With the advent of digital cameras, this task has become indecently simplified. It is no longer necessary to develop, print and even scan, even the most budget models Be sure to write the shooting date in EXIF, and non-budget ones also write the location coordinates - all that remains is to copy the files from the memory card and use any viewer program you like. What if you had several generations of photographers in your family, even amateurs?

This article will discuss what to do with old negatives, slides and prints. I note that I did not open America and any more or less qualified user can easily do all this himself.

1. Equipment

Purchasing a professional film scanner was not part of the author’s plans: in addition to negatives and slides, the archive contained about 4,000 photographic prints, which require a flatbed scanner, ideally with automatic feed. Of course, it is better to scan the original negative than the positive printed from it, but it was impossible to figure out for which photographs the negatives were preserved. Toad and common sense did not allow me to buy two scanners for what was essentially a one-time job.

As a result, for 5990 rubles. I purchased a mid-range Epson Perfection V350 Photo flatbed scanner equipped with an AFL (Auto Film Loader). Optical resolution of 4800 DPI allows you to scan negatives and slides. Of course, the dynamic range for this money is not the same as that of professional film scanners, and the speed leaves much to be desired, but...

In addition to the scanner, you will need a photo tank for washing old 35 mm films and a couple of clothespins for subsequent drying. You also need disk space: ~9000 photos scanned in adequate resolution (JPG of maximum quality) took 45 GB from the author. If someone decides to store data in a loseless format (TIFF/PSD/etc.), then even more.

2. Software

4. Background correction. In terms of meaning, this is analogous to Levels correction in Adobe Photoshop. It works well, some frames can be “extracted” immediately at the scanning stage. The “high” level is almost never used: if the frame is initially dark, trying to apply a filter will reduce the contrast to unacceptable levels.

5. Removal of defects. The most controversial filter. In pictures with a large number of uniformly filled areas (sky, calm water, furniture) it really allows you to remove a large number of defects. In photographs with a large number of faces of a small size relative to the frame area (group portraits, demonstrations), parts of the face may be mistaken for a defect, with all that it implies. He especially doesn't like the eyes :) The filter is resource-intensive and increases scanning time.

Sync Picasa Web Albums and Disk Catalog

After the first files from the scanner appear in the directory, you need to set up synchronization with Picasa web albums. In the album properties, select “Enable synchronization”:

After turning on the synchronization mode, do not forget to specify the size of the photos. For Reserve copy need to install " Images in original size" This will not affect the viewing speed, but it will greatly affect the synchronization speed (depending on your Internet connection speed). You can also turn on the “ private"if you don't want (I, for example, don't want:) for your photos to be publicly available. In “private” mode, you can grant viewing and editing access rights to the ones you select. Google users(Google account required).

That's all. Now, if you have the desire and time, you can digitize everything that was filmed in the pre-digital era. The scanner scans, Picasa automatically uploads photos to the web, and you don't forget to do it from time to time. backups to other media.

Don't forget about backup!

Additional Information:

- : A wonderful resource with articles on film scanning.
- in the same place: “Why you shouldn’t scan films on a tablet” (I completely agree, but...)

Original type. Scanning can be done in transmitted light (for originals on a transparent backing) or reflected light (for originals on an opaque backing). Scanning negatives is particularly challenging because the process is not simply about inverting the color gradations from negative to positive. To accurately digitize color in negatives, the scanner must compensate for the color photographic veil on the original. There are several ways to solve this problem: hardware processing, software algorithms for transitioning from negative to positive, or lookup tables for specific types of film.

Optical resolution. The scanner does not take the entire image, but line by line. Vertically flatbed scanner A strip of photosensitive elements moves and captures the image point by point, line by line. The more photosensitive elements a scanner has, the more points it can remove from each horizontal stripe Images. This is called optical resolution. It is usually calculated by the number of dots per inch - dpi (dots per inch). Today, a resolution level of at least 600 dpi is considered the norm.

Speed ​​of work. Unlike printers, the speed of scanners is rarely indicated, since it depends on many factors. Sometimes the scanning speed of one line is indicated in milliseconds.

Color depth measured by the number of shades that the device is able to recognize. 24 bits corresponds to 16,777,216 shades. Modern scanners are produced with color depths of 24, 30, 36, 48 bits.

Dynamic range characterizes what range of optical densities of the original the scanner can recognize without losing shades in either the highlights or shadows of the original. The maximum optical density of the scanner is the optical density of the original, which the scanner also distinguishes from complete darkness. The scanner will not be able to distinguish all shades of the original darker than this border.

Batch processing - This is scanning multiple originals at the same time, saving each image in separate file. Program batch processing allows you to scan a certain number of originals without operator intervention, providing automatic switching of scanning modes and saving scanned files.

Zoom range - This is the range of amounts of original rescaling that can be done during scanning. It is related to the resolution of the scanner: the higher the maximum optical resolution, the greater the magnification factor of the original image without loss of quality.

By interface type scanners are divided into only four categories:

Scanners with parallel or serial interface, connected to LPT or COM port These interfaces are the slowest. Problems may arise due to a conflict between the scanner and the LPT printer, if there is one.

Scanners with a USB interface Cost a little more, but work much faster. A computer with a USB port is required.

Scanners with a SCSI interface, with their own interface card for the ISA or PCI bus, or connected to a standard SCSI controller. These scanners are faster and more expensive than the representatives of the two previous categories and belong to a higher class.

Scanners with modern interface FireWire(IEEE 1394) specially designed for graphics and video processing. Such models have been introduced to the market relatively recently.

For office and home tasks, as well as for most work on computer graphics the so-called ones are best suited flatbed scanners. Various models of this type are more widely available on sale than others. Therefore, let's start by considering the principles of construction and operation of scanners of this particular type. Understanding these principles will provide a better understanding of the meaning technical characteristics, which are taken into account when choosing scanners.

A flatbed scanner is a rectangular plastic case with a lid. Under the cover there is a glass surface on which the original is placed to be scanned. Through this glass you can see some of the insides of the scanner. The scanner has a movable carriage on which a backlight lamp and a mirror system are installed. The carriage moves through the so-called stepper motor . The lamp light is reflected from the original and, through a system of mirrors and focusing lenses, enters the so-called matrix, consisting of sensors that produce electrical signals, the magnitude of which is determined by the intensity of the light incident on them. These sensors are based on light-sensitive elements called charge coupled devices(CCD, Couple Charged Device - CCD). More precisely, on the surface of the CCD a electric charge, proportional to the intensity of the incident light. Next, you just need to convert the value of this charge to another electrical quantity- voltage. Several CCDs are located side by side on one line.

