Physics Talk

Roger Breton · Guest

Hello everyone,

I'm considering buying an LCD projector.

I'd like to decide which light output is OK for my application, watching
movies in my living room.

Say there is 4 meters between the projector and the projection screen,
knowing a lamp total luminous flux in Lumens, is there an equation that
would give me Luminance? Assuming a perfectly reflecting screen?

I remember that movies are viewed in commercial theaters at around 45 cd/m2.
I'd like to find out what luminance to expect from lamps between 1000 to
2000 Lumens.

I don't want to overbuy but I don't want to underbuy either.

Prices are coming down though.

Thank's in advance for your kind help,

Roger Breton

Dieter Michel · Guest

Hello Roger,

last time I measured this, the factor was 0,37 cd/m2 per lux
for a halfway diffuse reflecting magnesia block and 0,29(4)
for a gain=0,8 projection screen. AFAIR there should be a factor
of pi between illuminance and luminance in case of ideal diffuse
reflection, but i will have to dig that out again.

Tom Lianza · Guest

Hi Dieter and Roger,

In general I agree with Dieter's analysis in general, although I'm not
sure where he gets some of his assumed numbers.

The Ansi Lumen spec is measure of total screen lumens striking the
screen. To compute the illuminance at the screen you take those lumens
and divide by the area, which Dieter does to get 617.3 Lux. That is the
surface illuminance. The luminance is a function of the screen
characteristics. For a perfectly diffuse 100% reflecting Lambertian
surface, you simply divide the illuminance by pi, which yields a perfect
diffuse reflectance of 196.5 . The screen gain calculation is the ratio
of a perfect diffuser and the actual screen luminance. If the screen
gain is .8 ( which is a good number for a naturally diffuse screen) the
resulting luminance is approximately 157.2 which is approximately a
factor of three times greater than 48 cd/m2 DCI specification. Which
also happens to be Dieter's "gut feel" number.

It is highly unlikely that you can get the ambient room illumination in
a home (unless it is a well designed theater) much less than 25 lux
striking the screen. This is because the ambient is measured with a full
screen white illuminating the room and then measuring the reflected
light back onto the screen + any light sources. So you need either a
small screen in a large room, or a room with dark colors. This ambient
(worst case) would limit the dark on the screen to about 8 lux (very
worst case). In an actual image, you could use the 18% grey assumption
and this number dips to 1.4 lux. If you take the ratio of this number
to the peak white, you get a static contrast ratio of about 110:1. This
is the big difference between home viewing and motion picture viewing:
You need to manage the light reflecting in the room far more than the
actual luminance of the display. Theaters should have a native static
contrast ratio greater than 500:1 . This is the tough part of a design.
Because the measure is a ratio, it is independent of screen luminance.

My own experience with installers of home theater has indicated that
most home owners hate the image after being adjusted to 48 cd/m2. This
is because they can not get the room dark enough for glare free viewing.
This is why a plasma display or high contrast LCD can give such a better
viewing experience. The contrast ratio is so much higher in ambient light.

Regards,
Tom

Dieter Michel wrote:

Dieter Michel · Guest

Hi Tom,

Gernot Hoffmann · Guest

Dieter Michel schrieb:

Dieter Michel · Guest

Hi Gernot,

Tom Lianza · Guest

Hi Dieter,

This is a difficult measurement to make without a low flare LUMINANCE
spot meter. It can be done with a shrinking black box. First display a
full screen black. Then project progressively smaller black boxes while
maintaining the same measurement position in the center of the screen.
It is very important to maintain a glare stop in front of the meter so
it is only "seeing" the center area through the glare stop. If you
don't do this you are generally measuring the flare in the spot meter
optics. You then plot the data as luminance vs. relative area of the
black box. You subtract the luminance of the 100% area black (which is
due to the projector contrast ratio) and then extrapolate the curve to
0% to arrive at the full screen white flare luminance. For a well
behaved projector system, this function should be very linear. Now that
I think of it, you could probably do this with a fixed illuminance meter
in the center of the field as well. I have never tried it, but I should
give it try when I get some time in the lab.

It's important to note that this process can be greatly affected by the
image processing electronics in the projector. It is definitely not
valid for a CRT projector due to beam limiting electronics. On a modern
LCOS or DLP projector it should yield a pretty good estimate of room
flare, if you take care to limit the field of view of the measurement to
include only the spot of interest.

With regard to your question about the black velvet screen illuminance.
That is a great way to get a rough estimate of the flare issue. The
problem is that screen "gain" is a function of the absolute reflectance
of the material and the geometric properties of the material. That is
why you can have a high gain screen which looks gray when viewed under
diffuse illumination but measures a gain much greater than one on axis.
High gain screens radiate back in a very directional fashion. I gave
a presentation of this at an ICC meeting where I built a screen that had
a split screen with a gain of 1 material on one side and and a high
gain 1.8 or so, on the other side. Prior to projection the high gain
side looked very gray relative to diffuse white gain of .8 screen. When
the projector was turned on, the results where quite obvious as the
relative luminance of the left to right image was swapped. A high gain
screen will be more prone to room flare issues than a low gain screen.

Regards.
Tom

Dieter Michel wrote:

Dieter Michel · Guest

Hi Tom,

thank you very much for your very detailed answer!

Roger Breton · Guest

Yes, please.

And thank's everyone with your contributions.

I know better what to expect.

Roger

Dieter Michel · Guest

Hi Tom, Roger and all,