[David Mendicino]

I understand that being of lower gain, it will reduce the absolute black level of the image by "absorbing" its' light.

How, it it actually does, can this increase shadow detail?


I mean, if there is a threshold in which blacks start to meld together to form the solid, grey image, how does darkening this point help the blacks below that point that were already missing?


for example:


Lets say my projector can produce black 5, 6, 7, 8 (as subjective levels), and I see them in my image, but black 0,1,2,3,4 I can't see at all as they are all combined to form a solid grey.

Now, if using the greyhawk, I make black 5 more like black 2, in terms of darkness, what happens to 0,1,2,3,4? What I am trying to say is this.


If you have shadow detail made up of an array of shades of black and grey, and yoru projector can only reproduced up to a certain point....with the greyhawk, you are only taking the limited reproduction of what you had, and darkening it, while still eliminating those the projector could not produce, correct?

 

[KFung]

David,


Not quite. Your projector (when calibrated correctly) should reproduce the full black to white range of your source. The trick is that digital projectors produce "full black" as dark grey (e.g. a value of 5 to use your example).


IMO, the loss of shadow detail occurs when light reflected back by the walls of the room wash out the lower levels. For example, if you had black as 5 and values of 6,7,8, etc. on the screen all would be well. However if you consider that the room reflects back maybe a value of 3, you get values of 8,9,10,11. This kind of compresses the grey scale in the dark values and you lose shadow detail.


A grey screen partially corrects this by absorbing light and reducing the effect and amount of light reflected by the walls. So where you had a back reflection of 3 in the previous example, this might fall to 1 with a grey screen. This will also help your colour saturation because there is less light reflected by the walls to wash out pure colours.


Regards,


Kam Fung

 

[jackmay]

I will take a stab at a theory.


The eye has a non-linear response, I believe logarithmic. Lets use log base 2 for a way of describing the effect.


Suppose the eye will see black when the light level is 2 Lambert and that is the lowest light level the projector will put on the screen. Log2(2) = 1. The next quantization level up would add another 2 Lambert and the output would be Log2(4) = 2 or twice as much light perceived by the eye. A big difference.


Now suppose there is a minimum light level of 16 Lambert which looks grey. The eye sees Log2(16) = 4. If we jump to the next quantization level of 2 Lambert higher, the eye perceives the next level as Log2(18) = 4.17, a small perceived difference.


By attenuating the light and getting it down to where the light sees the light level as black, small differences in grey level are much more apparent.


So the goal is to get the black level down to the place where it looks black for the ambient light conditions in the viewing room.


The expensive, inflexible way is to attenuate the light to the correct level with a low gain screen.


A better way is to have a variable attenuator of the projector output or use a bias light to make the eye adjust to see the lowest light level as black. Then the black can be set for different view conditions in the room.


An attenuator can be built by using two polarized sheets or lens and rotate one of them to get the desired light level. The bias light would just be an ordinary light probably behind the screen which is varied with a dimmer.


The low gain screen market exist. We need someone to market variable attenuators for projectors and maybe even a good bias light system.

 

[David Mendicino]

Hmm...let me ask this in another way.

First, isn't it incorrect to say the projector can do the ful range of black to white? That is, if the lowest the projector can produce is grey, then any other detail which the source presents which is darker than the grey just gets crushed into that grey.

Isn't this the reason why when things get dim on a digital projector, it starts to take on a hazy look of grey?


Let me try to make this an audio analagy, which is poor intself.


Lets say you can produce a signal down to say, only 500HZ, and the source has signals down to 20 hz. So in order to get that low, you come up with some device that takes that 500hz and manipulates it so that it sounds like 20hz. If you ahve done that, don't you throw away all the signal that was below the 500hz that you couldn't here?

And in that case, wouldn't you also lower all frequences, even the 20,000hz? This would apply to the lowering of the white value, in terms of projectors, whould it not?

 

[jackmay]

Quote:

Originally posted by David Mendicino And in that case, wouldn't you also lower all frequences, even the 20,000hz? This would apply to the lowering of the white value, in terms of projectors, whould it not?

The white levels will be reduced along with the dark levels. The ear does not adjust its frequency response as the lower frequencies occur.


The eye does adjust the range of light values it sees as the intensity goes up and down. When you walk into a dark movie theater, you can't see much detail until you eye adjust its range of light levels it can see to a lower level.


The problem with DLP and other projectors is that they have a small range of light variation. A contrast ratio of 1000 to 1 is only 30 dB. That is far less than the dynamic range of the eye.


What we need is a projector that work more like the eye. The most practical way to do that is to put a light level modulator inside the projector or even over the lens.


A light modulator might be a LCD panel with with no pixels (actually one large pixel). The light coming off of a DLP imaging forming chip would reflect off of or pass through the LCD panel. The LCD panel could also be between the bulb and the DLP in a light tight box so no stay light would leak our and all light would go through the LCD modulator panel.


The LCD panel would adjust the total picture light level up and down on each individual frame for the entire frame. The output would then be (2^image bits)*(2^frame intensity bits). This would produce a light output that matches the way the eye sees.


With a 1000 to 1 contrast ratio for both the DLP image and LCD frame intensity, we would then have a contrast ratio of 1,000,000 to one or a dynamic range of 60 dB.


Electronics would take the present video signal and calculate frame intensity level and calculate what to feed to the image and what to feed to the frame intensity devices.


I think all of this is within the present state of the art at a reasonable cost. Its just a matter of someone deciding to building a projector that is designed for video instead of for PowerPoint presentations.


The guy that owns Runco lives in Foster City where I live (a town of 25K people). Maybe I could go pound on his door to get him to start producing video projectors instead of PowerPoint projectors. But then again, I would never pay the price he sets on his equipment.

 

[David Mendicino]

NOw that I remember, I think such a concept was put forth before.


It couldn't be just ONE pixel, though, because that would not take into account differences within sceens. If half the screen is supposed to be white, and the other half dark, what are you gonna do?


I think someone also said use the reverse of a welders mask..that is, when the light goes down, it gets opaque, instead of opaque when bright.

 

[jackmay]

Quote:

Originally posted by David Mendicino NOw that I remember, I think such a concept was put forth before. It couldn't be just ONE pixel, though, because that would not take into account differences within sceens. If half the screen is supposed to be white, and the other half dark, what are you gonna do? I think someone also said use the reverse of a welders mask..that is, when the light goes down, it gets opaque, instead of opaque when bright.

Since the last response, I did a little research on what the eye can see. They eye can't see from total black to total white. In any one image the eye can see a range of about 10,000 to 1.


From seeing nothing to eye damage is about a 150 dB range. A more practical range from first seeing color to daylight, the dynamic range is about 100 dB.


I don't think HDTV can encode light levels over a 100 dB range so a projector does not need to handle that range. Maybe a 60 dB dynamic range will be a much as we need for the present HDTV standard. There is no need for such a high dynamic range that your eye will require minutes to adapt between the light levels of different scenes.


Our ultimate video projector then needs a dynamic range of 10,000 to 1 for the individual picture pixels and a light modulator raise light level up and down from frame to frame.


It would be possible to have two DLP in series in the light path to give a 1,000,000 to 1 contrast ratio across the screen, but I think that would increase the cost significantly and maybe require constant tweaking to keep the DLP's aligned.


It does say that it should be possible to build a high quality video projector now, we just may not be able to afford it for a long time.