Measuring contrast in black and white images. Criteria and methods for integrated assessment of image quality in raster graphic formats

When it comes to measuring certain image parameters, an unpleasant subtlety immediately arises. Humans and computers perceive images differently. A person isolates objects from noise, can see something in low light, and a computer understands a picture as a set of coordinates with corresponding brightnesses. And when a person and a computer are asked about some distinctive features of an image, they will immediately disagree. There must be some way to ensure that the conclusions they draw are similar.
Let's look at the methods that are used to analyze contrast in black and white images, and we’ll try to choose something more or less objective.

Method one

Contrast is determined by the ratio of the difference in brightness of observation object 1 and background 2 to one of these brightnesses. The range of displayed values ​​is from 0 to 1.
In fact, it shows nothing. Let's move on.

Method two

Method three

Method four

Method five

There is one subtlety with calculating contrast. There are two objects nearby, one has brightness 10, the other 20, using the first and third methods we get 0.5, in the second - 0.3. Brightnesses are 100 and 200, using the first and third methods we get the same 0.5, using the second - again 0.3, but with brightnesses of 10 and 20 you may not see the difference.

Contrast, in my opinion, is more objectively calculated using the Vorobel method, if the quality is poor and there is a lot of noise, take into account the areas of objects, and average the brightness values ​​of the objects from them.

Now let's see this in action:

There are three images in a row - normal, with an equalized histogram, and ideal. The selected areas were analyzed in .bmp format, brightness range 0 - 255.

The contrast of a normal image is K = 0.67.
- contrast of the equalized image K = 0.88.
- contrast of an ideal image K = 1.

This is the story, thank you for your attention!

Editor-in-chief - Vladimir Krylov, Ph.D.
Deputy editor-in-chief - Mikhail Nikulichev, Ph.D.

The first part of the article is devoted to the characteristics of modern LED screens that affect image quality - brightness control using PWM methods, time-division image formation and screen refresh rates. The second part of the article discusses - dynamic range brightness, color rendition and contrast of screens, drivers and modern systems control of LED screens, electromagnetic compatibility and industrial screen interference.

LED screen – complex electronic device, containing a large number of components. The image quality and performance characteristics of an LED screen depend both on the parameters of the components used in the screen and on the capabilities of the control system for this screen.

In terms of image quality, the following screen characteristics are important:

• screen resolution (so-called spatial resolution), in the case of LED screens, usually expressed as the distance between pixels (pitch size);
• maximum brightness (measured in Nits);
• dynamic range of brightness, expressed in the number of brightness levels that can be displayed on an LED screen (this characteristic is also called radiometric or energy permissions);
• frame rate, expressed in the number of frames displayed per second (fps) (this is the temporal resolution);
• frame update rate (refresh rate), measured in Hertz (this is also a time resolution);
• spectral resolution - how many spectral components form an image;
• color uniformity across the entire screen;
• white balance and the ability to customize it;
• linearity of brightness perception is a subjective characteristic of image quality, which is expressed in the ability to distinguish similar brightness levels with the eye, both in dark areas of the image and in bright ones;
• screen image contrast;
• characteristics of changes in screen image quality depending on the viewing angle;

In addition to image quality, we also note the following operational characteristics of the LED screen:

• Availability of a LED screen condition monitoring system;
• development of software (software) of the control system (the ability to build networks of LED screens, including networks containing both LED and LCD screens, the ability to control screens via the Internet, the presence of a built-in information security subsystem);
• level electromagnetic radiation in the form of industrial radio interference created by the LED screen.

Let's take a closer look at some of the above characteristics.

Formation of an image on an LED screen and control of brightness

Pulse width modulation (PWM) and refresh rate

The original image for display on the LED screen is formed in the form computer file, most often in the form of a video in some format (*.avi, *.mpg). This file is decoded by the control computer (or video controller), then converted into a special digital stream supplied to the driver chips DC, which in turn provide transmission electric current through an LED, which causes radiation in a certain spectrum.

To form different levels The brightness of LED radiation is measured using the pulse-width modulation technique - PWM (Pulse-width modulation). The essence of this technique is that, depending on the required brightness level, current is not constantly supplied to the LED, but only for some time (depending on the required brightness level), then it stops being supplied, then it is supplied again, etc. For example, to generate a brightness of half the maximum, it is necessary to pass the current half the time of a certain cycle, at a quarter of the brightness - a quarter of the time, etc. In other words, the LED operates in the “on-off” mode, and the on-time is proportional to the required brightness level.

From this technique it follows that the image on the LED (and therefore on the screen) is formed cyclically. The time of the minimum cycle during which the LED sequentially “turns on” and “turns off” is called the refresh period (refresh time). The inverse value is more often used - refresh rate.

