Most multimedia projectors has a contrast ratio of 2000 to 1..
I am wondering if this is excessive since the psychovisual limitation of the human eye is just 100 to 1.. This means that if black is represented by a value of 1 then the minimum value which the human can detect as white is 100.. Normally, we would double this value and use 8 bits (256 level) to represent the dynamic range of the human eye sensitivity.

But what I don't understand that there are some commercial products that would even use 12 bits video data(YUV)?? Wouldn't this be an overkill?

How about a contrast ratio of 2000 : 1?? Too much?? In any case, the video data that feeds into the projector is very likely to be 8 bits data type (YUV)..

I think 2000:1 it's an overkill, since the average contrast ratio for a LCD projector is 600:1 and I've seen that even 300:1 is good, imo.
But maybe the quality differences between 600:1 and 2000:1 (even if subtle and hard to notice to the human eye) are present and the companies think that this little difference is enough to release 2000:1 multimedia projectors..

...and there is a (sometimes huge) difference between the numbers advertised and those measured by independent comparisons.

Even then, I am surprised that for an input of 24 bits RGB data, one could get a contrast ratio of 2000:1 . Sure for the transformation of RGB color space to YUV color space, there are a total of 256*256*256 different levels of Y.. But is it logical that this transformation would produce so many levels of YUV...

Then again, if the transformed resultant 256*256*256 levels of YUV are transformed backed to RGB color space, we get pow(256*256*256, 3) levels of RGB! This doesn't seemed to be logical!

As for a contrast ratio of 2000:1 , this corresponds to a 12 bits resolution.
Some manufacturers of DVD players claims that they are using a 12 bits video DAC ! Can we really tell the difference between 12 bits Y from a 8 bits Y?

I remembered doing a composite video project. We digitized the input composite video into 8 bits data samples and decoded it fully in the digital domain (DSP processing) for computer photography applications.. It seems to me that even for a 8 bits data samples, still images, we can't really picked up noises on the decoded images.. (printed on high resolution color printers)

For moving video, it is even more likely that these quantization noises are not noticeable at all!

I would think that this would help in a lighted room, where the 2000:1 ratio would be degraded by other light sources. The contrast in a lit room with that kind of specs would still be very respectable.

Our eyes are not very sensitive to small changes in Y. For example, if Y has a range of 0 - 255, we could not tell the difference of 128.0 form 128.5! It would not be noticeable! (The human psychovisual has much greater frequency sensitivity on the Y than on the U, V components!)

In a bright room on the other hand, would only add values to the Y component.. It would raise the floor dark value.. I still do not know how having greater contrast ratio would result in "sharper" images.. Adding brightness to Y could cause some of the Y information to saturate at maximum values thus losing some important details..