Limited by Dynamic Range and Latitude

Assume the full well capacity for dual gain sensors are $FWC_1, FWC_2$ respectively. And pre-amp noise are $N_1, N_2$ photons. Post-amp noise are neglegible for modern sensors.

Assume $S$ photons are collected. SNR_1 = P ISO 100 200 400 amplifications are 1, 2, 4 W2/W1 Since it’s pre-amp noise then SNR only depends on S/N, which means N_1 > N_2.

We also notice that the noise floor is much cleaner given the same shutter speed, thus N_1 * 1 * 4(to compensate exposure) / FWC_1 > N_2 / FWC_2 * 4 FWC_2/FWC_1, which means the larger the FWC the noisier. And this inequality does no say anything useful.

If quantization error is smaller than noise, then DR is FWC/N = SNR at 100 white. 18% grey SNR then can be calculated as such. Otherwise the DR is limited by quantization. However, we know that DR is lower in second gain which means FWC_1/N1 > FWC_2/N2

s3 slog3 640*20

SLog3 ire41/50 ire95max

car log: take pictures mechanical shutter 5ms = 1/200 A7R4 14mm 1.8 shutter app

![[Pasted image 20231210180712.png]] ![[Pasted image 20231210180809.png]]

1/8000 f/5.6 iso100 ![[Pasted image 20231230124314.png]]

1/8000 f/22? 16x 4ev

moon veil: A7R4 apsc 400mm long fl f5.6 capture moon texture 1/200 A7CII 135mm 1.8 or 85mm 1.4 capture cloud 1/200 16x 4ev