What is the scanner resolution. The main technical parameters of the scanners. Scanning in the reflected light

Above, I explained why for scanners it is incorrect to use the term point per inch (DOTS-PER-INCH - DPI). In various types of input and output devices that are connected to computers, the resolution is measured in different units corresponding to this type of device. The resolution of printers and computer photophonation devices is measured at dots per inch, which are considered to be real dots that are applied with these devices on paper or film. These points are usually oval or round, and in some devices, their size can really change.

In such specialized printers, as printers with a thermal separation of the dye, even the intensity of individual points may vary. Thanks to the nature of the dyes used, the amount of the portable dye can be varied from 0 (no dye) to 255 (full color). In all cases, regardless of variation, these devices depict the points and their permission to measure the correctly using such a criterion, so the number of dpi points.

Computer displays are measured in other units, pixels. The resolution of the monitor can also cause confusion due to such a term as the point position (dot pitch, the distance between two pixels of the same color on the screen). With respect to the computer display, it is also easy to imagine such units of measure as pixels per inch (Pixels-Per-inch - PPI), although their number in reality may not be as important as you used to count. Your computer knows nothing about the number of pixels per inch on the display monitor. Everything that worries it is how many pixels in length and is placed in the height on the screen. The same monitor can be configured to 1024x768 pixels, 1600x1200 pixels or (like me) at 1920x1440 pixels. Higher resolution does not make an image that you see, clearer. It simply makes less separate elements. With an increase in the resolution of the screen, the elements such as the menu, dialog boxes and cursors are becoming smaller, freeing the additional space for other elements, such as drawings. The size dialog box, say, 400x400 pixels when resolving the monitor 640x480 can occupy almost the entire screen. At a higher resolution, 1920x1440, it will remain equal to exactly 400x400 pixels and will not become more clear. It will just be three times less.

As mentioned above, many people come foolishly using settings operating system Windows to change the "resolution" of the monitor to 72, 96 or 120 dpi. These settings do not change anything on the screen, except for font sizes, allowing you to vary them to create a comfortable look external view. The unproductive images will look on the screen in the same way at any of these "permissions."

Points per inch or lines per inch

When working with printers or computer photonoaming devices, do not confuse if you encounter the term line (or strings) on the photo (Lines-Per-inch - L PI). As you should be known, photos are printed from the extravagne halftone. Since in many types of output devices, the size of the points cannot be changed enough to describe all the various shades and colors present in the photo, halftone is used to create an optical effect. Plots of the image on the screen are divided into tiny points of different sizes, which human eye perceives pumped, resulting in white, black, pastel and all other colors of the image. These points are built up due to various combinations of printer points using large points like the matrix. For example, if one of these large points, called cells, has dimensions of 8x 8 printer points on each side, it contains from 0 to 64 points. When printing, our eye can see any of the 64 shades, which are described by these combinations.

The size of the cells used determines the resolution of the semolone screen, which is measured in the lines per inch. Suppose the printer can give 1200 dpi. If you split these 1200 dpi on cells of 8 points, we obtain that the maximum number of rows per inch, which can output this printer, is 150 (1200 divided by 8). The number of lines on an inch, or a raster lineture, defines a visible photo resolution, as well as the number of tones that can be transmitted. For example, if you use a greater cell size of 1010 points, it will give 100 shades, but the permission will fall up to 120 lines per inch. If you do not actively work with halftones, you probably will not need all this information, but in any case it will help to avoid confusion with dots per inch and lines in an inch. In fig. 3.2 For example, 16 combinations of points are possible for a square cell of the printer with a side of four points.

At the beginning of this chapter, I mentioned that the scanner resolution is correctly measured in samples (and not points) per inch. These samples are collected in the sensor array, as mentioned above in this chapter. The resolution of the scanner in the X direction (original width) is determined by the number of individual sensors reading each line. The required number of sensors calculate quite simply: the 35 mm film scanner can read the image from the slide or negative, the width of which is approximately equal to one inches; Consequently, for a scanner with a resolution of 4000 samples, an inch requires exactly 4,000 individual sensor elements.

For larger films, the required number of sensors increases. For a film scanner capable of taking a film of 6 cm wide (approximately 2.25 inches), the image with a resolution of 4000 elements per inch is needed, a number of 9000 sensors are needed. Tablet scanners, as a rule, scan reflective originals with a width up to 8.5 inches (21.5 cm), so to obtain a resolution of 3200 samples per inch that requires 27 200 sensors. Of course, tablet scanners that are used to work with the film can perceive and diapoys for negatives, which are much smaller than the scanning box itself - from 1 inches (2.5 cm) width (35 mm film) to sheets of film 4x5 inches (10x13 cm), therefore, when scanning a film, not all sensors of the tablet scanner are used. As you can guess, the development of arrays of sensors with such a huge number of individual elements is not easy.

