AI Technology Transforms Creative Production
AI technology has advanced dramatically in recent years, bringing sweeping changes to the creative industry. Particularly noteworthy is Sora 2, an AI model that generates high-quality video from text prompts. Video production — which once required specialized knowledge, expensive equipment, and significant time — is becoming more accessible and efficient thanks to Sora 2. However, simply entering a prompt is not enough to produce professional-level work with this powerful tool. A strategic approach and systematic workflow are essential.
This article details the specific video production techniques developed by acclaimed DP (Director of Photography) and director David Sheldrick, based on his own music video production experience using Sora 2. He applies a format originally designed for efficient single-day shoots — honed before COVID-19 — to the new tool of Sora 2, systematically building high-quality video content. The process spans from creative ideation, through rendering and music selection, all the way to meticulous editing. Through Sheldrick's practical expertise, we explore how to maximize Sora 2's potential and develop video production skills applicable to business contexts. This is essential reading for business professionals interested in new forms of visual expression in the AI era.
Unleashing Creativity: The First Step in Sora 2 Video Production — From Ideation to World-Building Gathering Ideas and Learning Prompts via the Sora 2 Explore Page Defining the World: From Period Setting to Visual Elements ChatGPT Integration: Refining Prompts and Creating Style Presets Systematizing Multiple Creatives: The Marie Antoinette Style as a Concrete Example The Execution Phase: Optimizing Sora 2 Rendering and Music Selection Applying Style Presets and Efficient Rendering Strategy Adding Dynamism: Inserting Dance Sequence Prompts The Importance of Music: Tips for Music Selection with Artlist.io Bringing It to Life Through Editing: From Sora 2 Assembly to Final Polish Building the Timeline from the "Sausage": Organizing Materials and Basic Structure Syncing with Music: Cutting to the Beat and Timing Adjustments Speed Adjustments and Transitions: Techniques to Expand Visual Expression The Reality of Editing Time: The Path to a Completed Assembly Conclusion
Looking to optimize community management?
We have prepared materials on BASE best practices and success stories.
Unleashing Creativity: The First Step in Sora 2 Video Production — From Ideation to World-Building
The journey of creating video with Sora 2 begins with building a strong creative foundation. This is not a casual brainstorming session — it is an extremely important phase that determines the quality and direction of the final work. As David Sheldrick emphasizes, before entering serious assembly work, you should invest at least half a day, and ideally a full day, exploring and experimenting with various creative directions. Trial and error at this early stage is the key to smoothly progressing through later steps and producing more refined results.
Gathering Ideas and Learning Prompts via the Sora 2 Explore Page
One of the best ways to find new creative inspiration is through the "Explore" page on the official Sora 2 website, which Sheldrick highlights. This page showcases diverse videos generated and shared by users worldwide, along with the prompts used to create them. It serves not only as a vast visual library for discovering ideas, but also as a practical learning resource for crafting effective prompts. By observing what words other users employ to achieve specific visual expressions or moods, you can rapidly improve your prompt engineering skills. You can learn through real examples how specific descriptors like "cinematic shot," "detailed skin texture," or "dynamic camera movement" affect the quality of generated footage. This research process is indispensable for finding the right style and expression for your project and acquiring the language needed to realize it.
Defining the World: From Period Setting to Visual Elements
Running in parallel with idea research is the definition of "World Building." Before drilling into the details of individual scenes (Location 1, Location 2...), you first need to clarify the overall atmosphere, style, and period setting that will permeate the entire work. This forms the foundation for establishing the tone and manner of the entire video and delivering a consistent experience to viewers. For example, is the video a historical period piece, a futuristic sci-fi story, set against a pastoral landscape, or inside a grand palace? Defining these fundamental settings naturally determines the direction of specific visual elements such as costumes, art direction, props, lighting, and color design. In Sheldrick's example, the clear theme of "18th-century Marie Antoinette" serves as the core around which all subsequent creative development unfolds. This world-building process is critically important not only for determining visuals, but also for deepening narrative and thematic elements and giving the work its own distinct identity.
