Understand the components, a two piece puzzle

Understand the components, a two piece puzzle

The manufacturers just need to associate two components to build their screens, a bezel and the most important thing a panel.



The panel is a 0.4” to 0.8” wide rectangle with on one side the screening surface and on the other side an electronic card. Between the panel and the bezel, two to twelve neon tubes in charge of the backlighting, liquid crystals through which the light travel and colour filters.

The process is as follows: neon tubes emits white light. This light try to goes through the liquid crystals. The liquid crystal cells, three by pixel, are controlled by a transistor, which is controlled by the motherboard located at the back of the panel. To each electrical tension emit by the transistor correspond a crystal position.



If they are vertically placed, the neon tube light is stuck, and the result is a black dot on the screen.



If the liquid crystals are horizontal, light goes through and creates a white dot.

Each pixel is made of three liquid crystal cells next to each others. In front of each cell is placed a red, blue or green filter. When all crystals are horizontally placed, the light goes through. The first cell emit a red light, the second one a green light and the last one a blue light. The eye gather all three lights and see one: white. This is the additive lightening process. Then, in order to recreate all available colours, it is necessary to apply various electrical tensions on the transistors. If the transistors command 256 red, 256 blue and 256 green, the maximum amount of available colours will be: 256 x 256 x 256 = 16.7 million of colours.

At the end of the day, the manufacturer options are rather limited. They need to select the right panel, maybe customize electronic components, choose the default settings and integrates all elements in a bezel.

The panel choice is obviously essential. From this choice depends, the number of colours on screen, the vision angles, the maximal amount of light available and the pixel reactivity. The manufacturer have to target the fastest liquid crystals. If liquid crystals are too slow, the picture shown won’t be the picture transmit by the graphic card and expected by the spectator. To make their choice, manufacturers can choose between three different panel technologies and dozens of available model.

Chronological history of the past two years

Chronological history of the past two years

As the last LCD comparative test on Hardware.fr is two years old, a summary of the last episodes is necessary.

During 2001, 15” LCD screens tested had response time between 25 to 60 ms. Surprisingly, the screen who won everyone’s approval was a 40 ms, the Belinea 10 15 35. With unpretentious performances on the paper, this screen was excellent during tests (at the time). It was by far the most reactive screen, and the one with the least afterglow. Performances were still far from CRT displays, but this screen was the best.
We already noticed that the response time communicated by the manufacturer wasn’t accurately representative of the exact pixel reactivity.
Since that time, we have worked on this question and we will explain further in this article why.

Already in 2001, the tested screens used the three panel technologies: TN, IPS, VA.

IPS, VA, TN

Chronological history of the past two years

IPS, In Plane Switching

Since 2001, the IPS technology has little or none progressed. Even if these monitors characteristics are increasingly attractive (some of the model have a 16 ms response time), in practice they are the least reactive and the worst for gaming. The lack of sharpness is a lot more present than with the VA or TN screens. Unless maybe if you have the last top-of-the-range products (like the NEC 2080UX+, we just received one but didn’t have the chance to test it yet tested), these screens are not well adapted for a play use. The TN and VA screens are less expensive and have better results for games and movies.
Colors rendering have a little bit improved in the last two years, but slower than TN and VA screens. However, they have an enjoyable color rendering, wide vision angles, and screens using this technology have remarkable ergonomics.

Still, we feel that during the last two years, the technology has stagnated. However, after the IPS, the Super-IPS, the Advanced-IPS, a new range of panel is arriving: the Super Advanced-IPS. Is it going to be the right one? The answer is on its way.

VA, Vertical Alignment

Fujitsu Siemens initiated the VA technology under the MVA designation. Samsung is now using it under the PVA designation, Sharp under ASV and CMO under Super MVA. In 2001, the response time was 25 ms and now the response time is … 25 ms. The only exception is the Samsung SyncMaster 193P. This screen is the first 20 ms. Reactivity is indeed a little bit better than with the 25 ms VA, but some improvements are still necessary. TN screens remains distinctly more reactive. Nevertheless VA screens outweigh TN screens on most of the remaining area. Black color rendering is often superior; vision angles are as wide as TN angles are narrow and colors are more accurate. These screens display 16.7 millions colors and the TN doesn´t.