The electrical signal at the output of the CCD is an analog quantity (i.e., its change is similar to the change in the input quantity - light intensity). Next, the analog signal is converted into digital form, followed by processing and transmission to a computer for further use. This function is performed by a special device called analog-to-digital converter(ADC, Analog-to-digital Converter - ADC). Thus, at each step of moving the carriage, the scanner reads one horizontal strip of the original, divided into discrete elements (pixels), the number of which is equal to the number of CCDs on the line. The entire scanned image consists of several such stripes.

Rice. 119. Diagram of the design and operation of a flatbed scanner based on a CCD (CCD): the lamp light is reflected from the original and, through an optical system, hits a matrix of photosensitive elements, and then to an analog-to-digital converter (ADC)

Color scanners now typically use a three-row CCD matrix and illuminate the original with calibrated white light. Each row of the matrix is ​​designed to perceive one of the basic color components of light (red, green and blue). To separate colors, they use either a prism, which splits a beam of white light into colored components, or a special CCD filter coating. However, there are color scanners with a single-row CCD matrix, in which the original is illuminated in turn by three lamps of basic colors. Single-row, triple-illuminated technology is considered obsolete.

Above we described the principles of construction and operation of so-called single-pass scanners, which scan the original in one carriage pass. However, three-pass scanners are still found, although no longer commercially produced. These are scanners with a single-row CCD matrix. In them, with each pass of the carriage along the original, one of the basic color filters is used: for each pass, information is removed from one of the three color channels of the image. This technology is also outdated.

In addition to CCD scanners based on a CCD matrix, there are CIS (Contact Image Sensor) scanners that use photocell technology.

Photosensitive matrices made using this technology perceive the reflected original image directly through the scanner glass without the use of optical focusing systems. This made it possible to reduce the size and weight of flatbed scanners by more than half (down to 3-4 kg). However, such scanners are only good for extremely flat originals that fit tightly to the glass surface of the working field. In this case, the quality of the resulting image significantly depends on the presence of extraneous light sources (the CIS scanner cover must be closed during scanning). In the case of volumetric originals, the quality leaves much to be desired, while CCO scanners give good results for volumetric (up to several cm in depth) objects.

Flatbed scanners can be equipped with additional devices, such as a slide adapter, automatic document feeder, etc. Some models are provided with these devices, but others are not.

Slide adapter (Transparency Media Adapter, TMA) is a special attachment that allows you to scan transparent originals. Transparent materials are scanned using transmitted rather than reflected light. In other words, the transparent original must be between the light source and the photosensitive elements. The slide adapter is a mounted module equipped with a lamp that moves synchronously with the scanner carriage. Sometimes they simply illuminate a certain area of ​​the working field evenly so as not to move the lamp. Thus, the main purpose of using a slide adapter is to change the position of the light source.

If you have a digital camera (digital camera), then you most likely do not need a slide adapter.

If you scan transparent originals without using a slide adapter, you need to understand that when the original is irradiated, the amounts of reflected and transmitted light are not equal to each other. So, the original will miss some of the incident color, which will then be reflected from the white coating of the scanner lid and pass through the original again. Some of the light will be reflected from the original. The ratio between the parts of transmitted and reflected light depends on the degree of transparency of the original area. Thus, the light-sensitive elements of the scanner matrix will receive light that has passed through the original twice, as well as light reflected from the original. The repeated passage of light through the original weakens it, and the interaction of the reflected and transmitted beams of light (interference) causes distortion and side video effects.

An automatic document feeder is a device that feeds originals into the scanner, which is very convenient to use when streaming images of the same type (when you do not need to frequently reconfigure the scanner), for example, texts or drawings of approximately the same quality.

In addition to flatbed ones, there are other types of scanners: manual, sheet-fed, drum, slide, for scanning barcodes, high-speed for streaming documents.

Handheld Scanner is a portable scanner in which scanning is carried out by manually moving it over the original. The principle of operation of such a scanner is similar to that of a tablet scanner. The width of the scanning area is no more than 15 cm. The first scanners for widespread use went on sale in the 80s of the 20th century. They were manual and allowed scanning of images in shades of gray. Nowadays such scanners are not easy to find.

Sheet or roller scanner(Sheetfed Scanner) - a scanner in which the original is pulled past a stationary linear CCD or CIS matrix; a type of such a scanner is a fax machine.

Drum scanner(Drum Scanner) - a scanner in which the original is fixed on a rotating drum, and photomultipliers are used for scanning. In this case, a dot area of ​​the image is scanned, and the scanning head moves along the drum very close to the original.

Slide scanner(Film-scanner) is a type of flatbed scanner designed for scanning transparent materials (slides, negative films, X-rays, etc.). Usually the size of such originals is fixed. Note that some flatbed scanners have a special attachment (slide adapter) designed for scanning transparent materials (see above).

Barcode Scanner(Bar-code Scanner) - a scanner designed for scanning product barcodes. Its operating principle is similar to hand scanner and connects to a computer or a specialized trading system. If there is appropriate software Any scanner can recognize barcodes.

High-speed scanner for working with documents(Document Scanner) is a type of sheet-fed scanner designed for high-performance multi-page input. Scanners can be equipped with input and output trays with a capacity of over 1000 sheets and input information at speeds in excess of 100 sheets per minute. Some models of this class provide two-sided (duplex) scanning, backlighting the original in different colors to cut out the colored background, compensation for background heterogeneity, and have modules for dynamic processing of different types of originals.

So, a flatbed scanner is best for home and office use. If you want to do graphic design, then it is better to choose a CCD scanner (based on a CCD matrix), since it allows you to scan three-dimensional objects. If you plan to scan slides and other transparent materials, you should choose a scanner that has a slide adapter. Usually the scanner itself and the slide adapter that goes with it are sold separately. If you cannot purchase a slide adapter at the same time as the scanner, you can do so later if necessary. It is also necessary to determine the maximum sizes of scanned images. Currently, the standard format is A4, corresponding to a regular sheet of writing paper. Most household scanners are designed specifically for this format. Scanning drawings and other design documents typically requires A3 size, which corresponds to two A4 sheets joined along the long side. Currently, the prices of scanners of the same type for A4 and A3 formats are getting closer. It can be assumed that originals that do not exceed the A4 format will be better processed by a scanner oriented to the A3 format.