The resolution in the second direction y (length of the scanned image) is determined by the distance that the sensor passes between the lines. It's called stepper increase. This scanner will be incapable of a sufficiently small movement of the sensor vertically so that the vertical resolution corresponds to the horizontal resolution. That is why you can meet such scanner specifications as the resolution of 2400 SPI horizontally at 1200 SPI vertically. In most cases, seeing such specifications, you can say that the seller is extremely honest. In the manufacturers' environment, there is a tendency to try to adjust the value of the vertical permission to the value of horizontal, even if for this they have to distort the numbers a bit.

For example, the scanner claimed that its resolution vertical ability is 2400 SPI, in fact it may not give so many rows in one inches. The array itself is probably wider than 1/2400 inches in this direction, so the only way to allegedly move the sensor to such a small distance is to make the lines to some extent overlap. Pondering this topic, it can be understood that the CCD sensors have one more potential advantage over the elements of the KSI in the direction of Y. If the size of the CCD sensor is less than the horizontal stroke of the scanning, it will also be less and in the vertical direction, making it possible Stepper increase. The physical dimensions of the KSI elements, on the other hand, more, so moving them vertically with a tiny step is much more complicated.

So, if we talk directly, the horizontal and vertical resolution of the scanner may not be quite as described. The permission also affects the optics, which is used to focus the image, the size of the scanned area and the shape of the sensor.

A more truthful measure of sharpness is a frequency-contrast characteristic, which takes into account that simple fact that optical systems (including the built-in scanners) focus all the main colors of the spectrum at all points. Scanner developers understand this, and also know that the human eye is more sensitive to one colors (for example, green), and to others (for example, blue) less. Therefore, optics and sensors are developed in such a way as to optimize the sharpness of the green channel, in a sense at the expense of the blue channel, since in this case the general picture will seem "clear." In essence, in the scanners up to 60% of the picture form an image of a green channel, 30% - the image of red and only 10% is the image of the blue channel. All of these factors contribute to the clarity of the final image, which absolutely cannot be associated with the "raw" value of the resolution in samples on an inch (or dots per inch), which the scanner sellers are so liked.

Information for specialists!

If you really are interested in this material or you just wanted to familiarize yourself with all items that are taken into account when calculating the real definition of scanned image, find the two new ISO specifications in the Google search system (International Standards Organization - International Organization Standardization) designed for scanners. IS016067-1 relates to standards for measuring the resolution of scanners designed for reflective originals, and those who are interested in scanning films will more interest the IS016067-2 standard. There is another standard, ISO 21550, which relates to a dynamic range. It should be noted that any suggested standards before becoming official international standards must pass the stage of work projects, projects for consideration by the Committee and several more stages.

Scanner - device for entering graphic raster information in computer. The list of scanner applications is almost endless, the following varieties of these devices have developed and produced:

• high-quality drum scanners that are capable of processing both transparent and opaque images - from 35 mm films to 16 feet materials by 20 inches high (over 10,000 TND) resolution;
• tablet desktop scanners of universal destination;
• compact document scanners designed exclusively for optical reading and recognition of documents;
• special photoskanners that work by moving a photo on a fixed light source;
• scanners of slides or negatives working with transparent images;
• manual scanners for use on a small table space.

• a - Tablet (Flatbed) Epson Perfection 3490 scanner;
• b - document scanner (Pass-Through Scanner) kodak i30;
• in - Movie Scanner (35 mm Film Scanner) Nikon Coolscan 5000 ED;
• mustek manual scanner.

Device and functioning of scanners

The scanner is a device that converts the visible image into the binary signal stream, in other words, the transformation of optical analog data into electrical digital.

The image is placed in front of the carriage, which consists of a source of lighting and array of sensors.

Light from the tube enters the sensors, which read optical data (for example, CCD), then prisms, lenses and other optical components passes. Like points or lups, these elements can differ very much in quality. High-quality scanner uses precise glass, enlightened optics with color fix filters. In cheaper models, plastic components are applied to reduce costs.

The light intensity reflected or passed through the image and the sensor collected is converted into voltage proportional to the light intensity.

Sensors scanners

The image sensor is usually implemented by one of three technologies:

• photoelectron multiplier (FEU or Photomultiplier Tube - RMT) - technology inherited from the drum scanners of the past;
• charging device (CCD or CHARGE-Coupled Device - CCD) - Sensor typical for desktop scanners;
• contact Image Sensor - CIS - More new technologywhich integrates features and allows you to create more compact scanners.

