How to choose a projector: rating-anti-rating of the main parameters

Today we will try to form some rating of parameters and their deviations, which are obtained when testing projectors. The main task of this article is to show what some deviations from the reference, how fatal each of them can be and at what values the image will be really bad, and at what values it will be good.

In parallel with the rating, let’s talk about image theory. This is for those who really want to understand cause-and-effect relationships and learn to clearly interpret test graphs. In this case we guarantee to choose the optimal projector exactly “in ten”.

Color gamut goes out of the competition

Summary: having a wide color gamut is good, but not critical if the gamut falls short of the exact sRGB triangle. And if the coverage is much larger than sRGB, it can be a problem at all.

First of all we would like to single out one parameter which will go into the “not contest” – it is the color gamma. Such a special position is deserved by the fact that this parameter can practically not be improved either by hardware or software manipulations, including calibration.

Usually the color gamut is limited by the hardware capabilities of the projector. That’s why manufacturers always use this parameter to the maximum. Although there are exceptions.

In our tests, the color gamut, as well as several other characteristics, can be found in the CIE charts. These are usually presented at the beginning of the test.

What is the color gamut? It’s simple – in terms of dry definitions, color gamut is responsible for the number of colors a projector can project. In reality, this means how saturated and vivid the colors will be.

The color gamut is represented graphically as a triangle. The vertices of this triangle are the primary RGB colors – red, green, blue. The colors inside the triangle are reproduced by the device, while those outside are not. Test pattern results are always shown as a triangle with white edges.

There is a reasonable question: what can be compared with? How it should be?

We have to compare with different color spaces. Like 220 in the socket, on the internet, and in computers in general, the standard is the sRGB color space. However, just like 380 for electric stoves, we also have exceptions. For example, the AdobeRGB space for photographers and printers. Our reference triangle shows the sRGB space.

Unfortunately I can’t show you the impact of a reduced color gamut in a specific photo. Most of the monitors you’re reading this on don’t display 100 percent sRGB. So if I artificially narrow down the range of a photo, you won’t necessarily see it. The monitor has not displayed such extreme colors before.

But believe me, in real life it’s not so dramatic. Even the fact that the projector with color gamut in the above chart can not reproduce a good part of green tones, does not spoil the picture at all. Movies or images very rarely have the most saturated colors.

There are situations where the color gamut of a projector is greater than sRGB. This does not mean that you should rejoice. What leads to such “overkill” with colors, we have already briefly described in one of the reviews: “all the palefaces will be red, like Indians – you could light up, and on the foliage in the neighboring grove as if all the green of the country had been poured”.

Is there a cure for this problem? There is, but not a panacea. This is a color management system or CMS. In simple layman’s terms, it is a system that ensures that graphics are displayed uniformly on various output devices, including projectors.

The system is based on color profiles, which specify the capabilities of the output device. And based on the color profile, it adjusts the values before the signal from the video card. For example, if our projector can output redder than it should in sRGB if the image has that color profile , then the “redness” of the output will be reduced.

However, CMS has two problems:

1. Get the device profile. Sometimes they come with the hardware, but that’s not common and in the case of very expensive models. If you don’t have such a profile, you will need a colorimeter or spectrophotometer, which can be bought starting at 7000 Dollars.
2. CMS only exists in operating systems. If you connect the projector to a DVD/Blue-ray player, you cannot use it. Moreover, even if you use Windows/MacOS, you need to use programs supporting CMS, and there are not many of them. For example, even the Windows desktop doesn’t allow you to set the image using CMS..

First place. Channel gamma curves – the old hatter’s secret tricks.

Summary:

1. Gamma variations greater than +/- 0.2 can be fatal to the image.
2. The worst case is when the gamma value in the shadows is greater than 2.2, and in light areas less than 2.2. It’s almost guaranteed to fail shadows and knock details out of the lighter areas of the image.
3. Worst is when the channel gamma curves don’t match each other much. It can lead to serious color distortions.
4. The distortions of the channel-by-channel gamma curves by degree of negative influence from the worst to the least important can be ranked as follows: green, red, blue. And in terms of importance, it’s about 60%, 30% and 10%, respectively.
5. S-shaped gamma can be effective in some situations, but is definitely not suitable for permanent work.

Gamma curves are probably one of the most important tools for analyzing image quality. The gamma curve affects the accuracy of the display of light variations, the overall contrast of the image and determines how distinguishable details in shadows or highlights are in the image.