The parameters listed above do not exhaust the entire list, but at this stage In our consideration, we can only use them for now. When choosing a scanner, three aspects are decisive: a hardware interface(connection method), optical-electronic system And software interface c (the so-called TWAIN module). Next we will look at them in more detail.

At first glance, the idea of ​​​​creating a flatbed scanner with an optical resolution of more than 600 ppi, which is not designed to work with transparent originals, seems rather dubious - after all, for the vast majority of originals scanned in reflected light, 300-400 ppi is more than enough. However, we should not forget that a significant proportion of the originals scanned both at home and in the office are images printed using the printing method. Due to interference phenomena that occur when digitizing rasterized images, noticeable moire appears on the resulting image, which is quite difficult to combat without compromising the quality or size of the image. To combat such phenomena, special algorithms embedded in scanning control programs are used. As a rule, the moire suppression function works by scanning the original at an excessive (that is, greater than user-specified) resolution and then software processing the resulting image. This is where the advantage of high-resolution scanners will be obvious in the truest sense of the word.

Main technical parameters of scanners

Resolution

Resolution, or resolution, is one of the most important parameters characterizing the capabilities of a scanner. The most common unit of measurement for scanner resolution is number of pixels per inch (pixels per inch, ppi). Ppi should not be equated with the more well-known unit dpi (dots per inch- 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 permission. The optical resolution can be calculated by dividing the number of photosensitive elements in the scanning array by the width of the tablet. It is easy to calculate that the number of photosensitive elements in the scanners we are considering, which have an optical resolution of 1200 ppi and a Legal tablet format (that is, a width of 8.5 inches, or 216 mm), should be at least 11 thousand.

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

In table 1 shows the required resolution values ​​for solving the most common problems. As you can see, when scanning in reflected light, a resolution of 300 ppi is sufficient in most cases, and higher values ​​​​are required either to scale the original to a larger size, or to work with transparent originals, in particular with 35 mm transparencies and negatives.

Table 1. Resolution values ​​for solving the most common problems

Application	Required resolution, ppi
Reflected light scanning
Illustrations for Web pages
Text recognising
Line art for printing on a monochrome printer
Black and white photo for printing on a monochrome printer
Color photo for printing on inkjet printer
Text and graphics for faxing
Color photo for offset printing
Transmitted light scanning
35mm film, photo for web pages
35mm film, photo for printing on an inkjet printer
60mm film, photo for web pages
60mm film, photo for printing on an inkjet printer

Many manufacturers, in an effort to attract buyers, indicate in the documentation and on the boxes of their products an optical resolution value of 1200 * 2400 ppi. However, a figure twice as large for the vertical axis means nothing more than scanning with half a vertical step and further software interpolation, so in this case the optical resolution of these models actually remains equal to the first figure.

Interpolated resolution is an increase in the number of pixels in a scanned image through software processing. The interpolated resolution can be many times greater than the optical resolution, but remember that the amount of information obtained from the original will be the same as when scanning with optical resolution. In other words, it will not be possible to increase image detail when scanning at a resolution exceeding optical.

Bit depth

Bit depth, or color depth, determines maximum number values ​​that a pixel color can take. In other words, the higher the bit depth when scanning, the greater the number of shades the resulting image can contain. For example, when scanning a black and white image with 8 bits, we can get 256 shades of gray (2 8 = 256), and using 10 bits we can get 1024 shades (2 10 = 1024). For color images, there are two options for the indicated bit depth - the number of bits for each of the basic colors or the total number of bits. The current standard for storing and transmitting full-color images (such as photographs) is 24-bit color. Since when scanning color originals the image is formed according to the additive principle from three basic colors, each of them has 8 bits, and the number of possible shades is slightly more than 16.7 million (2 24 = 16,777,216). Many scanners use a larger bit depth - 12, 14 or 16 bits per color (the total bit depth is 36, 42 or 48 bits, respectively), however, for recording and further processing of images, this function must be supported by the software used; otherwise, the resulting image will be written to a 24-bit file.

It should be noted that higher bit depth does not always mean higher image quality. Specifying 36- or 48-bit color depth in the documentation or promotional materials, manufacturers often keep silent about the fact that some of the bits are used to store service information.

Dynamic range (maximum optical density)

As you know, darker areas of an image absorb more light falling on them than lighter areas. The optical density value indicates how dark a given area of ​​the image is and, therefore, how much light is absorbed and how much is reflected (or passed through in the case of a transparent original). Typically, density is measured for a standard light source that has a predetermined spectrum. The density value is calculated using the formula:

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

For example, for a section of the original that reflects (transmits) 15% of the light incident on it, the density value will be log(1/0.15) = 0.8239.

The higher the maximum perceived density, the more dynamic range of this device. Theoretically, the dynamic range is limited by the bit depth used. Thus, 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) = 2.4. For a 10-bit image it will be slightly more than 3, and for a 12-bit image it will be 3.61.

Essentially, this means that a scanner with a higher dynamic range can better reproduce dark areas of images or simply dark images (such as overexposed photographs). It should be noted that in real conditions the dynamic range is less than the above values ​​due to the influence of noise and crosstalk.

In most cases, the density of opaque originals scanned for reflection is less than 2.0 (which corresponds to an area of ​​one percent reflection), and a typical value for high-quality printed originals is 1.6. Slides and negatives may have areas of density greater than 2.0.

Light source

The light source used in the design of a particular scanner greatly affects the quality of the resulting image. There are currently four types of light sources in use:

1. Xenon gas discharge lamps. They are distinguished by extremely short switching times, high radiation stability, small sizes and long service life. But they are not very efficient in terms of the ratio of the amount of energy consumed and the intensity of the luminous flux, have an imperfect spectrum (which can cause a violation of color accuracy) and require high voltage (about 2 kV).
2. Fluorescent lamps hot cathode. These lamps have the highest efficiency, a very smooth spectrum (which can also be controlled within certain limits) and a short warm-up time (about 3-5 s). The negative aspects include not very stable characteristics, rather significant dimensions, a relatively short service life (about 1000 hours) and the need to keep the lamp constantly on while the scanner is operating.
3. Cold cathode fluorescent lamps. 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 luminous flux, which negatively affects the spectral characteristics). For the listed advantages, you have to pay for a rather long warm-up time (from 30 s to several minutes) and higher energy consumption than hot cathode lamps.
4. Light emitting diodes (LED). They are usually used in CIS scanners. Color LEDs have very small dimensions, low power consumption and do not require time to warm up. In many cases, three-color LEDs are used, which change the color of the emitted light at high frequencies. However, LEDs have a rather low luminous flux (compared to lamps), which reduces scanning speed and increases the noise level in the image. A very uneven and limited radiation spectrum inevitably entails a deterioration in color rendering.