Technology of photoelectronic multipliers

FEU - Technology of high-performance color drum scanners sensors that are usually used to prepare color printing matrices. Expensive and heavy in service, they were the main input devices in the computer to the appearance of desktop scanners.

The original image here is carefully fixed on the cylindrical drum, which begins to rotate at high speed. The carriage with sensors and illuminators begins to move along the image. You can control the resolution or size of the image by selecting the speed of the carriage, the optical force of the lenses and the radius of the drum.

FEU scanners have two light sources, one for scanning in the reflected light, the other - for transparent originals. The backlight light splits into three beams that pass through the light filters (red, green and blue), and then enter the photomultiplier tube, where light energy is converted into an electrical signal. FEU scanners have much higher photosensitivity and more low level Noise than CCD scanners, and, therefore, are capable of good transmission of tones, being less susceptible to errors in refraction or focusing of light than their tablet colleagues.

However, drum scanners are slower and more expensive than CCD scanners. Currently, they are usually used only in specialized high-performance applications.

Charging device (CCD)

The instrument technology with a charge association, which underlies the desktop scanners, was previously used for a long time in devices such as telefix and digital cameras. CCD - solid state electronic devicewhich converts light in electric charge. The desktop scanner sensor, as a rule, has an array (ruler) of thousands of CCD elements placed on a movable carriage. The reflected light of the scanner lamp, passing the light filters, is sent to the CCD array through the system of mirrors and lenses.

Contact Sensor (CIS)

This is a relatively new technology of sensors, which began to appear in the market of tablet scanners In the late 1990s, the scanners of this system use compact banks of red, green and blue LEDs in combination with a ruler of the CCD sensors placed extremely close to the original image. As a result, a scanner was obtained, which is smaller, easier, cheaper and more economical than a traditional CCD-based device, but this technology is still far from excellence.

Scanner efficiency indicators

The mechanism of the sensor is not the only factor that sets the effectiveness of the scanner. The following indicators are important aspects of the device specification:

• resolution;
• discharge depth;
• dynamic range.

Resolution scanner

The resolution describes the accuracy of the device and is usually measured at inch points (TND). A typical resolution of an inexpensive desktop scanner in the late 1990s was 300 x 300.

A typical tablet scanner uses a CCD element for each pixel, so that for a desktop scanner having a horizontal optical resolution of 600 TDD and the maximum document width of 8.5 ", an array of 5100 (5100 \u003d 600 x 8.5) is required of the CCD elements in a block known as the scanner .

The head is installed on a carriage that moves along the original image. Although the movement seems continuous, moving occurs by discrete steps (in the share of inches), and information is read in each pause. In the case of a tablet scanner, the head is controlled by a stepping motor - a device that turns the axis at this angle (and not more) each time an electrical pulse is filed.

The number of physical elements in the CCD array determines the sampling interval of the X direction X, and the number of inch stops sets the sampling of the U. direction. Although they are usually referred to as the "resolution" scanner, the term is not completely accurate. Resolving ability (the scanner's ability to identify all the details of the image) is determined by the quality of electronics, optics, filters and motor drive, as well as the sampling rate (digitization).

By the end of 1998, the maximum density of the PZS elements in the ruler was 600 per 1-inch. However, the visible allowing ability can be increased using a technique known as interpolation, which consists in a software or hardware calculation of the intermediate values \u200b\u200bof the signal and insert them between real data. Some scanners make it more efficiently, others are less. Naturally, formulating the requirements for the resolution of the scanner, we should not forget about its coordination with the parameters of the information output device.

Consider how it would be possible to evaluate the requirements for resolving scanners depending on the quality of the output image.

Color polygraphy

Here, the equipment that reproduces various levels of color uses the method referred to as halftone processing. The setup devices used in offset printing - printing technology of glossy logs are capable of withdrawing 133 rows / inch. As experience shows, to obtain high-quality printing, the scanner permission must be 1.5 times higher, that is, about 200 tnt.

Jet printer

When scanning for subsequent output to the printer, the scanner resolution must comply with the resolution of the output as close as possible, taking into account the relative dimensions of the original and the output image. If they are the same, no adjustment is required. If, however, the output image must be printed in a different size (greater or less than the original), the scanner resolution must be appropriately adjusted.