The gamma reference value is 2.2. In our graphs the gamma value is presented in the following graphs. I would like to point out that this is not a gamma curve! This is the gamma value in the different areas of the histogram shadows, midtones, etc. .. .

The y-axis shows the gamma value. On the X axis is the lightness level of the image: in other words, the dark shadows on the left and the lightest parts of the image on the right. If the value is less than 2.2 it means that the area of the image with a given luminosity will be lighter than it should be. If the value is greater than 2.2, then the image will be darker.

Thus, the worst case, when the shadows have a gamma value greater than 2.2, and the light areas of the image are less than 2.2. In this case it is most probable that the details in the shadows and highlights will be indistinguishable.

Picture A shows a normal image. In photo B, the gamma has been changed so that it is greater than 2 in the shadows.2, and in the light areas of the image it is more than 2.2. the fly turned into a black spot, and the florets into a solid bride’s dress, which was fatal for the image. Nevertheless, agree, the image has become more contrast.

Picture B shows the opposite situation. Yes, the image has less contrast, but all the details, all the tones are preserved. Well, how better to see a faded picture with all the details or a contrast, but one spot?

The fatal deviation for the gamma curve value at which there can be fatal consequences for the image is approximately 0.3 and + 0.2. Because the plus deviations are more noticeable to the eye.

However, the gamma curve itself is “the average hospital temperature”. Much more interesting are gamma curves per channel, which as an average form the final curve.

As you know, an image consists of three base colors. But these colors are not equal in importance to the image.

This is what the graph of the channel-by-channel gamma curves looks like. The black curve is a reference which corresponds to gamma 2.2. the remaining lines are the curves of the channels. The curves are colored according to the color of the channel.

Just like in the previous graph, if the curve goes above the reference, it means that the image will be lighter and vice versa.

It is important to understand that the channel gamma curves affect not only the lightness of the image, but also the color. The more will be the stratification, the difference between the channels curves, the more serious color distortions will be observed in the image.

In practice, color distortions are much less striking than light distortions. Notice the following photo. The original image upper left corner on the right is decomposed into variations in lightness and variations in chromaticity.

You have to agree that in the case of lightness, the image contains a lot more information, details of the original image. The color variations in B/W also contain a lot of information, but it’s B/W. But in the color version bottom right corner it is hard to see anything at all. I’m guessing that if you hadn’t seen the original image, you wouldn’t have guessed that the.

You can see from this that “if color distortions are not striking to the eye red faces, bluish grass, etc. , you can get a realistic impression of the real colors in the image.d. the accuracy of lightness is much more important for the quality of the image”.That’s why I decided to put gamma curves in the first place in our ranking.

Let’s look at the left half of the photo. On it the original image and separately each of channels of the image is presented. Note, that the green G channel is closest to the original color image. The red channel is quite light and the blue channel is darker. This is the case in about 80% of cases

This all illustrates the rule: “the final image depends 30% on the red channel, 60% on the green channel and only 10% on the blue channel. That’s why the deviations in the blue channel are not as fatal as in the green or red channel. With one BUT – if there are no blue or blue objects with a lot of detail in the image.

This picture shows the effect of the gamma curve deviation for each channel. The pictures are RGB order. The top row shows the full impact. In the bottom row, the color changes were discarded, retaining only the lightness changes. The bottom left corner shows what shape the gamma curves were set for each channel. The blue line is the reference. The red line is the gamma curve of the modified channel. We set a small deviation, corresponding to about – + 0.2 in the area of light shadows.

Note the two rightmost images. The change of gamma curve of the blue channel practically does not lead to visible changes of the image. Both in terms of changes in image brightness and in terms of changes in colors.

Let’s look at the effect of changing the gamma curve of the red and green channel by lightness. In the case of the green channel, we see a much stronger influence. The picture became lighter and less contrasty. Especially well this can be seen on the body and butt of the spider. With the red channel, the changes are also noticeable, but not as dramatically pronounced. The color deviations presented on the top row are clearly visible. In terms of color, everything is subjective. Some people will dislike the reddish hue more than others, and some will dislike the greenish one. But if you zoom out of this picture, the green tint is much more noticeable than the red.

Another interesting phenomenon which can be observed in the channel gamma curves is the S-shaped gamma. We’ve already talked about this in part in one of our reviews, too.

To recap briefly. With a shape like this, you can achieve additional contrast in the middle tones of the image. But nothing is free. The price of this increase in contrast is the complete destruction of detail in the light areas of the image.