Scanning speed and warm-up time

During testing, the time required for a “cold” start and recovery from power saving mode was measured.

To evaluate the performance of the tested scanners, we measured the time required to complete several of the most common tasks. The countdown began from the moment you pressed the Scan (or similar) button in the application from which the scan was made, and ended after this application was again ready for work (that is, it was possible to perform any actions, such as changing settings or scanning areas).

Flatbed scanners:

how they work, how they differ and how to decipher the characteristics

A person choosing a scanner comes across so many numerical characteristics, names of “patented and only ours” technologies and simply mysterious phrases that it is easy to get confused. Testing in the computer press is usually carried out according to randomly selected criteria, while opportunities that are important for a particular person are left behind the scenes. Some scanner sellers make arguments that are lethal to the uninitiated person, such as “scanner A has no halftone breaks, and scanner B has a slightly higher mechanical resolution.” How to “translate” their statements and whether they make sense at all, which functions in your case are worth paying for, and which ones will remain “in reserve” - that’s what this article is about. Below we will try to consider design principles, used in flatbed scanners, from the point of view of electronics and optics, and we will evaluate the importance of the main characteristics of scanners, both expressed in numbers and given in advertising and manuals, and less obvious, but no less important. We will not consider hand-held and broaching (sheet-fed) scanners separately due to the commonality of the technologies used, the same value of numerical characteristics and the limited use of them for working with full-color images. All information provided by default applies only to the price category “up to 10,000 USD”.

A modern scanner functionally consists of two parts: the scanning mechanism itself (engine) and the software part (TWAIN module, color management system, etc.). During the purchasing process, they often forget that the scanner will not be able to work without its own driver, since it is not a standard Windows device. The reliability of operation and functionality (more precisely, the lack thereof) of TWAIN modules of especially cheap scanners are very reminiscent of the drivers of “rootless” video cards, with the difference that for scanners there are no “universal” drivers from the chip manufacturer or from the Windows package. If you are running WindowsNT, be doubly careful!

Mechanism:

The scanning mechanisms of flatbed scanners are produced by a very limited number of manufacturers, who supply them under OEM agreements to other companies. They complete them with their own set of software and sell them under their own brand. The "added" software can be very good indeed, but there are some pitfalls when purchasing such a scanner from a third party.

Firstly, the price is usually higher than the “original” model, although the mechanism remains the same.

Secondly, the inevitable loss of time for interaction between the mechanism manufacturer and the company selling the scanner under its own brand leads to some obsolescence of the model by the time it is released for sale, sometimes quite significant. Thus, some models of scanners from very well-known companies use “sample” mechanisms from 1993!

Thirdly, the “renamed” model is not always compatible with new versions of drivers from the manufacturer of the mechanism; in this case, new drivers will be available only after they are “refined” by the seller, in the worst case, if the seller has stopped selling scanners under its own brand, never .

Flatbed scanners, especially those intended for something other than a gift or toy use, when external simplicity are very interesting and quite complex opto-electronic-mechanical devices. However, their design is well established, production is well established and technologically not something out of the ordinary, so usually flatbed scanners in the price range up to $10,000 (including such famous names as AGFA, Linotype-Hell and UMAX) are made in Taiwan.

To understand the meaning of the characteristics, you need to imagine the design of a typical flatbed scanner (the design of expensive models is slightly different).

The original is placed on a transparent fixed glass, along which a scanning carriage with a light source moves (if a transparent original is scanned, a so-called slide module is used - a cover in which a second lamp moves parallel to the scanning carriage of the scanner).

The optical system of the scanner (consists of a lens and mirrors or prism) projects the light flux from the scanned original onto a receiving element that separates information about colors - three parallel lines of an equal number of separate photosensitive elements that receive information about the content of “their” colors. Three-pass scanners use lamps of different colors or changing filters on a lamp or CCD matrix. The receiving element converts the light level into a voltage level (still analogue information). Next, after possible correction and processing, the analog signal is sent to an analog-to-digital converter (ADC). From the ADC, the information comes out in a binary form “familiar” to the computer and, after processing in the scanner controller, through the interface with the computer it enters the scanner driver - usually this is the so-called TWAIN module, with which application programs already interact.

Light source:

In older designs, it is a regular fluorescent lamp (similar to conventional fluorescent lamps). Disadvantage: poor stability of lighting characteristics and limited service life. Modern models use a cold cathode lamp, which has better parameters and a significantly longer service life. How does the lamp affect the scanning result? It is quite obvious that when the characteristics of the illumination source of the original change, the light flux incident on the receiving matrix changes, carrying information about the original being scanned. If the properties of the lamp change “beyond recognition” after 2-3 months of operation, talk about correct color rendering there is no need for a scanner anymore.

In general, the characteristics of the light flux change even when the scanner warms up. In this regard, the design of current Epson models is somewhat alarming - the only known brands of scanners with three separate lamps of different colors, because each lamp can “float” in its own way.

The quality of the lamp is difficult to assess. Make sure, at the very least, that you are using a cold cathode tube (if so, be sure to indicate this in the description). Scanners aimed at professional work with color contain, in addition to a built-in self-calibration procedure based on the intensity of the light flux from the lamp, also circuits for maintaining flux stability when temperature changes.

By the way, an indirect sign of suitability for “full-color” work can be the initial warm-up time of the lamp after the lamp has been automatically extinguished when the scanner has not been used for some time (by the way, usually the warm-up time and the waiting time before the lamp goes out can be changed, but somewhere inside settings files).

Optical system:

The luminous flux from the original is projected onto a CCD (charge-coupled device) matrix, which converts it into an electrical signal. Typically a single focusing objective (or lens) is used that projects the full width of the scanning area onto the full width of the CCD sensor. The requirements for the quality of optics for such a task are very high; it is especially difficult to ensure acceptable quality of projection of the edges of the work area for color originals. The focusing quality and resolution of the optics can easily be assessed visually by scanning a special test target or protective areas of a banknote.

