A digital camera has an image processing chip that converts light (analog) into digital information. A single point of processing or conversion is a pixel (from picture element). Most people obsess over resolution as measured in megapixels, concluding that 20 megapixels must be better than 12, for example. However, the size of the sensor is very important in such comparisons.
Imagine taking a box and arranging 20 cups inside (say, 4 x 5). If you get a bigger box, you have options for bigger cups (which hold more than the smaller ones) or more cups of the original size. Either will hold more liquid than the original cups.
In so many tech areas, we think smaller is better. You’d rather carry a little camera than a giant one. However, when it comes to pixel size, larger pixels gather more light (just like those larger cups hold more liquid). Moreover, in electronics, components tend to interfere with each other (referred to as Signal to Noise — more signal, less noise is the goal). Cram too many, too small pixels into a space and you not only capture less information but you increase the noise from adjacent pixels. Therefore, larger pixels are better. Ironically, a specific camera’s pixel size may not be easy to find. (See www.snapsort.com and look closely at Specifications.)
Comparing two cameras that have the same image sensor, more pixels means smaller pixels, which probably means lower image quality. Ironically, that 12MP camera might be better than 20MP (at that point, compare other features).
Comparing two cameras that have the same resolution, say 12MP just for example, the one with the larger sensor will also have larger pixels and better image quality.
Point and shoot cameras include compacts, cellphones, and superzooms. (Point and shoot cameras have a fixed lens that can’t be replace.) The vast majority of point and shoot cameras use the smallest processor there is. They vary in how densely they pack small pixels.
It used to be you could call anything that wasn’t point-and-shoot a DSLR (digital single lens reflex). (DSLRs use interchangeable lenses.) Every camera that is not point and shoot uses a larger image processor than all point and shoots.
The only point and shoot exception I know of is the Fujifilm X-S1. It uses a sensor that is slightly larger than other point and shoots, but far smaller than any non-point-and-shoot.
What to look for and why
Now you know what
Above, the smallest green box is the size of most point and shoot sensors (1/2.3). The slightly larger dark blue box (2/3) is the size of the Fujifilm sensor. It can convert more light into information with less noise. When you get out to the much larger Four Thirds and larger, you’re comparing DSLRs (and equivalents). 35mm is the gold standard of film and slides.
More Effective Light Gathering
Imagine cramming millions of photo sensitive dots on a postage stamp size chip and you can appreciate that each of these must be microscopically small. Increase the surface area of the sensor by about 2.5x (to 24x36mm), and each pixel can be much larger; hence, it can capture more photons (light particles) in less time. That enables the sensor to provide greater signal purity: a stronger signal-to-noise ratio so less gain or amplification is required at high ISO levels. The resulting images are “cleaner”. They exhibit less digital noise, without obvious “graininess” or colorful speckles that can and less need for Noise Reduction processing that can obscure intricate detail.
When compared to the “typical” 1.8×23.6mm sensor used in many DSLRs, a full-frame 24x36mm chip is huge. While the difference may not be impressive at a glance, it’s worth noting that the surface area of the full-frame sensor is roughly 2.5 times larger. Assuming the same number of pixels on each chip, you can assume that the size of each photodiode will also differ significantly.
As a bonus, larger pixels also provide a wider dynamic (tonal) range: more detail in both the bright and the dark areas of a photo. That’s useful because it minimizes excessively bright highlights (blowout or clipping) and very dark shadows that can be entirely devoid of texture or detail. This can be important particularly in high contrast illumination when a scene includes both very light tones and very dark tones. For instance, imagine the interior of a cathedral, with light streaming in through a few stained glass windows but illuminating only a third of the interior. A sensor with a wider dynamic range will provide images with more visible detail in the entire scene.
Note the following comparisons: Snapsort rates the HS50 higher than the X-S1, despite this difference.
Even the very popular SX50 has pixels less than half the size of the X-S1. (This is just one variable, albeit an important and obscure one.)
One more oddity: Consider two point and shoots that are as close to identical as possible except that one has the smaller sensor. If the lenses are identical, the smaller sensor & lens yield a much greater zoom than the larger sensor with the same lens. (A variable called crop factor is at play here. See the Y-axis of the first graphic.) That’s why the Fujifilm HS50 EXR has a 42x (1000mm) zoom while the nearly identical X-S1 is “only” 26x (624mm). The HS50 will get you closer but the image quality will likely be lower, especially in less-than-bright conditions, where more light gathering with less noise is of increasing value.