First a few examples of the problem as I see it online every day. I am using the FamilySearch.org Memories program for my examples because the collaborative nature of the program gives me a good cross-section of the genealogical community.
MyThird-Great-Grandfather, George Jarvis, has 98 Memories on FamilySearch.org. I am not particularly worried about duplicates, but I would like to point out that the following uploaded photos included in his Memories are all very likely reproduced from exactly the same original photograph.
Visually speaking, the best of these images is a tie between the first one above and the last one. But let's look at the first image a little closer. Here is a zoomed in copy of the first image.
It looks OK. Here is the reality. The original photo is the "control." Whatever quality and clarity possessed by the original is fixed. You cannot get any better reproduction than the original photograph. Modern digital image processing may be able to "enhance the image and bring out some details hidden in the original, but we are essentially stuck with whatever information is conveyed by the original document, photo or whatever.
A good analogy is a glass filled with water. When the glass is full, that is all the water we can get out that glass. If I put the water in the glass into a different container, I will either get all the water in the glass or less. Every time I change the container, I will lose some water, either to evaporation or because of the size of the container. Here, the original photograph is the original container. Each succeeding copy of the original by scanning etc. decreases the quality to the copy.
Here is a zoomed in version of part of the best of this string of photos.
What we are seeing here is the loss of information from the original photograph. The squares are pixels, or distinct image elements that make up the digital image. The optical resolution (or detail) of the image is measured in dots per inch of DPI. Some archivists also use a measure called lines per inch or LPI. In the commercial world of photography, the same resolution or amount of detail is measured in Megapixels or MP. In all three, the implication is that "higher numbers are better." This is not always true for a variety of very complex limitations imposed by the way light is physically transmitted and further by the way our eyes perceive details.
In all three types of measurement, the resolution depends on the distance between two distinguishable radiating points. If I take a digital photograph or a scan of an original photograph, the amount of detail preserved (for this discussion, the resolution) will depend on a number of factors: the quality of the lens, the resolution of the original image, the resolution of my device's sensors, the output I choose, the quality and resolution of the display when I look at the final image.
If I go to Amazon.com or some other retail outlet and I look at a medium priced digital camera today, I will find something like the following.
Nikon Coolpix L340 20.2 MP Digital Camera with 28x Optical Zoom and 3.0-Inch LCD (Black)
This camera costs about $140.00. If I look at a flat bed scanner on Amazon, I might find the following:
This particular scanner costs about $65.
How do they compare? If I had the original George Jarvis photo, which one would give me the highest resolution reproduction i.e. the best quality? How do you know the answer? If you are the average genealogist and not particularly interested in either scanning or photography, you probably rely on the opinion of some "expert" who will tell you immediately that the flat bed scanner will give you a better quality digital reproduction. But I would guess that the "expert" has probably not tested those exact same devices. Neither have I and I really couldn't say which would give you a better quality output. The scanner is definitely cheaper and would probably give you a "good" image reproduction, but then again, you can't use the scanner to take pictures of your family at picnics and so forth.
Are we forced to buy both? How does 20.2 MP compare to 4800 x 4800 dpi? Can they be compared?
The whole concept of digital images is that they are made up of discrete dots usually called pixels. In fact, you can compare the two. The Nikon 20.2 Megapixel camera takes an image with a maximum of 20.48 pixels. The Epson Scanner, theoretically, has 40,800 x 56,160 pixels or a total of 2.2 billion. But there is a significant difference. The number given for the scanner is spread across about 8.5 x 10 inches of the bed of the scanner. But the real issue is the resolution used to make the scanned image on the scanner. You can select the "DPI" for your scan. Most scanning is done at about 300 to 600 dpi. Quoting from an article called "Digital Images and Genealogy"
If we took a 4" x 6" photo for instance and scanned it in at 600 dpi, we would get a digital image that is 2400 x 3600 pixels in size which is 8.6 megapixels. If we took an 8.5" x 11" photo and scanned it at 600 dpi we would get a digital image that is 5100 x 6600 pixels in size which is 33.6 megapixels. Given that many scanners today can go to at least an optical resolution of 3,200 dpi, scanning an 8.5" x 11" page at that resolution would result in a 27,200 x 35,200 pixel size image which is 957.4 megapixels.Even this explanation takes some qualification. Again, you cannot get more information (resolution or detail) than is contained in the original. The real answer is simple however. Cameras and scanners are both specialized tools that have some overlap in actual use. But the scanner makes higher resolution images of the same photo than using a camera, even a very high end, expensive camera would make. But the scanner is mostly physically limited to flat, portable items like paper and photos. Cameras can make images of those objects and many other multidimensional ones as well. Bottom line, you need both.
We can also look backwards, an 18 megapixel camera that takes photos with 5184 x 3456 pixel dimensions will resolve a 6" x 4" photo at 864 pixels per inch. If it took a photo of an 8.5" x 11" page, it would be resolving it at 406 pixels per inch.