Let's look at an example. Let the refresh rate of the LED screen be 100 Hz. If we need to ensure full brightness - 100%, then we constantly supply current to the LED throughout the refresh period, equal to in this case 1/100 s = 10 ms. If a brightness of 50% is required, then during this time we supply current for 5 ms, do not apply it for the next 5 ms, apply 5 ms again in the next cycle, not 5 ms, etc. If 1% of maximum brightness is required, then current is applied for 0.1 ms and not applied for 9.9 ms.

In addition to this technique, modified PWM methods are used: Scrambled PWM (Macroblock), Sequential Split Modulation (Silicon Touch), Adaptive Pulse Density Modulation (MY-Semi). The essence of these techniques is to “spread” the “on” time of the LED over the entire refresh period. So the formation of 50% brightness at a refresh rate of 100 Hz may look like this: 1 ms - LED is on, 1 ms - off, 1 ms - on, 1 ms - off, etc. That is, for 50% brightness we can say that the refresh period decreased by 5 times and became equal to 2 ms. Accordingly, the refresh frequency increased and became 500 Hz. But these figures are valid only for the formation of 50% brightness. For each brightness generation scheme there is a minimum brightness - 1 pulse (some minimum time) of turning on the LED and the rest of the time it is turned off.

Thus, the clear cyclicity inherent in traditional PWM is distorted when using modified methods, since, depending on the brightness level, periods with less time (and therefore a higher refresh rate) can be distinguished. You can, for example, say that for a given LED screen, the refresh frequency varies from 100 Hz to 1 kHz. This means that we display the minimum brightness on the LED screen with a refresh period of 100 Hz. And when high brightness levels are formed, periods (“on-off” of LEDs) with shorter duration can be distinguished.

So, for modified PWM methods, such a concept as refresh rate can be interpreted ambiguously. However, if we consider the refresh period as the minimum time during which the image is updated for all brightness levels, then this value does not depend on the PWM generation scheme.

Interlace scanning or time division of LED screens

In some cases, the LED screen design provides for an image formation method in which current cannot be supplied to all LEDs at once. All screen LEDs are divided into several groups (usually two, four or eight), which are turned on alternately. That is, the image formation methods described above are applied in turn to each of these groups. In the case of two such groups, image formation is similar to that used in analogue television interlaced scanning.

This method is used mainly to reduce the cost of LED screens, since its implementation requires fewer LED drivers (two, four, eight times - the number of times corresponding to the number of alternately switched groups), which make up a significant part of the cost of the LED screen. In addition, the time division method is almost inevitable when high resolution(i.e. small pitch) LED screen, since in this case it is extremely difficult to ensure placement large quantity drivers and their heat sink.

It should be understood that when using this method, the maximum brightness of the LED screen is reduced, and the refresh rate is also reduced (by the number of times corresponding to the number of groups).

Let's assume that we are making a time division between two groups of LEDs. One group is supplied with current according to the required brightness and the PWM method used. The other group is disconnected from the power source at this time. After the refresh period, the groups change - now the second is energized, and the first is turned off. Therefore, the total period during which all information on the LED screen is updated doubles.

The concept of refresh rate in this case becomes even more blurred. Strictly speaking, the refresh period as the minimum time during which the image is updated for the entire LED screen increases. However, if for each group we consider only the period during which the image is formed by the PWM method, then the refresh rate is the same.

LED screen refresh rate and the human eye

The refresh rate primarily affects the perception of the image by the human eye. The image, figuratively speaking, constantly “flickers”, although at a fairly high frequency. Human perception of light images is a psychophysical phenomenon and is designed in such a way that individual flashes of light are summed up over time. This summation occurs over a certain time (10 ms) and depends on the brightness of the flashes (Bloch's law). If the light “flickers” quickly enough, with a frequency above a certain threshold (CFF - Critical Flicker Frequency), then the human eye perceives this light in the same way as if it were constantly on (Talbot-Plateau law). The CFF threshold frequency depends on many factors, such as the spectrum of the light source, the location of the source in relation to the eye, and the brightness level. However, it is safe to say that under normal conditions this frequency does not exceed 100 Hz.

Thus, if we consider the perception of an image on an LED screen generated by the PWM or modified PWM method by the human eye, then an image with a refresh rate of 100 Hz and 1 kHz will be perceived the same.

Screen refresh rate and video camera

However, not only the human eye, but also video recording equipment, which has characteristics different from the eye, can act as a perceiving system. This is especially true for LED screens installed in stadiums, sports facilities or concert venues, from which video is usually broadcast. Exposure time, or shutter speed, in modern video cameras can vary from seconds to thousandths of a second.