Suppose it is necessary from the scanned postage mark with a size of 1 x 1.5 "Print on an inkjet printer, which has a print resolution of 600 TDD, and the image must be increased and compose in the amount of 2 x 3". If the brand was scanned at a resolution of 600 td, from the scanned image would have 600 pixels vertically (1 "multiply by 600) and 900 pixels horizontally (1.5" multiply by 600). An increase in the image to a size intended for printing (2 x 3 ") decreases the actual resolution of up to 300 TDD (900/3 \u003d 300, since 900 horizontal pixels will be located in 3"), and also in the vertical dimension. This is only half of the resolution of the printer, and the quality of the output will be lower than the optimal one. For better quality The printed image that actually uses 600 TsD, the scan must be carried out at 1200 TD.

Conclusion to the monitor

Such calculations can also be made if the size of the output image is smaller than the original. Suppose you need to scan a photo of 4 x 5, which will be displayed on Web page half-size, 2 x 2.5. Computer monitors usually have permissible ability 72 or 90 TDD. Photo scanning at 72 TDD gives an image of 288 x 360 pixels. The reduction in this size 2 times would give an image with a vertical resolution of 144 TDD, which is twice as needed. In this example, the original image could be from scanned at 36 tands without losing the quality of the result.

The relations used in these examples are described by the following formula:

SR \u003d (DR X DW) / OW.

where Sr is the perfect scanner resolution, TTD;

DR - resolution of the output device, TTD;

DW - width with which the image will be printed or displayed in inches;

Ow - width of the scanned original, in inches.

Interpolation

Despite the fact that in the scanners specifications may indicate permissive abilities in 2400.4800 and 9600, it is necessary to understand that they are actually not capable of distinguishing such a level of details. The actual optical permission of the CCD in the most modern scanners at best - 600 x 1200 TDD, and any higher indicators are based on interpolation.

An indication of the inhomogeneous resolution (for example, 600 x 1200 TDD) necessarily implies hardware interpolation, since the receiving data at 600 TDs on one axis (x) and 1200 on another (Y) will not clearly lead to the "square" of the image. At 600 x 600 rts, such a scanner will lower the resolution of 1200 TDs along the Y axis to 600 (usually it is done by increasing twice the step of the engine, which moves the head), and at 1200 x 1200 - will interpolate the measurement X. In this case, the integrated circuit chip in The scanner generates additional data using points that actually shot by the scanner and predicting the most likely color and brightness of intermediate pixels.

Color scanners

The heads of one color scanners contain a single fluorescent tube with three CCD, equipped with colored filters, while others have three colored tubes and a single CCD unit. The first produce a complete color image for the only pass, while the second is for three passages. However, since the end of the 1990s, single-pass devices make up most color scanners.

These scanners use one of two methods: either the splitting of the beam, or the CCD with color filters. In the first design, the light passing through the prism is divided into three primary colors, each of which is read by the corresponding CCDs. This method is considered the best for the processing of reflected light, but to reduce costs, many manufacturers use three CCD array, each of which is covered with a filter film so that it perceives only one of the primary colors. Being technically less accurate, this method usually produces results that are difficult to distinguish from those for the scanner with the splitting of the beam.

Discharge depth

The discharge (bit, color) of the scanner depth characterizes the amount of information contained in one pixel output image. The easiest scanner (black and white scanner on 1 bit) uses to represent each pixel "1" or "0". To reproduce the halftone between black and white, the scanner must have at least 4 bits (for 16 \u003d 2 4 halftone) or 8 bits (for 256 \u003d 2 8 halftone) for each pixel.

The most advanced color scanners support at least 24 bits, which means fixing 8 bits of information on each of the primary colors (red, blue, green). The device for 24 bits may theoretically fix more than 16 million different colors, although almost this number is much less. This is almost photographic quality, and therefore is mentioned usually as "full color" scanning ("True Color" Scanning).

Recently, an increasingly increasing list of manufacturers offers scanners with a discharge depth of 36 or 30 bits. Although few application programs Machine graphics are capable of processing images with a depth of more than 24 bits, this excess permission allows you to make useful graphics editing operations in both drivers and applications.

Dynamic range. The dynamic range is essentially similar to the discharge depth, which describes the color range of the scanner, and is defined both by the functioning of the Scanner ADC and the purity of light, the quality of color filters and the level of any interference in the system.

The dynamic range is measured in a scale of 0.0 (absolutely white) to 4.0 (absolutely black), and the only number given for a specific scanner, says how many shades the module can distinguish. Most colored tablet scanners with difficulty perceives subtle differences between dark and light colors at both ends of the range and has a dynamic range of about 2.4. This is certainly a bit, but usually enough for projects, where perfect color is not an end in itself. To obtain a larger dynamic range, use the top quality color tablet scanner with an increased discharge depth and improved optics. These high-performance modules usually provide a dynamic range between 2.8 and 3.2 and are well suited for most applications that require high-quality color (for example, offset printing). The drum scanners, often providing values \u200b\u200bfrom 3.0 to 3.8, are closest to the dynamic range limit.

