I am wondering about the specified contrast ratio of some recently released FPs such as the JVC HD2, the Sony VW60 and specifically the Epson 1080UB, which has a listed contrast ratio of 50.000:1.
When HT magazine tested the JVC HD1 Geoffrey Morrison, the reviewer, pointed out the fact that the luminance meter used for the test, a Minolta LS-100, could not measure accurately below a level of 0.001 ft-L and another poster here at AVS has stated the same, so how can Epson claim that its new FP reach up to 50.000:1 ?
If we accept the premise that the Epson unit would be capable of reaching a black level of 0.001 ft-L, wouldn't that then imply that it should theoretically reach an onscreen brightness of 50 ft-L, a figure impossible to achieve unless a rather small screen with a positive gain was used during testing in a completely darkened room ?
And how about the JVC contrast ratio of over 30.000:1 ? With a black level of 0.001 ft-L the theoretical onscreen brightness should be over 30 ft-L, again, a very high figure...
Are there any luminance meters capable of accurately measuring below 0.001 ft-L ? Say, at a level of 0.0005 ft-L ?
-------------------------------------------------------
Marcos

The Epson and Sony claim their figures using a variable iris, so that is the big fudge there. Native contrast for them is well below 5K:1.

The JVC HD1 quotes native contrast at 15K:1 and has been measured to achieve it. To double this, all the HD2 needs to do is reduce black level by half, ie from 0.01 to 0.005 (for example), so it is achievable.

For supre low level black, they reduce the size to elevate the black and measure and then interpolate.

The Epson and Sony claim their figures using a variable iris, so that is the big fudge there. Native contrast for them is well below 5K:1.

The JVC HD1 quotes native contrast at 15K:1 and has been measured to achieve it. To double this, all the HD2 needs to do is reduce black level by half, ie from 0.01 to 0.005 (for example), so it is achievable.
------------------------------------------------------------
Ah, but the measured black level of these FPs is at or near 0.001 ft-L, not 0.01 ft-L... I know that 0.005 ft-L is easily achievable by quite a few current FPs, but 0.001 ft-L is a different story altogether : it is five times "darker" blacks.
The JVC HD1 and the Sony VW60 have been tested and measured at such low black level, with resulting onscreen brightness at around 16 ft-L, implying contrast ratio(s) of 16.000:1.
So, again, if HT magazine tests the JVC HD2 and cannot possibly measure its black level reliably below 0.001 ft-L, are we to expect its full-on onscreen brightness to be around 30 ft-L ?
I wonder what instrumentation these manufacturers are using to test luminance in FPs as well as how accurate they are below 0.001 ft-L.
By the way, HT magazine found the native contrast ratio of the Sony VW50 at around 1.900:1 and the VW60 at about 3.500:1.....

So, again, if HT magazine tests the JVC HD2 and cannot possibly measure its black level reliably below 0.001 ft-L, are we to expect its full-on onscreen brightness to be around 30 ft-L ?
I wonder what instrumentation these manufacturers are using to test luminance in FPs as well as how accurate they are below 0.001 ft-L.

Black level is not fixed, it relates to screen size. You would normally measure this directly in the light path, not off-screen. Measure close to the projector, and both white level and black level will be elevated to a level where the meter reads accurately. There is no way you can derive light output from contrast ratios, unless the actual measurement is stated (i.e. as lumen output). When you have lumen output, you can calculate on-screen luminance if you know the size and gain of your screen (note, however, that a lot of screens are pretty far off from their stated gain!)

I use a light meter that goes down to 0.01 LUX, and position it close to the projector facing the lens (2 to 3ft away). The black level is then high enough to get an accurate reading well above the meters limits, and provided the white level isn't overloaded by being that close (not been a problem yet - still well within limits of the meter), I can get pretty accurate and repeatable on/off contrast readings.

Conversely I've seen some reviewers position the meter quite a way from the projector (12ft for example) and they get inaccurate black level readings, and then wonder why the contrast reading reduces as the lamp ages (black level reading remains the same because it's at the limit of the meter, but the white level reduces as the lamp dims). They conclude that as the lamp ages it reduces contrast...