TN, Twisted Nematic

Mostly used in middle-ranged and entry-level screens, the TN technology made its way to the top with several technological breakthroughs. They have been most frequent than on IPS or VA screens.

Without reaching the competition level, TN vision angles became wider. Other improvement, the last panel range (12 ms) is generating a credible black colour. But most of all, response time on TN screens is noticeably faster than IPS and VA screens.

However we are still in a similar position as we were in 2001. The fastest panel in all our tests is a 20 ms even if 12 ms panels exist. This panel is even beginning to date. Build by a Korean manufacturer, BOE-Hydis, this panel has been used in the Iiyama AS4314-UTG, the Hercules ProphetView 920 Pro, the Acer AL1731...

Understanding response time

Understanding response time

Reading screen characteristics doesn’t take long, but as you will see below they shouldn’t be fully trusted.



Response time is the first LCD screen’s characteristic. Expressed in milliseconds, this measure represent the amount of time necessary to a pixel to turn from white to black and then going back to white. Tested screens two years ago, had a 60 to 25 milliseconds (ms) response time for the most efficient screens. Nowadays, screens performances are reaching 12 ms. The 8 ms screen displays are already expected in September.

In theory the faster this response times is, the smoother and clearer the movements are on your screen.

This theory would be verified if the response time was constant and valid for all colours change. But a 25 ms white/black/white screen might need 120 ms to process light grey/ dark grey/ light grey.

Explanations

The response time as all LCD screen characteristics are defined by the ISO norm 13406-2. And as we will see here and later, this norm deserve quickly a deep review!

To turn a pixel from white to black, the electrical tension goes from 0 the maximum limit. Highly stimulated, the liquid crystals quickly take their new position. The amount of time required to process white to black is called rise time.

To get back to white, the electrical tension is cut. Crystals get back to their initial position easily and quickly. This process is called fall time.

The response time is the total of rising + fall time.

First trouble induce by the ISO norm, the rise and fall time are measured on 80 % of the whole signal. The extreme 10% are cut from this measure. This measure is at some point justified but simplifies and flatters the result. So if one system takes more time to start or stabilize the pixel, it won’t be noticed in the results.



In the previous graphic, the rising time would be 28.5 – 12 = 16.5 ms. However by looking at the signal, it seem logical to consider a measure of 40 ms closer to the reality.

The following issue is the response time between colours. To turn a pixel from light grey to dark grey, the electrical tension start from a low electrical tension to a strong electrical tension. The electrical stimulation is noticeable but no as strong as when it need to rise from 0 to the maximum limit. So crystals are slower to move and in practice the rising time is longer.

To go back to light grey, the process is reversed. There again, crystals will take longer to get back to their positions if the electrical tension is completely cut off. Fall time is also longer

For now the ISO norm doesn’t compel the manufacturer to consider intermediates response time. So the result is that manufacturers communicate on the same response time for different panel technologies and different crystals movement speed.

So we are now in a situation where LCD screens equipped with a TN AU Optronic panel 16 ms are a little bit quicker than the 16 ms LG-Philips and the Samsung 12 ms. These panels are unmistakably quicker than the 16 ms IPS panel and slower than the TN Hydis 20 ms. How is it possible not to be hopelessly confused?

This norm ISO-13406-2 should really be reviewed soon! (one time)

Understanding the brightness and the contrast ratio

Understanding the brightness and the contrast ratio

The contrast ratio is measured by opposing the lighter and the darker pixel produced. In order to get a good contrast ratio result, white has to be as white as possible and black as black as possible.

So with a measuring tool, we have to measure the white level: let’s say that this measure is 250 cd/m². Then we measure the black and this time it is 0,5 cd/m².

The formula to establish the contrast ratio is:

Black / White = contrast ratio.