Let's consider an LED screen in which the image is formed by the traditional PWM method with a refresh rate of 100 Hz. A static image is shown on the screen. Let's also assume that we are shooting an LED screen with a video camera at a shutter speed of 1/8 s, i.e. exposure time 125 ms. During this time, light from 12.5 refresh periods will reach the photosensor. When we take a series of frames with a given shutter speed, the difference is luminous flux, falling on the photosensitive element does not exceed the flux generated by the LEDs during 0.5 refresh periods, i.e. no more than 4% of the total flow. The difference is formed due to the fact that the video camera and the LED screen are naturally not synchronized and each frame taken by the video camera falls into different times relative to the beginning of the LED refresh cycle. Thus, the video image from the camera will show a fairly smooth picture from the LED screen.

Now let's reduce the shutter speed with which we shoot to 1/250 s, the exposure time is 4 ms. This time is 2.5 times less than the refresh period. Now the relationship between the start time of the video camera frame and the start of the PWM cycle will be significant. Some frames may fall at the beginning of the cycle, others in the middle, and others at the end. Thus, a significant error in the light flux is formed in different frames. That is, the image played on the video camera will randomly change brightness and will “float”. In addition, the brightness of the image will decrease, which, however, is typical for all objects shot at a short shutter speed. If you further reduce the shutter speed, then black frames are more likely to appear (when the beginning of the video camera frame falls on the part of the PWM cycle where the LED is “off”) and the image from the camera will begin to flicker.

Thus, if we want to shoot an LED screen with a video camera, on which the image is formed using traditional PWM, then the refresh rate must be comparable to or exceed the shutter speed at which the camera shoots.

In the case of using modified PWM methods, the same reasoning can be carried out. Due to the “spreading” of the LED turn-on time over the PWM cycle at high brightnesses, the image captured on a video camera will be more stable than when using traditional PWM. But at low brightnesses the situation remains the same - the picture will either change brightness or flicker. Since a real image contains, as a rule, different brightness levels, the image captured on a video camera will also have errors, although of a different nature.

So, when shooting video, it is impossible to avoid image distortion with arbitrary shooting parameters. You can always find the shutter speed at which the video will be distorted. The situation is similar to shooting an analog TV with an analog camera. Due to differences in scanning frequency, when shooting in this way, diagonal black stripes are visible on the TV being shot.

More important for video recording of an LED screen is the issue of uniformity of the image captured on the video camera. The LED screen is a modular design, consisting of several blocks, the image on which is directly formed by various controllers. If these controllers do not synchronize the start of the PWM cycle, that is, the start of the cycle in different parts of the LED screen occurs at different times, then the following situation may occur during shooting. On one part of the LED screen, the beginning of the video camera frame may coincide with the beginning of the PWM cycle, and on another, for example, on the middle. If the shutter speed is comparable to the refresh period, then in one area the image will be lighter and in another darker. In this case, the entire image on the LED screen will be divided into rectangles of different brightness, which represents greater discomfort for the viewer.

The cost of increasing the refresh rate of LED screens

Regardless of the method of generating PWM, the circuits that implement them have common features. The PWM generation circuit has a certain clock frequency F pwm. Let's say we need to generate N brightness levels. In this case, the refresh rate F r cannot exceed F pwm /N.

To illustrate, here are some examples:

These figures assume that there are independent PWM generation circuits for each LED, that is, the PWM circuit is implemented directly in the LED screen drivers.

In case of use simple drivers and generating PWM on the LED screen controller, it is necessary to consider how many drivers are connected in series and served by one PWM circuit. If one PWM circuit serves M drivers with 16 outputs, then the refresh frequency cannot exceed F pwm /(N*M*16), which leads to a significant reduction in the refresh frequency or the need to significantly increase the clock frequency.

In the case of using time division (interlaced scanning), as we have already said, the refresh rate decreases in proportion to the division coefficient.

So, to increase the refresh rate of LED screens, the following options are possible:

• use of “smart” drivers;
• increasing the clock frequency of the PWM generation circuit;
• reducing the number of brightness levels (color depth).

Each of these methods has its own advantages and disadvantages. So, smart drivers are more expensive than conventional ones, increasing the clock frequency increases power consumption (and therefore heat dissipation, the need for heat dissipation to avoid overheating), reducing the number of brightness levels reduces image quality.

Refresh of LED screens: Conclusions

Often such a parameter as the refresh rate of LED screens is used for marketing purposes as one of the indicators of image quality. It is assumed that the higher the refresh rate, the better the LED screen, all other things being equal. However, sometimes figures are given that mislead potential buyers. For example, specifying a refresh rate of several kilohertz, as we have seen, can mean either the use of modified PWM methods, for which the refresh rate is different for different brightness levels, or a decrease in color depth.

