Instagram story viewer> @obsidian.research> Posts

obsidian.research

@obsidian.research

298

followers

0

following

Obsidian | Research Division: Project ████
⌖future relic⌖
conscious pattern emerging | origin: unknown.
@obsidian__is

POSTS STORIES REELS TAGGED

Research #010 | @apetown101

New test from Letlotlo Mangope
Red in Between by @obsidian.research

3

0

6 months ago

Research #009 | @hartemus

“Throughout my artistic life and career, I have had a few mentors, and I must admit that in the beginning, when I was working in 3D, I also dreamed of being a DP, a Director of Photography. I have had this passion for light for a very, very long time. In my career, I also had the opportunity to be hired as a Senior Lighter, and my sense of light developed alongside my sense of narrative and visual storytelling.”

“So I am doing my best to reflect my passion for light in the way I use artificial intelligence and AI, but I must admit that I still have not found what I am looking for in the current tools. While working on my next short film, Barefoot, I really wish I had more powerful tools. I have done a few tests with higgsfield, with Nano Banana, and with different approaches, but I feel that we are still far from what is possible and what would be required for a feature film.” by @obsidian.research

3

0

6 months ago

⟟ Archival fragment recovered
⟟ Data corruption index: ██.██%
^• by @obsidian.research

3

0

6 months ago

Research #008 | @hartemus by @obsidian.research

4

0

6 months ago

Research #007 | @balderasmx1 by @obsidian.research

2

0

6 months ago

Research #006 | @apetown101 by @obsidian.research

3

0

7 months ago

Research #005 | @balderasmx1

Sharing a new clip where I re-stylized video locally using a combined 𝗖𝗼𝗻𝘁𝗿𝗼𝗹𝗡𝗲𝘁 approach inside ComfyUI stacking guidance to keep structure stable while pushing a new look. by @obsidian.research

3

0

7 months ago

Letlotlo Mangope | @apetown101
Creative Technologist & Animator

every frame treated as constructed scene. not a generated image. ai is a force multiplier. cgi drives layout and framing, ai drives texturing lighting and finishing . it’s a beautiful blend born of intentionality with the image making process not randomness. by @obsidian.research

3

0

7 months ago

Research #004 | @genmagnetic

“I set out to work with the very latest video model available Seedance 2.0 and see if I could make a hilarious beer ad, with memorable characters and perfect comedic timing, and this is what I came up with.” by @obsidian.research

3

0

7 months ago

Research #003 | Short and music by @hartemus

“When I made this project, it came from a mix of Parisian nostalgia and a memory from my childhood. When I was young, I always tried to climb as high as possible onto the rooftops so I could look at the stars, and feel closer to the cosmos. I had the impression that I could almost take flight.

So this film test is also a small piece of my nostalgia, of everything that shaped me in France: the architecture, refinement, beauty, elegance, and humanity.

As for the process, I always try to build my vision through a technical workflow in which the tools remain in service of my discernment. After years of working with Photoshop, photography, the visual effects industry, and my love of poetry, I decided to build my own arsenal of AI tools to serve my creative process more effectively.

That is why I used Midjourney, Nano Banana, and Sora. In general, this is how I use them: I use ChatGPT to help me refine the prompts I need for each platform. I use Midjourney when I want creativity and interesting creative iterations. I use Sora to steer the visuals back toward something more human and relatable. Then I use Nano Banana for its more technical, colder, and more precise qualities, especially when I want to create different points of view.

At the time, I used Kling for the animation because I felt it offered the best balance between image/pixels quality and prompt adherence.” by @obsidian.research

3

0

7 months ago

Research #002 | An AI Horror Experiment by @0x0f0000

“This spot was an exercise in exploring complex weavy.ai workflows with a focus on creating shots from storyboard frames, keyframe concept art and also creating additional coverage shots from a limited set of keyframes.” by @obsidian.research

3

0

7 months ago

Danilo Šteher | @danilosteher
Obsidian Head of Automotive CG

“As for my approach, I’d say the key is intent and purpose. You have to respect the logic of what you’re showing, and the only way to build on top of the deliberate engineering and aesthetic decisions of car design is by following the same principles. Treat every decision with the same level of intent.”

Form follows function. by @obsidian.research

3

0

7 months ago