Applied to our case, the result is:

250 / 0.5 = 500 : 1 .

These results are representative of flat screens.

Now it is tempting for the manufacturer to extricate its screen by submitting a higher contrast ratio.

To succeed in this operation, there are two possibilities:

The manufacturer can improve the black colour, but this operation is very difficult. They need to improve the panel filter or work on the liquid crystal movements to facilitate the backlight blocking by placing them even more on a vertical position.
The other possibility is to rise up the white colour of the lightning by adding neon tubes or replacing them by brighter neon tubes.

Under those circumstances it is possible to reach a 500 cd/m² white.

If this “improved white” is used in our formula, the result will be:

500 / 0.5 = 1000 : 1

With such a contrast ratio, this screen should be perfect from a graphic designer point of view. But in fact this screen will only be annoyingly brighter than the other screens.

Actually, the recommended brightness should not exceed 110 cd/m². CRT screens brightness is 90 cd/m². All 250 cd/m² screens (and most of the LCD screen are) are too bright. This brightness is annoying, it is a strain on the eyes and you will almost certainly reduce it manually.
In a nutshell, contrast ratios are currently measured in a non-representative way. To be accurate, this measure should be based on the same white level for all screens, for example 110 cd/m². Let’s hope that the upcoming ISO norm incorporate this kind of requirement. Henceforth, don’t pay a lot of attention to this characteristic.

This norm ISO-13406-2 should really be reviewed soon! (two time)

Understanding vision angles

Understanding vision angles

Regarding the technology, vision angle vary from 140° for TN, to 176° for VA and to 160° for IPS°. This is probably the only characteristic to be trusted, or at least partially. Indeed, TN screens have smaller vision angle than IPS screen, themselves slightly less open than VA screens. But it doesn’t mean that you can take this characteristic at face value.




The vision angle can be established in two different ways. Either the manufacture choose the angle which provides contrast ratio of: 10 : 1, or 5 : 1 for the user. In the first case: 10 : 1 the picture is terrible, and in the other case the picture is even worse.

So you will never get a nice picture from any if these two maximum recommended angles.

However the future range of OLED screen will bring some serious improvements. The tested prototypes provide total vision angles. Even in a vertical position of the picture, colours remain intense.

Understanding 16.7 or 16.2 millions colors?

Understanding 16.7 or 16.2 millions colors

All screens pretend to display 16 millions of colours. Still, after the comma, the figure changes: 2 or sometime it is better: 7.

On the paper, screens are able to display 256 shades of red, 256 shades of blue, and 256 shades of green. The combination of all three provides 16.7 millions of different shade. This is what VA or IPS screen does.

TN screens are more economical. In fact, they only display 64 red, 64 blue and 64 green. The maximum amount of colours is 262 144. In order to reach 16 millions of colours, the manufacturers are using the dithering process. The process is to display two close colours so quickly that only one is seen.

So instead of recognising all 0,1,2,3,4 shades and so on until 255, TN screens only knows the shades 0,4 ,8 ,12 ,16 ,20 … until 252.

Displaying shade 2

Manufacturers have several possibilities:

If the manufacturer wants to work on one pixel only: at T0 the pixel is white, at T1 the pixel has the shade 4, then the pixel comes back to shade 0 at T2, then shade 4 at T3, etc. The eye mixes the two shades and only sees the colour 2.

T0	T1	T2	T3

The necessity of using two periods to get the right shade is the inconvenience of this process. It is also possible to use the four next pixels (in a square shape). Two of them will have the shade 0 and the two others the shade 4. So from your point of view the colour will be 2.



Displaying shade 1

The first approach is to display shade 0 for T0, T1, T2 and shade 4 for T3. So you will see (0+0+0+4) / 4= shade 1. But three periods will be necessary to get shade 1. There again this process is too long.

T0	T1	T2	T3

The other possibility is using one pixel with the shade 4 and three pixels with the shade 0.



