Description
This animation wasnt created with AI. It was created with deforum, open source software that runs within stable diffision web ui. Most of you have seen deforum animations. They are characterized by generally one of two styles: one with forward and lateral movement, frame blending, and melodic music with slow morphing. The other style is characterized by stationary, rapidly changing frames which create an overall effect of shifting between scenes with a theme that stays consistent. Those who have used deforum, will attest that the animation you see here is not just technically complex, but it involves the coordination and synchronization of the whole gamut of parameters within deforum. Each frame is itself an independent image created with a specific model, steps, cfg value, prompt, seed, sampler etc. The animation itself is coordinated by a combination of several paramters which utilize a certain amount of data from the previous frame, depending on the values assigned. (for a directional paramter, each frame requires a velocity value. this value must manually entered for every single frame. deforum has no actual integration using the audio as a guide. it is up to.the user to determine a fitting prompt to match, and to assign parameters that will deliver the desired effect. For every second of animation, the FPS will reveal how many different parameter values there were. Just determining velocity values to accurately match audio is a feat in itself for layered audio with multiple drums and a high bpm. From there, there must be accurate and detailed prompt writing and knowledge to create a procession of prompt interchanges that allow for the visual changes desired. Deforum isnt actual Ai. Its uses text-to-image generation and gives additional capability to influence each image using previous frame data and parameters. It will only create what it is told to create, in the way that it is told to create it. It requires both a deep understanding of language, and how all parts of the whole work together. Once i saw how deforum actually functioned, and how much control it afforded (around a dozen different interlinked parameter types, with hundreds or thousands of unique instances per animation, with a possible range of at least 100 different values for each parameter instance) it became clear to me that the possibility of creation was far beyond what i had seen made. I envisioned something that would push its capabilities in all aspects. A colorful, complex, rapid, coordinated, constantly changing but without ever noticably morphing spectacle for both the eyes and the ears. Most deforum animations are 6-8 FPS with frame blending of 2-3 frames. Almost all of my creations are at least double that framerate, with no frame blending. This requires twice as many parameter values and six times the processing of most standard deforum animations, but it allows for an animation with so much going on, in such a fluid way, that most people would never assume how much technical and conceptual understanding and precision is involved. Diving in absorbed thousands of hours and thousands of dollars. It wasnt easy, but it was some of the most fun Ive ever had. I created hundreds of animations, with a look and feel that I have not seen personally represented in any other creations. At least that Ive seen. I only released a handful because I witnessed the general perception of anything created using Ai tools, and didnt have want something I had put in so much work to, to just be dismissed. Im minting some now because Im sick. I have metastatic cancer and I am in dire need of funds. I am currently living out of my truck while receiving my treatment, and I need to change that quickly. These are by design of my own vision and intelligence. I hope you enjoy them