The process of scanning graphic images has many pitfalls. Resolution? Color depth? OCR? This section will demystify some of the complexities of scanning and help you produce better scans without creating huge, unworkable files.
"Scanning Resolution" refers to how finely spaced the dots that comprise a scanned image should be. A typical resolution
is 600 dots per inch (600 dpi), which means that for every inch across the image, the scanner reads 600 tiny dots of
color (or "Pixels").
Choosing The Optimum Resolution
The most common problem encountered when scanning an image for printing, especially with today's hi-res scanners, is that of choosing the ideal scanning resolution.
- Too low a resolution will produce blurry printouts;
- Too high a resolution will take much longer to scan and will create a huge, unmanageable file that takes up enormous amounts of disk space.
Don't Scan More Than You'll Print!
Beyond a certain point, scanning at even higher resolution will not result in higher print quality. The reason for this is that while scanned images potentially contain millions of colors, most color printers use only four individual colors, known as CMYK. This gives them effectively only eight colors to work with:
- The pure colors Cyan, Magenta, Yellow & Black;
- The mixed colors Blue (=cyan+magenta), Green (=cyan+yellow) and Red (=magenta+yellow);
- The color White (unprinted paper).
Printers Use "Dithering"
To achieve what looks like a full rainbow palette of colors, printers use a process called "dithering" to mix those dots of color. For example, to create the purplish-blue of this Rainbow Lorikeet's face might require four dots of cyan, three of magenta, two of yellow and one of black. The dots are both small enough and close enough together that we perceive them as that shade of purplish-blue.
As a result, color printers generally only achieve an effective resolution of around 1/4 of their actual resolution when printing full-color images because the dots of color are scattered over a larger area. The exception is when they are printing any of those eight "pure" colors, which they can therefore output continuously without "dithering". (eg. The solid black lines in this Rainbow Lorikeet image, or pure black text, or a pure magenta border).
Scanning resolution – Rule Of Thumb
A fairly good general rule for choosing the scanning resolution based on the type of image and expected use is therefore as follows:
- For black and white images, such as logos, scan at the same resolution at which you will be printing. For example, if you have a 600 dpi laser printer, scan the logo at 600 dpi.
- For color images, such as photos, scan at around 1/4 the resolution of your printer. Thus, if you are printing to a 1440 dpi color printer, scan at about 360 dpi.
Enlarging or Reducing
Note that if you will be printing an enlargement, you should increase the scanning resolution accordingly, or decrease it when printing smaller.
For example, let's say you were scanning a 6" × 4" photo and printing to a 600 dpi color printer. If you want the printout to be the same size as the original, you could scan at 150 dpi (= 1/4 the resolution of the printer). If instead you wanted the printout to fit onto a full A4 page (11.69" × 8.27"), which is about double the size, you should also double the scanning resolution to about 300 dpi. This doesn't need to be exact – you could always scan at a little higher resolution just to be sure.
Scanning Color Depth
"Colour Depth" or "Colour Mode" refers to how much color information should be scanned or saved for each dot (pixel). The higher the color mode, the better the quality. The tradeoffs are that it takes a longer time to scan and that the finished graphic file takes up more disk space.
Common Color Depths
There are four common color depths used by scanners:
- Full color (also known as 24 bit or 32 bit color);
- Indexed color (also known as 8 bit color);
- Greyscale (also known as Black & White Photo);
- Monochrome (also known as B&W, Bitmap, or 1 bit).
With today's computers being both fast and high capacity, it's quite feasible nowadays to simply scan everything in Full Color mode, which is the native mode of the scanner. The scanned images can then be converted to lower quality modes later if required.
Color Depth – Rule Of Thumb
A fairly good general rule for choosing the color depth based on the type of image and expected use is therefore as follows:
- Color photos & pictures should be scanned in Full Color mode;
- Sepia photos, or old black and white photos that have faded or yellowed and for which you want to retain the character, should also be scanned in Full Color mode;
- Black & white photos & pictures, or ones that you will be printing in black & white, should be scanned in Greyscale mode;
- Black & white logos should be scanned in Monochrome mode;
- Black & white text that you want to reprint should be scanned in Greyscale, as this reproduces subtle details better than Monochrome mode.
Scanning Text / OCR
Some scanners also have a mode called "Text" or "OCR" (Optical Character Recognition). This differs in that it does two things: Firstly, it scans in the best mode to give a clear image with good contrast, then it proceeds to "Read" the scanned dots and attempts to convert the picture of the page into actual computer text.
For example, a single letter "G" would take dozens of bytes to store all the dots that make up its scanned image. By using
recognise those dots as the letter "G", it then only takes a single byte to store the value "71", which is the
computer's code for the letter "G".
OCR has two key advantages. Firstly, the converted text page requires much less storage than the scanned image of that page. Secondly, the resulting text can be edited in a word processor, sent via e-mail, etc.
NOTE: When scanning from newspapers or other thin paper with printing on both sides, use a dark backing sheet behind the page you are scanning so that the printing on the other side doesn't show through.
Saving Scanned Images
There are dozens of graphics file formats in use today, each with their own particular strengths and weaknesses. See my separate page on Graphics File Formats for more details on some of the more common ones used on the Internet.
Scanning For The Intended Use
Assuming that it is generally either photos or logos that are being scanned, and that they are generally being used either for printing or for the Internet (web sites, e-mails, etc), this simplifies the choices somewhat.
Saving Scanned Images – Rule Of Thumb
A fairly good general rule for choosing the file format based on the type of image and expected use is therefore as follows:
- Printing Use:
- TIF is generally best.
- Internet Use:
- JPG is best for photos;
- GIF or PNG are best for logos.
The reason TIF format is suggested for print use is that the resulting file is as close as possible to the original scanned image, whereas JPG and GIF have each undergone compression and have lost some image quality. TIF is also very much the defacto standard for use in the printing industry, although PNG is now fast gaining acceptance.