// BLOG · VIA CACHEPHOTO MAGAZINE

FROM THE MAGAZINE.

Post-production and AI: the latest stories from Cachephoto Magazine, Daniel Blanco A.M.C.’s digital magazine on cinematography, photography and AI. Each story opens on cachephoto.com.

AI · Post · Sep 12, 2026

Higgsfield inside Blender: you move the camera yourself, with your phone

Higgsfield's free Blender plugin connects Claude, 3D models and Seedance 2.5 in one pipeline. You record the camera move with your iPhone and the video comes out with that operating. No more crossing your fingers over a prompt.

Read on Cachephoto ↗

AI · Post · Sep 12, 2026

Gemini Omni Flash: swapping the background and the actor of a real shot just by describing it

Google's video model inside Higgsfield rebuilds your shot around what you ask for — no roto, no tracking, no green screen — and Nano Banana 2 Lite makes the references for under 50 centavos an image. What I use it for, and what I don't.

Read on Cachephoto ↗

AI · Post · Sep 12, 2026

Higgsfield Genjutsu: changing wardrobe in post with no roto, no tracking, no reshoot

Video-to-video that keeps your camera, light, grain and cut, and changes only what you ask. I tested it on my own FX5 footage: a jacket, the weather, race cars and tattoos. This doesn't look fake anymore.

Read on Cachephoto ↗

AI · Sep 12, 2026

Seedance 2.5 now generates in 1080p: what changes versus 2.0 and what each shot costs

A month ago I wrote that 720p wasn't for professional work. 1080p landed on Higgsfield and I compared it against Seedance 2.0 with my own generations: skin texture, focus that arrives late like a real focus puller, and 135 credits for 15 seconds.

Read on Cachephoto ↗

AI · Sep 12, 2026

Seedance 2.5 from scratch: character sheets, prompts with Claude and 30 seconds without a cut

My complete workflow for Seedance 2.5 on Higgsfield: turnarounds in GPT Image 2, a prompt system trained on your project, extend video shot by shot, and what happens when the model over-directs.

Read on Cachephoto ↗

AI · Post · Sep 12, 2026

Higgsfield Genjutsu: video-to-video that swaps wardrobe, weather and cars without a reshoot

Higgsfield's video-to-video model changes only what you ask for and keeps the camera, light, grain and cut of your original take. I've already tested it on my own FX5 footage.

Read on Cachephoto ↗

AI · Aug 8, 2026

Seedance 2.5 now has a real cost-per-clip breakdown

Seedance 2.5 can now be compared by real cost: an 8 second 720p clip can run from $2.55 to $5.76 USD depending on platform.

Read on Cachephoto ↗

Post · Aug 5, 2026

Sony FX5: X-OCN opens in Resolve and Premiere Beta

The Sony FX5 post workflow is coming together: DaVinci Resolve 21.0.4 imports X-OCN and Adobe adds beta support.

Read on Cachephoto ↗

AI · Post · May 17, 2026

HOOKS: I ran Higgsfield's Hook Generator for a week

A week with the Hook Generator in Higgsfield's Marketing Studio. How useful it actually is for creators and brands, and what happens when you ask it for volume without losing your voice.

Read on Cachephoto ↗

AI · Post · May 16, 2026

Netflix expands its VOID AI model to more studios

Netflix's generative tool opens up to external productions. Implications for post and for jobs in LATAM.

Read on Cachephoto ↗

AI · Post · May 15, 2026

Higgsfield Canvas: from moodboard to final render

Higgsfield Canvas as a previs tool: how to chain moodboard, art direction and final render without the DP giving up creative control at any step.

Read on Cachephoto ↗

AI · May 14, 2026

Higgsfield MCP + Claude: the full review

Connecting Higgsfield to Claude via MCP. What gets unlocked for audiovisual production when the model can orchestrate tools instead of just generating.

Read on Cachephoto ↗