It should be understood that high values refresh frequencies and, at the same time, color depth most likely suggest that this refresh in an LED screen is achieved at certain (high) brightness levels.

In the case of interlaced scanning, the frequency corresponding to one PWM cycle for one group of LEDs can be indicated, while the actual screen refresh frequency (which affects perception) is several times lower.

More informative, apparently, is to indicate the color depth and PWM clock frequency, with the possible addition of a range of screen refresh rates (for example, 200-1000 Hz) in the case of using modified PWM methods. If time division is used in the LED screen, then it is necessary to explicitly indicate this method of image formation (for example, time division = 1:1 - no time division, time division = 1:2 - PWM works on half the screen at the same time, etc.).

For eye perception, this parameter of the LED screen is generally insignificant. For frequencies above 100 Hz, the human eye will not see a difference in image quality. Therefore, it is necessary to understand whether it is necessary high frequency refresh and is it worth paying for it.

In case active use LED screen during video shooting, this indicator becomes significant, but you should also pay attention to the uniformity of the image when shooting video. For such LED screens, it may be better to conduct test shootings than to rely only on such a parameter as the refresh rate.

Tone correction in Photoshop

Sofia Skrylina, teacher information technology, St. Petersburg

Tonal correction of an image means lightening, darkening or increasing the contrast of the entire image or parts of it. This article will discuss methods for diagnosing image tonality and tools for tone correction of photographs.

Diagnosis of image tone

Before you begin image correction, you need to analyze the image and determine the tonal range, which will help you choose the right tools to correct the flaws in the original image. For these purposes, an image histogram is used.

A histogram illustrates the distribution of pixels in an image. This is a graph that shows the number of pixels at each color intensity level. Axis X tone gradations are located in the range from 0 (black or shadow) to 255 (white or light), and along the axis Y— the number of pixels of each level. The histogram allows you to determine whether an image contains enough detail in the shadows (left side of the graph), midtones (middle), and highlights of the image ( right side). In Fig. 1 shows an example of reading a histogram.

Rice. 1. Examples of reading a histogram: a - a very light photograph, the graph is shifted to the right, into the area of ​​highlights; b - photograph with a full tonal range, the graph is plotted at all levels of light intensity; c — dark photograph, the graph is shifted to the left, into the shadow area

To open the palette Histogram(Histogram), run the command Window(Window) -> Histogram(Histogram). This palette is not a correction tool; it is intended only for image diagnostics. To determine the tonal range, a histogram of the combined RGB channel is used, and advanced viewing is used to display statistics (Fig. 2).

Rice. 2. Histogram palette with statistics

Dropdown list Source(Source) becomes available for multilayer documents: you can evaluate the tonality of the current layer or the total image, taking into account all layers. Options Level, Counter And Percentile display statistics for the area under the mouse pointer (Fig. 3).

Rice. 3. Histogram palette for the selected layer with statistics of the current position of the mouse pointer on the chart

In the palette Histogram(Histogram) the following statistical information is provided below the graph:

In Fig. 3 the histogram occupies the entire tonal range. Graph height and parameter value Average(113.86) show that the image has quite a lot of highlights - which means the photo is exposed correctly. The deviation is insignificant (58.68), so the image does not have sharp light transitions. From all this it follows that this image does not require tone correction.

Please understand that there is no perfect histogram! Each image is different from the others and has its own unique pixel distribution graph. Moreover, it is not always necessary to make corrections when a histogram shows a clear shift towards highlights or shadows. For example, it is logical that a picture taken at night or in outer space has low level brightness And the histogram in this case will correspond to an underexposed photo (Fig. 4).

Note that the peak of the histogram is shifted to left side graphics - this indicates the content of a large number of shadows and a very small number of highlights in the image. This is also indicated by the parameter values Average(26.89) and Median(11). But this photo is not underexposed, it was taken in natural conditions. And it would be wrong to correct such a photo just for the “correct” appearance of the histogram.

Here is another example of an exception to the rule (Fig. 5). The winter landscape is the exact opposite of the previous example. The peak of the histogram is shifted to the right (towards the area of ​​highlights), and the graph has few dark areas. Parameter values Average(169.30) and Median(169) are close to maximum brightness. But, despite the histogram readings, this image does not require correction, its brightness is natural.

Levels

Dialog box Levels(Levels) is called by the command Image(Image) -> Correction(Adjustments) -> Levels(Levels) or the keyboard equivalent of Ctrl+L (in Mac OS - Command+L). The window displays a histogram of the image. But, unlike the palette Histogram, in this window we can make corrections by manipulating three sliders: - shadows, - midtones, - highlights (Fig. 6).