Theoretically, the 24-bit scanner offers a range of 8 bits (256 levels) for each primary color, and the difference between two of 256 levels is usually not perceived by the human eye. Unfortunately, the smallest of the meaning bits are lost in noise, while any tone fix after scanning is even more narrowing the range. That is why it is best to first install any brightness and color fixes at the scanner driver level to the final scan. More expensive scanners with a depth of 30 or 36 bits have a much wider range, offering more detailed shades, and allow the user to make tone corrections ending with a decent 24-bit image. 30 bits scanner takes 10 data bits for each color, while scanners are 36 bits - 12 bits. The scanner driver allows the user to choose which 24 bits from the source 30 or 36 bits to save, and which is not. This setting is done by changing the "Gamma Curve" curve and is available when setting up tones (Tonal Adjustment Control) TWAIN drivers.

Scan modes

Among the overall diversity of image representation methods in the computer are the most common are:

• strike graphics (Line Art);
• halftone image (Greyscale);
• color image (Color).

Strike graphics - The easiest format. Since only black and white information is saved (in the computer, black is presented as "1" and white as "0"), only 1 data bits are required to save each viewed image. Strike graphics are most suitable when scanning drawings or text.

Halftone image. While computers can save and extract images in halftons, most printers are not able to print various shades of gray colors. They use a method called halftone processing using a dot raster that simulates halftone information.

Images in shades of gray - the simplest method of saving graphics in the computer. A person can distinguish not more than 255 different shades of gray, which requires a single data byte with a value from 0 to 255. This type Images make up the equivalent of black and white photography.

Full-color images - the most voluminous and most complex, persistent and processed in personal ComputerUse 24 bits (8 to each of the main colors) to present a full color spectrum.

Scanner designs

According to the applications, personal and production scanners distinguish between technical sales - manual, tablet and projection devices.

Resolution

Permitting ability, or permission, is one of the most important parameters characterizing the scanner capabilities. The most common unit of measurement of the resolution of scanners - number of pixels per inch (pixels Per Inch, pPI). Should not identify PPI with a better known unit dPI (dots Per Inch - The number of dots per inch), which is used to measure the resolution of raster printing devices and has a slightly different meaning.

Distinguish optical and interpolated resolution. The value of optical permissions can be calculated by separating the number of photosensitive elements in the scanning line to the width of the tablet. It is easy to count that the number of photosensitive elements in the scanners considered by us having an optical resolution of 1200 PPI and the format of the Legal tablet (that is, a width of 8.5 inches, or 216 mm), should be at least 11 thousand.

Speaking about the scanner as an abstract digital device, you need to understand that optical resolution is sampling frequency,only in this case the counting is not in time, but by distance.

In tab. 1 shows the required values \u200b\u200bof the resolution to solve the most common tasks. As you can see, when scanning in the reflected light, in most cases, it is quite sufficient to resolve in 300 PPI, and higher values \u200b\u200bare required either to scale the original to a larger size, or to work with transparent originals, in particular with 35-millimeter diaposims and negatives.

Table 1. Permissive ability values \u200b\u200bfor solving the most common tasks

Application	Required resolution, PPI
Scanning in the reflected light
Illustrations for Web Pages
Text recognising
Strike printing on a monochrome printer
Black and White Photo For Printing On Monochrome Printer
Color photo printing on a jet printer
Text and graphics for transfer by fax
Color photo for offset printing
Scanning in the passage
35 mm film, photos for Web pages
35 mm film, photo for printing on an inkjet printer
60 mm film, photo for Web pages
60 mm film, photo for printing on an inkjet printer

Many manufacturers, seeking to attract buyers, indicate the documentation and on the boxes of their products the value of the optical resolution of 1200 * 2400 PPI. However, halvening a large number for the vertical axis means nothing but a scanning with a half vertical step and further software interpolation, so that in this case the optical resolution of these models actually remains equal to the first digit.

Interpolated resolution is an increase in the number of pixels in the scanned image due to software processing. The magnitude of the interpolated resolution can increase the value of the optical resolution many times, but it should be remembered that the amount of information obtained from the original will be the same as when scanning with optical resolution. In other words, increase the detail of the image when scanning with a resolution exceeding the optical, will fail.

