Last Modified: January 3, 2006.
If you can afford a decent digital body today, you should probably forget about film. The only exception I can think of is if you're primarily interested in black and white photography. If that's the case, be sure to read my notes on digital black and white photography.
The image of the Snowy Egret on the right was taken with a Nikon D2x. Click on it to obtain a larger version, and even that larger image is still scaled down by a factor of three.
For reference, I own and have experience with three digital cameras: one point-and-shoot and two Nikon SLR bodies.
The Sony DSC-S85 is a high-end digital point-and-shoot (or was high-end when I bought it). It originally cost about $800 and can take photos with 4 million pixels. That camera is now out of production, although the price dropped to about $500 before Sony replaced it with new models. From what I can tell, equivalent cameras today cost about $300.
My other cameras are a Nikon D100 body and a D2x body that can use almost all my old Nikon lenses. The D100 has a resolution of 6 million pixels and cost about $1700. Don't be fooled into thinking I made a mistake and only got 2 million pixels for more than doubling the price; the Nikon pixels are "better" as I shall describe later in this document. The D2x cost around $5000, and has a 12 million pixel resolution. I have only had the D2x for about a month, but I used the D100 heavily for about two years. The funny thing is that in the month that I've had the D2x, I have already taken more photos with it than I did in the entire two years with the D100.
I still have more experience with film cameras, but by now I and starting to feel very comfortable with the digital cameras. If you'd like to see some of my digital work before reading on to make sure I can walk the walk as well as talk the talk, I've got some example web pages. The best shots by far are taken with the D2x:
Birds at the Palo Alto Baylands
Here are a few others:
Costa Rica 2002 volunteer work
Costa Rica 2002 as a tourist and
Burning Man, 2002
were taken with the Sony.
The Nikon SLR has now made a few trips and you can see my first digital SLR attempts on these pages from:
Costa Rica 2003
Palo Alto Baylands and
Burning Man, 2003.
As a first approximation, a digital camera is basically identical to a standard film camera except that in place of the film, the incoming light shines on a CCD (charge-coupled device) that senses the incoming colors electronically. Then, instead of recording on film, the data from the CCD is written to some sort of electronic memory. This may be something like a memory card, a small disk drive, or even an optical CD. In the field, instead of changing rolls of film, you'll change to a new memory card or a new CD when you fill one with photos.
But unlike standard film, after you have copied the data from your memory card to a computer or something you can erase that card and reuse it. In fact, you can review each photo as soon as you take it, and erase it immediately, in the field, if you do not like the results.
Depending on the cost of your camera, it may be equipped with most of the same options that a standard 35mm camera has: zoom lens, macro setting, flash, exposure control, auto and manual focus, exposure locks, et cetera.
In addition, some are equipped with various electronic filters that can be applied to the photo to produce things like black and white, vignetting, solarization, sepia prints, et cetera. In my opinion, these features are worthless--why not do your digital retouching at home on your computer where you can see what you're doing and no doubt have far more powerful image processing programs?
Often the camera (this is true of digital point-and-shoots, especially) has the ability to record short MPEG movie clips, including sound if you want. I didn't think this would be too important, but have actually found a few nice uses for this feature.
Here is the best example I know of. I was in Costa Rica and one of the people in our group was a professional photographer who has helped produce films for the BBC, et cetera. We saw a nice line of leafcutter ants and I wanted to get a photo, but the photographer was filming, and I got in his way. He's a really nice guy, but he started joking about what he would have said if he were a
The other type of filter that might be useful is one of those split level filters. These are half clear and half some other color, normally gray. Then if you've got a very bright sky and a dark foreground you can put the split right on the boundary and it'll be easier to get a good initial exposure range by darkening the sky relative to the foreground. Digital cameras have the same problem that normal cameras do in the sense that they can only record light in some range. Anything brighter than one end of the range will record as white and anything darker than the other end of the range will record as black. By using a split level filter, you can reduce the range of light that the camera sees, allowing you to record details both in the sky and in the forground. Without the filter, depending on how you expose, you'll get a detailed foreground and a burned out sky or a detailed sky and a blackened foreground. Use the dark part of the filter on the sky and the clear part on the foreground.
On the other hand, if you've got an unchanging scene, you can put your digital camera on a tripod and take the same shot with different exposures and combine those photos later with software like Adobe's Photoshop to yield a much
But the uniformly-colored filters can easily be replaced with a computer program, since all they do is remove light from the photo. Similarly, many of the special effects filters (like the ones to add highlights or rainbows) can probably be done even more efficiently in an image processing program like GIMP, Adobe Photoshop or many others. In fact, Photoshop has built-in filters that imitate all the standard color filters: the 81A, 81B, et cetera.
My Sony camera, for example, can record resolutions up to 2272 by 1704 pixels. For a 64 megabyte memory stick, this is about 30 photos, sort of like a roll of film. But instead of $5 for a roll of film, I needed to spend around $100 for each stick. Today, the prices are, of course, lower and will continue to get lower. On the newer Nikon, the story is even worse. The highest quality image burns up 9.6 megabytes in "raw mode" per photo. I use a 1 gigabyte microdrive and it'll hold 103 photos at this highest quality and resolution.
There are a couple of ways around this. First, take lower-resolution photos. Almost nobody has a computer screen (at least today) that can display 2272x1704 images (or 3008x2000 in the case of the D100). But I like the highest resolution since I often crop drastically. A nice compromise for me is to shoot at 1600 by 1200 and I can get about 64 shots on a memory stick at that resolution. Of course at the lowest resolution I could have gotten 900 shots per stick.
Also remember that you get to erase the bad shots right away, so you probably need only the storage capability for half of what you'd take on regular print or slide film. Or if you're a bad photographer like me, you can erase a much higher proportion of the shots as soon as you take them.