In Fig. Figure 7 shows a histogram of a dim image. Please note that the graph is not distributed over the entire tone interval, but only over part of it. There is not a single pixel of brightness level to the left or right of the graph.

Therefore, during correction it is necessary to expand the tonal range. To do this, assign zero brightness to the darkest pixels, that is, move the black slider to the right to the bottom of the graph, and assign maximum brightness to the lightest pixels, that is, move the white slider to the left to the bottom of the graph (Fig. 8).

Rice. 8. Correction of a dim image in the Levels window is carried out by shifting the black and white sliders to the bottom of the graph

Simultaneously with the change in the location of the sliders, the histogram in the palette changes (Fig. 9), which shows us that as a result of the correction, the image now has pixels over the entire tonal range (strip graph).

Rice. 9. Changes in the Levels window entail changes in the Histogram palette

Notice the triangle with an exclamation point in the palette window Histogram, which appears during the correction. It warns that brightness levels have been removed as a result of their redistribution across the entire scale. Therefore, level dips have formed, which are clearly visible if you click on this icon (Fig. 10).

Rice. 10. The result of increasing the contrast in the Levels window and the changed appearance of the histogram after correction

Lightening and darkening an image

To lighten an image that is too dark or darken an image that is too light, you need to change the position of the gray slider, that is, the gamma of the image. By default, gamma is 1. For a dark photo, the slider moves to the left (gamma value greater than 1), for a light photo - to the right (gamma value less than 1).

There are examples of very light or dark images in which not only the histogram peak is shifted towards highlights or shadows, but the entire graph of brightness levels is not distributed over the entire tonal range. To correct such an image, it is enough to assign zero brightness to the darkest pixels (for light images) or assign maximum brightness to the lightest pixels (for dark images). In other words, move the black slider (for light images) or white (for dark images) to the bottom of the graph. In this case, the shift of the gray slider occurs automatically, but if necessary, to enhance the effect, the gray slider can also be shifted towards highlights or shadows.

In Fig. Figure 11 shows the original light image of Eltz Castle and its histogram. The histogram is not distributed over the entire tonal range, and its peak is shifted to the right.

To correct this image, the black slider is moved to the bottom of the graph, and the gamma value is slightly reduced (Fig. 12).

Correcting tonality using curves

Dialog box Curves(Curves) opens with the command Image(Image) -> Correction(Adjustments) -> Curves(Curves) or the keyboard equivalent of Ctrl+M (in Mac OS - Command+M). This window allows you to make corrections using 14 different points in the tonal range of the image (from shadows to highlights). The tonal range is represented as a straight diagonal line (Fig. 13).

To display a histogram of the image at the same time as a straight line, select the checkbox Histogram(Histogram). Clicking the mouse with the Alt key (on Mac OS - Option) anywhere in the coordinate system changes the grid spacing, which can also be done using the two buttons at the bottom of the window (see Fig. 13).

To correct the image in the mode of changing the curve using points (button), you need to add points to the graph, and then bend the curve.

To add a point to the graph, simply click the mouse at the desired location on the straight line. If you need to delete control point, then you first need to select it by clicking the mouse, and then press the Backspace key (in Mac OS - the Delete key). You can also click on it while holding down Ctrl key(on Mac OS - with the Command key). Cannot delete curve endpoints!

Attention! If you are not satisfied with the result of the correction, click Alt key(on Mac OS - Option key) - button Cancel (Cancel) will change to a button Reset (Reset). Click it - this will allow you to cancel a failed correction without closing the window. Then try again. Moreover, in addition to tone correction windows, the action of this key applies to most dialog windows!

Lighten or darken an image

To lighten or darken an image using the dialog box Curves(Curves) you need to set a point in the middle of a straight line and drag it up (to lighten) or down (to darken) to make the graph convex or concave, respectively. In Fig. Figure 14 shows the original dark image of a lizard and its histogram in the window Curves.

To brighten the image, the straight line is transformed into a convex curve (Fig. 15).

Rice. 15. The final brightened photograph and an example of correction in the Curves window

Increasing image contrast

To increase the contrast of the image, it is necessary to make the correction line look like the letter S. To do this, add at least three points to the line (Fig. 16).

During the correction, you need to move the top point up and the bottom point down (Fig. 17).

Tone interval correction

So far we have looked at examples of images that could be corrected as in the window Levels(Levels), and in the window Curves(Curves), because the correction was carried out on the general tonal range. Because the dialog box Curves allows different sections of the curve to be adjusted independently of each other, this tool provides more options than level correction.

