For media company Black Mixture, AI isn’t just a tool — it’s an entire pipeline to help bring creative visions to life.
Founded in 2010 by Nate and Chriselle Dwarika, Black Mixture embraces an innovative approach to integrating AI into creative workflows. Initially, the company established its visual reputation in motion design and production, utilizing traditional artistry methods. Over time, it has evolved to include educational components, guiding emerging artists in incorporating AI tools into their projects. This includes curating custom AI resources and providing tutorials on its YouTube channel.
In recent years, Black Mixture has enhanced its creative workflows through GPU-accelerated features. The company currently implements advanced AI models and features running locally on systems powered by NVIDIA GeForce RTX 4090 technology, allowing artists to perform complex AI tasks efficiently.
Using generative AI, Black Mixture is able to create or enhance video footage and image assets. Recently, the RTX AI Garage has highlighted various apps and workflows the company uses, along with helpful AI tips and tricks.
Moreover, a new release of NVIDIA Broadcast version 2.0.2 is out, featuring up to 15% performance improvement for GeForce RTX 50 Series and NVIDIA RTX PRO Blackwell GPUs.
An AI for Presentation
Iteration stands as a crucial element in the generative AI content creation process, often requiring the generation of hundreds of images per project. While foundational models like ChatGPT and Google Gemini can quickly produce singular images, artists such as Nate and Chriselle Dwarika demand heightened speeds and creative control. With on-device AI tools accelerated by RTX, achieving this speed and control is now feasible.
Generating a standard 1024×1024 image in ComfyUI takes 2-3 seconds on my RTX 4090. When you’re batch-generating hundreds of assets, that’s the difference between an hour, an entire day or more.” — Nate Dwarika
Through the use of generative AI, the Black Mixture team can draw inspiration from a broader array of high-quality images compared to manual methods, allowing them to refine visuals according to their clients’ preferences.
Nate Dwarika employs the node-based interface ComfyUI in his text-to-image workflow, which offers immense flexibility in organizing nodes. This effectively combines different traits from various generative AI models, all while being accelerated by the GeForce RTX GPU.
The FLUX.1-dev model, an image generation model from Black Forest Labs, exemplifies this. Deployed within ComfyUI on an RTX AI PC, the model harnesses CUDA accelerations in PyTorch, significantly speeding up an artist’s workflow. An image that may take about two minutes to generate on a Mac M3 Ultra, only consumes 12 seconds on a GeForce RTX 4090 desktop GPU.
FLUX.1 Kontext, another offering from Black Forest Labs, allows users to generate images starting from a reference image or text prompt, guiding edits with suggestions, streamlining complex workflows without extensive fine-tuning.
The stunning dragon visual showcased below illustrates the power of FLUX.1 Kontext, enabling rapid iterations based on the artist’s initial concept, allowing for additional inputs like poses and depth maps within the same prompt window.
Quantized variants of FLUX.1 Kontext are designed to lessen VRAM requirements, enabling more users to operate it locally. These variants leverage hardware accelerations in NVIDIA GeForce RTX Ada Generation and Blackwell GPUs, offering more than double the performance due to NVIDIA TensorRT optimizations. They come in two quantization levels: FP8 and FP4, suited for different GPU series.
Dwarika also utilizes Stable Diffusion 3.5, which supports FP8 quantization for the GeForce RTX 40 Series. This support results in 40% reduced VRAM usage, nearly doubling the image generation speed.
Equipped with these visual AI tools, Black Mixture produces photorealistic, original visuals for clients both time- and cost-effectively.
Nate and Chriselle Dwarika complement these innovations with traditional content creation tools, all enhanced by RTX. For video production, Nate captures footage using OBS Studio, employing NVIDIA’s built-in encoder (NVENC) within his GeForce RTX 4090 GPU. As NVENC operates separately, resources are freed for intensive ComfyUI workflows, ensuring high-quality encoding alongside processing demands.
The final touches and edits occur in Adobe Premiere Pro, which integrates numerous RTX-accelerated AI features, such as Media Intelligence and Enhance Speech. This combination allows for rapid final exports using GPU-accelerated encoding.
Looking ahead, Black Mixture plans to launch an Advanced Generative AI Course.
Next-Generation Streaming
The recent NVIDIA Broadcast version 2.0.2 provides a 15% performance enhancement for NVIDIA GeForce RTX 50 Series and NVIDIA RTX PRO Blackwell GPUs. This boost enables the utilization of AI features like Studio Voice and Virtual Key Light on compatible GPUs.
Studio Voice enhances microphone quality while Virtual Key Light improves lighting, contributing to better visual presentations. The Eye Contact feature now offers improved stability, making interactions more engaging even when a user glances away from the camera.
Interested users can download NVIDIA Broadcast.
Each week, the RTX AI Garage blog series presents community-driven AI innovations and resources for users keen on building creative workflows, productivity apps, and more using AI on their PCs.
Join the conversation around NVIDIA AI on platforms like Facebook, Instagram, TikTok, and stay informed by subscribing to the RTX AI PC newsletter. Connect on Discord for discussions related to the latest in AI technology.
Follow NVIDIA Workstation updates on LinkedIn and X.
See notice regarding software product information.