ChatGPT Integration: Refining Prompts and Creating Style Presets
To translate a defined world into concrete instructions that Sora 2 can understand — that is, into detailed prompts — Sheldrick recommends leveraging large language models (LLMs) like ChatGPT. By entering basic ideas and keywords into ChatGPT with instructions like "Expand this prompt and make it more detailed for use in Sora 2 video rendering as a preset," you can efficiently generate richer, more specific prompts. For example, a simple concept like "18th century, Marie Antoinette style" can be expanded into a detailed prompt incorporating elements such as fabric textures (silk, lace, brocade), a color palette (pastels, gold accents), lighting atmosphere (soft natural light, chandelier sparkle), camera work (elegant dolly shots, close-ups), and specific emotions or moods (decadent, romantic, whimsical). The detailed prompts generated this way are saved and used as "style presets" in Sora 2, becoming a powerful tool for maintaining a consistent visual style throughout the project.
Systematizing Multiple Creatives: The Marie Antoinette Style as a Concrete Example
Once an overarching style is established, the next step is to define and systematize multiple specific scenes and elements — the "creatives" — within that world. This is similar to pre-planning locations and shot types in music video production. Sheldrick believes this structured approach is equally effective when using Sora 2, and presents the following concrete examples of "Marie Antoinette" style creatives. This structuring is arguably one of the most important aspects of Sheldrick's workflow.
Creative 1: Hair & Makeup Close-up — A shot focused on an enormous 18th-century-style wig, white powder makeup, and the model's expression.
Creative 2: Shots Inside a Grand Palace — Filming in spacious hallways and ballrooms, capturing architectural style and lavish interior decor.
Creative 3: Hunting Scene — Depicting the noble pastime of the era, including horses, period costumes, and natural scenery.
Creative 4: Garden — Envisioning perfectly manicured hedge mazes and geometric gardens reminiscent of Hampton Court Palace in London.
Creative 5: Horses — Horses as an aristocratic motif, focusing on elegant movement and detailed harness work.
Creative 6: Kintsugi Model — An original visual inspired by "kintsugi," the Japanese art of repairing broken pottery with gold, which also functions as a metaphor for giving old things new value.
By placing multiple specific creative elements under one large style umbrella, you can maintain overall consistency while generating diverse shots. Each creative has its own prompt, but is situated within a unified world through the overarching style preset. This systematic approach streamlines rendering work in Sora 2 and simplifies material management during the later editing process.
The Execution Phase: Optimizing Sora 2 Rendering and Music Selection
Once the creative direction is set and the world and specific scene elements have been systematized, the next step is the "rendering" phase — actually generating footage with Sora 2. At the same time, music selection, an indispensable element for bringing any video work to life, plays an important role at this stage. David Sheldrick's workflow demonstrates specific strategies for progressing through these processes efficiently and effectively. Rather than simply entering prompts and waiting, the work requires leveraging presets, running multiple renders, and keeping music integration in mind.
Applying Style Presets and Efficient Rendering Strategy
The detailed style prompts created with the help of ChatGPT are saved and managed using Sora 2's preset feature. As Sheldrick demonstrates, you can register a prompt as a style preset simply by pasting it in from the "Manage Presets" menu. This preset serves as the foundation for maintaining a consistent visual tone throughout the project. For example, once you set a preset containing elements like "18th century, Marie Antoinette style, pastel colors, soft lighting, cinematic," you no longer need to write out those details every time you enter a prompt for an individual creative (hair & makeup, palace, garden, etc.).
When executing renders, it is important to "render each creative multiple times." Because AI generation involves probabilistic elements, the same prompt produces slightly different results each time. For each creative (e.g., the hair & makeup close-up), Sheldrick applies the style preset, enters a basic scene description prompt (e.g., "close-up of a Korean K-pop model getting her hair and makeup done"), and runs this repeatedly. This collects a large number of video clips, some close to the intended image and others with unexpectedly attractive variations. This "cast wide and pick the best" approach is an efficient material-gathering strategy that takes advantage of the characteristics of AI-generated content. You can then select the best clips from what was generated, or combine multiple clips for richer, more dynamic visual expression. Presets are also not fixed — you can fine-tune them as you review rendering results.
Adding Dynamism: Inserting Dance Sequence Prompts
A characteristic feature of Sheldrick's workflow is the frequent insertion of a "second prompt" about dance sequences in addition to the basic scene prompt. This is a technique for adding movement and energy to the generated footage. For example, even in a hair & makeup scene, alongside basic descriptions like "wearing a huge 18th century Marie Antoinette wig, white powder makeup," he adds prompts specifying diverse dance styles and camera work: "bold camera shot of ethnically diverse K-pop couture fashion while dancing in unison, dancing in a Queen's bedroom, crunk dancing, street dance, dancing with attitude, dynamic dance, movement, dynamic music video camera work." The intent is to add unexpected dynamic movement and expression even to scenes that might otherwise be static, creating the kind of energy found in music videos. By specifying concrete genres like K-pop, crunk, and street dance, or adding abstract directions like "attitude" and "dynamic movement," Sheldrick encourages diverse interpretations from Sora 2 and enriches the variety of generated footage.