In Fig. 18 shows an image of Nevsky Prospekt. The histogram is shifted slightly to the left, indicating that the image has predominantly dark pixels.

If you try to fix a flaw in the window Levels(Levels), moving the white slider to the bottom of the graph, we will not get the expected result. The image becomes lighter, but the changes are noticeable in the light areas (Fig. 19). And if you try to increase the gamut of the image, then along with the houses the sky and decorations for Victory Day fade.

Rice. 19. Correction in the Levels window on the overall tonal range does not give the desired result

In this case, we need to lighten only the dark images of houses, leaving the light areas of the sky unchanged. To do this in the window Curves(Curves) you should determine the tone interval that needs to be protected from influence and the interval that is subject to correction. If, without closing the dialog box, you move the mouse pointer over the image, then a point appears on a straight line corresponding to the brightness value of the selected pixels.

In our case, the interval for correction is the lower part of the straight line - the diagonal of the two lower squares. It is in this area that the pixel brightnesses of the dark fragments of houses are located. The remaining areas (points on a straight line located in the two upper squares) must be protected from exposure. To do this, add several points in this interval (Fig. 20).

Rice. 20. The lower interval is subject to correction, and the upper one is protected from influence

To lighten image fragments, you need to make part of the correction curve convex (Fig. 21).

HDR Toning

Photoshop CS5 has a new feature - Toning HDR(HDR Toning), which allows you to stylize a single photo as an HDR image. But it can also be used to correct tonal spacing, affecting the shadows and highlights of an image. Moreover, this function allows you to detail parts of the image, which is very convenient at the final stage of correction. So, in our case, the photograph of Nevsky Prospekt became flat during tone correction, in addition to lightening the necessary areas. Increased detail (+105%) and image saturation (+30%) in the window HDR Toning made the photo much more attractive (Fig. 22).

This window opens with the command Image-> (Image) -> Correction(Adjustments) -> HDR Toning(HDR Toning).

Exposure adjustment

Dialog box Exposition(Exposure) is designed to adjust the tone of HDR images, but also supports 8-bit images. To call it use the command Image(Image) -> Correction(Adjustments) -> Exposition(Exposure).

Tone adjustments can be made by changing three parameters:

• Exposition(Exposure) - designed to correct the light portion of the tonal scale with minimal impact on the darkest fragments;
• Shift(Offset) - replaces shadows and midtones with minimal impact on highlights;
• Gamma correction(Gamma Correction) - changes the gamma of the image.

In Fig. Figure 23 shows a dark photograph of a squirrel, which is confirmed by the histogram of the image.

In this case, to correct the image it is necessary to influence separate areas photographs in different ways: the snow needs to be lightened much less than the squirrel, and for its face the contrast needs to be increased. These tasks can be successfully completed in the dialog box Exposition(Exposure) - fig. 24.

Quick tone adjustment

In addition to the functions discussed above, Photoshop has tools that allow you to instantly correct the tonality of an image. They don't require extensive adjustments, and some of them allow you to adjust your image with just one click!

Correcting pitch using the dialog box Brightness/Contrast

This easy-to-use dialog box
has only two sliders - Brightness(Brightness) and Contrast(Contrast). The correction comes down to changing their position (Fig. 25).

Rice. 25. Brightness/Contrast Dialog Box

Auto-correction of image levels

Automatic tone and color correction is carried out in the dialog box Levels(Levels) or Curves(Curves) by clicking the button Auto(Auto), and its setting is in a dialog box that opens by clicking on the button Options(Options), - see fig. 6 and 13.

To automatically correct the image by tones, the commands are also used Autoton(Auto Tone) and Auto contrast(Auto Contrast) from the menu Image(Image).

With half of the given command examples Autoton And Auto contrast coped successfully, except for photographs of Eltz Castle, Nevsky Prospekt and a squirrel. Before making manual corrections, try automatic correction levels, because if the result is successful, automatic commands will significantly save you time.

Correcting tone using pipettes

Dialog boxes Levels(Levels), Curves(Curves) and Exposition(Exposure) contain three pipettes: black, gray and white - see fig. 6, 13 and 24.

For color images, all three eyedroppers are in the windows Levels And Curves are used to remove color shift, that is, for color correction. And for tone correction you can use black and white eyedroppers, but only for halftone images. Color images are tonal-corrected using all three eyedroppers in the window Exposition.

The principle of working with eyedroppers is as follows: you need to select the desired eyedropper, and then simply click it on the area of ​​​​the image that should be black, gray or white. Please note that quick adjustments using pipettes are not always feasible. The image should contain the intended black, neutral or white areas. For example, a photo of a sunset will most likely not contain neutrals or whites.

