I've been working on a 3D render of Mac Miller's iconic mirror used in the album artwork for "The Divine Feminine". I wanted to challenge myself to make close up renders in Blender to accompany a different camera angle that I'm deciding to animate. I'm torn because I LOVE the depth of field on the flowers, but I wonder if it might be too much. What are your thoughts?
This is stunning work, Jerry! The craftsmanship on Mac Miller's mirror is spot on — it's such an iconic piece. My vote is for "tons of flowers" — the depth of field makes it feel alive and cinematic, which perfectly matches the dreamy atmosphere of The Divine Feminine. Can't wait to see the animation!
The depth of field looks incredibly premium, and if you’re aiming for a conversion-centered aesthetic, those blurred flowers act as a perfect natural frame to keep the focus exactly where it belongs—on the mirror.
Process note:
🌑 Echo is a black screen where every sound sends a wave of light through the scene in real depth.
🎞️ Starting footage: one licensed stock clip of a temple corridor from invideo's stock library. The sound is original, synthesised for this piece. The voiceover is my own.
💬 What I asked for: a black screen where each ping lights the scene by distance, built from a depth map, with controls for each wave, afterglow, wave thickness, contour lines, outlines, colour and a final reveal.
🛠️ What was built: a depth map (Depth Anything Video), two "Echo" custom effects on the depth map (a Multiply layer for light and an Add layer for glow), an "Echo colour" effect on the footage, and keyframes for six waves and the reveal. The first build was done by Claude, connected to invideo Editor through invideo's MCP server.
🎚️ Refinement: I asked invideo's own agent (Agent Two Lite) in the editor to raise Afterglow from 0.14 to 0.3 on both Echo effects. It made the change on the timeline, and that conversation is shown in the video. I also tried a longer afterglow and a gold colour, both shown.
✅ All building and the final cut were done in invideo Editor.
Generative image and video models don't keep a face: the same woman drifts from one shot to the next, and garments lose the details that make them recognisable. on-model-pipeline generates fashion stills and clips with a fixed cast.