This method is very efficient. But it still has two main issues: the screen flickering is perceptible and the shade 253, 254 and 255 can’t be displayed. With the dithering, all colours are displayed from 0 to 252, so the result is 253 red x 253 blue x 253 green = 16.2 millions of colours.

Understanding dead pixels

Understanding dead pixels

Alarming news, manufacturers are allowed to sell us defective screens. And most of the times you can’t even complain about it. This might sound redundant, but the origin comes from the ISO norm 13406-2.

The ISO norm 13406-2 defines four different screen ranges. The highest and strictest is the class 1. This class forbids any malfunctions. This category is the best. The worse, “Class 4”, authorizes up to 549 malfunctions, and 1344 defective pixels and sub-pixels for a 20” screen. Fortunately, no manufactures refer to this class.

Almost all manufacturers base their products guaranties on Class 2.

On the top of the class category, three kind of malfunctioning have to be identified: permanent white pixels, black pixels, and color pixels (red, green, blue).

Regarding the ISO norm, here is the calculation of malfunctioning acceptance by screen size:



So in practice, a manufacturer doesn’t have to replace your 17” screen if you have two white pixels, 2 black pixels and 5 color pixels dead.

Nevertheless, some manufacturers are flexible on that area. They even change your screen if the dead pixel is located to a very irritating place (the middle of the screen by example). Ask the reseller about the possible option for dead pixels. Ideally, ask the reseller to switch on the screen in front of you.

Last possibility; buy your screen from a money-back guarantee shop. Otherwise, with on line retailer you have by law up to seven days return you product.

So one more time: This norm ISO-13406-2 should really be reviewed soon! (three time)

Upcoming technological breakthrough and new technologies

Upcoming technological breakthrough and new technologies

At the end of July is expected the first 17” flat screens with a 10 ms panel. And in September, the 8 ms new displays should be released. 19” screens are always a little bit late. The first 12 ms is currently released and is going to be really widespread in September.

This response time progress goes along with improved new filters. These filters are better at capturing diffused light (colours are brighter and black is deeper), and at rising up vision angles (even if they remain too small for entry level products).

Prices aren’t going in the right direction. After a spectacular price fall during 2003, prices increased of 30 % from November to May 2004. There is currently a stabilization phase, but manufacturers already announced rising prices in August. These rising prices are due to the growing success of LCD panels. LCD panels are indeed used for flat screens but also laptops and LCD TV, two huge and very successful markets.

However LCD screens aren’t going to be the next generation screens. In fact, the LCD successor is already on its way and NEC will release a screen this summer. This 20” screen is called Lumileds and costs 7500 €. The innovation is the substitution of neon tubes by a 48 matrix diodes. The result is a deeper black and an improved harmonized colour. Areas usually close to neon tubes are as bright as the distant ones. This expensive technology is targeting the most demanding graphic designers.

Later on are expected the OLED screens. Sony, Kodak, Epson and Toshiba are the most passionate about this technology. Diodes aren’t just replacing neon tubes but also are the liquid crystal cells. So at the end of the day there are not 48 diodes, but 3 (red, green, blue) x 1024 x 768 = 2 5359 296 diodes for a 15” screen! For 17 and 19 screens the total amount of diodes rises up from 4 to 6 millions for 20” screens.



Later on are expected the OLED screens. Sony, Kodak, Epson and Toshiba are the most passionate about this technology. Diodes aren’t just replacing neon tubes but also are the liquid crystal cells. So at the end of the day there are not 48 diodes, but 3 (red, green, blue) x 1024 x 768 = 2 5359 296 diodes for a 15” screen! For 17 and 19 screens the total amount of diodes rises up from 4 to 6 millions for 20” screens.

Regarding the biggest sizes, LCD screens are just behind Plasma screens for TVs. But plasma screens loose their gas quickly when the screen is permanently switched on (their brightness is reduced by half in two years), for this reason LCD screens last longer. The other side of the coin is a higher price. For your information the biggest screen yet are made by Samsung. Their sizes are respectively 57” for the LCD screen and 80” for the Plasma screen.