Using adjustment layers

All the tools discussed in this article make irreversible changes to the image layer. To avoid losing the original photo, it is better to experiment with duplicate images or layers. You can also save correction results as snapshots in the palette Story(History). But remember that after closing a document with multiple snapshots, only the current snapshot will be saved. Therefore, photographs should be used only to select the most successful correction result.

Another way to perform image correction without losing original photo— create an adjustment layer. Adjustment layers allow you to go back and make subsequent tonal changes without deleting the data from the image layer or making permanent changes.

To create an adjustment layer, use the button with the circle icon in the palette Layers(Layers). Clicking on it brings up a pop-up menu in which you can select the name of the tonal correction tool: Brightness/Contrast, Levels, Curves or Exposition. After settings in the palette Layers(Layers) an adjustment layer appears, which, like a regular layer, can be turned off or deleted at any time. Therefore, in this case there will be no irreversible changes in the image. Moreover, you can create several adjustment layers to choose the most successful correction result. So, in Fig. 26 palette Layers(Layers) contains three adjustment layers with different tonal correction tools. The adjustment is made using curves.

Image histogram, dialog boxes Levels And Curves, in addition to tone correction, are used to diagnose and remove color shifts, that is, for color correction, which will be discussed in one of the upcoming issues of our magazine. 

One of the most important characteristics of a TV when choosing is the contrast value of the image on the TV screen. If you choose a TV based on picture quality, be sure to pay attention to the contrast value of different models.

By definition contrast equals brightness ratio in the most bright point screen to the brightness of the point where the darkest image is. In other words, we divide the white level by the black level and get contrast. Only the values ​​of these levels can only be obtained through a special test of the TV using specialized instruments. That's why to a simple user you have to trust either the manufacturers or various reviews on sites where TVs are tested. Who to trust more and how to check the contrast, and we’ll talk further.

We said that the contrast is one of the most important characteristics TV. Therefore, manufacturers try to maximize this value to improve sales. The manufacturer can measure the brightness of a pixel in the laboratory when applying a signal that is never used in real conditions. Then measure the brightness of this pixel in the absence of a signal, which is impossible during normal viewing. After this, the contrast value is calculated. And the values ​​measured under such conditions are included in the product passport. Because of this, we see today that the contrast values ​​of many TVs are simply off the charts. All this is possible because there are no mandatory rules in the world for measuring the contrast of displays.

Separate static (natural) and dynamic contrast. Natural contrast depends only on the capabilities of the display, while dynamic contrast is obtained as a result of the use of additional technologies.

Static contrast is measured by the brightness of points in one scene (brightest and darkest). When measuring dynamic contrast, technologies are used to increase it. The TV itself, when playing a video, adjusts the contrast depending on the scene, which at the moment shown on the screen. That is, the backlight in the LCD matrix is ​​adjusted. When showing a bright scene, the luminous flux from the backlight increases. And when the scene changes to dark (night, dark room, etc.), then the backlight begins to reduce its luminous flux. It turns out that on bright scenes, due to the increase in light from the backlight, the black level is poor, and on dark scenes, the black level is good, but the luminous flux will decrease. This is hard for us to notice because in bright scenes even backlit black appears completely black. And in dark scenes, the brightness of light objects seems sufficient. This is a feature of human vision.

This backlight control scheme increases contrast, although not as much as the manufacturers claim. And indeed, many TVs with dynamic contrast have superior image quality to devices that do not have such an adjustment scheme.

But still, displays with high natural contrast will be valued higher. This can be demonstrated by displaying a picture of white text on a black background. On a screen with high static contrast, the text will indeed be white and the background will be black. But a display with high dynamic contrast, if it shows a black background, then the letters will already be gray. Therefore, when playing regular video on a screen with increased natural contrast, the picture will be as close as possible to the real image. For example, there will be bright street lights against the evening sky. And against the backdrop of daylight bright sky a black car will really be black. This is the image we see in cinemas.

As real as possible, in terms of contrast, the image was on the screens CRT TVs. But with the advent of the HDTV era, these television receivers gave up their place in the market to other devices. Today, high natural contrast values ​​are achieved using LCOS home projectors. The first place among these devices is occupied by JVC devices with their version of D-ILA. Next we can mention Sony with SXRD technology. In third place you can already put plasma TVs.

LCD TV manufacturers have introduced several technologies in recent years to achieve the level of contrast that is possible in other models. Best results The use of LED backlighting with local dimming improves contrast. In this case, it is impossible to adjust the backlight of each pixel and each LED is not controlled individually, but the result is still good. But manufacturers have abandoned the most effective type of backlight, when LEDs are located over the entire screen area. Such production turned out to be expensive. Today, so-called side lighting is mainly used. Here the LEDs are located at the top and bottom. Local dimming schemes have also been developed for side lighting. TVs with such backlighting show fairly good results in terms of contrast.

