Systematiq Ai

On Episode 19, we're cracking open our vault of Midjourney tokens, sharing game-changing phrases that transform ordinary prompts into extraordinary generations. We break down the remarkable power of specific tokens, demonstrating how singular words can elevate everything from tattoo designs to architectural details. Then we dive deep into the art of color manipulation, exploring how techniques like color blocking and tonal transitions can dramatically enhance image composition. The episode unveils professional-grade strategies for mastering lighting and shadows, with us sharing our personal favorite tokens for achieving cinematic results. A major highlight includes our approach to facial expressions, revealing how specific token combinations can transform lifeless AI portraits into emotionally charged masterpieces. We wrap up with an extensive discussion on token hierarchy and strategic prompt placement, providing you with actionable strategies for immediate implementation. --- ?? Midjourney Fast Hour [00:00] Getting Started [03:45] Community Love [09:00] Token Power [13:15] Detail Magic [26:30] Color Play [34:45] Shadow Dance [41:15] Facial Expressions [49:30] Photo Magic [54:15] Special Effects [59:45] Wrap-Up cc: Drew Brucker Rory Flynn Midjourney

RUNWAY’S NEW “EXPAND” FEATURE → USEFUL. Serious "quality of life" update. You can now expand a video + outpaint. To landscape or portrait. PROCESS: 01. Gen initial video. 02. Download video. 02. Open Gen-3 Turbo. 04. Select “Expand Video.” 05. Select Portrait or Landscape AR. 06. Generate. MJ PROMPT Ex: wide angle documentary photography, captured in a unique composition, the best photo of this journalist's life, vibrant award-winning photography, epic scenery, photojournalism --ar 5:3 --quality 2 --stylize 200 --v 6.1 RW Initial Prompt (5:3 AR): the camera follows behind a woman walking away from us toward the horizon, a fire is blazing in the sky, natural dynamic flame motion, showing the conflict of the human condition, 35mm style RW Expand Prompt 1 (9:16 AR): the camera follows behind a woman walking away from us toward the horizon on a street, a raging fire is blazing in the sky, natural dynamic flame motion, showing the conflict of the human condition, chaotic environment, 35mm style RW Expand Prompt 2 (5:3 AR): the camera follows behind a woman walking away from us toward the horizon on a street, a raging fire is blazing in the sky, while a rundown city surrounds her on both sides, black smoke fills the sky, natural dynamic flame motion, showing the conflict of the human condition, chaotic environment, 35mm style, QUICK THOUGHTS: + This was needed + It matches style/aesthetic well + Text prompts still have heavy influence + Your text prompts influence the “outpainting” + So w/e you want to generate “net new” + Make sure it’s in your prompt + Meaning if you have a closeup of face + When you expand you’ll “add background” + So prompt for what you want in the BG + This will allow for multiple variations Hope that makes sense. Have fun. PS: Runway Cheat Sheet in comments section #runwayml #minimax #midjourney #ai #innovation?

Controversial Take… Image Curation > Prompting…in Runway. This is not talked about enough. Using the wrong images = headaches. Using the right images = saved credits. Here’s a few tips ↓ 01. FOR STARTERS + Reference Image = Huge output influence + Poor image selection = Poor vid generation. + Static Image motion = Better vid motion + Higher Res = Sharper definition 02. DO’S & DON’TS → Use images with motion baked in → This gives runway a starting point → Instead of imagining “new” motion → Give it 50% of the motion. → So RW can pick it up and run with it. MJ Static Prompt: Ex. Car with spinning wheels vs static wheels RW Prompt: the camera follows a car driving to the left 03. WALKING EXAMPLE IMG 1:? + The feet are in motion + And It leads to a smooth transition IMG 2:? + The feet are static + And it has a hard time starting the motion Good MJ Prompt:? editorial photo, front view closeup, orange hiking boots [WALKING TOWARDS US] on a dirt path in the alpine forrest, [BOKEH MOTION], late November setting, wet leaves and bare trees, high contrast photography, contrasting colors, 35mm high precision photo --ar 16:9 —q 2 —p Bad MJ Prompt:? editorial photo, front view closeup, orange hiking boots on a dirt path in the alpine forrest, late November setting, wet leaves and bare trees, high contrast photography, contrasting colors, 35mm high precision photo --ar 16:9 —q 2 —p RW Prompt:? slow motion shot of orange boots walking towards us on a hiking trail in an alpine forest, dynamic natural motion 35mm style 04. PROMPTING FOR MOTION + EXPRESSION Focus on: + Character/scenery motion + Character expression. MJ Prompt:? [BOKEH MOTION], medium closeup of a womans face [LOSING HERSELF IN THE MUSIC], tulum jungle rave atmosphere, [SUBTLE NEON LIGHT STREAKS] painting her face, 16mm, asymmetrical composition, moon light, silhouettes, techno --ar 16:9 —q 2 —p RW Prompt:? hand held slow motion shot of a woman cheering in a tulum jungle rave, silhouette lighting, 35mm style Hope this helps. Have Fun. ? PS: If you liked this…There’s a Runway Cheat Sheet (LINK IN COMMENTS) PPS: Testing this new “video carousel” format…no idea if it’s useful or not. (LMK) #runwayml #minimax #midjourney #ai #innovation

