XW

Basic Photography I: How to Buy a Camera

Introduction

The greatest hurdle for me in getting started with photography was in the very beginning, when I was overwhelmed by the choices of cameras and all the different numbers and terminology, such as ISO, F/#, or 55 mm focal length. I’ve learned a lot through experience and articles, so I’d like to share what I learned for those who are just starting out.

The most basic tool you need in photography is the camera, so we’ll be covering some important camera specifications in this article. There are many types of cameras, ranging from inexpensive, compact, point-and-shoot (P&S) cameras, which can be as low as $75 when on sale, to more expensive and bulky DSLR cameras (\<$1000) suitable for the serious amateur.

Compact cameras are easier to carry around and thus open up more opportunities for shooting. As some people say—well, I don’t exactly remember what they say, but it goes something along the lines of, “A bad photo is the one you don’t take.” “The best camera is the one you carry with you.” Nowadays, cell phone cameras are a viable alternative for snapping spontaneous shots, even for print (my old Razr had a 1.3 MP camera; it was not that useful). On the other end of the spectrum, DSLRs generally offer significantly greater color depth, less noise, faster response, and greater sharpness and technical flexibility with high-quality interchangeable lenses. DSLRs are especially useful if you want to make large, high-quality prints. Between these two extremes is a growing market of hybrid cameras that fall somewhere along the tradeoff curve.

Which type you choose depends on your needs and tastes. If you’re making slideshows and taking photos just for the sake of memories, a P&S is usually good enough, although you may occasionally wish you had something better for those wow moments. If you’re into capturing every detail of your trip, making large poster prints, or photography is part of your trade, then a DSLR is most likely a non-negotiable investment.

For me, I didn’t get a camera until college, and I ended up using it for most of grad school as well. It was an Olympus Camedia C-750UZ, a P&S with 10X optical zoom, and was advanced for its time. It had 4 megapixels, but my friends around me (ahem Jess/Albert/Basile) were playing with fancier toys, so of course, I had photo-envy.

Of the many reasons I wanted a better camera, the most important were:

  1. The ability to shoot in RAW. My pictures were in JPEG so the color range was severely limited1, and I could not change the exposure in post-processing (the editing phase, analogous to analog film development) without losing large amounts of data.
  2. The C-750UZ was slow to turn on, slow to focus, and slow to shoot. I missed many a beautiful shot because by the time I took the camera out, turned it on, and tried to focus on the subject, the moment had passed.

I spent months looking at DSLRs. During the course of my research, I found Snapsort to be the most useful. It has an excellent feature for comparing specifications between cameras, which ultimately persuaded me to choose the Nikon D5100 over the Canon T3i Rebel.

One of the reasons for choosing either a Nikon or a Canon is that they are the most popular brands, have great reputations, and both have a large collection of lenses available. If you need to borrow a lens, it’s not too hard to find someone else with one. If you’re looking for a good entry-level DSLR, try sticking with the large manufacturers, such as Nikon, Canon, Pentax, Sony, and Panasonic.

Important Considerations

Depending on what your needs or desires are, the importance you attach to the various characteristics of the camera will vary. Personally, the specifications related to image quality rank at the top of my list, because I feel that those set the fundamental limits of the camera. If my battery doesn’t last very long (within reasonable limits, of course), I can always purchase a spare as backup, but if my sensor is noisy, there’s not much I can do about that apart from buying a new camera or a very expensive lens (i.e., a wider aperture allows me to reduce my ISO setting). Post-processing only helps so much.

I will not go into much detail concerning lenses; many entry-level DSLRs come with a starter lens (kit lens) that is reasonably good and versatile. My Nikon D5100 came with an 18-55 mm zoom lens2, although I eventually bought more specialized lenses (specifically, a prime lens and a telephoto lens) for better image quality and versatility. I still use the kit lens when I want to take wide-angle shots.

Features

SD/SDHC/SDXC support: SDHC cards are the standard these days, so make sure to get a camera that supports SDHC cards (or even better, SDXC). If you get the microSD or miniSD cards, you’ll need adapters. (If you have a MacBook Pro, there’s a native slot for SD cards). After you get the camera, you’ll need to get a card. I would say to get at least 8 GB, which corresponds to ~300 photos for an 18 MP camera at its highest quality setting. I would recommend 32 GB if you can afford it. Also note that the cards come at different speeds. Make sure you get a card that supports ~45 MB/s or higher, so that the camera’s internal buffer can be rapidly written to card.