When choosing a TV in a store It is difficult to evaluate the quality of the display contrast. External bright lighting interferes; screens can have different coatings: anti-glare or glossy. The passport does not always contain the true contrast value, because manufacturers measure it in laboratories and by applying special signals to the screen. Even after reading several reviews on the Internet, it is not always clear what the real contrast value is. After all, everyone measures it in their own way.

Eat several contrast measurement techniques. First, a black field is supplied to the input and the brightness is measured, and then a white field is applied and the brightness is measured. It produces good contrast, but real viewing There will never be a completely white or completely black picture. At the same time, when displaying a regular video signal on the TV, video processing is turned on, which also makes its own changes. More accurate readings are given by the ANSI test, when a checkerboard field with white and black fields is shown on the screen. This is more consistent with the normal image. But in this case, the white fields will affect the measurement of the brightness value of the black fields. So one correct method There is no contrast measurement.

So the recommendations for choosing a TV with good contrast remain the same. If you'll mostly be watching movies in a darkened room, then plasma is your best bet. In a well-lit room, an LCD TV with LED backlight due to its high brightness. Between these models you can put an LCD TV if there is a reserve in light output. And you need to remember the main thing, any TV needs correct setting. Adjust the brightness and contrast of the device correctly to obtain the highest quality image.

Additionally:

Screen brightness

The brighter the screen, the less you have to strain your eyes to comfortably see the image. This is especially true if you have to watch TV in bright daylight. When viewing 3D images, the screen brightness is even more important role due to the need to use glasses. Any 3D glasses (passive polarized or with active shutters) darken the images that are perceived by the eyes.

Minimum sufficient for comfortable viewing TV in most cases has a brightness of 450 cd/m2. As the screen diagonal increases, the brightness indicator in the TV passport also increases. If for 19-inch LCD TVs the brightness can be 250 cd/m2, then for 36-inch TVs it is no less than 500 cd/m2. For rooms with variable lighting, TVs often use a built-in ambient light sensor, which itself regulates the brightness of the screen backlight.

Brightness plasma TVs can range from 1000 to 2000 cd/m2, which is significantly higher than other types of TVs. Some plasma TV manufacturers do not even consider it necessary to indicate this characteristic. It is only necessary to note that an excessive increase in brightness further increases the already considerable energy consumption of plasma TVs.

To check the brightness, when broadcasting a story with normal lighting (for example, news), turn the brightness value on the TV first to minimum and then to maximum. At minimum, the picture should darken noticeably to the eye, and at maximum, the image should also noticeably lighten. At the same time, you can clearly see how much brightness this TV has.

Image Contrast

The contrast value shows how many times one area of ​​the image is brighter than another. In the TV passport, the contrast is usually written in the form, for example, 800:1, which shows the ratio of the white level on the screen to the black level. Until now, LCD TVs lag behind in terms of contrast plasma panels. Among TVs with small screen sizes, the minimum sufficient contrast ratio is 600:1. LED TVs have a higher screen contrast (up to 1200:1).

But when you come to the store and look at the technical specifications, you can see the stated contrast ratio of 6000:1, 7000:1 and even 10000:1. You shouldn't be surprised by such high numbers. This is the so-called “dynamic contrast”, which is provided special technology. When displaying a brighter image, the brightness of the matrix backlight increases, and in dark scenes, the backlight brightness decreases. Indeed, in scenes with high brightness, dark areas of the image are not so important, since our eye perceives them as very dark, so increasing the brightness of the backlight does not distort big picture. It’s the same in dark scenes - our eyes perceive light areas differently, which makes it possible to reduce the brightness of the backlight.

To measure dynamic contrast, take the white level at the brightest backlight and the black level at the lowest backlight. This is how we get such large values. But at any given time, the screen contrast does not exceed the static contrast value. Dynamic contrast only works when the picture is changing. For large LCD TVs, where all the disadvantages of low contrast are especially visible, contrast values ​​range from 1000:1 to 1600:1.

The static contrast value of plasma TVs can reach 30,000:1 or even more, and the dynamic contrast has already exceeded 1,000,000:1. This is due to the ability of plasma to completely extinguish its pixel to a perfect black color.

Lamp or LED resource

This parameter shows how long the backlight lamp in an LCD TV or light-emitting diodes in an LED can work while maintaining its performance characteristics. Today, the lamp life is approximately 60,000 hours, and LEDs - up to 100,000 hours. In translation, this will turn out to be about 7 years of continuous operation, so when choosing a TV, you can not pay attention to this indicator.