Blinded By Light

The full frame protagonists are at it again. It’s the same stupid argument. Larger sensor means more TOTAL light gathered therefore lesser noise. Stop the bullshit. Please!!!

In fairness, it’s quite easy to be mislead by this kind of misinformation. If noise is inversely proportional to the amount of gathered light then it makes sense that a larger sensor would result in cleaner photos. Unfortunately, just looking at the total amount gathered light is being very short-sighted. It does not give us the whole picture (no pun intended).

Allow me to explain it again for the nth time. But before that, please read the following articles because they explain this concept in greater detail.

1. https://dtmateojr.wordpress.com/2014/02/28/understand-your-lens-part-3/ — Concentrate on understanding the effect of focal length on light intensity because a lot of people tend to ignore this bit. They are too preoccupied with just the aperture opening maybe because they are more familiar with “fast” lenses without even understanding what “fast” really means.

2. https://dtmateojr.wordpress.com/2014/03/08/debunking-the-myth-of-full-frame-superiority/ — If there is one thing that you’d want to fully understand here, make it the “thought experiment” on dividing a full frame sensor which also explains how shutter curtains work.

3. https://dtmateojr.wordpress.com/2014/06/10/debunking-the-myth-of-full-frame-superiority-part-2/ — This is a good counter-argument to the fact that no two digital sensors are exactly the same even if they are of the same type. A D7000 sensor for example is almost every bit the same as the D800 sensor but because of the improved processing the latter may produce better photos. And so I used film as an example because the same film emulsion will always behave the same way regardless of format (size).

4. https://dtmateojr.wordpress.com/2014/05/19/megapixel-hallucinations/ — Some full frame protagonists insist on comparing ENLARGED APS-C images to full frame equivalents in terms of noise. Of course an enlarged APS-C photo will, for the lack of a better word, enlarge everything including noise. This article debunks that by showing the MATH behind resampling as well as showing samples of real SNR measurements of APS-C and full frame sensors.

5. https://dtmateojr.wordpress.com/2014/04/21/rain-can-teach-us-photography/ — explains what happens in a sensor and why PIXEL size and NOT sensor size matters in greater detail.

6. https://dtmateojr.wordpress.com/2014/05/09/understanding-exposure/ — are for the equivalence clowns who think that they could get away with the bullshit by manipulating ISO. In short, you can’t.

Now if after reading those articles you still need a bigger cluebat then read on…

The biggest mistake that full frame protagonists make is that they equate a sensor to a solar panel. In a solar panel, total light gathered is everything. In a solar panel, every “sensor” contributes to the total energy produced. Of course the bigger the better. Photography though is far from being like a solar panel. Camera sensor pixels are independent of each other. That’s why within an image you will encounter darker parts that are more noisy and blown up parts that have been saturated by light. Each individual pixel receives its own independent number of photons. Pixels can’t share their photons with other pixels. Well sometimes adjacent pixels do “share” photons but this is an undesirable effect called “sensor bloom”. You can see why looking at noise as a result of the total light gathered is wrong. Noise should be examined at the pixel level because this ultimately defines the efficiency of your sensor.

While it is true that a larger sensor gathers more light compared to an APS-C or M43 for the same exposure by virtue of the larger area, this argument is not photographically sound. Photographic exposure is all about LIGHT PER UNIT AREA and not just total light. Saturating a pixel only requires a fixed number of photons. Anything more than that is just wasted light because as soon as a pixel clips then “no data” is presented for processing into an image. An APS-C sensor for example requires half the total amount of light required for a full frame sensor. If you force the same amount of light to both a full frame and APS-C then the latter will oversaturate, i.e. overexpose. It’s like pouring two liters of water into a one liter container. It does not make sense. It is photographically disastrous and plainly stupid. Therefore you get the same noise-free image in an APS-C for half the total amount of light hitting the sensor. Again, you get the same noise-FREE image for HALF the TOTAL amount of light. Again, it’s all about LIGHT PER UNIT AREA and NEVER just total amount of light gathered. It’s all about light intensity.

A smaller area requires lesser incident light. A smaller sensor requires a smaller lens-projected image circle. Smaller image circle is what defines a “crop” sensor or crop lens (e.g. Canon EF vs EFS lenses or Nikon FX vs DX lenses). You crop a full frame image circle just enough to illuminate a smaller sensor. Makes sense?

Unfortunately, there are those that remain blind and they resort to other stupid arguments such as printing at the same size or enlarging a cropped image. Of course a larger sensor is capable of larger prints but this has got nothing to do with light. But let’s be stupid for a minute and assume that a cleaner print is the result of more light gathered during exposure. What happens then if you print at a smaller size? Did you just throw away the light? If not, then where did the light go? If print size has got anything remotely related to light then projecting the same amount of light into a smaller print is like pouring two liters of water into a one liter container so we expect the smaller print to be overexposed, right? But it doesn’t. Because print size has got nothing to do with light and therefore has got nothing to do with noise. The apparent increase in noise when you enlarge a print is NOT the effect of light but the effect of resampling (refer to megapixel hallucination article), i.e. resolution.

In conclusion, total amount of light is just half the truth. The other half is sensor area. Combining both, we get light per unit area otherwise known as photographic exposure. Exposure is what ultimately dictates noise. Smaller area requires lesser light therefore the same exposure results in the same noise profile for different sensor sizes of the same type.

I hope this is the last time I will ever write about this topic. It’s getting long in the tooth and very boring really.

I promise to write a happier article next time. Really. I promise that. 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s