You measure the lux at the screen and multiply by the area in square meters which will give you the lumen value. You can then use the contrast figure to calculate the black level and convert the lumens to ft lamberts (divide the lumens by the screen area in square feet) to give a figure for the actual black level.

So the very low numbers you see aren't necessarily measured but calculated, so you don't need an expensive piece of lab equipment to get these high on/off contrast figures.

HTH

Gary

....a Minolta LS-100, could not measure accurately below a level of 0.001 ft-L and another poster here at AVS has stated the same, so how can Epson claim that its new FP reach up to 50.000:1 ?...

Perhaps large transnational companies like Epson, Sony, etc.. have the resources to invest in R&D labs that have better equipment than your typical PJ reviewer? The quoted numbers have "your mileage may vary" issues, and I'm sure they are portrayed in a way most favorable to the manufacturer, but I don't doubt they have the ability to measure these things accuratly.

You just measure close to the projector so your meter doesn't need to be able to measure very low numbers that measuring at the screen would require.

Gary

Gary covered it, but just wanted to mention that if a person has a meter which reads off the screen (instead of in the light path pointed toward the projector), they can put a piece of screen material or paper in front of the projector and measure off that. Or put the projector very close to the screen for a small image with brighter levels.

Sometimes this can swamp the sensor at the high end, but then there are other methods if CRs get too high. Kind of like how the length of a house can be measured with only a 10' long tape measure. Just split things up. For estimating the on/off CR of a CRT front projector I used a digital with a known on/off CR (I had measured it), then put neutral density filters in front of the digital until the absolute black level looked the same as the CRT's absolute black level and measured how much the ND filters had dimmed things by measure white with and without them there (then just ran the appropriate mathematics). I don't have any sensors that can measure 700k:1 on/off CR directly, but by splitting things up I was able to estimate that as the on/off CR of that CRT in the mode it was being used in (with some gamma changes down low to help shadow detail when setup with high CR, since CRTs tend to crush shadow details normally if the absolute black level is set too low with the Brightness control).

For a digital with very high CR a person could measure 20 IRE (or %stim or whatever somebody wants to call that) and 0 IRE close to the projector or with a very small image and then 100 IRE and 20 IRE far from the projector, or off the screen with a large image. Then just multiply the 2 ratios for the whole on/off CR (100 IRE to 0 IRE), since the 20 IRE cancels out in, like if it is y in the following equation: x/y * y/z = x/z.

--Darin

That's not very accurate at all :( What you could do is make a time exposure with film in a batcave. Then experiment until the exposure of a reading you can measure results in the same amount of exposed light. Calculate EV offset, and you have a very accurate measurement. This is how I bet companies do it.

Given the inaccuracies of film exposed over this sort of time frame I would trust the accuracy of the ratings on ND filters more readily.

ted

That's not very accurate at all :(Why do you claim it isn't very accurate at all? I realize it isn't going to give many significant digits and just reported one at 700k:1 on/off CR. There are of course issues with perception and color balance, but I bet this was in the ballpark (especially compared to people guessing at what a CR with a CRT setup like this could be). BTW: With this setup, we could see the screen after going to a blackout after just a few seconds.

--Darin

You're telling me that a CRT projector would saturate a 35mm film exposure in even 30 seconds with 0 IRE?

Why would you think I would suggest that?

The linearity of negative density at those exposure times is likely to be suspect and thus not valid for repetetive comparisons - unless film has changed since the last time I ventured into a darkroom.

ted

Your eye isn't accurate. Your eye is never accurate in telling you which is brighter or darker. How much is the 700k:1 off by? What's your margin of error?I of course didn't put a margin of error or know, but the implied margin of error is plus or minus 100k:1 (plus or minus one in the first digit). As far as the brighter or darker, we were switching almost instantly (covering one projector and then the other to judge), so at least we were doing the best we could there. It wasn't like we were trying to judge over some longer time. We covered all except the very center of the screen with black velvet and did measurements and comparisons there.

And I don't have a film setup. Are you suggesting taking a snapshot from the CRT's absolute black image, then going into a lab and doing something with the film, then going back to the house with the CRT and trying a different amount of exposure, etc.?

--Darin