demobook

Midjourney: Interpreting Descriptive Prompts in Conversational Mode

MidjourneyTry it →Watch full video →Wade McMaster - Creator Impact ·

Demo summary

The video shows Midjourney V8 successfully identifying the 'Statue of Liberty' from a descriptive prompt without the landmark being named explicitly.

Step-by-step

  1. Enable conversational mode
  2. Enter a descriptive prompt that describes a subject without naming it explicitly
  3. Submit the prompt to generate images
  4. Review the generated images and the system's interpretation of the subject
  5. Disable conversational mode to compare results using the same prompt

Options

  • Toggle conversational mode on or off to change how the model interprets descriptive text

Tips

  • Use conversational mode if you want the system to explicitly identify and label the subject of a descriptive prompt

Highlights

I'm actually really impressed. Midjourney in the past would have absolutely failed this test completely.

All demos from “Midjourney V8 is NOW LIVE!! Here's how to use it.

  1. 0:341:40Navigating the Midjourney Alpha Sidebar and SettingsThe creator demonstrates the new sidebar interface on alpha.midjourney.com, showing how to toggle models, adjust HD mode, and use the image upload tab for style references.MidjourneyAI Image Generator
  2. 2:380:48Testing Negative Prompting in Midjourney V8 Conversational ModeThe user demonstrates how Midjourney V8 handles complex negative instructions like 'don't include dogs' by switching between standard, raw, and conversational modes.MidjourneyAI Image Generator
  3. 3:560:45Interpreting Descriptive Prompts in Conversational ModeCurrentThe video shows Midjourney V8 successfully identifying the 'Statue of Liberty' from a descriptive prompt without the landmark being named explicitly.MidjourneyText to Image
  4. 4:451:05Comparing Detail Coherence: Midjourney V7 vs V8A side-by-side comparison of V7 and V8 generating a portrait with specific details like different colored eyes and a red mole to test model intelligence.MidjourneyAI Image Generator
  5. 6:371:14Text Rendering and Paragraph Generation in V8The creator demonstrates Midjourney V8's ability to render specific text on a forehead and a full paragraph of handwritten notes compared to V7's gibberish.MidjourneyAI Image Generator
  6. 9:291:08Real-time Render Speed ComparisonA live demonstration of the rendering speed difference between Midjourney V7 and V8, showing V8 completing a grid significantly faster.MidjourneyAI Image Generator
  7. 11:070:30Generating 2K Resolution Images with HD ModeThe user shows how to enable HD mode to generate 2048x2048 resolution images directly in Midjourney V8.MidjourneyAI Image Generator
  8. 12:531:53Using Quality Settings (--q 4) for Enhanced DetailThe video demonstrates the impact of the quality 4 setting on image detail and texture using Photoshop to compare the results against standard quality.MidjourneyAI Image/Photo Enhancer
  9. 15:512:22Adjusting Stylize, Weird, and Chaos ParametersThe creator demonstrates how varying the --stylize, --weird, and --chaos parameters affects the aesthetic output and variety of Midjourney V8 grids.MidjourneyAI Image Generator
  10. 18:232:04Applying Mood Boards and Style ProfilesThe user demonstrates applying saved mood boards and personalization profiles to prompts in the Midjourney Alpha interface to change the artistic style.MidjourneyAI Style Transfer
  11. Watch “Midjourney V8 is NOW LIVE!! Here's how to use it.” →

Text to Image

  1. 1:070:50Generate anime images with Midjourney NijiThe video shows how to use Midjourney's Niji model and the expressive modifier to generate stylized anime characters and backgrounds for video source material.FILM CRUX
  2. 0:510:24Generate and modify images in Midjourney webThe user demonstrates entering a text prompt for 'a cat with a hat' in the Midjourney web interface and shows how to adjust settings like aspect ratio and variety.The AI Advantage
  3. 3:560:45Interpreting Descriptive Prompts in Conversational ModeCurrentThe video shows Midjourney V8 successfully identifying the 'Statue of Liberty' from a descriptive prompt without the landmark being named explicitly.Wade McMaster - Creator Impact
  4. 20:262:20Running prompts in Midjourney via DiscordThe presenter demonstrates pasting a generated prompt into Discord using the '/imagine' command to generate a 4-image grid in Midjourney.Jonathan Mast
  5. 0:540:46Generate neon zombie-fied mutation art in MidjourneyThis clip shows the results of the 'neon zombie-fied mutation' prompt, which applies oversaturated colors and undead features to subjects using Midjourney.Wade McMaster - Creator Impact
  6. 2:250:49Create Mad Max dystopian aesthetics in MidjourneyThe creator shows how the 'Mad Max style' prompt generates warm, dystopian visuals with rusted vehicles and post-apocalyptic attire for various characters.Wade McMaster - Creator Impact
  7. 4:120:39Generate original Game Boy graphics in MidjourneyThe creator demonstrates the 'original Game Boy graphics screen cap' prompt to produce low-resolution, pixelated, monochromatic green-scale imagery.Wade McMaster - Creator Impact
  8. 5:260:46Generate Hung Gar Kung Fu inspired poses in MidjourneyThe video shows how the 'Hung Gar Kung Fu inspired' prompt changes character poses to traditional martial arts stances and adds billowing Shaolin-style robes.Wade McMaster - Creator Impact
  9. 6:120:41Create samurai ukiyo-e tattoo art in MidjourneyThe creator demonstrates a complex prompt for 'samurai ukiyo-e inspired tattoo art on plain white' to generate Japanese-style tattoo illustrations.Wade McMaster - Creator Impact
  10. 6:530:55Generate revenge style movie posters in MidjourneyThis clip demonstrates the 'revenge style movie poster' prompt, which emphasizes red color schemes and intense, vengeful facial expressions.Wade McMaster - Creator Impact
  11. 7:390:41Create embroidery patch assets with MidjourneyThe user shows how to generate pill-shaped and oval embroidery patch graphics using Midjourney prompts to create assets for a variable data print project.Midjourney Experience
  12. 1:282:31Generate retro images using the PTS promptThe user shows the process of pasting the generated prompt into an AI image tool to produce 1950s-style visual results.Paper Trail Studio - Amanda