This photo illustrates the effect of the S gamma. Details, in particular the droplets on the petals, are completely indistinguishable. The background is much more expressive.

When the goal is for the projector to be readable in bright light, the S shape of the gamma curve is a great and perhaps the only solution.

If one of the projector modes has a gamma curve shape like this, it’s probably a good thing. But if all modes have this shape, then it’s definitely very bad. For everyday use, these modes are not suitable categorically.

Second place. Stratification of RGB levels: parasitic shades wander across the screen.

Summary: The worst thing is when the channel levels differ from each other, and especially when their values change throughout the brightness of the image. This means that different parts of the image will have different stray tones.

A difference of 10 percent from the reference value can already be a cause for concern.

For our second place in the rating, I would put such a parameter as RGB levels. In our reviews they are presented in charts like this.

RGB levels are one of the key tools for measuring color fidelity. The values of the channels must be identical and at 100% level.

In this case, the grayscale will be completely neutral, and therefore the color calibration is also excellent. You can tell if a particular grayscale tint is present because of the deviation of the graphs from 100% of the value.

Interpreting these graphs is quite difficult. You have to have a good knowledge of color theory to be able to tell what would be the hue of e.g. 110%R 96%G 116%B at 50% lightness.

And it will be like this. A circle corresponds to 128R 128G 128B 100%R, 100%G, 100%B at 50% brightness level and a square to 140R 123G 149B. When these two colors are side by side, the difference is obvious, but when there is no gray reference, believe me, the eyes get used to it, and the brain begins to perceive this purple-gray as a completely neutral color. And color blind people won’t see the difference at all, because in grayscale these colors differ by only 1 point, which is almost indistinguishable to the eye.

In the case ofRGB levels, you can’t say that one channel is more important than the other. They are all relatively equal.

You should remember that a certain hue is only attributable to a certain part of the image, depending on the brightness level.

Let’s take the following image as an example.

Initially, it is completely neutral black and white . That is, throughout the luminance R=G=B.

In the second image, I increased the red level in shadows, the green level in midtones and the blue level in highlights. Because of this, the details of the structure, the trees, turned red, the dark part of the clouds green, and the light clouds bluish.

In a color image, the situation will be slightly different. And the degree of the negative effect will depend on the particular image.

I applied a similar change to this image. And here’s the trouble, the water turned green, the stones in the foreground turned reddish. And the sky, which should have turned blue, looks pretty decent. The thing is, it was originally yellowish. Yellow is the opposite of blue. That’s why the sky has no abnormal color.

The worst thing about theRGB levels is our example, where the levels of each channel change depending on brightness. This, as we have already found out, means that the hue will change depending on the brightness.

If the levels are not at 100 percent, but are nevertheless stable throughout the brightness, it just means the wrong color temperature, which is the third place of our anti-rating.

Third place. Color temperature: it’s good when it’s neither warm nor cold.

Summary:

1. If the color temperature of the projector is warmer than 5500K, the image will be noticeably yellow. It will be very conspicuous. At the same time, the allowable deviation into the cold area can reach 8500K.
2. If the color temperature is unstable throughout the brightness, the worst thing is when the shadows have a warm hue and the highlights have a cooler hue.
3. The tint is much more noticeable than the color temperature deviations.
4. Do not forget that the real color temperature will be compensated by the quality tint of the screen and the ambient light temperature.

In third place of our anti-rating was the color temperature of the image.

Imagine the color of a white sheet of paper on the street. But if we go into a room lit by incandescent light bulbs, we will see the color of the sheet as slightly warmer, more yellow. But we will still perceive it as white, because its real color has not changed.

That’s how our brain works. It automatically brings the objects to some baseline, compensating for the peculiarities of the ambient light. The problem is that this mechanism does not fully work with projectors. Therefore it is important for the projector to have a color temperature which is taken as this very basis.

Classically, color temperature is measured in Kelvin. 6500K is considered the reference. This is the temperature of normal daylight. Less is warmer, or more yellow, more is cooler, or more blue.

In our tests, color temperature information is presented in similar graphs:

As always, the values are given as a function of image brightness. First let’s see what the image looks like at different color temperatures.

In the first image the image temperature is normal. This is how I designed it. On the second and third made adjustments to the plus and minus. Agree, the cool color temperature of the last image, though not correct, but more tolerable than the second option.

So you have to be more wary of unnecessarily warm color temperatures. It’s much more conspicuous than the cold.

