>>4495920>>4495919I asked Grok cause its been decades since I've done this kind of calculations. Feel free to check it by yourself. Here's relevant bits.
"how big in mm dimensions a 100 megapixel 3:2 sensor had to be assuming diffraction limited f1.2 lens of focal length same as sensor diagonal in mm"
Horizontal width = 12,909 × 0.00066 mm ≈ 8.52 mm
Vertical height = 8,606 × 0.00066 mm ≈ 5.68 mm
Diagonal = √(8.522 + 5.682) ≈ 10.25 mm
"what would be the practical ISO of sensor with such small pixel size and Bayer filter?"
Base / native ISO — roughly ISO 25–50 (very low to keep read noise from dominating tiny signals).
Practical clean daylight ISO — up to ISO 100–200 at best (still noisy at 100% view).
Usable low-light ISO — ISO 400–800 would already look quite bad (heavy chroma noise, loss of color accuracy, posterization in shadows).
...
In short: Without aggressive binning + computational tricks, this sensor would feel like using a ~2012–2015 compact camera at high megapixel modes — base ISO ~50–100, really struggling above ISO 400. With all modern phone-style processing turned on, it could deliver surprisingly usable results up to ~ISO 3200–6400 in brightened low-light scenes, but only at heavily binned resolutions (6–25 MP) and with typical AI-smoothed look (reduced fine detail and texture).This is exactly why no one builds 100 MP sensors this small for "photography-first" cameras — the pixel-level performance would be too compromised without relying on smartphone-style heavy processing.