Finally, if you carry a laptop on your vacation, you can just download the photos from your camera at the end of each day and get away with one or two memory cards, even at the highest resolution. I recently spent three weeks in Italy with two 64MB Sony memory sticks, took all my photos at the highest possible resolution, and never came close to running out of memory, even in the most scenic places. At the end of each day I just downloaded all my photos to the laptop and erased the memory sticks.
But resolution isn't everything. The CCD that collects the light in the smaller digital point-and-shoot cameras like my Sony is very small compared with the size of normal 35mm film. The CCD in the big SLRs is now approaching and in some cases equaling the size of a 35mm slide. Even if the resolution of a 35mm digital SLR is the same as in a point-and-shoot, the results will be much better. The reason is that the area of the CCD devoted to each pixel will be much larger, so random variations in the number of photons collected for each pixel will be statistically smaller. If you look closely at digital photos from a point-and-shoot, you'll see that if you look at something that is pretty uniform, like the sky, that if you stand back, the average color is pretty good, but if you look at the individual pixels, they vary a lot. Thus big enlargements will suffer unless you do some good filtering.
To see the quality possible with a 35mm camera with a decent-sized CCD, take a look here. The photo of the Barn swallow is particularly good, I think. This was taken with a Nikon D100.
Also, when a digital camera is advertised as having "4 million pixels", that's really more like 1 million pixels as we normally think of them, since four of their pixels are used to gather what amounts to one pixel on your computer screen. Usually, they collect on four adjacent spots on the CCD two green samples, a red sample and a blue sample. (You need extra green for technical reasons.) Each pixel on your computer screen has a red, green and blue component.
Digital cameras can increase the effective resolution with software, so the situation is a little better than what I described in the previous paragraph. And Sigma has something called the "Foveon" technology that also does better, but it's still expensive and it's not clear they'll be able to make a business of it.
This has a
From a reader:
"The reason why the display does not go on at the same time you are taking the picture is because the LCD screen on the back and the CMOS chip are so closely put together that they create a magnetic field and can distort images. The field, for some reason isn't strong enough in a CCD , which is what most point and shoot digital cameras use, however this is not the case with the higher end CMOS chip."
The comment above makes some sense, since the 35 mm SLRs have to use the same lenses they did with film, so the CCD has to sit exactly where the film did, relative to the lens. If you're designing a point-and-shoot, you can use a lens that focuses closer to the front of the camera to keep it away from the CCD. The 35 mm digial SLR manufacturers could make the back of the camera fatter, but maybe it would be so fat that it's not worth the trouble.
You can, of course, use the old flashes in manual or semi-automatic modes, but with the introduction of TTL, all my brain cells that remembered how to do this died.
The digital CCD is smaller than a standard piece of 35mm film, so you are effectively cropping out the center of each shot. You paid gobs of money for lenses that could produce a good image over the entire 35mm film area, and you're now using only 2/3 of it or so. This makes it very difficult to take wide-angle photos. If you compare full-frame shots taken with the same lens on a film and my digital 35mm cameras, digital camera effectively multiplies the focal length by about 1.5. So if you slap on a 300mm lens, it's like using a 450mm lens on a film camera.
This seems cool if you use telephoto lenses a lot, but if you want the equivalent of a 20mm wide-angle shot on a digital camera, you'd need to purchase a 14mm lens, which is not a cheap prospect at all.
When I think of a pixel, I think of one spot on a computer screen that can display any color; a combination of red, green and blue. But on a camera, the advertized pixel count is simply the number of receptors. Each receptor can only (except on the Sigma Foveon cameras) receive one color -- red, green or blue. So at first glance, it looks like you should divide the advertised number by 3. But it's worse -- you really need to divide by 4, since the green receptors don't work as well and almost every digital camera uses two green receptors for each red and blue.
What the digital cameras then do is to take the data from all these red, green, green and blue receptors and expand it to be a much larger number of pixels with clever extrapolation algorithms.
On a digital SLR, however, it isn't quite that bad, however. A typical pixel on a computer screen has 8 bits of accuracy in each of the channels. In other words, it can distinguish 256 levels of red, green and blue for a total of 236x256x256 colors -- around 16 million. But the receptors on the Nikon, at least, can distinguish 12 bits of color for 4096 different shades of red, green and blue. This extra information makes the extrapolation more efficient, but it is hard to know the exact relationship between the advertised number of pixels in a digital camera and the number of "good" screen pixels you'll get on your computer.
The penalty for storing images in raw mode is simply the size. On my D100, each raw-mode image requires 9.6 megabytes of storage which can add up fairly quickly. If you get a 1 gigabyte IBM microdrive, it holds 103 raw-mode images on the Nikon D100. By contrast, it would hold more than 300 of the highest-quality jpeg images, or thousands of the low-quality images.
Here are some additional notes on RAW mode that you may find helpful.
If you decide to take raw mode photos on vacation, you'll need to take a laptop along to transfer your images every evening. This has the added advantage that you can toss out the ones that are bad, but not so bad that they don't look OK on the monitor on the back of the camera body. Obviously, for a day trip, 100 photos is usually plenty, especially since you can blow away the obvious errors just after they're taken.
There is also a limit to how fast the data can be read out of the CCD to the buffer. On my camera it's about 1/3 second per shot. The bottom line is that I can take at most 3 shots per second, for up to four shots. Then I have to wait about 4 seconds for each additional shot. This doesn't seem too bad, but for flying birds or fast-moving sporting events, it's a pain in the butt, where a Nikon F5 can take 8 shots per second for a whole roll of 36.
Want to send me mail? Click: ( firstname.lastname@example.org)