Written by Jeff Hui at Voltagedivide blog.
It might be fun to explore how much the amount of megapixels in a camera affect the quality of the photo. It is frequently used as one of the sole metrics to evaluate the performance of a digital camera. Is this a reliable number to go by, or is there more to picking out a “good” camera than just looking at the amount of pixels?
To understand why the amount of megapixels has been associated with performance, we need to understand what makes a camera work. The most popular type of camera is the single reflex lens, or the SLR. The aptly named camera works by reflecting the photo image off a mirror, into a pentaprism in order to be seen in the viewfinder. When the shutter button is pressed, the mirror folds upward, allowing the light to pass through and onto film for fractions of a second. Film is made up of silver halide crystals, a structure that when exposed to light will rearrange.
Photography was based on the principle that if you expose the film to light for just the right amount of time, only certain parts of the crystal will rearrange. When the film is taken to develop in a solution called developer¸ the lighter grains in the crystal react, turn black, and stay that way. The film is then put into a bath known as fixer which removes the untouched crystals. This creates a translucent photo negative, which is then reversed in a similar process to create the photos. Two adjustments that could be made to change how pictures come out are the aperture and the shutter speed. Changing the aperture of a camera changes the hole size in which light is allowed to pass through.
The size is dictated by the f number (no unit), and a lower f number will let more light in. The f number represents the ratio of the focal length to the diameter of the entrance pupil. The setting can be incremented in intervals of √2 (1.4, 2, 2.4, 4, etc). The shutter speed is how quickly the shutter opens. This number is expressed as a fraction of a second.
With the digital age came the birth and popularization of the DSLR. It featured a camera sensor and hardware to replace film. The CMOS (Complementary metal–oxide–semiconductor) is the predominantly used sensor today. The CMOS works by detecting light digitally with a bunch of photo diodes. The photo diodes each represent a pixel, and work in a similar fashion to the grains in the silver halide crystals in film.
With the introduction of the sensor came one more way to (digitally) adjust how a picture will turn out; the ISO. The ISO speed is the sensor’s sensitivity to light and is best left at a lower value as setting it high may leave the picture looking “grainy”. The ISO adjusts the gain on the whole sensor itself. Gathering light information does not come without noise. The bigger the photo diodes are, the better the individual diode will perform in gathering useful information and the less it will be affected by the noise.
The figure on the left outlines the steps the scene takes once the shutter button has been pressed. The optics (or lenses) work together to refract the light to prepare it for detection in the CMOS sensor. The light information travels from the CMOS to the image signal processor (ISP) where the processor “guesses” the correct pixel color using neighboring pixel colors. Today, the ISP has been programmed for white balance correction, noise reduction and many other features but it falls outside the scope of this piece. Finally the data file is saved as a JPEG in the internal memory.
Mo Pixels, Mo Problems
Now that we have a functional understanding of what a modern camera does, we can begin to explore how significant MP is. The amount of megapixels in a camera is the amount of pixels (Mega is meaning 106) present to capture light. The more pixels we have, the more detailed our information can be. By that logic, having MORE is obviously better right? It’s not quite that simple.
In the last 5-10 years, the trend in tech has been to make things more form fitted and slim. A by-product of this design trend is a decrease in size for all components. As silicon technology improved, the ability to make smaller pixels also did. This allowed manufacturers to create even smaller cameras with the same amount of pixels; the very ones we see in smartphones today. More importantly, it allowed manufacturers to create the same size cameras with a considerably higher MP count.
Recalling how an image signal processor works, we remember that it takes the information from the CMOS and identifies the color according to its neighboring pixels. Smaller pixels don’t receive as much light information as larger pixels. The less light information each pixel can capture, a bigger percentage of that information is noise.
An analogy I once heard to explain why bigger pixels are better was to imagine each pixel as a bucket on a piece of sidewalk (the sensor). If the goal is to collect as much water as possible, then the larger the buckets you have, the more water you’ll be able to collect. With smaller buckets, more water would hit the sidewalk (creating noise) instead of being collected or useful.
This is the reason why entry level DSLRs will sometimes outperform smartphone cameras despite having fewer megapixels. This also explains why DSLRs with a larger MP count continue to perform better. A large camera allows for a large amount of MPs while still being able to maintain a large sized sensor to decrease pixel density. With a largely fixed sensor size, the only ability to improve lies in creating more pixels and effectively restricting pixel size. Even though a higher ISO can help cameras with smaller pixels, the improvement in quality comes with an increase in noise as well.
When you use the digital zoom function on your smartphone, you’re merely eliminating the amount of pixels on screen, and that results in a “pixelated” and lower quality picture. This doesn’t happen on cameras equipped with optical zoom, where the lenses used are moved to achieve the zoom. On smartphone cameras, you forfeit that ability for, once again, form factor. You’ll also notice that smartphone cameras tend to have subpar performance in low light. Because of the pixel density and the ISO setting needed, the camera sensors are so sensitive to light that more noise gets mixed in with the information. This, once again, gives us the iconic grainy look.
Apple iPhone are regarded as having one of the better phone cameras nowadays, yet they have not opted to increase their sensor size since the release of the iPhone 5 (with a 6% increase from the iPhone 4s to 1/3″). They also have not changed their 8MP camera since that bump in sensor size. Has Apple achieved a healthy balance between pixel density and sensor size? With an (proposed) increase to 12MP camera with no indication of a bigger sensor size, there is a possibility we won’t see an increase in quality.
So when looking for a camera, be sure to do your research on your camera sensors. More MP can be better but not the largest factor in camera performance. Manufacturers do not advertise sensor size as much as I would like, but it’s sometimes a good bet to look at a specifications sheet for it if you are really concerned about the quality. Although everything points to smartphone cameras being restrictive, I feel like for the most of us, it is still our go to camera device. On trips, I’ll have a DSLR in my bag but still opt to use my Galaxy mobile camera out of convenience. So if you’re like me, maybe you wouldn’t mind the drop in quality in exchange for convenience.
With this information, (hopefully) you can approach camera/smartphone shopping with a more critical and technical point of view.