The Megapixel Myth.
Don’t believe the marketing hype!! It was mostly a ‘thing’ in the early 2000’s and not so prevalent anymore, however it has led a lot of people to still believe that more pixels is best, which isn’t always the case. It all depends on what you want to take photos of, and when it comes to light painting and photography in low light conditions, there’s a sweet spot.
Before I got into light painting (or more to the point Cas got into light painting because I kept dragging her out to take photos at night), I did a lot of astrophotography. I learnt a lot about camera sensors and why some are better than others at shooting in low light conditions.
What is light?
It’s a question we should answer first in order to understand the megapixel myth. If you’ve ever delved into quantum physics, you may already know that light can take two forms, waves and particles known as photons. This is most famously demonstrated by the two slits experiment.
Light photons are fired at a partition containing two slits.
When observed, the light pattern on the back panel looks like what you would expect to see if you fired paintballs at the slits. They either go through one or the other and end up on the back.
However, when not observed, the theory states the photons pass through both slits, then interfere with themselves causing a wave pattern. Much like you would observe if sea waves passed through two harbour openings next to each other.
For the rest of this blog, let us think of light as photons, tiny particles that hit our camera sensors.
Does size matter?
Well… if we’re talking about camera sensors and low light photography, maybe… but this is where the megapixels come in. It’s really the ratio between the sensor size and the amount of pixels that sensor has that we are worried about.
How do sensors work, an example.
Let’s assume the following for our little thought experiment, based on a popular pastime of students, beer pong!
The ping pong ball = a photon of light
The table = the sensor
The red cups = the sensor pixels
Scenario 1: Small sensor/high megapixel
Lets say you have a small table (or asp-c sized sensor) that has a really high megapixel count. To fit that high amount of red cups on the table will mean the cups are very small. There is a small chance of capturing all, or even most of the ping pong balls you throw at them.
Scenario 2: Larger sensor/high megapixel
Imagine the number of cups you have stays the same, but the table is now bigger, which means you can fit bigger cups on the table. You will capture more ping pong photons than scenario 1.
Scenario 3: Large sensor/small megapixel
So, we have a bigger table, but we’ve reduced the amount of cups, therefore we can increase the size of the cups to take up the same amount of table area as scenario 2. Bigger Cups = more chance of capturing your ping-pong ball photon!
Noise to Signal Ratio
A problem shooting at night is the introduction of unwanted noise in our photos. Sensors like the ones I described in scenario 3 are more likely to have a better noise to signal ratio.
This means that for the amount of light (signal) applied to the sensor, the noise is less noticeable. This is because for a given amount of light, the large sensor/small megapixel is able to collect more information to process into an image.
In conclusion, what does this mean in the real world…
The bigger cups, or ‘light wells’ in a sensor make it better at capturing the light that’s available to it, essential in low light photography like light painting and astrophotography. My recommendation would be to look at cameras with sensors that have the following megapixel ratings for low light photography.
Full frame 35mm camera sensor - 20 to 30 megapixels
ASP-C Sensor (assuming a crop factor of 1.5) - 13 to 20 megapixels
This gives a really good balance between light capturing performance and pixel count, so you can still capture plenty of detail. Especially if you are introducing astrophotography into your light painting shoots. You want that balance to capture stars, introduce less noise and have more dynamic range.