The most powerful models of flatbed scanners feature interchangeable lenses: when operating in normal mode, the optics operate similarly to single-lens mechanisms; when switching to the second, “enhanced” mode, another lens is used, which projects only part of the width of the scanner desktop onto the full width of the CCD matrix. Thus, an area of ​​smaller width is projected onto a constant number of receiving cells of the CCD matrix and the optical resolution increases accordingly. Usually the documentation indicates the number of cells of the CCD matrix. The latest matrices of 42-bit scanners have 10,600 cells (although in single-pass scanners the matrix has three parallel lines of receiving cells - one per color, the number of elements in one is indicated). By dividing the number of cells by the width of the scanning field, we obtain the optical resolution. Note that some professional flatbed scanners have more than two (up to 5) switchable lenses, but this is already in the “above 10,000” category.

For scanners operating in the territory of the former USSR, the security of their mirrors is of great practical importance, optical system and CCD matrices against dust and insects. Even small specks of dust and fluff directly on the matrix or lens lead to noticeable defects.

Resolution: optical, mechanical, physical and miscellaneous.

Optical: number of elements in a matrix line divided by the width of the workspace. Determined by the matrix and the width of the working area, the smaller of all given resolution figures. But it may not be given at all! The first candidate for use as an example: in the characteristics of the HP ScanJet 5100 “Resolution, Optical: 600dpi Hardware Super Sampling”. About the ScanJet 6100 model, which is about twice as expensive, it is simply written “Resolution, Optical: 600dpi”.

Mechanical: the number of times the CCD matrix “reads” information, divided by the length of the path traveled by the scanning carriage during this time. Sometimes it is also called optical (“optical resolution 300×600”), but in reality this is not so (optical will be 300, and 600 is also a real resolution, but of the mechanism, not the optics). As a rule, the mechanical resolution is set by the manufacturer to be 2 times greater than the optical one (sometimes equal to it or 4 times greater), and since the CCD matrix cannot scan with a resolution higher than the optical one, and the scanned square must remain square, the missing “width” the points are calculated (interpolated). Interpolation not only does not provide a visible increase in quality when scanning full-color originals, but can also degrade clarity and significantly reduce scanning speed.

Physical resolution, true resolution, real resolution: everything that is somehow determined by the scanner mechanism.

Interpolation- an arbitrarily selected resolution, up to which the scanner program supposedly undertakes to “calculate” the missing points itself (for example, produce 16x16 points, having received 3x3 points from the scanner). The value of this indicator is questionable and it has absolutely nothing to do with the scanner mechanism. Note that it is sometimes better to enlarge originals such as engravings by scanning with interpolation resolution, while scaling a color image is usually always better done in Adobe Photoshop and scanned with a resolution equal to optical (that is, for a scanner with the specified “optical” - in fact physical - resolution 300×1200dpi should be set to 300 dpi). If you need to scan a full-color image with a resolution less than optical, then it is better to set the resolution to a multiple of optical (that is, for a 300×1200 dpi scanner, set it to 300 dpi or 150 dpi, but not 200 dpi!) or the nearest higher and scale it in Adobe Photoshop.

Important: the main task when scanning a full-color image, get the maximum REAL information at the scanner output. The information from an individual cell of the CCD matrix is ​​real, but the result, for example, of scanning with a resolution of 2/3 of the optical one is interpolation by the scanner driver or controller of information from three hundred cells into two hundred pixels.

What optical resolution is needed for your work:

For text recognition programs of regular size (not microfilm) 200-300 dpi, for working with graphics it is a little more difficult to determine. The maximum resolution at which it still makes sense to scan can be calculated using the formula “ to ensure a good quality margin, the scanning resolution should be 1.5-2 times higher than the resolution of the file fed to the printing device multiplied by the scaling factor" If the original is printed using the offset method (this is all printed products) and raster moire suppression is performed not by the scanning driver, but in Adobe Photoshop, set the scanning resolution to 2 times higher. Scanning at a higher resolution will simply be a waste of time. The lower limit of scanning resolution is determined by the capabilities of the computer on which the scanned image will be processed (a raster file of a full-color A4 image with a resolution of 300 dpi has a size of more than 20 MB), and the visual perception of the finished print. For example, raster files for printing large full-color posters for outdoor advertising are prepared with a resolution of 50-100 dpi, not only because of the huge size of these files (hundreds of megabytes), but also because further increases in resolution no longer improve the perception of the poster.

note: The resolution of a full-color file for printing on a color printer is by no means the same as the printer’s print resolution! Since each point of a full-color image with “8 bits per color” can have 256 gradations for each color, and a point printed by a conventional printer is either present in a given place or it is not. In practice, for printing at a 1:1 scale, the resolution of the original raster image should usually be from 150 to 300 dpi. At the same time, an image printed from a 300 dpi file can be visually assessed as excellent. A printer with single-color dots uses its 600, 1200, or 1440 dpi to render halftones, so its halftone resolution will be equal to single-color divided by 16 (a gross simplification, but generally true). For sublimation and other Contone technologies, each printed dot can have a certain number of shades (for sublimation, any dot can be any of 16 million colors and its halftone resolution is equal to one color).

A scanner with an optical resolution of 600 dpi will allow you to scan a 10x15cm photograph with a number of dots sufficient to print it on a magazine spread. By scanning at 3048 dpi optical resolution for a street billboard, you can enlarge your original by 50 times or more.

It is recommended to prepare a file for output on film sent to a printing house with a resolution 1.4 times higher than the output lineature (some experts recommend a file resolution 2 times higher than the lineature, but not even higher).

By the way, if you meet a person working in color at a publishing house, sincerely ask him to explain the meaning of the concept of lineature (here called the conventional term “halftone resolution”). A knowledgeable person will feel the need to immediately and thoroughly drink beer to discuss such a conceptual issue - lineature can also be a given input parameter...

Number of bits per color (color depth, bit depth)

Regular quantity binary information about the color of one point of a full-color image in a computer - 24 bits for each point, 8 bits for each of the main ones RGB colors, which gives over 16 million color options for this point. The eye cannot distinguish finer shades, and output devices usually do not reproduce them. Why scanners and graphics packages are they 48-bit? Technological answer: The CCD matrix in higher-bit scanners is usually more sensitive and has less intrinsic noise, the analog-to-digital converter is of better quality and has less intrinsic noise, and so on.

Mathematical answer: because at each stage of information conversion - during gamma correction, color synchronization program operation, image processing in a graphics editor, color separation when printing - the low-order bits cease to contain useful information. Expensive 36-bit (and higher) scanners use so-called downloadable gamma correction curves; in them, information about the color of a point is corrected without recalculating in the driver the data received from the scanner output, which is lost helpful information in the least significant bits, and inside the scanner, perhaps even at the stage of analog-to-digital conversion. Some 30-bit models use similar technologies, and, according to the manufacturer, the data from them contains as much useful information as “regular” (apparently without hardware gamma correction) 36-bit scanners provide. And one more thing: a scanner that operates with data of higher bit depth can have a larger dynamic range and can “distinguish” more details in the image, especially in the shadows (here by details we mean not small strokes, but gradations of saturation or brightness - “ polar bear in a snowstorm»).