Bigness

The bit, or the depth of color, determines the maximum number of values \u200b\u200bthat the pixel color can take. In other words, the higher the bittenness when scanning, the greater the amount of shades may contain the resulting image. For example, when scanning a black and white image with a bit of 8 bits, we can obtain 256 gradations of gray (2 8 \u003d 256), and using 10 bits - already 1024 gradations (2 10 \u003d 1024). For color images, two variants of the specified bit are possible - the number of bits per each of the base colors or the total number of bits. Currently, the standard for storing and transmitting full-color images (for example, photos) is a 24-bit color. Since, when scanning non-ferrous originals, the image is formed by an additive principle of three basic colors, then each of them accounts for 8 bits, and the number of possible shades is a little more than 15.7 million (2 24 \u003d 16 777 216). Many scanners use greater bittenness - 12, 14 or 16 bits per color (full bit amounts, respectively, 36, 42 or 48 bits), however, for recording and further image processing, this feature must be supported by the software used; Otherwise, the resulting image will be recorded in a 24-bit file.

It should be noted that higher discharge does not always mean more high quality Images. Indicating 36- or 48-bit color depth in the documentation or promotional materials, manufacturers often silent that part of the bits are used to store the service information.

Dynamic range (maximum optical density)

As you know, darker sections of the image absorb more light falling on them than light. The magnitude of the optical density shows how dark this section of the image is and, therefore, what amount of light is absorbed, and which is reflected (or passes through the transparent original). Typically, the density is measured for a certain standard light source having a predetermined spectrum. The density value is calculated by the formula:

where D is the value of the density, R is the reflection coefficient (that is, the proportion of reflected or passing light).

For example, for a portion of the original reflecting (transmitting) 15% of the light falling on it, the density value will be log (1 / 0.15) \u003d 0.8239.

The higher the maximum perceived density, the more dynamic range This device. Theoretically dynamic range is limited to the bit used. So, an eight-bit monochrome image can have up to 256 gradations, that is, the minimum reproducible shade will be 1/256 (0.39%), therefore the dynamic range will be equal to log (256) \u003d 2.4. For a 10-bit image, it will already be a little more than 3, and for 12-bit - 3.61.

In fact, this means that the scanner with a large dynamic range allows you to better reproduce dark areas of images or simply dark images (for example, converted photographs). It should be noted that in real conditions the dynamic range is less than the above values \u200b\u200bdue to the effect of noise and cross-interference.

In most cases, the density of opaque originals scanned to reflect does not exceed 2.0 values \u200b\u200b(which corresponds to a portion with one-surrender reflection), and a typical value for high-quality printing originals is 1.6. Slides and negatives may have sections with a density above 2.0.

Light source

The light source used in the design of a scanner, a considerable degree affects the quality of the resulting image. Currently, four types of light sources are used:

1. Xenon gas discharge lamps. Their extremely small time of inclusion, high stability of radiation, small size and long service life are distinguished. But they are not very effective from the point of view of the ratio of the amount of energy consumed and the intensity of the light flux, have a non-ideal spectrum (which can cause a violation of the accuracy of color reproduction) and require high voltage (about 2 kV).
2. Fluorescent lamps With a hot cathode. These lamps have the greatest efficiency, very smooth spectrum (which can be used under certain limits) and low heating time (about 3-5 s). Negative parties include not very stable characteristics, quite significant dimensions, relatively short service life (about 1000 hours) and the need to keep the lamp constantly included during the scanner process.
3. Fluorescent lamps with cold cathode. Such lamps have a very long service life (from 5 to 10 thousand hours), low operating temperature, smooth spectrum (it should be noted that the design of some models of these lamps is optimized to increase the intensity of the light flow, which is negatively reflected on the spectral characteristics). The listed advantages have to pay a fairly long warm-up time (from 30 s to several minutes) and more high than that of hot cathode lamps, power consumption.
4. LEDs (LED). They apply, as a rule, in CIS scanners. Flowododides have very small dimensions, low power consumption and do not require time to warm up. In many cases, three-color LEDs are used, with a large frequency of changing the color of the emitted light. However, the LEDs have a rather low (compared to lamps) the intensity of the light stream, which reduces the scanning speed and increases the noise level on the image. A very uneven and limited spectrum of radiation inevitably entails deterioration of color reproduction.

Noise

As mentioned above, a 24-bit discharge scanner is theoretically able to reproduce even quite dark originals. However, in practice, some factors prevented by the applied image receipt technology, and first of all regular and casual noise. Consider these noise more.

Increased fragments of the original (right) and its scanned image (left). A random noise is noticeable on the left fragment

Increased fragments of the original (right) and its scanned image (left). On the left fragment are noticeable manifestations of regular noise in the form of vertical strips

Casual noise is manifested in the form of "snow", granularity or chaotic foreign dots in the image and occurs as due to instability of work semiconductor devices (with temperature changes and over time) and as a result of distortions made by electronic components. Such noise in the dark areas of the image is most noticeable, since with an equally noise level, the signal / noise ratio will be much smaller on them than in light areas. To minimize random noise, the calibration procedure is performed before scanning, during which the threshold values \u200b\u200bare measured and the base voltage offset for each sensitive element.

