>> |
Anonymous
>>265581 >>265585
I posted that giant shitlong post (2 actually) about sensor physics and noise vs. sensor size. This is a comparatively easy topic.
When you're talking about film, ISO (or ASA) refers to the sensitivity of the film. As you probably know, film is made of up grains of silver halides which turn black after a certain number of photons have hit them. A faster film (ASA 400, say instead of ASA 50) will have larger grains. Because the grains are larger, they can absorb aalllll the photons in a larger area. So they don't need as much light coming in the lens to make a proper exposure. Obviously larger grains means less detail in the image. Using smaller grains means each one sees fewer photons and therefore needs a longer exposure, but the detail is better. This topic is actually significantly more complicated than that - the image is latent before developing, comparing dynamic range to chemical sensitivity, flip-point distributions, quantum noise, and so on - but that's waaay too detailed for here).
So that explains why fast films are grainy and slow films are smooth. But how does that apply to a digital sensor?
In your camera's CMOS or CCD sensor, each pixel is represented by a little photon gathering well. These wells can store up to a certain number of photons before overflowing (blowing out to white), and it's the job of the shutter and aperture to keep the number under the sensor's maximum. Ideally, something that is white in your image will fill up the sensor's wells completely, and something black will have no photons in it. This is using your sensor to its maximum efficiency.
Now, because you can't increase the sensitivity of these wells to photons (a "full" well will always require the same number of them, unlike film), you should theoretically always need the same amount of light, right? If there's more, then you can close down the aperture, but if there's less, you're screwed?
|