The following is a guest post by Barry Wheeler, Digital Projects Coordinator, Office of Strategic Initiatives
- What is resolution?
- What resolution should I look for when I buy a scanner?
- What resolution should I use when using my scanner?
These are questions we hear frequently when speaking to people about their digital conversion projects. Unfortunately, the questions are hard to answer. The material can get very technical and can be difficult to apply. So I’ll try to answer the first question now and the second two questions in a follow-up blog post later on. As always, feel free to ask questions or make comments.
Here’s a sample scenario: You’re in the store looking at a scanner and the box contains perhaps the cheapest legal sized desktop scanner available. In big print the manufacturer claims, “2400 x 4800 dpi.” So, you have a very inexpensive scanner with very high resolution. This is just what everyone wants. But is this the only consideration?
Remember this – resolution is important but it is only one measure of scanner quality. Other measures are also as important but won’t be discussed in this series of posts. For example, two other important characteristics are:
- the range and accuracy of different colors the scanner can capture
- the ability of the scanner to capture details in shadows and highlights.
If color accuracy and fine image detail are important, you have to consider much more than simple resolution.
Now, to begin to answer the first question – what is resolution? Let’s take a look at how a scanner works. Figure 1 shows a simple scanner in action. A hidden stepper motor and gear train pull the light and sensor assembly along a geared track. As it moves, a light shines into the light guide and is focused up toward the page. The motor moves the assembly along in very small steps – the manufacturer specifies a rate of up to 4800 steps for each inch the capture line is pulled across the document.
Figure 2 shows the main scanner parts from below. The stepper motor and gears are shown along with a portion of the light guide. The light guide is an optical plastic rod that spans the page width. The LED lights shine into the guide from one end; the baffles and mirrors guide the light onto the page and back onto the sensor.
Figure 3 shows a small section of the sensor. The sensor is a row of 2400 light-sensitive diodes per inch. Each time the motor moves the scanning head a single step, the light along a row is bounced off the page and onto this line of tiny light-capturing elements.
Now we can understand the manufacturer’s claimed resolution. Since with each step, the reflection is measured along the row, the manufacturer claims a maximum resolution of 4800 rows per inch. Since there are 2400 sensors per inch in the row and each sensor measures the reflected light at one dot on the page, the manufacturer claims a maximum resolution of 2400 dots per inch in each row. Thus, the manufacturer claims a maximum resolution of 4800 dots per inch x 2400 dots per inch.
But the International Standards Organization (ISO) does not accept this claim.
With each step, each of the diodes attempts to measure the reflected light at one point. Technically, each diode measures a sample of the reflected light at each point. So, the ISO claims a maximum sampling rate of 4800 dots per inch x 2400 dots per inch.
Does this make a difference? Aren’t the terms “resolution” and “sampling rate” just different words for the same thing?
No. These words make a huge difference – believe the numbers and you may be very disappointed. I’ll explain this later on in Part 2.