Regular noise occurs due to cross-interference (inspected with neighboring photosensitive elements), short-term changes in the base voltage in the CCD matrix, the effects of high-frequency electric fields, changes in the brightness of the light source, etc. Regular noise, in contrast to random, very well noticeable, since it is manifested in the form of horizontal, vertical or diagonal strips.

The scanner, as noted above, has the resolution determined by its constructive features. It can be a hardware (optical) or interpolation (reconstructed computational means). Permitting ability is the maximum characteristic defined by the technical characteristics of the scanner. However, when scanning an image, you can arbitrarily choose what resolution it should be done in this particular case. The installed scan resolution may be less than or equal to hardware (optical) to resolve the scanner, but may exceed it. In the latter case, we can only talk about interpolation resolution. With the installed interpolation resolution of scanning, in addition to the hardware itself, software transformations are involved. The latter can be good or bad: it all depends on the conversion and source image algorithm.

The quality of the resulting image received by them is the amount of memory, as well as the scan speed depends on the selection of scanning resolution. Image quality is primarily its clarity, smooth color transitions. In other words, a good scan result should not look markedly worse than the original.

The smaller the resolution, the less volume and time spent on scanning and vice versa. However, with the quality of the result, the situation is more complicated. It suggests an analogy with the choice of a fishing network. Which network to choose - with small or large cells, depends on the size of the fish you want to catch. The scanner is a device that extracts the information contained in the image. It is impossible to get information more than it was in the original, but its description can be made redundant. Redundant descriptions graphic information Usually expressed in excessively large volumes of relevant files. Ideally, we need to configure the scanner so as to extract as much graphic information from the original, or at least no less than you need.

The ability to correctly choose scan resolution comes with experience. However, experiments can be streamlined so that the experience came quickly. Images for simplicity can be divided into two main types: photos and pictures. Images of the type of photography (photographs, painting, etc.) are characterized by a large number of shades and smoothness of their transitions, and drawings (posters, drawings, engravings, etc.) - a relatively small amount of shades, in the presence of contours and increased contrast. Thus, not only photographs fall into the class of photographs, and the class of hand drawn graphics includes not only images created by a pencil, brush or pen. Sometimes there are images that are difficult to attribute to one or another type. In this case, try and so, and the case. Next, take several pictures of each type and scan them at various permissions. Start with the minimum value of 72 ppi, increasing it with some step to the optical resolution of the scanner. In the process of the experiment, you need to fix two permissions:

• starting with which the image quality becomes acceptable;
• starting with which the image quality is practically not changed.

Averaging the received data for each type of image, you will receive a resolution value to be installed when you first attempt to scan. When scanning the case is approximately the same as when using a professional camera, when it is necessary to manually set the shutter speed, aperture and focal length (Sharpness). An experienced photographer quickly estimates the shooting object and sets the desired parameters of its machine. However, a professional will make several pictures of the same object with a slightly different camera settings. Similarly, when scanning often has to take several attempts.

By installing the scan resolution, it should also be considered whether the image is increased in size when it is displayed on the monitor screen or when printing it. With increasing size (i.e., with stretching), the image quality, generally speaking, can worse. This case creates an image with some resolution margin. So, if it is supposed to increase the picture twice, then the resolution must be twice as much as something that was sufficient for the starting size. Solid side, if it is supposed to display a reduced image on a monitor or print, then permission may be resolved accordingly. Small images must have a small resolution. This situation often occurs in a Web design, where the same picture is often submitted in two versions: small (thumbnail, miniature) - low resolution, and large - high resolution.

If your computer has a sufficiently large memory and time for scanning for you are not critical, you can recommend a resolution setting equal to the hardware (optical) scanner permission. Then, if necessary, the resolution of the resulting image can be reduced by means graphic Editor. In Photoshop, use the Image\u003e Image Size command to do this. However, the increase in the resolution of the graphic editor does not improve the image quality. With a decrease in the resolution (downsample), pixels are removed from the image and, thus, the number of graphic information decreases. With an increase in the resolution, the graphic editor adds pixels using a certain interpolation algorithm to calculate their values \u200b\u200b(accounting for the values \u200b\u200bof neighboring pixels).

Fig. 123. Window installation window and image resolution in Photoshop

Generally speaking, optimize the final version of the image is better than a powerful graphic editor, such as Photoshop. Working with graphics from the point of view of a designer (artist) usually occurs in the space of a graphic editor, and not software scanner. But this does not mean that software Scanner (TWAIN interface) must be forever forgotten. Although they were created mainly to ensure that the user can work with the scanner, independent of his package graphic programsSometimes they can be effectively applied before Photoshop shows all its power.

