In Association with Amazon.com

Document Scanning

The results of scanning a 5" x 7.5" document at different color modes and resolutions are discussed below. The scans were saved in three common file formats to determine the corresponding file size and reproduction quality.

A Black and White, a.k.a. lineart, scan at 300 samples per inch saved in GIF format provides a fair quality reproduction with a modest file size. This combination can provide a good image for electronic transfer. Not all scanner software can generate GIF format images however, in which case it is usually best to generate a collaborative file using the grayscale mode.

Grayscale scans are more versatile than Black and White scans and can be produce collaborative, service, and archive images. A 150 spi JPG scan can function as a collaborative or service image, with a 300 spi JPG providing enhanced screen and print quality. A 300 spi TIF image can serve as an archival file if the original document has little or no color information.

Truecolor, a.k.a. 24 bit, scans can reproduce all of the color information in a document.

Image categories

  • Thumbnail: small image, e.g. 100 pixels x 150 pixels, used for content preview.

  • Collaborative: limited file size, e.g. < 200 KB, for efficient image transfer. Formats: GIF, PNG, JPG.

  • Service: good-quality image intended for short-term use. May be compressed for efficient storage and transfer. Formats: GIF, PNG, JPG, TIF.

  • Archive: lossless high-quality master image free of compression artifacts. Used for reproduction and long-term storage. Requires reprocessing as image standards change. Formats: TIF, PhotoCD.

Black & White (Lineart)

File Quality
Resolution
(samples per inch)
Format Size¹ Print² Screen Screen
Size³
Comment
100 GIF 24 KB Poor Bad 1.25x  
  TIF 49 KB Poor Bad 1.25x  
150 GIF 19 KB Poor Poor 1.9x  
  TIF 108 KB Poor Poor 1.9x  
300 GIF 61 KB Fair Fair 3.75x Collaborate
  JPG 123 KB Fair Poor 3.75x  
  TIF 433 KB Fair Fair 3.75x  
600 GIF 175 KB Good Fair 7.5x  
  JPG 384 KB Good Poor 7.5x  
  TIF 1,680 KB Good Fair 7.5x  

1 JPG images at 4:1 compression. TIF images are uncompressed.
2 on a 600 dpi laser printer. Scanning resolution should match printer resolution for best quality lineart prints.
3 on a 17" monitor @ 1024x800

IE float bug
300 spi
IE float bug
150 spi

After setting the color mode and resolution the most critical parameter for Black and White image scanning is the Threshold . In a Black and White image all the intermediate gray pixel values that appear on screen are rendered as black or white. The Threshold value determines the gray level that distinguishes a white pixel from a black pixel in the final image. Lower threshold values cause more of the gray samples in the original document to be rendered as white.

The effect of the Threshold value on a document scan is illustrated in this animated black and white screen capture.

Grayscale

  File Quality    
Resolution
(samples per inch)
Format Size¹ Print² Screen Screen
Size³
Comment
100 spi GIF 267 KB Poor Very Good 1.25x  
  TIF 385 KB Poor Very Good 1.25x  
150 spi GIF 632 KB Good Excellent 1.9x  
  JPG 210 KB Good Very Good 1.9x Service & Collaborate
  TIF 863 KB Good Excellent 1.9x  
300 spi GIF 2,612 KB Good Excellent 3.75x  
  JPG 852 KB Good Excellent 3.75x Service
  TIF 3,448 KB Good Excellent 3.75x Archive
600 spi GIF 9,830 KB Good Excellent 7.5x Enlargement
  JPG 2,865 KB Good Excellent 7.5x  
  TIF 13,793 KB Good Excellent 7.5x  

1 JPG images at 4:1 compression. TIF images are uncompressed.
2 on a 600 dpi laser printer.
3 on a 17" monitor @ 1024x800

150 spi GIF

This GIF format image scanned at 150 spi has excellent on-screen quality

IE float bug

Grayscale document scans can usually be improved by manually setting four parameters. Instead of being limited to a single threshold value in a Black and White image, grayscale images have two thresholds.

The White-Point sets the minimum threshold value for white pixels. Samples greater than the White-Point will be rendered as full white. The White-Point is usually set to a value slightly below the rightmost peak in the histogram.

The Black-Point determines the maximum threshold value for black pixels. Samples less than the Black-Point are rendered as full black. It is usually set to a value slightly greater than the leftmost peak in the histogram.

Documents with a large number of gray samples between the White and Black Points will benefit from Gamma correction. Lowering the Gamma value will lighten the midrange pixels and emphasize the darker tones, raising it will darken the midrange pixels and emphasize the lighter tones.

Most scans can be improved with some degree of Unsharp Masking. It should be the last step in the image editing process and should not be performed in the scanning program if the image is going to be edited.

The effects of the these values on a document scan is illustrated in this animated grayscale screen capture.

Truecolor (24 bit)

File Quality
Resolution
(samples per inch)
Format¹ Size Print Screen Screen
Size²
Comment
100 spi         1.25x  
150 spi JPG 582 KB   Excellent 1.9x  
  TIF 2,584 KB Poor Excellent 1.9x  
300 spi JPG 2407 KB   Excellent 3.75x Service
  TIF 10,337 KB Good Excellent 3.75x Archive
600 spi JPG       7.5x Enlargement
  TIF   Good Excellent 7.5x  

1 GIF format can not encode at 24 bit/pixel. JPG images were compressed at 4:1 compression. TIF images are uncompressed.
2 on a 17" monitor @ 1024x800

150 spi JPG

JPG format image scanned at 150 spi has excellent on-screen quality

Pages