Important: Obviously, a higher-bit analog-to-digital converter (for example, 36-bit) can be connected to the same CCD matrix as in a 24-bit scanner. In practice, a larger scanner will not necessarily have a larger ACTUAL dynamic range.

If a printing device uses CMYK dyes and can reproduce 256 shades of each of these colors for each full-color point given to it at the input, then the 36-bit color descriptions of this point received from the scanner, note, in RGB colors, will not be at all superfluous.

note: The bit width of the data transferred to the computer (namely to the scanning module) may be less than the bit width of the data inside the scanner.

Professional models usually have the ability to select the bit depth of the transmitted data (for example, 36 or 24 bits) and the dynamic range of 3D and higher. However, even in the price category “from 1000 USD” There are models (usually they are noticeably cheaper than “full-bit”) in which only 24 bits are transferred to the computer. This is explained by the presence of a certain “proprietary” algorithm for converting color information from the scanning bit depth (30 or 36 bits) to 24 bits at the output. Let us note, however, that such “improvements” have not been noticed in the products of the leaders of the publishing market.

By the way, in color mode, scanners with higher bit depth usually scan a little (10 percent) slower than previous models. This is understandable - the data has increased by 20 percent.

Range of optical densities, maximum density.

A parameter that not all sellers of household scanners have heard of and which is not always reported by the manufacturer for scanners under $500. Optical density is a characteristic of the original. It is calculated as the decimal logarithm of the ratio of light incident on the original to light reflected from the original (for opaque originals) or transmitted (for slides and negatives). The minimum possible value of 0.0 D is a perfectly white (transparent) original. A value of 4.0 D is an extremely black (opaque) original. When applied to a scanner, its range of optical densities characterizes the scanner's ability to distinguish nearby shades (this is especially critical in the shadows of the original). The maximum optical density of the scanner is the optical density of the original, which the scanner also distinguishes from “ complete darkness" The scanner will not be able to distinguish all shades of the original “darker” than this border. In practice, this means that an “office” scanner can lose all the details in both dark and light areas of even an ordinary photograph, not to mention scanning a slide, and especially a negative.

What types of originals and scanners are there??

Regular color photography and printed materials - up to 2.5D. Negatives and x-rays - 3.0-3.6D. Inexpensive flatbed scanners have a dynamic range of 2.0-2.7D, good 36-bit 3.0-3.3D, latest models- 3.6D. The range of optical densities of the scanner is not determined by the brightness of the lamp, as it might seem, but is related to the quality (as well as the type and bit capacity) of the ADC, CCD matrix and the operating algorithm of the scanner controller. In high illumination, the matrix “goes blind”, and the ADC has an upper limit, the voltage above which does not differ. In low illumination, the matrix has a sensitivity threshold and its own noise, and the ADC has the weight of the least significant digit, the voltage below which does not differ. If I'm not mistaken, the mathematical limit of the dynamic range for a scanner with a 30-bit ADC is 3.0D, 36-bit is 3.6D (decimal logarithm of the number of possible gradations for each color, which is equal to 2 to the power of the number of digits per color). In reality, some of the discharges are “eaten up” by transformations and noise.

IMPORTANT: Manufacturers may provide completely different optical density range data. Real range - determined by scanning the reference original.

Design range- a certain figure, apparently a compromise between the requests of the marketing department and real indicators. An unusually high value certainly applies here.

All scanners we have encountered so far that cost up to $1,000 produced 24-bit data and had a real optical range of 1.8-2.5D (in the documentation it may be 2.7D or even 3.0D).

The maximum difference encountered between the declared and actual dynamic range is 0.6D.

Dmax— maximum optical density. The dynamic range is less than this value by Dmin- usually Dmin=(0.1-0.2)D. (The scanner's ability to distinguish bright areas is also limited.)

note: It will not be possible to scan a negative with acceptable quality using a regular 30-bit flatbed scanner, even if a slide module is sold for it. Even a 30-bit scanner with best-in-class real dynamic range can scan color slides tolerably, but don't expect acceptable results with artistic black-and-white negatives shot by a professional photographer. For negatives you need a different class of scanner. In general, for use in printing, it is better for the owner of a scanner with a maximum optical density below 3.0D to scan negatives and slides that require additional color correction “on the side”, and save on the slide module, especially since they cost up to $700 for some models. Inexpensive slide scanners are not the answer - usually their performance and scan quality are no better than flatbed scanners.

Receiving element - CCD matrix

One of the most important components affecting scanning quality. The characteristic given in the documentation is the number of elements per line (per color). The number of elements divided by the width of the scanner’s working area is equal to the optical resolution (it is actually determined by these two parameters).

Unreported but extremely important matrix parameters:

• noise level— limits the dynamic range and the actual number of data bits containing useful data. In principle, nothing prevents you from connecting a 36-bit ADC to a cheap noisy matrix, but this is unlikely to improve the quality of the resulting image. True, it won’t get worse.
• sensitivity variation from cell to cell— even if the scanner provides calibration, it is performed using averaged values ​​from several cells.
• crosstalk level— a brightly lit cell affects its neighbors.
• color combination— in single-pass scanners, colors are separated by three lines of the CCD matrix.

Since the rejected matrices will obviously not be thrown away, but will be sold as substandard at a reduced price, guess which scanners they will end up in?

This year, entry-level scanners with a CIS receiving element appeared, but you shouldn’t expect any real advantages from them, other than the thinness of the scanner. In fact, this technology may not be entirely suitable for full-color work, despite its large bit depth.

Scan quality: presence of artifacts, sharpness, noise.

“30-bit 600×1200dpi scanners” cost differently. Because these numbers do not yet guarantee the actual quality of the scanned image. The differences between a high-quality mechanism and the “cheapest in the Moscow region” are akin to the difference between cameras. A DSLR with a five-element (without Zoom) lens can shoot on the same film as a fifty-dollar point-and-shoot camera with a plastic lens and a fixed focus, but pictures from a point-and-shoot camera can make you regret not only the money spent on it and printing photos, but also ruin the pleasure of your vacation.

Multi-colored repetitions around the contour of an object, colored spots, “cloudiness” and blurred images - all these unpleasant surprises are almost guaranteed to be found in radically cheap models.

