{"id":57559,"date":"2026-09-28T10:37:11","date_gmt":"2026-09-28T02:37:11","guid":{"rendered":"https:\/\/www.1ai.net\/?p=57559"},"modified":"2026-09-20T14:47:22","modified_gmt":"2026-09-20T06:47:22","slug":"ai%e6%8f%90%e7%a4%ba%e8%af%8d%e5%88%9b%e4%bd%9c%e7%ac%ac%e4%ba%8c%e5%8d%81%e5%85%ad%e8%8a%82%ef%bc%9a%e5%91%8a%e5%88%abai%e5%91%b3%ef%bc%8c%e7%94%a8%e4%b8%89","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/57559.html","title":{"rendered":"SECTION 26 OF THE AI TIP: SAY GOODBYE TO \"AI TASTE\" AND RECREATE THE IMAGE WITH THREE ARCHITECTURAL PRINCIPLES"},"content":{"rendered":"

And today, 2026, when we talk about Ai drawings and video generation, if you're still in the middle of a \"breathing hand\" or \"unnatural light,\" you're probably out of line. With the launch of the Nano Banana Pro, the maturity of Midjourney, and the eruption of video-generation models like DreamAi, Coven, Seach, etc., AI has come close to perfection on the \"character quality\"\u3002<\/p>\n

BUT WHY DO A LOT OF PEOPLE STILL SMELL LIKE \"AI\" AT ONCE<\/p>\n

The answer lies not in the luxuriousness of the hint, but in the absence of a \u201cconstructive consciousness\u201d\u3002<\/p>\n

MANY CREATORS, WHEN USING AI, TEND TO FOCUS ON \u201cWHAT TO PAINT\u201d (THE SUBJECT), WHILE IGNORING \u201cHOW TO PAINT\u201d (FORM). WHEN YOU DO NOT SPECIFY THE STRUCTURE, THE AI MODEL TENDS TO PRODUCE THE MOST FLAT AND STABLE \"BIG-TOP\" CONTOURS FOR INSURANCE PURPOSES. IT'S NOT WRONG, BUT IT'S MEDIOCRE, EMOTIONAL, BREATHLESS\u3002<\/p>\n

IN THIS SECTION OF THE CURRICULUM, WE WILL MOVE BACK TO THE NATURE OF VISUAL ART, LEAVING ASIDE THE CUMBERSOME PARAMETERS. USING THE HOTTEST TOOLS OF THE MOMENT, WE WILL ADOPT THREE BASIC MAPPING PRINCIPLES TO TEACH YOU HOW TO EVOLVE THE WORK OF AI FROM \u201cGENERATED PICTURES\u201d TO \u201cDESIGN PHOTOGRAPHS\u201d USING THE PHRASE \u201cMOBILIZATION\u201d IMAGES\u3002<\/p>\n

Chapter I: Breaking the deadlock - Prioritize control of the main position (triple approach and breathing sense)<\/p>\n

1.1 AI'S \u201cMIDDLE TRAP\u201d<\/strong><\/p>\n

Whether Midjourney or Dream AI, when you enter \"A pretty girl in the rain\", 90%, AI puts the girl in the center of the picture and the background is deformed\u3002<\/p>\n

\"eb53c8e9j00tlg7rx01ezd000v90nbp\"<\/p>\n

The biggest problem with this \u201cdocumentary picture\u201d is the \u201cobstruction of sight\u201d. The eyes of the audience are directly at the centre, and it ends at a glance, and it is impossible to move in the picture and naturally not to create a story. That's why a lotAI Video<\/a>LOOKS LIKE A PPT BECAUSE OF THE LACK OF SPACE ON THE MOVEMENT\u3002
\n
\n1.2 Triple (rule of Thirds): The golden rule that makes the picture \u201clive\u201d<\/strong><\/p>\n

Theoretical basis: The image is divided into three equals, forming the \"well\" type. Place the main subject on four lines or four cross-points (i.e. one third or two thirds of the picture)\u3002<\/p>\n

\"eef172faj00tlg7s400kyd000u0013wp\"<\/p>\n

Visual psychology: When the subject deviates from the centre, the negative space is left on the side of the picture. This space not only gives the image a \u201crespiratory sense\u201d, but, more importantly, it implies the relationship between the subject and the environment, leading the audience to imagine the subject \u201cwhere to look\u201d or \u201cwhere to go\u201d\u3002<\/p>\n

