Dopamine roulette | complete AI music video - YouTube | 2024
Daria Zorina
AI Music Producer
AI Artist
AI Content Creator
ChatGPT
Runway AI
Stable Diffusion
Production Features:
1. Challenge & Consistency:
This project was a speed challenge to test the AI’s capabilities. As a result, the video exhibits minimal consistency. I didn’t focus on ensuring precise matching of keyframes due to the emphasis on rapid execution. However, a basic level of consistency was achieved, making the video visually cohesive.
2. Generation Process:
About 60 video clips were generated from 25 keyframes. Around 20 variations of the song lyrics were created to find the right fit. The project used multiple rounds of AI generation for both video and music.
3. Main Challenges:
One of the biggest technical hurdles was the animation of the pupil turning into a card. The AI struggled with this transition, requiring manual intervention using Photoshop to refine the keyframes and further animate the sequence using multiple neural networks.
hardest shot
4. Advantages of Using AI:
The entire process, from concept to final cut, was completed in just 5 hours thanks to AI tools. The production included both horizontal and vertical formats. Without AI, this project would have taken significantly longer.
5. Timeline:
Although this project took 5 hours, similar projects for clients may take between 2 days to a week, depending on the consistency required and client approvals. Each stage, from music generation to final animation, involves multiple steps.
These videos can be produced as music videos or promotional content.
A song can be crafted to highlight your product in any language, or you can opt for an instrumental paired with a voiceover for advertising purposes.
This flexibility makes the videos suitable for various creative or commercial campaigns.
Workflow:
1. Concept Development
• Objective: To create a 2-minute cinematic video with rapid scene changes, inspired by a music video format.
• Process: The initial idea was to evoke drama and intensity through an evolving narrative, focused on dynamic visual storytelling.
2. Song and Lyrics Creation
• Step 1: Generate lyrics with GPT to match the dramatic tone.
• Step 2: Write 20 versions of the song to ensure the lyrics sync with the desired atmosphere.
3. Storyboarding and Character Design
• Step 1: Sketch key characters and location settings for scenes.
• Step 2: Create detailed keyframes for each major event within the video.
• Step 3: Finalize the storyboard, outlining the sequence and progression of the video.
4. Song Approval and Refinement
• Client Interaction: For clients, there would be a song approval step, where feedback is collected and revisions are made accordingly.
5. Keyframe and Character Generation
I created keyframes based on predefined characters and locations, ensuring that every element was consistent with the video’s narrative and aesthetic.
6. Animation and Scene Development
At this stage I started generating, using mostly runwayml, but not without other networks as well
7. Dynamic Edits
Create a rough draft of the full video with music, scenes, and effects.
Make last minute dynamic changes to enhance the storyline after assembling the draft video and overlaying the music
For example, add a last minute scene in the casino where trays of chips follow the character to add dynamics and meaning
• (Client Interaction): Send this draft to the client for feedback.
8. Final Video Assembly
Final compositing and adaptation to vertical format.
Pricing Consideration:
For a project like this:
• Base price: For this project, with a normal deadline of one week, I would estimate the cost at $880 (with the same level of consistency as the video) to $1600 for complete character consistency.
There was only one complex scene with the pupils that required additional work.
• Urgency (24-hour delivery): In cases where speed is required (completed within 24 hours), it would require quicker approvals from the client, and the cost would be double due to the rapid turnaround.