Image stabilization: this feature purportedly allows you to shoot 2-4 stops slower. This gives you more flexibility with your shot: with the appropriate tradeoffs, which I’ll cover in a future article, you can get better depth of field, better sharpness (its primary usage), or less noise while shooting handheld. There are very few reasons for getting a camera without image stabilization.

Numbers

Sensor size: this is what sets the fundamental limit of image quality. Bigger sensors are generally better, though they require a bigger body and bigger optical components (optics for short), so they will likely be heavier. Professional cameras, excluding medium format or larger, use full-frame, or 36 mm x 24 mm sensors. The Nikon D5100/Canon T3i and similar entry-level DSLR cameras have sensors that are about 2/3 the full-frame dimensions, or about 24 mm x 16 mm.

Resolution: resolution has a number of definitions in the field of optics, depending on the context. In consumer devices that have screens or sensors, it has come to mean the dimensions of the maximum image size that can be displayed or captured, in pixels. Made popular and controversial by the megapixel war3, this number indicates how many points a photo is made up of. Higher resolution means you’ll get a larger picture—a picture with more pixels. However, just because a camera produces an image with more pixels does not mean the picture quality is better. You can still have a large, fuzzy picture from a bad lens, bad focus, or movement during capture. Most of the time, it’s not the lens’s fault that a picture is not sharp4.

Here’s a concrete example regarding the folly of focusing on the MP count: if the smallest feature you can distinguish in a photo taken by a 16 MP camera is 2 x 2 pixels large, a 12 MP sensor that can distinguish 1 x 1 pixel details is clearly superior, assuming all else is equal. In the latter case, you can “resolve” up to 12 million points (12 MP divided by 1-pixel-large details), whereas in the former case, you can only resolve 4 million points (16 MP divided by 4-pixel-large details). This means the 12 MP camera contains more information, and therefore more detail, than the hypothetical 16 MP camera, by a factor of 3. The optimal situation is when the smallest spot that your lens can produce is the size of one pixel.

Practically speaking, your budget will limit the quality of your equipment. Because of the quadratic relationship between the linear dimension, which our minds associate with “size,” and the MP count, it takes 4 times as many MP to achieve a “2x larger” picture. Your photos will consequently take a lot more storage, and more time and memory to process. Larger files also mean you need to switch out your card or download your pictures more often, unless you get a bigger card. Don’t forget you also need more backup storage. However you look at it, you’ll be paying extra in the form of time or money.

This leads to the critical question, “How good is good enough?” This is where another definition of resolution comes into play. Pixel density, which we measure in points per inch (ppi)5, is calculated by dividing the number of pixels along one dimension of your photo by its physical length in inches. According to one expert, 150 ppi is pretty good, and 200-250 ppi is very good6. The table below lists the kind of sizes you can print for a given ppi. The sizes rounded to the nearest inch7. As you can see, even sub-10 MP cameras can get you large prints that look nice. The reason is the quadratic relationship between linear pixel density and MP count. A 2x increase in MP count gives only a 40% decrease in linear pixel density, so only large differences in MP number will significantly affect your print quality, assuming a given print size.

PixelsSample CameraImage Dimensions@300 ppi@225 ppi@150 ppi
6 MPNikon D703008 x 20007” x 10”9” x 13”13” x 20”
10.1 MPCanon EOS Rebel XS3888 x 25929” x 13”12” x 17”17” x 26”
14.2 MPNikon D31004608 x 307210” x 15”14” x 20”21” x 31”
17.9 MPCanon EOS Rebel T3i5184 x 345611” x 17”15” x 23”23” x 35”
24.1 MPNikon D32006016 x 400012” x 18”18” x 27”27” x 40”

When you get to large images, another principle comes into play: the larger the image is, the greater the viewing distance is generally8, and thus the less ppi you need, because at some point, your eye cannot discern adjacent pixels anymore. To illustrate this point, imagine a black and white checkerboard pattern on a very large sheet. As you back up, the squares look smaller and smaller, until at a certain distance, the sheet is becomes a solid block of gray; this is the reason why going beyond the resolution of the Retina display has negligible benefit.