\"4f425d83j00tlg7s9000fod000v9014ap\"<\/p>\n

The work of Elena Georgiou's cameraman, the fine use of \"negative space\"<\/p>\n

1.3 Operational operations: How to land in a hint<\/strong><\/p>\n

When using the Nano Banana Pro or Midjourney, simply entering \u201crule of trirds\u201d is sometimes not enough, and we need a more specific position description\u3002<\/p>\n

Common formulae: [subject description] + [location\/construction] + [environment\/negative spatial description]<\/p>\n

CASE A: VERY SIMPLE HUMAN PHOTOGRAPHY<\/p>\n

We're here to create a picture of a movie-sensitive caf\u00e9\u3002<\/p>\n

Misdemonstration:<\/p>\n

A portrait shoots of a woman drining cake in a Cafe, Sonny day, beautiful light.
\ncopy<\/p>\n

OUTCOME: FACE-TO-FACE, CENTER-TO-CENTER WITH A DEFUNCT CAF\u00c9, LIKE A NORMAL TOURIST PHOTO<\/p>\n

\"9dbcaef6j00tlg7si0177ad000v9000hfp\"<\/p>\n

Optimization of the hint (three-part):<\/p>\n

A socialistic woman works in a sleek, safe on the right verticical land<\/p>\n

(AI RESULT: THE PERSON SITS ON THE THREE-POINT LINE ON THE RIGHT SIDE OF THE PICTURE, LOOKING AT THE STREET OUTSIDE THE LARGE WHITE WINDOW ON THE LEFT. THE VIEWER WONDERS WHAT SHE'S LOOKING AT<\/p>\n

\"9cc6b01ej00tlg7st00zrd000v90hfp\"<\/p>\n

Tool-specific skills<\/p>\n

Nano Banana Pro: It understands very well the language of nature. You can describe it directly in a conversation type: \"Put the figure on the lower right corner of the picture and leave the top left corner white for the sky.\"<\/p>\n

\"b111031a2j00tlg7t00087d000pg00o6p\"<\/p>\n

DREAM AI\/CLIN (Video-generated): It is important to use a three-dimensional schematic when generating a video headline. For example, a description: \u201cSubject authorized at the left third\u201d. In this way, AI is more likely to generate mirrors (e.g., to the right, Pan Right) based on the white space, using space spares to create dynamic\u3002<\/p>\n

Chapter 2: Rejection of dissipation - Control of visual motion lines (lead and rhythm)<\/p>\n

2.1 What's a visual motion line<\/strong><\/p>\n

If the main position determines where the audience first looks at it, then the visual motion line determines where the audience looks at it second and third\u3002<\/p>\n

\"dec6caa8j00tlg7t70ead000v9000abp\"<\/p>\n

The Master's photography<\/p>\n

WHEN AI PRODUCES BIG SCENES (E.G. CITY, LANDSCAPE) IT IS PRONE TO \u201cAVERAGE DISTRIBUTION\u201d PROBLEMS: IT'S FULL OF DETAILS, HOUSES, TREES, PEOPLE, NO FOCUS. THE EYES OF THE AUDIENCE DON'T KNOW WHERE TO PUT IT, CAUSING VISUAL FATIGUE\u3002<\/p>\n

2.2 Construction of \"Invisible Road Signs\" in hints<\/strong><\/p>\n

We need to be inPrompt<\/a>, or directly describe the direction of the view\u3002<\/p>\n

Core keyword:<\/p>\n

Lead Lines
\nVanishing point
\nPerspective
\nDeep of field
\nVISUAL Flow
\n
\n2.3 Operational: Building film-level depths<\/strong><\/p>\n

CASE B: ARCHITECTURE AND SPACE<\/p>\n

Scene: You want to create a future museum\u3002<\/p>\n

General description:<\/p>\n

A futuistum interior, white walls, absolute art.<\/p>\n

\"bf963d93j00tlg7tg0121d000v9000hfp\"<\/p>\n

Actionline optimized description:<\/p>\n

Wide forwards lined by the passing lights guiding the eye to the center, a mussum, one point perceptive, deep down of the field, multi-moralist relationship.<\/p>\n

