail
The famous painter Van Gogh's painting is an expression of asymmetric balance
AI OFTEN IGNORES THIS WHEN GENERATING IMAGES, LEADING TO THE SUBJECT HAVING A BACKGROUND EMPTY
You can tell it how to balance the visual weight on both sides with a hint
2. Operational applications
Example of error:
a woman sitting by the window, chinese light
People are usually on the right side, left is pure blank, and the picture is light。
Improvements:
a woman sitting by the window at the right one-third
At this point, the curtains and shadows become a combination。
The shadows of the light make the picture stable, but asymmetric, and it looks more natural
Common balance control writing
Usage, hints, effects
balanced balance through contrast between light and shadow
balanced by small object on the object side
asymmetric balance through warm and cool tones
balanced by structural balances in context
Balance is the key to "stability."
WHEN YOU WEIGH THROUGH LIGHT, COLOR OR STRUCTURE, AI UNDERSTANDS SPACE RELATIONS
The image is no longer floating, nor is it biased
A complete hint structure template
The following logic can be used to control the metaphors of the diagram:
[subject], raised at [left/right/top/bottom] one-third of the conflict, drawing lines/light direction]
Example:
a man in a race walking at night, posted at left one-third, road reflectings leaving him, bled by bright street on opposite side, cinematic lighting, wet attack
THROUGH THIS STRUCTURE, AI UNDERSTANDS AT THE SAME TIME:
Who's the subject, where the light comes from, where the sight goes, how the picture balances。
Core logic of image control
Subject position determines the focus
Put the subject in a three-point position, make the picture natural and space
Visual motion lines determine the rhythm
The light, the road, the eyes guide the sight, the layers
The balance is stable
Weighted with light, colour or object to stabilize the picture
WHEN THESE THREE ARE COMBINED, THE IMAGE CREATED BY AI HAS THE LOGIC OF A PHOTOGRAPHER'S MIND
THE IMAGE IS THE MOST EASILY IGNORED PART OF THE AI IMAGE, BUT IT DETERMINES WHETHER THE WORK IS LIKE A WORK
WHEN YOU CLEARLY EXPRESS VISUAL INTENT IN A HINT, AI CREATES A LOGICAL, SPATIAL, EMOTIONAL IMAGE
IF YOU'RE JUST STACKING STYLE AND MATERIAL, AI IS JUST DRAWING MATERIAL
But when you control the picture, it starts to understand the picture
I hope you've seen some success