RUNWAY GEN-3 → CHEAT SHEET 300+ hours of Runway testing… Distilled into a simple…”Beginner’s Cheat Sheet” If you're finding it hard to get started. This will help. (Hi-Res Download → https://lnkd.in/eSaHEGwD) HERE’S WHAT IT INCLUDES: + (18) Prompt Formulas? + (107) Cinematic Tokens? + 6-Step Quick Start Guide? + Creator Control Panel Diagram? + Camera Control Diagram + Camera Movement Diagram? + Keyframes Diagram? + [14] Cinematography Categories? + [15] Shot Types? + [13] Camera Motions? + [12] Camera/Lens Types? + [9] Color Grading Styles? + [14] Composition Techniques? + [8] Balance Techniques? + [12] Lighting Techniques? + [8] Camera Speed Techniques? + [8] Cinematic Styles? + [8] Animated Styles QUICK THOUGHTS: + If you’re just getting started with vid models + Try text-to-video first…before img-to-video + You’ll get a better feel of what you can control + When you use img-to-vid + You have less control because you img can limit + What the video tools can do. Happy Monday. PS: Working on a more comprehensive training for Runway. (Coming soon) #runwayml #minimax #midjourney #ai #innovation

Minimax Outtakes V3 → NAUGHTY PIKACHU. (This post provides zero value) Friday’s are for testing/experimenting. (…and being dumb) Lots of errors and inconsistencies. >> This post is purely for entertainment purposes<< MIDJOURNEY PROMPT: flash photo, Batman and real life pikachu having fun at a house party, 1980s --ar 3:2 --stylize 75 --v 6.1 MINMAX PROMPT STRUCTURES: 01. The camera [motion/direction] as Pikachu and Batman [action] 02. The camera [motion/direction] revealing [outcome] MINIMAX PROMPT EXAMPLES: 01. The camera pans down to reveal empty beer bottles everywhere 02. The camera pans up to reveal Spiderman crawling on the ceiling upside down 03. The camera pans down to reveal the joker crawling on the ground, before he looks at the camera and waves awkwardly 04. The camera pulls out revealing a Wonder Woman walking into the scene, blowing smoke into Pikachus face comedically 05. The camera pulls out revealing Deadpool throwing Pikachu out of the party QUICK THOUGHTS: + Minimax is just fun + It’s a great way to learn the gen-ai vid tools? + It responds well to camera/character prompts + It also does well with emotion + The prompt coherence is surprisingly strong + And it has oddly humorous outputs + However, the content moderation is lax + Be careful using it for commercial purposes There is zero point to this video. Other than saying…just play… You’ll learn better that way. Have fun. #minimax #runwayml #midjourney #ai #innovation?

RUNWAY → ADVANCED CAMERA CONTROL Testing on static product/object shots. It seems to be the best use for this feature. (Currently) PROCESS: 01. Create image in Midjourney 02. Upload in RW GEN-3 03. Select Camera Control 04. Add Text Prompt 05. Adjust Camera Motions CAMERA MOTIONS + Horizontal: Move Left or Right + Pan: Arc Left or Right + Vertical: Move Up or Down + Tilt: Arc Up or Down + Zoom: In or Out + Roll: Left or Right MIDJOURNEY PROMPT: Editorial photo of a white coupe concept car , sleek rims with gold accents, deep contrast and contouring, tuner performance intricate detailing, neon accents, futuristic sleek and clean minimal design, dark environment, neon accent lights on the walls --ar 5:3 --p --stylize 200 --v 6.1 RUNWAY PROMPT STRUCTURE 1: The camera [ACTION + DIRECTION] on a [SUBJECT + ENVIRONMENT] RUNWAY PROMPT STRUCTURE 2: The camera [ACTION + DIRECTION] and [SHOT DESCRIPTION] of [SUBJECT + ENVIORNMENT] RUNWAY PROMPT EX 1: The camera slowly pushes in on a car in a studio RUNWAY PROMPT EX 2: The camera raises into the air and tilts down revealing a high angle shot of a car in a studio environment QUICK THOUGHTS: + Works best w. static character/object shots + When using Advanced Camera Control + Ambient movement is limited + So if you want more character/object movement + Use Img-2-vide w/o camera control + Text prompting makes a huge difference + The text needs to match the camera motion selected + The object permanence is good (not perfect) + Camera motions “weight” is sensitive + Higher value = more significant movement + Zoom-in loses some fidelity + Zoom-out keeps fidelity better (imo) This is still a new feature for Gen-3. The ambient motion will improve over time. Have fun. #runwayml #minimax #midjourney #ai #innovation?

