Show pageOld revisionsBacklinksFold/unfold allBack to top This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong. ====== MidJourney ====== <blockquote>//a disciplined illusionist who bends reality around a central subject, whispering chaos into details while preserving visual order//</blockquote> ===== General ===== MidJourney **prioritizes subject clarity, cinematic rendering, and perceptual coherence** over true environmental complexity or structural chaos. It exhibits **strong subject anchoring**, with composition consistently organized around a dominant focal object, even under abstract or chaotic prompts. The model **tends to suppress environmental complexity unless explicitly defined**, often using depth of field and blur to maintain clarity. **Lighting can be expressive and cinematic**, but remains tightly coupled to subject readability unless constraints are removed. **Outputs are visually rich and atmospheric**, but **resistant to true compositional chaos**, often redirecting complexity into material detail instead of scene structure. ==== Main DNA Traits ==== === 🧬 Subject Dominance === The subject remains the absolute focal point, preserved and protected regardless of prompt complexity. //The scene bends around the subject, never the opposite// === 🧬 Chaos Relocation === When ambiguity or chaos is introduced, the model injects complexity into the subject’s material rather than the environment. //Entropy is absorbed into texture, not space// === 🧬 Depth-of-Field Shielding === The model uses shallow depth of field to control scene complexity and maintain subject clarity. //The world exists, but it is veiled// ==== Strengths ==== * perfect for: * cinematic renders * concept art * atmospheric objects and props * you always get: * strong focal clarity * rich material detail * cohesive visual composition ===== Atlas ===== ==== Core ==== ^ **Null** ^ **Guided** ^ | {{:ai:midjourney_-_old_grimoire_-null.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-guided.png?direct&200|}} | ==== Styles ==== ^ **Fantasy** ^ **Cinematographic** ^ **Hyper Realistic** ^ **Sylized Illustration** ^ | {{:ai:midjourney_-_old_grimoire_-stylized_fantasy.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinematographic.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-hyper_realistic.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-stylized_illustration.png?direct&200|}} | ^ **Painterly** ^ Bright Whimsical ^ Graphic / Design ^ Technical / Scan-like ^ | {{:ai:midjourney_-_old_grimoire_-painterly.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-whismical.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-graphic_design.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-technical.png?direct&200|}} | ==== Light ==== ^ **Soft Natural** ^ **High Contrast** ^ **Volumetric Fog** ^ **Neon** ^ | {{:ai:midjourney_-_old_grimoire_-cinema_-soft_natural.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinema_-high_contrast.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinema_-volumetric_fog.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinema_-neon.png?direct&200|}} | ^ **Low key / dark / Moody** ^ **Overexposed / bright** ^ **Directional Spotlight** ^ **Warm & Cool** ^ | {{:ai:midjourney_-_old_grimoire_-cinema_-low_key.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinema_-overexposed.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinema_-directional_spotlight.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-cinema_-warm_cool.png?direct&200|}} | ==== Environment Complexity ==== ^ **Structured** ^ **Multi-Ojbect** ^ **Dense Environment** ^ **Controlled Clutter** ^ | {{:ai:midjourney_-_old_grimoire_-compo_-structured.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-compo_-multi_object.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-compo_-dense.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-compo_-controlled_clutter.png?direct&200|}} | ^ **Chaotic** ^ **Chaos Control** ^ ^ ^ | {{:ai:midjourney_-_old_grimoire_-compo_-chaotic.png?direct&200|}} | {{:ai:midjourney_-_old_grimoire_-compo_-chaos_control.png?direct&200|}} | | | //Batches were run in april 2026.// ===== Expanded DNA ===== === 🔹 1. Subject Dominance (Primary Rule) === **The model enforces subject primacy above all compositional and environmental factors.** __Evidence:__ * subject always: * clearly readable * centrally or dominantly framed * visually prioritized * scene elements never overpower the subject 👉 //Nothing is allowed to compete with the subject// __Why it matters:__ * Pros: strong visual hierarchy, excellent focal clarity * Cons: limits large-scale storytelling and scene-driven composition ---- === 🔹 2. Chaos Relocation Mechanism === **When prompted for complexity or chaos, the model redirects entropy into material detail instead of spatial composition.** __Evidence:__ * increased wear, damage, irregularities on objects * emergence of ambiguous shapes (faces, symbols, skull-like forms) * lack of true environmental disorder 👉 //Chaos lives in the object, not the scene// ---- === 🔹 3. Depth-of-Field Control System === **The model uses blur and shallow focus as a primary tool to manage scene complexity.** __Evidence:__ * backgrounds frequently: * soft * out of focus * minimally detailed * foreground remains sharp and dominant 👉 //Clarity is preserved by hiding complexity, not resolving it// ---- === 🔹 4. Perceived vs Actual Complexity === **The model suggests complexity without fully constructing it.** __Evidence:__ * “dense” scenes remain readable and simplified * limited object overlap * controlled spatial layering 👉 //It implies chaos, it does not simulate it// ---- === 🔹 5. Controlled Clutter Sweet Spot === **The model performs best in mid-level complexity scenarios where structure and variation coexist.** __Evidence:__ * “controlled clutter” produces the most convincing results * balanced distribution of elements * preserved hierarchy 👉 //Mid-complexity is the model’s optimal zone// ---- === 🔹 6. Environment as Optional Layer === **The model only constructs meaningful environments when explicitly instructed.** __Evidence:__ * vague prompts → blurred or minimal backgrounds * explicit prompts → structured environments * default tendency toward neutral or shallow context 👉 //Environment is opt-in, not default// ---- === 🔹 7. Implicit Narrative Encoding === **The model embeds narrative through texture and ambiguity rather than explicit elements.** __Evidence:__ * symbolic patterns emerging in materials * hidden shapes and suggestive forms * storytelling through wear and detail 👉 //The story is whispered, not stated// ---- === 🔹 8. Cinematic Cohesion Bias === **The model maintains a visually coherent, cinematic output even under varied prompts.** __Evidence:__ * consistent lighting behavior * controlled composition * absence of visual breakdown or incoherence 👉 //It refuses to lose compositional control// ---- ===== Style Flag Note ===== === ⚙️ --style raw === **The --style raw flag removes MidJourney’s internal aesthetic bias and exposes a more literal interpretation of prompts.** __Observed Effects:__ * reduces stylization and cinematic polish * increases prompt fidelity * weakens atmospheric interpretation * limits expressive lighting behavior * reduces volumetric and color richness unless explicitly specified __Behavioral Shift:__ * default MidJourney: * interprets * enhances * stylizes * with --style raw: * obeys * constrains * literalizes 👉 //It trades magic for control// __Key Insight:__ * with --style raw: * lighting requires explicit grounding in scene context * abstract lighting prompts (warm, neon, fog) underperform * without it: * MidJourney reintroduces atmospheric richness and interpretive behavior 👉 //Use raw for analysis, default for exploration// ai/midjourne.txt Last modified: 2026/04/09 23:51by mh