Chand Raat never really looks the way you imagine it in your head. It’s louder, more crowded, slightly messy… and somehow better because of that.
I wasn’t trying to create those overly polished AI images you usually see. You know the type perfect lighting, flawless faces, everything staged. That doesn’t feel like Chand Raat at all. So I tried something different this time.
The idea was simple: keep everything a little imperfect. Slight blur, awkward angles, people moving in the background. The kind of photos you’d actually take without thinking too much.
Here’s how it turned out.
🌙 Mehndi Stall: First Stop
Prompt Used:
9:16 raw handheld selfie: Subject A (male, reference image 1) and Subject B (female, reference image 2) at a mehndi stall, Subject B seated getting henna applied while Subject A leans in for a spontaneous selfie; keep both faces exactly the same, Subject A in a maroon kurta pajama (no glasses), Subject B in a deep red kurti with jeans, bangles, earrings, mehndi on hands; natural smiles, slight motion blur, busy stall with people and vendors behind, warm small lights, realistic skin texture, no filters or cinematic effects.
🌙 Somewhere Between Stalls (Gajra Moment)
Honestly, this wasn’t even meant to be a “main” shot. It’s the kind of moment you’d usually ignore but it ended up feeling more real than the rest.
Prompt Used:
9:16 raw candid shot: Subject A (male, reference image 1) holding a jasmine gajra while a vendor ties another on Subject B (female, reference image 2)’s wrist; Subject A in maroon kurta pajama (no glasses), Subject B in deep red kurti with jeans, bangles and mehndi; natural handheld feel with slight motion, busy stall with flower baskets, petals and people behind, warm lighting, real skin texture and fabric details, no filters or cinematic effects.
🌙 Ending Where Everyone Ends Up: Food Stall
Prompt Used:
9:16 raw candid shot: Subject A (male, reference image 1) and Subject B (female, reference image 2) eating gol gappay at a busy Chand Raat stall, vendor preparing puris behind; keep both faces exactly the same, Subject A in maroon kurta pajama (no glasses), Subject B in deep red kurti with jeans, bangles, earrings, mehndi on hands; natural candid expressions, slight motion blur, people and steam in background, bright stall lights with steel containers and stacked puris, realistic skin texture and fabric details, no filters or cinematic effects.
🧠 What Actually Worked (After Some Trial and Error)
At first, the images looked too good. Which sounds weird, but that’s the problem. Everything felt staged like a photoshoot pretending to be candid.
So I had to dial things back.
One thing that made a big difference: removing cinematic language completely. No dramatic lighting, no “perfect composition.” The moment I stopped trying to make it look like a movie scene, it started looking more like a real phone photo.
Another thing adding movement. Not too much, just enough. Slight blur in hands, people passing behind, fabrics not sitting perfectly. Real life isn’t still, especially not on Chand Raat.
And honestly, the background matters more than you think. Random details like tissue papers, baskets, people waiting… those things carry the realism. Faces alone won’t do it.
Also, this might sound small, but explicitly telling the model what not to do helped more than expected. No filters, no HDR, no studio lighting. Without that, it kept drifting back into that overly processed look.
✨ Final Thoughts
If the image looks perfect, it probably won’t feel real.
Chand Raat isn’t neat. It’s crowded, a little unplanned, sometimes even uncomfortable… but that’s exactly why it works.
And once you stop trying to “perfect” the output, the results start feeling less like something generated and more like something that actually happened.
Try Google Gemini Here:
This blog post and AI prompts were created by Shahbaz Ahmad.
Follow me on TikTok @Dudefrompak for more ready-to-use prompts.
📢 Join Our WhatsApp Channel
Get daily AI photo editing prompts, tools, and tips directly on your phone.
Join Now on WhatsApp 🚀
