The hardest problem in production image editing isn't making a new lo… — Dheepan Ratnam · VideoMakers

Dheepan Ratnam@Dheepanratnam on XHire DheepanView original post

The hardest problem in production image editing isn't making a new look. It's changing one element on a 2560×1440 plate without redoing the rest. I ran Ideogram 4.5 Precise Edit across three real studio workflows at native 2560×1440 resolution its Amazing 🔥 More details 👇

Ideogram 4.5 Precise Edit tests: SKU localization, a 1971 cinema period conversion, and an 8K plate upscale.

398 views
8 likes
1 reposts
1 replies
0 bookmarks

Posted Oct 29, 20260:30Landscape · 1920×1080

Animation Kids & family

Tool test Screen recording Tech & gadgets Fashion & beauty Cities & architecture

This video is Dheepan Ratnam’s work, shown from their public post. See it on X.

More from Dheepan Ratnam

Testing Seedance 2.5 Video model Blending Bollywood with explosive Brazilian Carnival Samba across Lisbon, Seville, Jaipur, and the Taj Mahal. Seamless motion, timeless frames, and zero cuts. 🎬 Seedance 2.5 @ 720p single 30s shot next - Afro-Arabian Oasis

3.1KX

cinematic rhythm in white and gold ✨ kayadu 🤍 Seamless choreography, timeless frames, and zero cuts. AI cinema just keeps getting better. 🎬 seedance 2.5 @ 720p looks fire 🔥 single 30s shot

3.9KX

Sunset beach → Art Deco ballroom → sandstone palace—one classical Bollywood fusion, three worlds. Seedance 2.5 turns text (plus 1 reference image) into AI choreography across 3x15s generations. Sound ON 🔊👇 is this Sara Arjun?

4.4KX

Testing Seedance 2.5 on a full South Indian group dance sequence 🎬✨30s continuous single-shot 4K cinema look. The way it maintains character consistency a while matching the beat is next-level

29KX

Picture this: a misty temple, lanterns glowing, and a choreography that fuses Bollywood’s cinematic drama with the bold artistry of Japanese Kabuki finished with the razor-sharp elegance of a folding fan dance. 🔥 Sunday vibes

19.1KX

Someone asked if Ai can speak tamil here we go AI video just learned Chennai Tanglish one single take Speaking natural Chennai Tamil/Tanglish on a sunlit film set with every eyebrow lift, hand gesture, and breath pause locked to the millisecond. what do you think? Nano banana image + seedance

10.6KX

Is this Ai? Testing 30-second single-take character consistency with seedance 2.5 uncut version 👇 Look at how far we've come on skin texture, room acoustics, handheld camera walk-out to the balcony, and micro-expressions from 0:15–0:30. Only giveaway left? Watch 0:07 when she goes to pour tradi

66.6KX

30 seconds. One continuous single-take. 4K cinematic. Bollywood group routine. 🤯 I just tested Seedance 2.5 for a dance sequence and the results are wild timing, footwork, and camera motion… all generated from one prompt with music. If you’re curious how close AI video is getting to What should

18.2KX