Posts

Vignette no. 1

-
Image
Image prompt made with Qwen: An abandoned underground maintenance corridor beneath a Bangkok temple complex, pitch black except for a single failing emergency light flickering near the ceiling. The air is thick with humidity and faint steam. Moss-covered concrete walls show cracks seeping moisture; corroded copper pipes run overhead, one bent and leaking just outside frame. A single fresh water droplet has just fallen — unseen — onto the damp stone floor, creating a small ripple in a shallow pool surrounded by cracked tiles. Reflections in the water subtly distort the dim red glow above, and concentric ripples are the only sign of motion in an otherwise frozen, silent space. No figures, no text, no visible source — only evidence: rust trails, pooled condensation, and the echo of absence. Cinematic lighting, ultra-high contrast, deep shadows, volumetric haze, 35mm film grain, atmospheric perspective — as if something is listening from the dark beyond the light. Image generated with Gemi...
Post a Comment
Read more

When Veils Part 3m 37s

-
Image
 I refined some glitches that annoyed me the first time around and changed my titling to my new concept of using a QR code that provides static information for the credits. I prefer the collaboration aspect of using my last name and AI. The idea sprung up one day when I was looking at a law firm sign. Film festivals submitted to date: Rome AI Festival 2026   Rome, Italy Berlin AI Festival 2026  Berlin, Italy An expressionistic journey that was initiated with a paint texture on acetate. By allowing the AI generative model to interpret the image, a story emerged. With my direction a coherent and expressive narrative was formed. All video and sound was created with generative AI models Video generated by: Midjourney Kling PixVerse Most of this video was created with Midjourney. Allowing it to randomly choose action with some detailed prompting. Utilizing their expand feature helped with the initial texture to human scenes. Kling was used to stitch some components using ...
Post a Comment
Read more

Cinema Sound Study no. 1

-
Image
  A long sequence of AI generated scenes that started with an image blur. I wanted to test Runway's Gen4 model for it's pareidolic judgement. It was rather boring in it's interpretations, but after using it to test my new subscription to Krotos Studio Pro and experiment with the AI generated sound effects it offers, I found it to be much more worthy of watching. The entire sequence is a continuous expand on the previous clip. Sometimes I get curious what would the man do if I kept going? It's an open world in a way, but like Vidu, I think it would eventually get into an infinite loop.
Post a Comment
Read more

Midnight Hustle

-
Image
  Essentially a music video. It's a light-hearted take on my cab driving experiences in Baltimore, Maryland in 1990.  Most of the 70's looking stuff was generated with Perchance Stable Diffusion. It provided the most accurate grittiness, especially the row house segments. They look very authentic. Unfortunately the AI model was updated with new training and these images are impossible to duplicate now. The music was created with AIMusic. The lyrics were created by AI after supplying a good amount textual ideas of what I wanted to project. I did a little tweaking and generated the music. This was a long process of casual collecting of video clips. Putting it all together was rather boring to me, but I felt I needed to justify all the accumulated clips with something.
Post a Comment
Read more

Pareidolia Study no. 5

-
Image
  This is the fifth of a series of experiments on the psychological phenomenon called pareidolia. According to Wiki: Pareidolia is the tendency for perception to impose a meaningful interpretation on a nebulous stimulus, usually visual, so that one detects an object, pattern, or meaning where there is none. Pareidolia is a specific but common type of apophenia (the tendency to perceive meaningful connections between unrelated things or ideas). Common examples include perceived images of animals, faces, or objects in cloud formations; seeing faces in inanimate objects. I have been fascinated with this phenomenon for most my life having incorporated it in my art since the early 1980's. Initially, I experimented with creating textures with various photographic processes and pigments on various materials to achieve interesting patterns that could be imagined into interesting works of art. Many of these textures I made on a frosted acetate film and stored in 5x7 negative sleeves. This f...
Post a Comment
Read more

Pareidolia Study no. 4

-
Image
  An analog texture I made in the early 90's was upgraded using Leonardo's creaive upscaler. I then used Midjourney to create extended videos up to 20 seconds. Starting with the upscaled texture and moving forward in the timeline: The AI generative model was minimally prompted to move. It was left to decide what the image is and what it should do. Each direction was made n 5 to 20 sec. durations. This experiment allows for an in depth look at how AI recognizes patterns and how AI implies motion with the pattern. I fed the image of the woman that Midjourney created in the middle and added her to the beginning and end using Vidu. Video can be found here:
Post a Comment
Read more

When Veils Part

-
Image
  A texture I made in the early 1990's. I used Midjourney to create a series of videos from this texture with minimal prompting. I then created a story. I used Kling's  start and end frames feature for video generation. This allowed me  to direct the action. Music was created with Aimusic. I wrote a series of pleas to be left alone that serve as lyrics. Sound effects and background voices from characters were created with Elevenlabs.  It's an expressionistic gesture.  Maya Daren's "Meshes of the Afternoon" was coming to mind on occasion when I crafted this.  Video can be seen here:
Post a Comment
Read more