Keep It Bright (Official Music Video)

My second music video work. More character consistency and cinematic. Realism, enjoyable vibing music, well lip-synced. Balanced editing matching lyrics, character, environment. I'm proud of it. 😊

Comment 1

Inspiration

This is my second music video right after "Mood Ring Queen". I was creating a somewhat theme of this character and thought of a scenario where this character is leaving her home away from friends and family for higher studies. The girl has sadness in heart for leaving her loved ones but she's that cheerful person hides her sadness and tries to make her loved one feel happy. The lyrics and thought extended to reflect just that and portrayed in this music video. This time it did no take me full 1 month to make it, but about 10 days, roughly.

What it does

Resonates well with the music, lyrics, expressions, and visually entertaining.

How I built it

  1. The main character, face was initially produced on Grok Imagine.
  2. Different variations of character images were created using Flux Kontext, Qwen-Image-Edit, Seedream 4 Edit, Nano Banana, Higgsfield Soul, ImagineArt.
  3. Trained this character model lora for Qwen-Image using tools like Ostris.
  4. Generated storyboard using Grok.
  5. Generated keyframe images for image-to-video using Wan 2.5 as main model on Higgsfield AI. 5.1 Used Midjourney, Qwen-Image, ImagineArt for initial images. 5.2 Then image-to-image generation with character lora on Qwen-Image model, to have matching character face. 5.3 Upscale to near 4k resolution with high crips detail using custom workflow in ComfyUI using models like Wan 2.2 + Qwen-Image.
  6. Fed as image-to-video models for different models on Higgsfield AI. Models used:
  • Alibaba Wan2.5

Challenges I ran into

This time the hurdles were not as much since my last video "Mood Ring Queen" Wan 2.5 worked flawlessly as solution. No other models I tried came close to Wan 2.5 for lip-synced video.

Accomplishments that I'm proud of

Having a well edited video which is not a simple a stitched video of multiple visuals only video. I edited the video in Final Cut Pro heavily. Generate multiple shots with different angles, and situations to have rhythemic flow of the video. Kept extra audio generated from videos to keep the vibes of music video alive, which are not initially from audio song, but from video generated using models like Wan 2.5.

Little different in this video is using Qwen-Image-Edit with Multi-Angle Lora. The entire car scene was not possible without using specifically that LoRa. It's extremely helpful. It's my preferable model because it's open source and I can run locally keeping my budget unhurt.

What I learned

Learned better process of making a video. Storyboard, scripting scenes, sticking to selective models and process helps efficiently build on work and the outcome comes out better too. Enhanced prompting for creating AI music using tools such as Suno 4.5+.

What's next for Keep It Bright (Official Music Video)

Full album is already released on all major streaming platforms. Here's the link: https://distrokid.com/hyperfollow/gladiskatzer/glitch

Currently working on "Purr 4 Me" music video. I had released a short preview version with less work. But the full version will have all the bells and whistles of a professional music video. I'm giving all the knowledge I gained on to this next music video.

Link: https://www.youtube.com/shorts/qhwtTIi4TPY

Built With

  • fcpx
  • imagineart
  • midjourney
  • qwen-image
  • suno
  • wan2.5
Share this project:

Updates