|Virtual pixel is image improvement?|
Recently firms that sell large electronic LED screens, especially in southeastern Asia and in Russia, to advance their goods on the highly competitive market, started to declare that their screens use the technology of “virtual pixel”. They claim that “virtual pixel” doubles the actual resolution of screen, i.e. LED screen with the “usual” resolution 320x240 pixels in reality is converted into the LED screen with the “virtual” resolution of 640x480.
All world leaders in LED screen technology – Daktronics, Optotech, Barco, Lighthouse etc. - are concerned with creating new models with larger resolution in real physical pixels. In all their newest models these companies stopped using “virtual” pixels technology, although in previous models (3-4 years ago) “virtual pixel” was present. Why do leading developers and producers of LED screens reject “virtual pixel” technology?
Let us try to analyze this situation and to determine, where and when it makes sense to use technology of “virtual pixel” and is it true that “virtual” pixel' doubles the real resolution. Unfortunately, suppliers do not always provide real information on the use and functions of “virtual pixel”. They assure buyers that LED screen models with “virtual pixel” are not worse than similar models with real pixel.
In the majority of cases “virtual pixel” proves to be just a smart marketing trick. There is nothing new in this technology and small advantages are balanced by deficiencies which suppliers naturally prefer to hush up. Let us figure this out.
“Virtual pixel” of video screen “Virtual pixel” of video screen
Consider a video screen with pixels that contain of light sources positioned as a square (irrespective of the type: LEDs, incandescent lamps etc.). Each light source radiates light of a certain wavelength (or narrow range) or in laymen terms produces colored light. A picture Fig. 1 is an example of a typical pixel.
When image is displayed on a video screen in a “normal” mode (Fig. 2) each pixel of the original image corresponds to a certain pixel on a screen. For example, if a pixel in the top left corner of an initial image had R, G, B color, the pixel in the top left corner of the video screen will look the same. It is taken for granted that color elements in a pixel are well balanced in brightness and colors and no additional correction is needed.
In a “virtual pixel” mode each pixel of the initial image corresponds not to a screen pixel but to a light source, i.e. part of the pixel. The initial image has a doubled resolution so that each pixel of an image corresponds to each light source on a video screen. For example, four pixels of the top left corner of the initial image (Fig. 3) shall be reflected due to “virtual” transformation in a one screen pixel in a way shown (Fig. 4).
Thus, in a “virtual pixel” mode one screen pixel contains information on four pixels of the initial image. The image projected on a screen has doubled resolution in each dimension compared to a “physical” resolution of a video screen. This usually leads people to conclude that screen resolution also doubles. Which is not exactly true. In fact, one screen pixel cannot hold and display all information from the initial four pixels. Part of the information gets lost. The result may be the following.
Let’s say that the initial image (with resolution twice higher than “physical” screen pixel resolution) looks as a horizontal green line (one pixel thick) on a black background. If the line appears on an even row of pixels, the video screen will display a corresponding green line. But if the line shifts to an odd row of pixels, it will simply disappear: the video screen will remain black. In other words, smaller details and sharp color borders shall be displayed with distortions (artifacts) which are not evident in an initial image.
Are there any advantages of “virtual pixel” technology? Yes. In some cases the overall displayed quality maybe improved though image details will be distorted. This technology works better with smooth color gradients or on patchy images when color distortions are not evident. In a way, we can talk about doubling screen resolution only for black color because all light elements with black color look the same, i.e. they remain unlighted.
The above description relates to the simplified implementation of “virtual pixel” technology. This approach may be modified. Usually modification is made by displaying some averaged value. Averaging can be both spatial and temporal. With simple spatial averaging a certain algorithm will create a mean average of the initial four image pixels and transfer this information to the screen.
With simple temporal averaging one of the four pixels of the initial image will be displayed on a screen at higher frequency (double or quadrupled). Spatial and temporal approaches may be combined. However, until now there is no clear answer to a question of “optical equivalent”: how does the human eye perceive screen image based on above two approaches.
In practice screens with “virtual pixel” technology usually operate on “standard” mode, i.e. the virtual pixel option is switched off if the control system allows to do it. This is done to avoid color and image distortions that were not corrected during image adaptation by designers. When “virtual pixel” cannot be switched off, it is possible to minimize image distortions by introducing various filters (e.g. “blur”) that smooth out the picture, blur the details and remove distortions. But this may negatively reflect on the overall image clarity.
Conceptually, “virtual pixel” is an attempt to smooth out digital image (interpolation algorithms) as displayed on a screen. However, there are no universal interpolation algorithms: different types of images require different algorithms. As a result, application of a “virtual pixel” mode becomes inexpedient.
An alternative approach may be the following: initial images in double resolution must be adjusted to physical screen resolution by software that uses an interpolation algorithm specially selected for a given type of images. Usually, all standard designer tools have a large selection of such algorithms. This approach allows to get predictable results: that is, the screen will display exactly the same image as can be seen on a PC monitor without any additional artifacts. All of the above relates only to screens with the uniform distribution of pixels and LEDs around the screen surface.
In case LEDs are grouped together as clusters, other “virtual” algorithms can be applied but the initial image may require an even higher (e.g. tripled) resolution compared to the screen pixel resolution. Again, this will not mean that the screen resolution is tripled by “virtual” technology. Some special algorithms may be developed for other cases when LEDs are located not as a rectangular but as a triangle: RGB.
“Virtual pixel” technology is easy to misuse which results in the loss of even more information. Actually, “virtual pixel” technology on large screens appeared a long time ago. It was first adapted on lamp screens (made of bulbs) where it was called bulb-mode. Lamp screens had significantly lower resolution and significantly bigger pixel size. Naturally, developers tried to smooth the image edges to improve image quality. Later, same approach was expanded on LED screens.