Paweł Krzemiński, Creative Director at TELEVISOR, takes a deep dive into the company's latest AI projects.
Q> Please tell us a little about your role at the studio and working as a televisor. What do you do and what excites you most about it?
Paweł> As Creative Director of TELEVSOR, I am responsible for guiding and choosing both the style and technical tools that best help achieve the agency and director's vision.
My role includes creative contributions to the production process, working closely with the team to develop innovative concepts and visual styles. I work to refine ideas so that each project has a unique and appealing aesthetic. This includes continuous research and experimentation with new technologies and techniques to push the boundaries of what is possible. What excites me most about this work is being able to bring an imaginative idea to life and see it resonate with an audience in an impactful way.
Q> Can you tell us a little about how this project came about and where the idea came from?
Paweł> The music video for the new arrangement of the iconic Polish song has already become a traditional marketing effort for the Tymbark brand. Each time, top Polish artists are invited to perform their songs. However, this work stands out in that it incorporates the use of artificial intelligence (Al). The literal translation of the song's title is “Everything Tells Me Someone Loves Me”, but I changed it slightly to make it more accurate to the project's specifications. This innovative project was made possible by the pioneering approach of Krzysztof Grazyper, director of DDB Agency and a huge believer in TELEVISOR Studio.
Q> It's a beautiful and nostalgic song. What does that mean?
Paweł> It's about the feelings you feel when love appears in your life. Everything looks beautiful and full of magic. You know what I mean, right? Hahaha.
Q> AI – What do you think about this? This is currently a very big topic being raised in many forums by many agencies and clients. Is it well documented in the briefs you receive?
Paweł> My view of AI in our industry is cautiously optimistic. AI is certainly a hot topic, frequently discussed on various platforms and increasingly mentioned in the briefs we receive from agencies and clients.
AI tools are becoming a valuable asset not only in the creative process but also in preliminary design work. However, while AI can increase efficiency and provide new insights, it cannot replace the human element of creativity that is essential to our work. Our team's unique artistic vision, emotional depth, and innovative thinking are invaluable to each project.
In our studio, we integrate AI in ways that complement our creative efforts, to handle repetitive tasks or to generate initial concepts that can then be refined and humanized. Use AI. This allows you to focus more on the creative aspects that truly define the final output. A balance between leveraging the strengths of AI and preserving the integrity of human creativity is what we aim to achieve with the project.
Q> What was your creative process like and how did you develop the idea and script?
Paweł> The idea to use AI in the project came entirely from the agency. This puts us in a great position to focus on selecting the best visual aspects of AI currently available. The agency was already working on using these new technologies, so there was no need to convince the client. This allowed me to directly work on researching and implementing cutting-edge AI tools to enhance the visual appeal of my projects.
Q> Can you tell me what the process was like working with the agency/client on this?
Paweł> Despite having a desired visual style and overall tone for the project, creating a video with Al is an ongoing creative process. The Al software is fed enough data for the expected results, so it returns a huge range of content. Next, it needs to be processed by the artist and his team of CG supervisors. This means selecting the material and adapting it to video. This process typically requires several cycles to reach the final desired result.
Q> Which tool did you choose to generate content with AI and why?
Paweł> We created the video using an exciting fusion of Al (Stable Diffusion) and NeRF (Neural Radiance Fields) technologies. This combination not only provides great entertainment, but also opens a new chapter in the future of visual art. The Stable diffusion software allowed us to precisely control the often confusing Al results using specific tools and video materials. NeRF, on the other hand, allows you to scan 3D space after capturing a shot and transform the data into glitchy holographic vision. This innovative technology crossover has allowed us to unleash our creative imagination and create authentic, eclectic and interesting expressions of each artist.
Q> How can you explain what a nerf is in the simplest terms?
Paweł> In the simplest terms, NeRF, which stands for Neural Radiance Fields, is a technology that uses artificial intelligence to create detailed 3D scenes from a few 2D photos. Imagine you take several photos of an object from different angles. NeRF uses these photos to generate 3D models that can be viewed from any angle, even those that were not photographed in the first place. It's the magic that turns flat images into fully explorable 3D worlds.
Q> What has new technology brought to this implementation?
NeRF allows you to animate cameras independent of their original recording position. This allows for smooth, dynamic camera movement within the scene, allowing for unique perspectives and cinematic shots not possible with traditional methods. A 3D representation of a scene allows you to explore different moments and perspectives within the same timeline, giving you a more flexible and comprehensive understanding of the action.
Q> Are you now able to do things that you couldn't do before?
Paweł> Creative exploration at a fraction of the cost. Stable diffusion lets you explore multiple creative paths from a single text prompt. For example, a prompt that describes a “city of the future” can generate a variety of interpretations, each with its own unique architecture, lighting conditions, and atmosphere. This allows creators to consider a wide range of visual ideas and choose the ones that appeal to them the most.
Q> What technical challenges did you face?
Paweł> The technical challenges ranged from ensuring high-quality input data to optimizing model training. The integration of AI has dramatically changed the post-production process. Rather than traditional post-shoot editing, we created a basic moving mock-up before shooting. AI generates large amounts of footage and provides multiple solutions for each scene. I cut the clip to the length I needed and chose the version that worked best for me. This hybrid approach maximized creativity and efficiency, resulting in high-quality visual content.
Q> Based on your experience implementing AI in your projects, how do you see the future as a studio in terms of the presence of AI?
Paweł> Creating a video with the help of Al might be a shortcut. I couldn't agree more. Al offers new possibilities and a completely different approach, but that doesn't mean less work. Creating videos with Al is not a shortcut, but it is a powerful transformational tool for idea creation and production. Creating a truly unique video requires a team of experts, their knowledge, and skills. I'm an optimist. I don't see this as a threat, but an opportunity for our industry to take on a completely new challenge.