Babe, wake up! New MJ:FH episode just dropped. On Episode 17, Rory Flynn and Drew Brucker break down the new external editor and retexture features, sharing their first-hand experiences as beta testers. They dissect real client applications with powerful workflows for transforming existing assets. ? Watch/Listen Apple: https://t.ly/ycOTO Spotify: https://t.ly/hzEmQ YouTube: https://t.ly/zvLdf #midjourney #aiimagegeneration #midjourneypodcast

RUNWAY → ADVANCED CAMERA CONTROL. Very solid for static shots. (Not so great with additional motion) PROCESS: 01. Create image in Midjourney 02. Upload in RW GEN-3 03. Select Camera Control 04. Add Text Prompt 05. Adjust Camera Motions CAMERA MOTIONS: + Horizontal: Move Left or Right + Pan: Arc Left or Right + Vertical: Move Up or Down + Tilt: Arc Up or Down + Zoom: Push In or Out + Roll: Left or Right MIDJOURNEY PROMPT: drone aerial shot, futuristic 5-star treehouse hotel exterior, nestled among the trees, eco-architecture luxury, sustainable, early morning shadows, bold, high-contrast, dramatic shadows, nature's symphony --ar 5:3 --p --s 200 --q 2 RUNWAY PROMPT STRUCTURE: The camera [motion] [direction], revealing [details] RUNWAY PROMPT EXAMPLE: The camera pans left revealing an immaculate Yucatan jungle canopy, white sand beaches, crystal teal water, epic scenery QUICK THOUGHTS: + This works well with images like this + Additional movement is hard + Ex. Characters + dynamic actions + Text prompting is important + It influences the output heavily + Focus on the “outcome” in the text prompt + Combining camera motions works well Hoping to get motion brush in Gen-3 soon. Have a weekend. PS: MORE INFO on RUNWAY TRAINING → SIGN UP FOR NEWSLETTER (LINK IN COMMENTS) #runwayml #minimax #midjourney #ai #innovation

MJ EDITOR + RUNWAY → Scenery Change. Love those vintage green-screen driving scenes. Tried to recreate a similar feel. Not perfect…but fun. PROCESS: + Gen initial image in MJ + Highlight/Edit scenery in MJ Editor + Animate in Runway + Udio for Music MJ ORIGINAL PROMPT: 1970s cinema car scene, shot from behind at a high angle , 2 men sitting in convertible facing away from us driving on a California street lined with palm trees, cool blue and orange color grading, 16mm vintage film --ar 5:3 --p --stylize 200 --v 6.1 MJ EDITOR PROMPT: 1970s cinema car scene, [environment + details], cool blue and orange color grading, 16mm vintage film --ar 5:3 --p --stylize 200 --v 6.1 RUNWAY PROMPT: A 1970s cinema-style scene of two people talking with flowing hair in a classic convertible, shot from behind. The car interior is steady, but the background simulates motion like a vintage green screen effect, with the illusion of the car driving forward. Slight film grain and warm retro faded colors UDIO PROMPT: vintage los angeles rock and roll, Soulful guitar riffs breezy feel. warm, layered sound. Rhythmic bass and light percussion, Hints of blues and psychedelia, depth and vintage authenticity. PROBLEM AREAS: + Scale issues with other cars + Perspective flaws on street/background + Windshield view fluctuates? + Random artifacts MJ QUICK THOUGHTS: + Editor matches style really well + Without SREF/IMG ref here + Focus edit prompts on the scenery + So change prompts for “environment” + Less focus on the “subjects” RW QUICK THOUGHTS: + Vintage aesthetics work well + On most img/video generators? + Used first frame gen for these videos + For first frame gens… + Prompt needs some aesthetic direction + When using first frame for continuity + So RW can create in the same style Have Fun. PS: Working on some more in-depth Runway training… Subscribe to the newsletter for updates → https://lnkd.in/e4dg_d-T #runwayml #minimax #midjourney #ai #innovation

Ai at Scale → Possible. If that even remotely interests you... >>Check out our Pod with Phillip Maggs<< #runwayml #midjourney #ai #innovation #chatgpt