#Midjourney AI

[xxluna-rougexx]

Hogwarts

stylized in art nouveau

[rootsinthefuture]

[jellyfisharcade]

Midjourney + Lightroom + clarity upscaler

Jellyfish Arcade!

[deepdreamnights]

I remade the trailer to Robot Monster, to serve as a trailer for a hypothetical remake of... you guessed it, Robot Monster.

Destruction has come, hu-mans, and its silliness will not protect you.

My thoughts and how-to process blog post under the fold.

I made Robot Monster's Trailer Remake primarily with Vidu and with Midjourney.

For most shots I started with a photoshopped midjourney gen (or stack of them), which was used either as a prompt image or starting frame.

Some shots, like the earthquake, were done with start-and-end frames.

Vidu has some quirks for my Roland Emmerich Christ-the-Redeemer shot. I attempted the image several times as a direct image as start frame, but it would reset to a new camera angle each time, rebelling against my inaccurate version:

As the AI could recognize the statue, but not it being in inaccurate surroundings. I eventually used it as a prompt and not a start frame got a good enough shot.

I tended to go for 8-second shots on quality mode, to give me more to cut around and edit. Almost no shots play without some cropping, speed adjustments or other edits in this, and anyone using AI for a larger project is going to find much the same.

While 90% of the shots are from Vidu a few I used Hailuo's Minmax to accomplish. Mainly things like a few low-motion Ro-man talking shots, the computer-communication device, and the motion title card for "electrifying", etc.

Vidu likes to move, a lot, and for stuff that needs subtle movement I sometimes find it helps to mix things up.

I've found that when image prompting for a character, like Ro-Alice, it sometimes helps to do a fullbody and portrait two-for-one. This helps keep the character design consistent, and you can kinda tell which Ro-Man shots I made before I figured this trick out.

I also reused shots of the dinosaurs from my other AI video projects for meta reasons.

Right now it doesn't make videos so much as it makes shots you can weave into videos.

I'm actually impressed at how well it understood the concept of Ro-Man, only giving him a full ape face or a weird tail or the like a couple of times.

My general approach to the concept was "What if you kept the premise the same but had a budget." Whereas in reality you'd never actually get that combo, since if they had money, they wouldn't have made Robot Monster.

It also let me play with a fanon idea I've had for awhile that the Ro-Men were the helmets, and the ape-creature was some biological organism used as a conveyance.

For the audio, I took the audio to the trailer and used Suno's cover-features to both clean up the sound and change the musical style. The back half of the original track was completely warped by the cover process, but I used another bit of trailer-style music to cover that bit, and to extend for the longer ending shot, since my version of the trailer is about 20 seconds longer than the original.

Some prompts utilized:

in a sci-fi lab in a cave, a furry alien monster wearing a spherical helmet with reflective faceplate walks around aubrey plaza in a white sleeveless slip-dress and dark pantyhose in a glass tube, the tube pulses with green light. She is in a glass cylinder, he is walking around it, with curiosity. The scene is menacing, slow movement, pensive. horror movie scene, the tone is tense and frightening. professional lighting and cinematography. Oscar winning, 2003, practical lighting, effects, and costuming.

the robot spider-robots with spherical heads walk around as though searching for something. horror movie scene, the tone is tense and frightening. professional lighting and cinematography. Oscar winning, 2003, practical lighting, effects, and costuming.

the alien ape-creature wearing a space helmet (the robot monster), in a modern city. He throws green lightning from his hand, disintegrating a policeman into ash. Monster-movie sci-fi scene, dramatic camera angles and lighting. Practical costuming and special effects. High budget and high concept.

slow motion fly-through footage, the air is full of slow-moving glowing bubbles. green electric sparks arc from one bubble to the next producing an ominous mood. The scene conveys spreading menace and fear. One long, unbroken shot. filmed on location, effects by weta digital, ILM, stan winston studios, believable and hyper-realistic. Shot on location. trailer shot. high-speed film

All-in-all, a fun project, and one that came along when I really, really, really needed something to concentrate on for long stretches of time.

Make something fun, folks.

[a-bluedream-posts]

(M04G4) Fubuki by namakxin

[marshmallowfairbanks]

Art: Mighty Oak

[lady-phasma]

Diversity in fantasy aesthetics

my favorites!

The start of this series of posts is here, explaining my purpose with these images.

If you don't like AI art then keep scrolling or block the tags. I don't use it for financial gain, I use it to make images from my words, and really only for Tumblr.

I didn't even tell it to hide a cute tattoo!

Please save/download/reuse any and all of these, reblog if you want.

[variablexstudio]

Series X1 - Plastic Supermodel

[yama-bato]

Josef Sudek

Josef Sudek style in Midjourney AI

[gacougnol]

In the style of josef Sudek

Made with AI (midjourney)

[paladinpeterparker]

“are you okay?” “yeah, i got my girls!”

i promised you all i’d give you the girls so…here they are! i feel like these are going to be a little more controversial than the boys because i’ve seen some very different opinions on what the girls look like! but this is my take! :)

made using midjourney ai

[xxluna-rougexx]

Hogwarts

An intimate journey through the lens of a photographer

Feel free to use them for your desktop wallpaper!

[rootsinthefuture]

Ikea catalogue from 2225

[jellyfisharcade]

Midjourney - Lightroom - Real-ESRGAN - Jellyfish Arcade

[deepdreamnights]

Burnout

That's ridiculous. I'm as ready to go as ever.

The image(s) above in this post were made using an autogenerated prompt and/or have not been modified/iterated extensively. As such, they do not meet the minimum expression threshold, and are in the public domain. Prompt under the fold.

Working with multi-prompting under 6.1. It produces good results, but ignores most of what's going on. So... kinda like 5.

Prompt: A red, glowing alien cloud with an eye in the center of it, floating over Antarctica. A man stands on desert ground, looking up at it, shot from behind him. The night sky is illuminated by orange and teal lighting, creating a cinematic atmosphere.:: A dramatic sunset sky over the Great Plains, with vibrant orange and yellow hues painting an epic scene of swirling clouds above a lone tree silhouetted against the backdrop. The photograph captures the intensity of nature's display in high resolution, with sharp focus on intricate details like ripples or cracks in cloud formations. It highlights how natural elements can create stunning landscapes that mesmerize and evoke emotion.:: beautiful woman in neon yellow-green body art, full-length body portrait glowing with light from within, intricate swirls of vines and flowers in a fantasy style with hyper-detailed black background and vibrant colors, a digital painting in the style of Artgerm and WLOP and Greg Rutkowski:: 3D cartoon in the Pixar style of a full body of a woman wearing a black and brown bodysuit with makeup shaped like an S symbol on her face against a red background. She has short hair and wears high heels. Red eyes. Rendered in the Pixar style. In the Disney style. A cartoon character with no shading. A full-body view where no feet are visible. Her hands are raised as if helping herself stand up after falling down. She looks confident but also very sad.

[a-bluedream-posts]

Frozen Kiss by Dantegonist