In real life, you’re not going to be pushing the limits of your camera by taking pictures of alternating squares or lines. Instead, you’ll get a general feel for the clarity of the shot via the texture of leaves, the fineness of hair, the ripples in the water, or the edges of flower petals. But the point is that as you print larger images, you may not even need to adhere to the lower limit of 150 ppi.

With all this talk about how MP is overrated when it comes to image sharpness, I would like to point out the other side of this coin. A higher MP affords more leeway in cropping so that the composition can still be adjusted after a shot is taken. You can lose a lot of pixels and still produce nice prints.

Noise: no matter how sharp your image is, noise, or graininess, will limit what details you can distinguish in a photo. That is because at some point, you have no way of telling the difference between what was in the scene and what was noise generated by your camera. A larger pixel size will generally give you less noise, assuming all of the supporting electronics are the same. Snapsort gives you an effective number for low-light performance in ISO, as measured by DxOMark. This ISO number is a measure of how sensitive you can set the camera before the noise exceeds a certain threshold. The higher this number, the better.

Color depth: measured in bits, it tells you how well the camera can distinguish colors. The more bits, the better. Every 1-bit increase is a doubling of the number of colors the camera can distinguish. Greater color depth allows for a more nuanced photo and greater flexibility in post-processing. This specification is computed by DxOMark.

Dynamic range: measured in EV, this metric tells you the range of brightnesses your camera can detect in your scene. That is, in a scene with both dark and bright areas, a sensor with a high dynamic-range means you can see the bright area without its being washed out at the same time as being able to see details in the dark areas. Once you capture those areas with your sensor, then you can correct them later during post-processing. Every 1 EV increase is a doubling of this ability to separate values. Greater dynamic range means less uncorrectable over- and under-exposure. This specification is also computed by DxOMark.

Image quality: a “summary” number specific to DxOMark, in arbitrary units. The higher the number, the better.

Video: unless you’re confining yourself to taking photos, having video capability is pretty useful. Support for 1080p at 30 frames per second is pretty standard these days.

Less Important Considerations

Battery life: most DSLRs’ battery life is decent covering a range of 400-600 shots. A lower battery life can be a bit annoying because you have to charge or swap batteries more often, but it is easily correctable: just buy a spare battery. Being prepared helps guard against missing a critical shot.

Size, weight: there’s not necessarily a better or worse here; a heavier body could be more stable but also more to carry; the ideal size of a camera depends on your hand size, although smaller is preferable for travel. A few millimeters’ or a few grams’ difference is not really that significant.

LCD: most of your shots will be taken as you look through the viewfinder, which gives your eyes a better sense of focus than an LCD. However, an LCD with a larger pixel count allows you get a better idea about the quality of the picture after the it has been taken, which is very useful too.

Shooting speed: most entry-level DSLRs will shoot around 4-5 shots per second for continuous shooting, so it’s really not that much of a difference. Also, the time it takes between pushing the shutter release and the opening of the shutter itself is comparable across many cameras; a smaller shutter lag of course is better.

Conclusion

There are many cameras out there, so picking one is not easy. Figure out what’s important to you, and make an informed decision about buying one. There is no “right” camera. Or rather, the “right” camera is the one you end up enjoying using. Have fun!


  1. JPEG stores 8 bits, or 256 levels, per color, vs. 10-bit, 12-bit, or 14-bit color for RAW, corresponding to 1024, 4096, or 16384 colors, respectively. You might wonder if the human eye can actually distinguish all those levels of a single color, but the color depth actually becomes extremely useful during post-processing.
  2. According the photography lore, a focal length of 50 mm on a normal 35 mm camera has a field of view close to that of the human eye. Since the D5100’s sensor is 2/3 of the normal size, the “normal” focal length for my camera is around 35 mm instead of 50 mm. Relative to that then, the 18 mm covers the wide-angle end of the spectrum, while 55 mm is zoomed-in. Thus, my kit lens covers a fair range on both sides of “normal.”
  3. The Megapixel Myth
  4. The Megapixel Myth
  5. DPI and PPI Explained
  6. The 300 ppi Print Myth
  7. Wikipedia has a list of standard sizes
  8. The Megapixel Myth