in pursuit of stolen ghosts | Duet in Latent Space is an application developed by Marlon Barrios Solano, showcasing a blend of concept, programming, generative AI, sound design, and performance. It was created during an artistic and research residency at Lake Studios Berlin in February 2024.
Latent space refers to a conceptual and mathematical space generated by algorithms, particularly those used in machine learning and generative models. This space is characterized by the high-dimensional representation of data where similar data points are mapped closely together, and dissimilar points are further apart. In essence, latent spaces serve as the underlying structure that models use to understand and generate new data instances based on learned patterns, without being explicitly programmed with specific rules for every possible outcome.
The intention to perform actions in latent space, especially with a model, is to perform within an unknown probability space, utilizing the generative capabilities of the model to create art or perform in real-time. This process involves interacting with the model in a way that it generates outputs based on the vast, unseen dimensions of the data it has learned from, yet doing so in a manner that is unpredictable and unique to each performance. The performer, by manipulating inputs (prompts) or guiding the model's generative process, improvises with these outputs, creating an interplay between the expected and the unexpected, the known and the unknown. This act of performing in real-time within latent space becomes a dance (or a drawing) within a probability space.
The concept of the "stolen ghost" is deeply personal to me, reflecting my feelings of loss and disconnection stemming from my Venezuelan heritage and subsequent migration to the USA and Germany. Our history, marred by multiple colonizations and the obliteration of our cultural ancestors, leaves a void where a rich tapestry of traditions and identities once thrived. This disconnection from my roots, exacerbated by the physical and emotional distances of migration, embodies the essence of the "stolen ghost," a profound sense of loss not just of home, but of the ancestral voices and legacies that once guided us. This app helps me to create and an imagine these ghosts...ad infinitum.
The project is an amalgamation of various elements including software development, a real-time text to image gerative AI space for performance as a live app that runs in the browser, a sound space crafted with Musicgen, and a drawing tool. Additionally, users/performer can immerse themselves in the experience by playing the soundtrack created specifically for the app.
thepursuitfstolenghosts_s.mov
- LCM Turbo: Leveraging the power of Fal.ai for (almost) real-time, instant text-to- and image-to-image conversion. The model creates images very fast as if it is creating the drawing together.
- Excalidraw: is a virtual whiteboard application that allows users to sketch hand-drawn diagrams and illustrations quickly and easily. It's designed to mimic the feel of sketching on paper, but with the convenience of digital tools. Excalidraw. The output of the Excalidraw is sent to Fal.ai guided by the prompt.
- Soundtrack: A specially composed soundtrack that users can play, enhancing the visual experience with a complementary sound space
- Live Application: Experience the Bauhaus Time Traveler in action here.
- Marlon Barrios Solano: For more about Marlon, his projects, and how you can support his work, visit his Linktree.
LCM Turbo allows you to transform text anmd visuals into generative visuals instantly. Follow these steps to set up and start using the app.
Node.js is required to run the application and manage its dependencies.
-
Clone the Repository
Get started by cloning the repository to your local system.
git clone <repository-url>
-
Install Dependencies
Enter the project directory and install necessary dependencies.
cd <project-directory> npm install
-
Configure Environment
Rename
.env.local.exampleto.env.localand input your Fal.ai API Key. You can get your API key from the Fal.ai Dashboard.# .env.local contents example FAL_AI_API_KEY=your_api_key_here -
Run the Application
With everything set up, launch the application.
npm run dev
Open
http://localhost:3000in your browser to see the app in action.
MIT License
Copyright (c) [2024] [Marlon Barrios Solano]
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.