If the color temperature of the projector is warmer than 5500K the image will be noticeably yellow. It will be very conspicuous. At the same time, the tolerance in the cool region can be up to 8500K.

As with RGB levels, color temperature can vary depending on the brightness of the image. The rule of thumb is: the worst thing is for the shadows to have warm colors and the highlights to have cool colors.

The left frame is the original. In the middle a case where the shadows are cool and the highlights are warm. The right frame, on the contrary. Yes, due to the fact that there are a lot of shadows in the image, the middle frame has become quite cold. But at the same time it looks quite natural. But the right frame is not. Pay attention to the color of the face and hands, it has become unnatural. Fairly warm and pleasant color of the right from us cheeks flows into a distinctly bluish forehead. It looks awful.

But that’s not all there is to look at in the color temperature area. Two colors cannot fully describe the color temperature. There is also the so-called tint, or shift, which is a literal translation. It is responsible for the change from green to pink.

On our charts, the tint information can be found in the color gamut charts.

The black line describes the change in color temperature in kelvins. The white dots, gray wedge points illustrate the color temperature values in Kelvin, taking into account the tint. The tint value is the deviation from the black curve.

This example shows the effect of the green tint all points of the grayscale wedge are above the reference curve . Notice how unnatural the image has become. Therefore, the distortion of the color temperature is much more dangerous than the distortion of the color temperature.

There are two blue ellipses on our graph. These are the confidence intervals dE=3 and dE=10. If the grayscale wedge points are out of the second interval, breaking away from the reference curve, it guarantees noticeable parasitic tinting of the image.

Don’t forget that the actual color temperature of the projector will depend on two other things:

1. Ambient Light Temperatures. If it is warm and the lighting is strong enough, the image will be warmer.
2. The quality or hue of the screen on which the image is being displayed. If it has any hue, that hue will be transferred to the image.

Fourth place. Uniformity of brightness: no darkness in the corners!

Summary: os

A brightness level deviation of more than 25%, as well as a significant difference between the brightness values of the corners can lead to a decrease in viewing pleasure.

This characteristic indicates how much darker the image corners are than the center.

In our tests we bring these images

Serious negative effects may occur in two cases:

1. The difference with the center reaches more than 25%. In this case the corner darkening will be really noticeable not only on the homogeneous background but also on any other image.
2. There is a difference between angles or different sides of the frame. Vignetting, which is what we are talking about, implies uniform shading on all sides when moving away from the center. If this darkening is uneven, for example, the right side is noticeably darker than the left side, then the image will be very unpleasant to look at. It will be very noticeable.

Fifth place. Color deviations are just as undesirable as deviations in behavior

Summary:

A characteristic for professionals who have requirements for color fidelity. The ideal values for working with color are 03. For general use it is acceptable 025.

Color deviations in primary and secondary colors is a characteristic we cite for professionals. This characteristic shows how much the colors we see on the screen will visually differ from some reference value sRGB colors . The ideal values for color are 03. For general use it is acceptable 025.

I will give you some examples to illustrate this:

dE=15.8

dE=22.0

dE=33.0

PS: what about contrast?! Score…

Yes, it is a characteristic that is mentioned in all the specifications and gets a lot of attention.

But you know, it’s like choosing a fast car solely by the number of horsepower. The Belaz 7560 has about 3,546 horsepower, but it clearly won’t go any faster than the Lamborghini Gallardo with its modest 540 hp. with.

Besides horsepower, there are a lot of other factors that affect the final result. Similar situation with contrast. It is foolish to look only at it when choosing a projector.

I repeat once again, contrast greater than 2000:1 is the delusion of marketers. Sophisticated systems for measuring it, etc.d. etc.n.

Contrast is the ratio of LIGHTING between the brightest and darkest points in the image. This characteristic depends most of all on the luminous flux. For example, if the white dot corresponds to 266.1 cd/m² and black 0.1751 cd/m² the contrast will be 1519:1. But if the black point becomes two times lighter, say 0.3 cd/m² the contrast will be 867:1.

The point is, you can’t tell the difference. Contrast is therefore the last thing you should look at. And only control that it is not too low. The critical value for a projector is about 250:1.

****

Hopefully, our ranking of the most serious disadvantages and advantages will help you choose the best option. There is probably no such thing as a perfect projector, so it’s important to know how to interpret the results of test measurements. And it’s important to understand which deviations are really serious and which are just the minutiae of life.