### Simple Scanner question

greenspun.com : LUSENET : Large format photography : One Thread

This may be a stupid question- when a scanner is described as having a resolution of, say 720x1440 dpi, what is that first dpi number refering to? Is the resolution different on the x axis and the y? Is 720 the true scan resolution and 1440 is inerpolated? Than

-- Tom Gorman (honeychrom@aol.com), August 25, 2001

Tom - the first number refers to the optical resolution, and the second, usually larger number refers to the stepping of the motor. In your example, 720 would be the optical resolution.

-- Michael Mahoney (mmahoney@nfld.com), August 25, 2001.

Aha.. OK, I understand what the optical resolution number tells me, but does the 'stepping of the motor' number tell me anything/mean anything?

Also, while we're on the numbers game, what exactly do the two numbers mean when used to describe a PRINTER (ie 2882x720 dpi)? I've noticed that some printer descriptions have two numbers, sometimes the first is larger, sometimes the second, but then some printers (ie. Epson 3000, 5000 etc.) simply say 'true 1440 dpi resolution.' Thanks

-- Tom Gorman (honeychrom@aol.com), August 25, 2001.

WAIT!!! When they say 720x1440, they're generally referring to the x and y resolutions of the scanner or printer. They're usually not the same due to the design of the scanner or the printer. For example, the horizontal resolution of a scanner is going to be determined by the spacing of pixels in its ccd whereas the vertical resolution is going to be determined by how accurate the motor which moves the scanning ccd is.

For example, my epson can print at 720x1440. It prints at 1440 dpi across the page and 720 dpi along the page. (although I probably have these directions reversed).

Generally, with scanners, they will give you a separate number for the interpolated resolution.

-- Nathaniel Paust (npaust@nmsu.edu), August 26, 2001.

Tom, the two numbers on a flatbed scanner represent the X & Y axis, true optical capability....but even that you have to be careful because makers tend to mix up the optical and interpolated resolutions to gain a marketing advantage. It is quite normal for flatbeds to scan higher optical across the axis which requires no stepper motor. However, with drum scanners, this does not hold true since there is only one scanning eye going through a rotating drum which holds the image. As for printing dpi, they are not to be confused with scanning ppi, pixels per inch. A printers dpi, dots per inch, is an indication of the number of dots the printer will lay down in both directions....for example, Epson 1440x720 refers to 1440 dpi going left and right, and 720 dpi going up and down...meaning the weak link is always the paper handling capapbility.... Roland is the only ink jet printer that prints true 1440 x 1440 dpi... in general, the more dpi, the more resolution. However, there is many other tricks to making a print look sharp including the number of ink colors, the dithering pattern and ink/paper combination.

DPI and PPI are not equal, but quite often these terms are mistanely intermingled. Most scanning software asks you what final dpi you want, and it reverse calculates the amount of ppi required to acheive such. So there is little need to determine ppi in todays applications.

-- Bill Glickman (bglick@pclv.com), August 26, 2001.

As others have already said, along one axis the number indicates the density of the CCD array. Along the other, it shows the smallest movement of the stepper motor. In the ideal case, this should give you the pixels per inch (PPI) resolution of the scanner.

However, from reading other threads in this forum, the ideal case does not always exist. Scanners may have optics which degrade the image quality, the stepper motor might not reliably step at the smallest increment, and CCD's may have some artifacts.

If you read through some of the archived threads on scanners, some forum members have tested scanners and found the true resolutions to be lower than the "CCD density/stepper motor" numbers. The manufacturers want to claim the highest numbers they can get away with, but they may be less that truthful!

-- Michael Chmilar (chmilar@acm.org), August 28, 2001.