Translator: A wide-angle shot, with a strong lead line from the top light leading to the centre of view, a museum with a sense of the future, a little perspective, a deep view, a very short picture\u3002<\/p>\n

\"04ad925ej00tlg7tq011ld000v90hfp\"<\/p>\n

[WASTE CASE C: DYNAMIC VIDEO GENERATION (SCH\/COVEN\/DREAM]]<\/p>\n

IN VIDEO GENERATION, THE MOTION LINE DETERMINES THE LOGIC OF THE MIRROR. IF YOU WANT THE VIDEO TO HAVE AN IMPACT, YOU HAVE TO DESCRIBE THE Z-AXIS MOVEMENT\u3002<\/p>\n

Phrasing policy:<\/p>\n

Step 1: Ventura Tip (for Nano Banana Pro \/ Midjourney)<\/p>\n

Target: Generate a static picture that looks like it's about to rush out, with a line pointing to a distance\u3002<\/p>\n

Low angle policy making a race car on an asphalt track, protective view. I'm not sure<\/p>\n

\"1efb2079j00tlg7tx014cd000v9000hfp\"<\/p>\n

Step two: Tusheng video alert (for respiration \/ conch)<\/p>\n

Target: Activate the motion lines in the static frame\u3002<\/p>\n

The car comes from behind at high level.<\/p>\n

Chapter III: More than symmetry - control of the image balance (allocation of visual weights)<\/p>\n

3.1 Balance symmetry<\/strong><\/p>\n

When it comes to balance, the first reaction of many is the extreme symmetry of Wes Anderson. Symmetry is indeed a beauty and represents order, dignity and a sense of labour. But beyond that, the higher is the asymmetric balance\u3002<\/p>\n

\"fae70c1aj00tlg7ua01m6d000u00140p\"<\/p>\n

Wes Anderson<\/p>\n

IN AI DRAWINGS, IF THE SUBJECT IS ON THE LEFT, AND THE RIGHT IS EMPTY, THE IMAGE WILL LOOK \"LEFT HEAVY, RIGHT LIGHT\" AND IT WILL BE PAINFUL. WE NEED TO \u201cREFILL\u201d THE WEIGHT ON THE RIGHT BY COLOR, BRIGHTNESS OR BODY SIZE\u3002<\/p>\n

\"44245042j00tlg7ug00ljd000v90dtp\"<\/p>\n

The film's screenshot<\/p>\n

3.2 Distribution of three dimensions of visual weight<\/strong><\/p>\n

Brightness balance: The brightness is heavier than the darkness\u3002<\/p>\n

Color balance: Warm\/high saturation is heavier than cold\/low saturation\u3002<\/p>\n

Balance of volume: Large objects weigh more than small objects\u3002<\/p>\n

\"277c9a90jlg7um00t8d000v90hgp\"<\/p>\n

3.3 Operational operations: building balance with light<\/strong><\/p>\n

[Waste Case D: Commercial Product Map]<\/p>\n

Suppose you make a poster for a black perfume. The subject is black, and if on the left side of the picture, the background is completely black, it is invisible; the background is completely white and it is too sudden\u3002<\/p>\n

Constructive policy: The subject is left (in the dark) and a bright object or light source is needed on the right to balance\u3002<\/p>\n

Phrasing:<\/p>\n

On the right side, a soft warm of light shooting a white silver column between the dark subject and light backstage, chiaroscuro lighting, Luxurious feeling.<\/p>\n

\"Light vs. Dark\" and \"Left vs. Right\" to force AI to think about the right and left weight of the picture instead of simply putting the bottle in the middle.\"<\/p>\n

\"2825a4edj00tlg7uv00txd000v9000hgp\"<\/p>\n

Chapter IV: Use of the \u2014 construction of the phrases \u201ccore thesaurus\u201d and \u201ca universal formula\u201d<\/strong><\/p>\n

In order to get you to use these techniques more quickly in your daily creation, we have compiled a generic diagram-character for tools such as Nano Banana Pro \/ Midjourney \/ DreamAi. It is suggested that the contents of this chapter should be collected as your dictionaries\u3002<\/p>\n