The Importance of Music: Tips for Music Selection with Artlist.io
In video production, music is not mere background noise — it is an extremely important element that determines the overall mood, rhythm, and emotional impact of the work. Sheldrick notes that AI-generated music quality is still developing, and recommends using high-quality stock music platforms. His personal favorite is Artlist.io, which he praises for the quality of its tracks.
Music selection typically happens once a reasonable amount of rendered footage has been gathered, or during the early stages of editing (assembly). In Sheldrick's approach, he first places a selected track on the editing timeline and then edits the footage using the music's structure (intro, verse, chorus, bridge, outro), rhythm, and build-ups (drops) as a blueprint. The music serves as the "blueprint" for video editing. Platforms like Artlist.io allow you to search for tracks by genre, mood, tempo, instruments, and more, making it efficient to find music that matches the project's world and intended atmosphere. For example, for an "18th century, Marie Antoinette" project, candidates might include tracks that blend classical and modern beats, electro-swing, or grand, dramatic orchestral music. The selected music becomes the guideline that determines cut timing, scene length, and transition timing, so careful and strategic music selection is essential.
Bringing It to Life Through Editing: From Sora 2 Assembly to Final Polish
Once the creative direction is solidified, diverse footage has been gathered through Sora 2 rendering, and the music that will form the heart of the work has been selected, the final stage of the process — "editing" — begins. In this phase, separate video clips and music are combined and elevated into a single cohesive work. The process Sheldrick calls "Assembly" goes beyond simply arranging materials; it demands numerous creative decisions regarding sync with music, the creation of rhythm, and visual storytelling.
Building the Timeline from the "Sausage": Organizing Materials and Basic Structure
As the first step in editing, Sheldrick uses a technique he calls the "Sausage." This refers to placing all generated video materials in a single row on the editing software timeline. At this point, the goal is not to worry much about sync with music or timing of cuts, but simply to gain a bird's-eye view of all available materials. Seeing this "sausage" state of the timeline lets you visually confirm what variations exist for each creative (hair & makeup, palace, garden, etc.) and how much material is available in total.
Next, based on the creative structure defined earlier (Creative 1, 2, 3...), materials on the timeline are roughly organized and grouped. For example, clips related to hair & makeup, shots inside the palace, and garden shots are each gathered together. This makes the subsequent detailed editing work significantly easier to manage. As Sheldrick demonstrates, the creative structure serves as a guiding framework for the overall composition during the editing stage as well. You begin to consciously think about which creative goes in the intro, which scenes to show at the build-up, and so on. This initial organization and structuring is the key to efficiently handling the large volume of AI-generated materials.
Syncing with Music: Cutting to the Beat and Timing Adjustments
Once the music track is placed on the timeline, the work shifts to cutting and arranging footage to match the music's rhythm and progression. This is one of the most important techniques in video editing and is essential for enhancing the viewer's emotional engagement. Sheldrick demonstrates paying particular attention to moments where the rhythm is emphasized, such as "bass hits" and "drops" in the music, and matching cut timing to those moments.
For example, suppose you want to place a shot of the model opening her eyes to coincide with the moment when the bass enters from the quiet of the intro. The clip generated by Sora 2 may not be exactly the right length for that moment. If the clip is too long, trim the unnecessary portion. In Sheldrick's example, he precisely adjusts the start point of the clip so the model's eyes open at the exact moment of the bass hit. He also demonstrates splitting a clip at any point where it contains an unintended jump (which occasionally appears in Sora 2 generations) and deleting the unwanted portion. This work of carefully listening to the music's beats, melodic lines, instrumental fills, and lyrical content, then determining cut points and scene transition timing to correspond with them, forms the core of the editing process.
Speed Adjustments and Transitions: Techniques to Expand Visual Expression
In addition to straight cutting, adjusting the playback speed of video clips is a highly effective technique for syncing with music. Sheldrick demonstrates speeding up clips (accessible via Command + R in Final Cut Pro on Mac) to match a short music phrase or beat. For example, if a 3-second music phrase has only a 5-second video clip, accelerating the clip to approximately 167% (5/3) makes it fit the length of the music exactly. Conversely, slow motion can be used to emphasize specific movements or create dramatic effects.
