cgardner
Offline
Image Upload: Off
|
 p.1 #4 · Techn. question: dpi? | |
The term "dpi" has it origins in offset printing where an continuous tone image was reproduced by two colors - white paper and black ink - by breaking the image down into a pattern of dots using a screen.
The original halftone screens where made of glass, etched with a pattern of lines 85, 120, or 133 lines to the inch with the lines filled with carbon black. Then two line screens were sandwiched together at 90 degree angles to form a checkerboard pattern with the same frequency. To create a halftone the glass screen was suspended a few thousands of an inch above a sheet of high contrast lithographic film. That air gap caused the screen pattern to degrade in the same way a diffuse light source created a fuzzy edge shadows. The gradation in light intensity which resulted is what determined how large a dot was exposed on the litho film. Producing an halftone in those days was a real act of craftsmanship. Kodak simplified the process by introducing a continuous tone screening process. Instead of a pattern of sharp lines the screen consisted of a piece of film exposed with a continuous tone pattern of dots which was put over the litho film on the film back of the copy camera, held in place by a vacuum.
Laser drum scanners and laser printers form halftone dots differently. The actual dots on the paper can still only be solid or white paper.
Each "hard" printing dot is actually formed by a pattern of smaller dots created by the laser beam. On a Hell DC3OO scanner used in the early 1980s a laser beam was split with mirrors into six separate beams. Those six beams passed thought LCD light valves which controlled the printhead which was connected to valve assembly via 6 tiny fiber optic strands. As litho film on a rotating drum passed over the printhead various combinations of the 6 fibers would expose the film to form the dot. So each printing dot was actually formed by a 6 x 6 matrix of smaller dots created by the 6 fibers in the print head. The original scanners where film based. You'd put a transparency on a clear drum, and the film on the imaging drum, set the scanning parameters and then separate lead screws would move the scanning head over the transparency at the same time the laser beam was creating the dots of the film. The electronics were quite primative. The DC300 used simple high voltage transistors to control the light modulators and actually had a 1 KB "core" memory consisting of little magnetic donuts suspended on wires. It was a challenge just to keep the darn thing running.
Digital imagesetters and laser printers still operate similarly in how the printing dots are formed on the paper because offset ink and laser toner are binary: either there is a solid dot or white paper. What controls the tone is how large the dots are an the dots are formed in way similar to the old DC300: a straight line printhead forms each printing dot from a matrix of smaller laser created dots. A laser printer which prints 300 dots per inch might have an printhead which creates each dot with an 8 x 8 matrix with a printhead density of 300 x 8 = 2400 imaging sites per inch.
Ink jet printers differ from imagesetters and laser printers, controlling the size of dots with the volume of ink which is deposited. But the printheads are similar in concept, a row of orifices with fixed spacing. The way inkjet resolution is specified (e.g. 2400 ppi) actually represents the resolution of the printhead x the number of colors. So 2400 ppi resolution might actually be four separate overlapping colors created with printhead orifices which are spaced 1/300th of an inch apart. The more colors, the higher the resolution rating becomes.
When the first digital imaging systems were created it was determined, by experimentation and observation that an oversampling rate of between 1.7 and 2 image pixels per printing dot was needed to convert a pixelated digital image into a halftone pattern which would be accepted perceptually as continuous one at normal reading distance. That threshold for offset printing is in the 133 to 150 dots per inch range for a B&W image. Thus was established the convention of using 300 pixel per inch digital images for offset printing reproduction. 300 ppi was the optimal 2/1 down-sampling ratio to the the 150 dots per inch screen ruling commonly used for color offset printing. 150dpi x 1.7 = 255ppi represents the minimum level of resolution which will typically acceptable.
That's an anecdotal account of the evolution based on my experience which dates back to shooting halftones conventionally in the photo labs at National Geographic and teaching lithographic process photography. NGS still had a glass screen on one of its cameras so I had an opportunity to use and understand how it worked.
How images are perceived and how much resolution is necessary for acceptable reproduction depends a number of factors such as the output device, viewing distance, and how critical the standards of the viewer are. Displays such as billboards can have very low output resolutions on the order of 30 dots per inch, but still be perceived as continuous tone by someone in whizzing by in a car at 70 MPH. The same photo will look different at reading distance when output on different types of printer. So the best way to determine what the best resolution is for a particular size image is to print samples using various methods, put them up on the wall at the distance and lighting conditions they will be viewed at and judge visually which looks best.
Chuck
|