4.1 Thesaurus of the three principles (Cheat Sheet)<\/strong><\/p>\n

A. RELOCATION OF THE MAIN BODY (BREAKING OF THE CENTRAL MAP)<\/strong><\/p>\n

Rule of trides
\nOff-center relationship (deviolation of the centre)
\nSubject on the left\/right side
\nNegative space (negative\/painted \u2013 minimum style)
\nWide shot (widen angle \u2013 there's space only if the picture is wide enough)<\/p>\n

ESTABLISHMENT OF VISUAL GUIDANCE (REFUSAL TO DISPERSE)<\/strong><\/p>\n

Leating lines
\nOne-point perceptive
\nVanishing point
\nDecth of Field (Deep - Fake background highlights)
\nLow angle \/ High angle<\/p>\n

C. CONTROLLING THE IMAGE BALANCE (UPGRADED)
\n<\/strong>
\nSymmetric balance
\nAsymmetrical barance
\nVision weight
\nContractor between light and dark<\/p>\n

4.2 Portfolio-capable formulae<\/strong><\/p>\n

IF YOU DON'T KNOW HOW TO WRITE, APPLY THIS FORMULA, WHICH SOLVES THE PROBLEM OF 90%:<\/p>\n

[Construction mode] + [subject description and location] + [environment and guidance line] + [photo\/style\/balance element]<\/p>\n

Examples of formula application:<\/p>\n

Wide shot, Rule of trirds relationship
\n[major] a lonely astronaut standing on the far left
\n[environment] looking at a massive planet on the right, leaving lines of the calter firing eyes to the horizon
\ncinematic lighting, balanced visual weight<\/p>\n

4.3 Special tips for video creators<\/strong><\/p>\n

FOR VIDEO TOOLS SUCH AS DREAMAI \/ CLIN \/ SEACH, IT IS ESSENTIAL TO INSERT \"DYNAMICALLY CONSTRUCTED\" IN THE HINT:<\/p>\n

Dolly In\/Out: Go along the line with the leating lines\u3002
\nPan Left\/Right: Co-opted with Rube of trirds, moved from place to subject or from subject to substance\u3002
\nFollowing (Tracking Shot): used in conjunction with Off-center, keep the subject moving in a non-centre position and increase the film feeling\u3002<\/p>\n

CONCLUSION: The graph is the \"attestation\" you left behind<\/strong><\/p>\n

When the pixels generated by AI become more and more authentic, the real difference between the \"generator\" and the \"generator\" is no longer whether you write \"8k, high Quality,\" but whether you have any director thinking\u3002<\/p>\n

The triangulation reflects your respect for the viewers<\/p>\n

The visual motion line reflects your control of the narrative rhythm<\/p>\n

The balance of images reflects your understanding of light aesthetics\u3002<\/p>\n

Also using Nano Banana Pro or Midjourney, when you begin to control these three points, you are no longer a user fed by algorithms, but an artist who controls them\u3002<\/p>\n

It's not just for the picture to look good, it's for your eyes and your judgment to be seen on the screen. This is your unique creation\u3002<\/p>","protected":false},"excerpt":{"rendered":"

And today, 2026, when we talk about Ai drawings and video generation, if you're still in the middle of a \"breathing hand\" or \"unnatural light,\" you're probably out of line. With the launch of the Nano Banana Pro, the maturity of Midjourney, and the eruption of video-generation models like DreamAi, Coven, Seach, etc., AI has come close to perfection on the \"character quality\". But why do a lot of people still smell like \"AI\" at once? The answer lies not in the luxuriousness of the hint, but in the absence of a \u201cconstructive consciousness\u201d. Many creators, when using AI, tend to focus on \u201cwhat to paint\u201d (the subject), while ignoring \u201chow to paint\u201d (form). When you don't specify the structure, the AI model is for insurance, and it tends to<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[149,144],"tags":[802,3149,956,5321,192,6481],"collection":[],"class_list":{"0":"post-57559","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"hentry","6":"category-jiaocheng","7":"category-baike","8":"tag-ai","12":"tag-prompt","13":"tag-6481"},"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/57559","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=57559"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/57559\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=57559"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=57559"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=57559"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=57559"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}