Speed adjustment is an important tool not just for matching duration, but for controlling the rhythm and energy of footage. Combining fast cuts and slow motion creates a dynamic sequence with varied pacing. Scene transitions also greatly influence the overall flow and atmosphere of a work. Using not only simple cuts but also fade-ins/outs, dissolves, wipes, or more creative effects in transitions can facilitate smooth visual flow or evoke specific emotions. Sheldrick's example focuses primarily on cut editing, but in practice these techniques are used in combination.
The Reality of Editing Time: The Path to a Completed Assembly
Sheldrick indicates that the assembly (initial editing) process takes a considerable amount of time. What he demonstrated covered only a few scenes from the beginning, and he notes that even going through the full "sausage" state, placing all clips to match the music, and completing a basic cut structure takes "an hour or two." He further reports that completing the full assembly ultimately took him about 4 hours in total. This varies depending on the volume of rendered materials, the complexity of the music, and the quality level required by the editor — but it demonstrates that even though AI generates the video materials, human creativity and time are still necessary to assemble them into something meaningful.
The editing process is not a mechanical act of stitching materials together. Which shots to choose, what order to place them in, when to cut, and how long to hold each — these countless choices greatly influence the final impression of the work. As Sheldrick demonstrates, following the defined creative structure while flexibly adding scenes to match the music's progression and adjusting cuts with rhythm in mind — it is through this process of trial and error that Sora 2's generated video fragments are transformed into a powerful work with a unified message and emotion.
Conclusion
The Sora 2 video production process demonstrated by David Sheldrick is a practical approach that masterfully blends cutting-edge AI technology with the systematic workflow cultivated through traditional filmmaking. His method does more than just show how to use Sora 2 as a tool — it illuminates the universal principles of producing high-quality video content: (1) building a clear vision and world, (2) structured creative development, (3) meticulous coordination with music, and (4) creative trial and error during the editing phase.
The first half of the process — researching on the Sora 2 Explore page, using ChatGPT to refine prompts, setting style presets, breaking down individual creative elements, and gathering materials through multiple renders — is a refined strategy for efficiently converting ideas into concrete footage. In particular, the structured approach of defining an overarching style and placing individual creative elements under it is highly effective for controlling the diversity of AI-generated content and producing a work with consistent coherence.
The latter half's editing techniques — building the "sausage," cutting to music beats, and adjusting speed — are essential steps for infusing human sensibility and rhythm into AI-generated materials, elevating a mere list of video clips into a story that speaks to the emotions. Using high-quality stock music from platforms like Artlist.io is key to raising the overall quality of the work and achieving a professional finish. The more than 4 hours Sheldrick spent on assembly suggests that while AI automates and streamlines parts of the production process, human creativity and time investment remain indispensable for guaranteeing final quality.
David Sheldrick's approach to Sora 2 is applicable not only to music video production but also to a wide range of business video needs: corporate promotional videos, product introductions, branding content, short-form social media videos, and educational materials. High-quality visual expression, once difficult to achieve due to time, budget, and expertise barriers, becomes more accessible by combining Sora 2 with a systematic workflow. In particular, focusing upstream efforts on concept development and world-building — then leveraging Sora 2 as a powerful visualization tool — can greatly expand the scope of marketing and communication strategies.
AI video generation technology, including Sora 2, is still evolving, but its pace of advancement is breathtaking. In the future, we can expect more intuitive operation, improved precision in controlling generated footage, and advances in AI-driven music generation and editing assistance. By learning from pioneering efforts like those David Sheldrick has demonstrated, we can ride the wave of this technological revolution and advance toward a future where we use AI as a creative partner. The key is not to be driven by the tool, but to wield AI with a clear sense of purpose and a strategic approach — maximizing the creativity that is uniquely human.
Reference: https://www.youtube.com/watch?v=0dhX84UkwFs
Streamline Event Management with AI | TIMEWELL Base
Are you struggling with large-scale event operations?
TIMEWELL Base is an AI-powered event management platform.
Proven Track Record
- Adventure World: Managed Dream Day with 4,272 attendees
- TechGALA 2026: Centralized management of 110 side events
Key Features
| Feature | Impact |
|---|---|
| AI Page Generation | Event page ready in 30 seconds |
| Low-Cost Payments | 4.8% transaction fee (industry-lowest level) |
| Community Features | 65% of attendees continue engaging after events |
Feel free to reach out for a consultation on streamlining your event operations.