newslounge.co · Dec 28, 2023

Good Morning. 

We're back with another edition of newslounge to wrap up the year. We hope everyone celebrating had a lovely Christmas and managed to survive their family gatherings unscathed. As 2023 draws to a close, let's take a moment to reflect on how this year has profoundly transformed the world and presented us with unexpected challenges.

Please take a moment to respond to our survey (at the bottom of this email). Your feedback is invaluable in helping us improve. Feel free to reply to this email with any comments or suggestions. :)

on Today’s menu:

• Blending Basic 3D Renders with AI

• Harnessing the Wind: A Dive Into Kim Byeong Jin's VFX Craft

• Videopoet and Midjourney v6

• The New Era of Clone Actors

• End Of The Year Thoughts

-Ardy

was this email forwarded to you? you can sign up here.

HEADLINE

Blending 3D Renders with AI to Create Stunning Visuals

In my latest project, I've been experimenting with the process of transferring the early stages of scene and environment development from a 3D software (Houdini, in my case) to ComfyUI. This allows me to explore various aesthetics and play with different visual concepts. I've utilized masks to gain individual control over each major section of the image, such as the car, ground, and pillars.

💡Here's a quick overview of my workflow for now. I plan to share more details in the upcoming issues and delve deeper into the process.

3D Rendering Phase: The process begins with rendering key CG elements - beauty, z-depth, masks. This stage forms the bedrock for AI integration.

Initial AI Generation: Utilizing the beauty pass as a foundation, I augment it with edge(generated with Canny Edge Preprocessor) and z-depth details. This composite image is then processed through a diffusion model, often enhanced with a custom LoRA based on reference photos. This step is crucial in creating a solid starting image for the AI's interpretation.

Element Refinement: The subsequent phase involves isolating and refining the main elements, like the Porsche, using the masks. By treating each element as a subgraph, I can focus on refining specific parts of the image with precision.

Upscaling and Finalization: In the final step, I upscale the image (from 1k to 4k) and add final details with prompt keywords.

This process can also adapt the AI model to different shot angles of the Porsche, ensuring stylistic consistency.

-AA

CASE STUDY

Harnessing the Wind: A Technical Dive into Kim Byeong Jin's VFX Craft

Kim Byeong Jin, inspired by a challenge from the Korean Game VFX Artist Community, starts with a blank canvas. The theme? Wind. What unfolds next is a fascinating journey from rough sketches to a full-blown battle scene, kind of like directing a mini-movie, but cooler.

The Toolbox of a Wizard:

• Adobe Substance Designer: This is where textures come to life, making sure that every swirl of wind looks as real as it gets.

• Blender: Think of it as the Swiss Army knife for 3D modeling and animation – essential and super handy.

• Embergen: This is where the magic of smoke and mirrors (well, just smoke) happens, adding that extra oomph to the visuals.

Beyond the Basics:

No typical video editing here. Kim uses the Unreal Engine’s level sequence editor like a maestro, tweaking and tuning to get every moment just right. It's all about making the impossible look easy.

Setting the Stage:

It's all about the mood. Kim plays with lights and shadows to make sure the VFX doesn't just shine, but dazzles. It's a delicate dance of brightness and darkness, ensuring every scene is picture-perfect.

Tackling the Tough Stuff:

Here's where things get real. Integrating and polishing FX elements is like solving a complex puzzle. Kim's secret? A mix of Photoshop wizardry and taking breaks to clear the mind. Sometimes stepping back is the best way forward.

Kim Byeong Jin's approach to VFX is a thrilling blend of art and science, where every detail is a piece of a larger, breathtaking picture. It's a peek behind the curtain of the VFX world, revealing the meticulous process and innovative spirit that drives this dynamic field.

🙋‍♂️Read the full interview here.

AROUND THE WORLD OF AI

Get Ready For A Fresh Wave in AI Art

Google recently unveiled VideoPoet, a large language model adept at numerous video creation tasks such as text-to-video, image-to-video, video stylization, video inpainting and outpainting, and converting video to audio, all with impressive temporal coherence.

Arc, a LoRA developed by Ismail Seleit, has been trained to generate stunning curvilinear architecture.

Midjourney v6 is here! This latest update is a big deal for anyone into digital art and AI

🔺Access: Available through Discord with a subscription plan.

🔺Cost: $10 per month for about 200 image generations.

Strengths:

• Photorealism: Tends to excel in creating more photorealistic images.

• Detail Orientation: Produces detailed and nuanced images, especially effective in photorealistic human and architectural representations.

Weaknesses:

• Cartoon Images: May not closely follow prompts in cartoon-style images.

• Text in Images: Less effective in incorporating text into designs.

Where the money is going?

Despite concerns about potential dangers, the arrival of a new AI-driven technological era has led major tech companies to compete fiercely in this field. This has made AI experts the most sought-after professionals across various industries this year.

💲💲Venture capitalists were quick to jump on the bandwagon...

In 2023, AI startups attracted a significant 26% of all venture capital funding, a substantial increase from just 11% in 2022.

The sector secured a staggering $17.9 billion in the third quarter alone, according to Pitchbook data shared with Bloomberg.

WATCH THIS

The New Era of Clone Actors

This is one of the first steps in a trend that could upend the way movies are made.

AI gives Hollywood producers the ability to capture a scan of a star and use it to do, well, pretty much anything.

As AI technology becomes more prevalent in Hollywood, actors are increasingly concerned about their digital doubles being used without their consent. This issue was a significant point of contention in recent union negotiations between actors and studios.

Bryan Lourd, CEO of the Creative Artists Agency (CAA), emphasizes the importance of individual ownership of digital presence. CAA's approach, particularly with the CAA Vault, is to ensure actors have control over their digital scans.

However, the legal landscape surrounding the use of these digital doubles is still murky. There is no federal law specifically protecting the likeness rights in this context, and state laws vary.

What are your thoughts on this?

Farewell to 2023

End Of The Year Thoughts

The year 2024 is shaping up to be a landmark year, especially in the creative industry, which is poised for unprecedented changes. As we navigate this exciting era, it's crucial to reflect on our personal goals and aspirations.

In the rapidly evolving landscape of technology, we're witnessing a significant shift. Major players like Google, Microsoft, and Apple are expected to expand their influence, potentially leading to the consolidation of smaller companies, especially those developing ChatGPT-related technologies.

A key trend to watch is the rise of on-device AI. This innovation is set to revolutionize how we interact with technology, as AI integration becomes a standard feature in the hardware of leading tech giants.

However, with great advancements come new challenges. The regulatory environment is still catching up with these technological leaps, indicating that we may see new laws shaping the future of AI and its applications.

Despite the uncertainties and the chaotic nature of the current tech landscape, 2024 promises to be a thrilling journey.

It's a reminder to stay focused on our personal growth and mental well-being amidst the flood of change. Let's embrace this new world with clarity and purpose, ensuring that we don't lose sight of our individual paths in the midst of these exciting developments.

Rate Today’s Edition