Scanner controller

The functional part of the scanner, which is difficult to understand due to the confidentiality of information about the algorithms used, has a huge impact on the speed of the scanner and the accuracy of color rendering.

Often, the manufacturer emphasizes the fact that its scanners (read: controller) use unique technologies. Apparently, the buyer should be imbued with faith in the unprecedentedly high image quality of this particular scanner, ensured by the presence of these technologies only in it alone. Indeed, each company has its own names of “proprietary” technologies. The best you can learn about them is the result they promise. Common words like " unprecedentedly clear images with bright and rich colors“It’s better to discard it right away. When working with full-color images, there is a reference point - professional publishing models. If the “machine intelligence”, which with its secret methods makes candy out of a blurry slide with shifted colors without human intervention, is not implemented in them, where does it come from in a scanner costing up to $1000, produced by a company that has never had anything to do with the development of equipment for professional work? with color?

The interface may vary.

Own (completely non-standard) interfaces, the scanner comes with its own unique card and works only with it. This card may not work in the computer after Upgrade or may fail.

SCSI(more or less, not always Fast SCSI-2). If you are going to use the scanner not with the card supplied, please note that easy compatibility is only with Adaptec controllers, and not UltraSCSI modifications. All other options can cause problems (I fully understand what ASPI-compliant means, but believe me, in this case it is better to “live with ISA” than with non-Adaptec for PCI.)

The non-Adaptec interface cards supplied with SCSI models do not promise to connect other SCSI devices, if only because they are not supplied with drivers (but for some you can find drivers yourself). However, such cards are directly understood by the scanner driver and provide the simplest and most convenient process for initially connecting the scanner and migrating to new versions of operating systems. Some of these cards do not require a dedicated interrupt.

Adaptec will allow you to connect anything, but requires some interruption and some fiddling to install. The data buffer size in tablet models varies from 64 KB to 3 MB.

LPT(and its variants, with EPP or Bi-Directional support or requirement).

Important: The scanner may require one of the high-speed parallel port options. While EPP is usually always available, the 8-bit Bi-Directional option required for Epson scanners is not implemented everywhere. A “pass-through” connector for connecting a printer does not guarantee that any printer will work with it.

PCMCIA(PC CARD) - this scanner and this Notebook may or may not work together, it’s better to try!

Software part

Modern programs running under Windows communicate with the scanner through a special program supplied with it - the TWAIN module (on Macintosh the scanning module is executed as a Plug-In for Photoshop). All programs that support the TWAIN standard (these are all known programs, both graphic and OCR), in theory, should work with any scanner that supports it (these are all modern scanners). In practice, some Russian text recognition programs may not work with a scanner that has not been previously tested by the developer.

IMPORTANT Note: Since the scanner TWAIN module is a regular program, this program may not work under some operating systems in general (and even versions of Windows 95 differ), or work very poorly. The general law of “driver quality” is valid here - drivers from unknown manufacturers do not work very reliably, and with the release of the next version of Windows, a new driver will be needed for normal operation.

Some useful properties not always found in TWAIN modules:

• opportunity automatic detection scan settings.
• preview window with selection of the scanned area and display of the results of the settings made and image correction in real time.
• smooth adjustments of brightness, contrast, gamma correction.
• selection of black and white points, preferably with an “eyedropper” and setting the value.
• printed raster suppression filter, multi-level or customizable.
• inversion (negative) and reflection (reversal) of the original.
• built-in color synchronization system with a set of profiles, allowing you to adjust the scanned image to suit specific device output or convert it to CMYK.
• Possibility of scanning over the network.
• various filters built into the driver for correcting sharpness and emphasizing the edges of the image. Inferior to those available in Adobe Photoshop (with the exception of the LinoColor program of Linotype-Hell scanners).

Functionality found in professional models:

• tonal correction with separate RGB/CMYK curves, separately in highlights, shadows and midtones.
• compensation " color shift» original, numerically specifying the subtracted color or indicating the reference color that the image point specified by the operator should have after scanning.
• automatic subtraction of the film color of the slide (does not replace color shift compensation due to possible own color distortions on the slide, but it won’t hurt either).
• possibility of batch and group scanning, automatic recognition framed slides.
• performing color separations with setting the appropriate profiles and printing parameters. Publishing packages are usually more difficult to set up color separation, but they perform it better than a scanner driver (an exception is the LinoColor program for Linotype-Hell scanners. But it also costs real money).
• printed raster suppression filter with the ability fine tuning operator

Calibration, characterization, color correction and colored targets

It is important to understand the difference between the two types of scanner calibration:

• a periodic calibration procedure of two or even one shade of gray is designed to compensate for lamp aging.
• scanner characterization—creating a scanner color profile for the color synchronization system.

The first one only slightly changes the shape of the correction curve and is not capable of making fatal changes to the information about the color of the point. Color profile devices can give advice like “ we will consider all 40% purely red areas to actually also have 10% blue, and leave all 50% areas unchanged" Are you trying to restore the correct shades of images processed in this way?

The characterization tools used in production are noticeably more powerful than those included with common types of scanners, so you should not immediately reject the factory profile and assume that a certain procedure involving a colored target will obviously give a better result. Modern pre-press scanners are usually supplied calibrated to the supplied standard profile at the factory (how? by “flashing” a correction table) or complete with an individual profile and provide quite acceptable color accuracy.

Ordinary photographs or slides themselves need color correction - colors even on film from different manufacturers are transmitted completely differently, and photographs from “express printing” usually have a radically disrupted color balance, since printing is carried out in automatic color correction mode by default.

Morality: There is no point in creating a scanner profile based on a color target on AGFA photo paper (note, the shelf life of these targets is 1 year) in order to scan a slide on FUJI film.

There is also no obvious benefit from the scanner driver subtracting the color of the blank film when scanning a slide if color correction will be done anyway.

It is better to scan images intended for repeated use without any corrections, “as is”. If you save an already corrected image and re-correct it, you will lose quality or may not be able to get an acceptable result at all.

REALLY IMPORTANT: when performing color correction based on the image on the monitor, you need to at least set it color temperature(5000K if this image will be printed on paper) and gamma (1.8).

