Mid-Journey V8- New Quality Unlock
No one has been talking about it. But recently, they have released their V8 model, However, they're saying that it's currently in testing, so it will get better over time. Users have gotten their hands on it, but they're saying good things such as its five times faster, mind-blowing 2K HD.
Some are saying bad things, such as it's a downgrade, they feel like they're using the V3 model, the quality picture is terrible, it cannot generate hands. Some are even saying they should stick to their Neo banana and Grok image generator. Shiiit. As a developer, you don't want to hear that. However, there is this one user on X that tests out the model by creating something kind of landscape-ish prompt, she fed Midjourney V8 alpha a prompt describing a scene, an emotional journey—a man witnessing first light over mountains, shifting from overwhelming vastness to quiet intimacy. Instead of cramming it all into one static image like older versions might, V8 understood the progression and delivered a triptych: wide cinematic establishing shot → tight emotional close-up → transformed return to scale. VIEW POST HERE
The v8 is currently good at adherence (especially --raw), finally reliable text rendering (including Chinese characters), cleaner results with longer/specific prompts, mood boards shine when you crank stylize or use --hd/--raw.
But, in overall in this, it all comes down to how you use the V8 model and your style in prompting. Because if you put bad prompt, the output will not look good. Versus the good one, you will get something good out in return.
The V8 focuses more on realistic, lifelike output.
Strong results come from clear, detailed prompts with no fluff, combined with multiple blended sref codes, high stylize settings, and personalization.
Using just one simple sref with short, vague prompts and no personalization leads to weaker, less polished results.
Here is some example prompts, along with the images of what V8 model has created,
Prompt from @michal_gren
"high contrast artistic photo of an athlete in dramatic lighting, smoke swirling from head, blurred movement around limbs, deep shadows, Nike-inspired styling, avant-garde sports editorial, cinematic natural light, vintage grain, soft contrast, documentary fashion sports vibe --chaos 10 --p 2v55sji"
MidJourney 8V Model
Prompt
"a human hand gently touching frosted glass, fingers softened by mist and condensation, dreamy diffused blur, shallow depth of field, abstract foreground haze, muted skin tones and hazy whites, ethereal fine-art photography, intimate and quiet mood --chaos 10"
Prompt from @harsha3d
"ancient stone monastery interior, monks seated at long wooden tables eating in silence, layered composition (foreground monks, midground server standing, background arched window glow), wide master shot high angle, 32mm lens, ARRI Alexa LF, cinematic 2.39:1, soft diffused daylight 5600K, candle practicals 2800K, minimal fill 8:1 contrast, subtle rim lighting, heavy haze, volumetric rays, deep focus f/5.6, warm amber tones, muted palette, Kodak Vision3 LUT, serene spiritual mood, painterly realism, no modern eleme"
That's it for the blog post. Will you be using this model? If so, let me know in the comments. By the way, here's a ChatGPT Midjourney Training Sheet I have created to get better analysis, more detailed, more enriched prompt.
See you soon enuh●
Comments
Post a Comment