The following table is given as an example of memory costs when scanning an image of a 4x4 inches size (11x11 cm) in various modes and at various permissions.

Image type	The scope of the image at various permissions
	100 ppi.	150 PPI	300 ppi.	600 ppi.
Color	469 KB	1 MB	4.12 MB	16.5 MB
Gray.	156 KB	352 KB	1.37 MB	5.5 MB
Artline	19.5 KB	44 KB	175 KB	703 KB

In conclusion of the conversation on the resolution of scanning remark, the circumstances that have to additionally take into account when choosing a permit. First, if the scanned image is intended for printing using a laser or inkjet printer, then the permission installed can be 3-4 times less than the resolution of the printer. This is true primarily for color or halftone (in shades of gray) images. For ARTLINE or HALFTONE image, scanning resolution should be selected, if possible, equal to the resolution of the printer. For example, if you have an ordinary jet printer with a resolution of 300 ppi, then. Try to first scan the image with a resolution of 75 PPI. If the result turns out to be unsatisfactory, increase the scan resolution by 2 times. Secondly, permission often has to be changed when scanning images from high-quality printed publications. The reason for the so-called Moire is the effect of the interaction of several periodic structures (in this case, discrete scanning structures and printed raster). Often, this side optical effect is eliminated by choosing a higher scan resolution. Moir suppression will be discussed in more detail below. Thirdly, when choosing the initial and, if necessary, the subsequent values \u200b\u200bof the scan resolution should strive to ensure that the selected resolution is multiple the optical resolution of the scanner shared by the whole degree:

Installed resolution \u003d optical resolution: 2 i, where i \u003d 0, 1.2, 3, ...

For example, if the optical scanner resolution is 600 ppi, then the installed scan resolution should be as close as possible to 600, 300, 150, 75 ppi. Such a choice contributes to the achievement of the greatest clarity of the scan result.

Since visibility (for scanner) colors and shades on paper is determined by the color of the lighting, the white color of the lamp seems to be neutral and universal (allowing you to see the points of any color). However, white lamps quickly lose the brightness and in scanners designed for prolonged intensive scans using green phosphoric lamps. In this case, the scanner manifests a tendency to suppress light green and light blue colors on the page. This effect is sometimes even used when processing forms (recognizable forms are printed on light blue or green blanks). But historically many papers in the field of insurance and health care were made on pink or red paper, so red lamps have received widespread lamps to suppress the appropriate background. Currently, all manufacturers of production scanners allow the device with a lamp of the required color or order an additional one (several) colored lamps (or light filters) for improved scanning in specific circumstances.

Basic technical parameters of scanners

Resolution

Permitting ability, or permission, is one of the most important parameters characterizing the scanner capabilities. The most common unit of measurement of the resolution of scanners - number of pixels per inch (Pixels per inch, abbreviated pPI). Do not identify PPI with a more common abbreviation dPI (DOTS PER INCH, number of dots per inch). The last unit is used to measure the resolution of raster printing devices and has a slightly different meaning.

Distinguish optical and interpolated resolution. The value of optical permissions can be calculated by separating the number of photosensitive elements in the scanning line to the width of the tablet. It is easy to count that the number of photosensitive elements in the scanner having an optical resolution of 600 PPI and the format of the Legal tablet (that is, 8.5 inches width, or 216 mm) should be at least 5100, and with a resolution of 1200 ppi - 11,000! Speaking about the scanner as an abstract digital device, it is important to understand that optical resolution is sampling frequency, only in this case the counting is not in time, but by distance.

IN table. one The required values \u200b\u200bof the resolution for the most common tasks are given. As you may see, when scanning in the reflected light in most cases, more than enough permission in 300 ppi, and higher values \u200b\u200bare required mainly to work with transparent originals, in particular 35-millimeter diaposims and negatives.

Many manufacturers, seeking to attract buyers, indicate the documentation and on the boxes of their products the value of the optical resolution of 600x1200 PPI (or 1200x2400, respectively). However, halvening a large number for the vertical axis means nothing but a scanning with a half vertical step and further software interpolation, so that in this case the optical resolution of these models actually remains equal to the first digit.

Interpolated resolution is an increase in the number of pixels in the scanned image due to software processing. The magnitude of the interpolated resolution can increase the value of the optical resolution many times, but it should be remembered that the amount of information obtained from the original will be the same as when scanning with optical resolution. In other words, to increase the detail of the image when scanning with a resolution exceeding optical, will fail.