It is also necessary to imagine the operation of color synchronization systems: the dot colors received by the scanner driver can first be changed by the scanner driver itself according to It’s not always-clear-for-what-cases-it’s-intended profile if the built-in control system is activated. Moreover, along the way, the driver may try to adjust to the monitor, also unclear which one, and make a preliminary correction for the printer, in the hope that the image will not be viewed, but will be printed without any color correction. Next, the data is transferred to the program from which the scan is performed. If its built-in color management system is activated, it may be AGAIN The received data was corrected according to the profile of an unknown scanner, then according to the profile of an unknown monitor when outputting to the monitor, and according to the profile of an unknown printer when printing. On top of all this there is also an operating system and special programs color synchronization, which can “correct” the data transmitted to the printer and monitor, and the possibility of automatic color correction in the printer driver or rasterizer.

It is important to understand that only one color synchronization system should make these corrections. If color correction for the printer is already selected in the scanner driver - application program and the operating system should send data to the printer without changes, and the colors on the monitor will be “wrong”.

In printing, by the way, color is often checked “blindly” - not by the monitor, but by the percentage of colors at a given point. We know what values ​​correspond to flesh color, grass, sky, and so on.

How to choose a scanner?

The answer is unexpectedly simple - for the task at hand. You just need to answer yourself - how will the scanned image be used, what programs will it be processed, what devices will it be displayed on, what are the requirements for image quality, what operating system will be used on the computer, what interface should the scanner be connected to.

If you are going to scan full-color images and then print them, look for signs in the scanner that they are intended for publishing and design work. In order to place color originals on WEB in 256 colors, a high resolution and a large dynamic range is useless, but a stable working TWAIN module and Descreen filter are highly desirable.

If you need to scan 35mm negatives with magnification for the entire page of an A4 color catalog on glossy paper, and you have allocated 500 USD for the purchase of a scanner. — better save that money for the scanning bureau.

At all, modern marketing Almost always, even at the development stage, a product is positioned for only one typical group of consumers, and if, for example, the main advantage of a VCR is the ease of use (apparently by housewives), it is unlikely to contain the functions necessary for editing material from a camcorder. Position your requirements for a scanner on the market for similar devices - almost always the product is aimed at a specific range of typical tasks and customers, and functions that are secondary to them can be implemented extremely poorly. Look for models that advertise properties that are useful for your task as the most advantageous, and not those that are clearly “unusual” for it. The raster moire suppression filter and the ability to work with crumpled A0 drawings in blue with an uneven background color are usually mutually exclusive.

A common mistake is trying to select a scanner for publishing work from a segment-oriented "typical middle american office use" products.

note: specialized slide scanners usually have a rather conventional design, similar to flatbed scanners. This means that their “slide orientation” in itself does not provide any advantage in scanning quality, while the price is comparable to the price of a flatbed scanner with a slide module, which has similar characteristics.

Advantages of low-cost slide scanners - high speed work and automatic slide feeder. But they are not implemented in all models.

Positive press reviews and prizes received by the manufacturer for several years are very good. Advice from an uninterested acquaintance who would be an expert is even better.

It is important, however, especially in relation to publishing tasks, to correctly interpret what you read and hear: the fact is that the vast majority of publishing houses in the world (not Russia) work for Apple platform Macintosh, and if you mechanically follow the enthusiastic reviews of “Apple fans” when choosing a scanner to work in Windows environment, you can miss quite a lot. Many publishing industry veterans with long traditions of working with the Macintosh pay incredibly little attention to the software running on their Windows devices.

In addition, testing in the computer press is usually carried out according to randomly selected criteria, while opportunities that are important for a particular person remain behind the scenes. I recommend that you carefully read the facts given in the review and leave aside the conclusions. When reading the last review, I was wondering what this person would say about a high-end flatbed scanner for $40,000? Probably something like this:

« There is no possibility of automatically sending data to a fax modem, in the text scanning test it showed the lowest speed of all, it uses an outdated SCSI-II interface, connection is difficult for a non-specialist, there is no interface card included in the package, the lamp takes a long time to warm up. True, there is a plus - good quality scanning photos, but does not include a text recognition system».

If you have the appropriate skills, it is useful to visually assess the quality of the scan. It is worth testing your ability to distinguish small parts, such as concentric lines and small text on a banknote. It seems unrealistic to check the correct color rendering on an unfamiliar computer, especially considering the possible distortions introduced by incorrectly configured color synchronization systems of the scanner, graphics editor or operating system.

I recommend immediately weeding out models from relatively small companies that have been offering scanners for less than two or three years (due to fears that in a year they will get tired of selling scanners and at the same time supporting those already sold), as well as models for which drivers cannot be freely obtained from the Internet. Remember that without a stable driver (TWAIN module) the scanner cannot be used for its intended purpose, you won’t be able to “attach” another scanner driver to it, and there are no “proprietary” drivers for the scanner included in the Windows package. This is especially true in anticipation of Windows98/NT5.

How bad can drivers be?

RADICALLY bad. They may not work at all (or “freeze every once in a while”) under one of the Windows variants or with some programs (in particular, Russian-language OCR).

The author had the opportunity to connect an “inexpensive” scanner from one of their most famous manufacturers, which with the driver from the standard package (in a colored box with a decent set of programs!) did not work in any way - neither under Win3.1, nor under Win3.11, nor under Win95, neither with Russian nor with their English versions, nor through the included programs, nor through OCR, nor through well-known graphic editors. A new version drivers solved the problem, but how could such a “gift” be put on sale?!

In the article Evgeniy Kozlovsky“To a gift horse...” in Computerra describes an even darker example of attempts to work with the Primax Phodox scanner.

Be sure to make sure you can get new versions of scanner drivers and software for free online. Some manufacturers do not make the latest version of the driver publicly available, but offer to send it free by mail to those who purchased a scanner in the last six months and for $50-90 to others. " Send by mail for money“to Russia - for this you at least need to have a credit card, and one that is actually used abroad. Getting the driver from where you bought the scanner can be problematic - usually you have to rewrite the entire CD-ROM to do this.

Personally, when choosing equipment (especially an unfamiliar group), I practice a subjective assessment of the product and manufacturer based on secondary characteristics. I immediately discard the manufacturer with clearly slippery advertising on the verge of deception, or statements “ perfect image quality and reliability" If the manufacturer is deceiving in at least one case, why look further?

An indicator of the quality of equipment is plastic. Try the body with your finger IBM laptop and remember the feeling. At the same time, you can touch neighboring models. So far we have not been able to find normal equipment in a case made of very bad plastic.

Additional information is provided by the packaging. A box made of coated cardboard with bright pictures contradicts the concept of “for the European market.” The ideal is recycled cardboard with dim markings, while the seal inside is not polystyrene foam, but three-dimensional cardboard elements!