In the video, David Sheldrick outlines his structured approach to creating Sora music videos, emphasizing the importance of planning, world-building, and using tools like ChatGPT for developing detailed prompts. He details the process from creative exploration and rendering footage to assembling the final edit, ensuring that the visuals align with the music’s energy.
In the video, David Sheldrick shares his process for creating music videos, particularly focusing on his work with Sora. He begins by discussing his background in directing music videos before the COVID-19 pandemic and outlines a structured approach to filming. This method involves shooting multiple locations in a single day, ensuring efficiency and consistency in the creative process. He emphasizes the importance of planning and organization when moving from one location to another.
The first step in his video-making process is the creative phase, where he dedicates time to exploring various creative directions. David suggests using the Sora website’s explore page as a valuable resource for inspiration, allowing creators to see what others are rendering and learn effective prompting techniques. He highlights the significance of world-building, where creators define the setting and style of the video, such as whether it is a period piece or set in a fantastical environment.
To aid in the creative process, David utilizes ChatGPT to help develop detailed prompts for Sora video rendering. He presents a diagram that outlines his overarching style, which is inspired by the 18th-century Marie Antoinette aesthetic. Within this framework, he identifies several creative elements, including hair and makeup close-ups, grand ballroom shots, and thematic visuals like gardens and horses. This structured approach allows for a cohesive visual narrative throughout the video.
Once the creative elements are established, David moves on to the rendering phase, where he inputs his prompts into Sora to generate footage. He emphasizes the importance of running multiple iterations of each creative prompt to gather a variety of shots. Additionally, he incorporates a second prompt focused on dynamic dance sequences, ensuring that the visuals align with the energy of the music. For music selection, he prefers using stock music from artlist.io, which he finds to be of higher quality than AI-generated options.
Finally, David discusses the assembly of the video, where he organizes the rendered footage into a timeline. He refers to this initial arrangement as a “sausage,” which serves as a basic structure for the final edit. Throughout the assembly process, he makes adjustments to the footage, such as speeding up clips and aligning cuts with musical drops. After several hours of editing, he completes the assembly, showcasing the final product that combines his creative vision with the dynamic elements of the music.