What common tool would one use to measure the number of Lumens a pj is putting on the screen? I assume some sort of photo light meter?? :confused:

:cool:

The AEMC CA813 is a good choice.

Hello jimwhite,

We use the Minolta LS-100 spot meter. It is quite pricey, but does an excellent job providing consistent and repeatable results.

Regards,

Dave

The LS-100 is not a bad instrument, but it is not what I would choose for this. It is much more expensive. The LS-100 has an f1' of 8% compared to the CA813 which is also an 8% f1' instrument. The CA813 will also not require compensation for screen gain since the LS-100 only measures reflected light.

Agreed on the CA813 ^^^

Here is where I got mine http://www.byramlabs.com

As umr said, the CA813 is the best choice. It is designed for the use we need and is the best instrument, at the lowest price, for measuring PJ output. In general, photographic light meters are not suited for projector measurement use.

NOTE: The next level of the CA813 link that derekjsmith pointed to is: http://www.byramlabs.com/product_info.php/products_id/5991. This meter is also recommended in Appendix A of the GetGray Caldisc ReadMe file (www.calibrate.tv). It can also be purchased direct from AEMC (www.aemc.com ==> www.repaircalibration.com), with NIST calibration, if needed.

Hmm, sorry for jumping in here .. but I have a Eye-One Display 2 on order, my first meter, I was expecting to be able to use that with ColorHCFR or CalMan when trying to achieve that 12-14 ftL figure people are talking about. Is this not something people are doing with their Spyder2's and D2's, you really need an additional meter?

Yes you do need an additional meter for measuring the correct Lumens. Both the Spyder and D2 read higher then the actual. My D2 reads almost twice as high as the Spyder.

Hmm, sorry for jumping in here .. but I have a Eye-One Display 2 on order, my first meter, I was expecting to be able to use that with ColorHCFR or CalMan when trying to achieve that 12-14 ftL figure people are talking about. Is this not something people are doing with their Spyder2's and D2's, you really need an additional meter?Those meters are designed for measuring and calibratiing the grayscale and color performance. They are not designed to measure lumen output (you don't calibrate for that anyway). The CA813 is designed to measure Lumen intensities of visual light sources and not intended for any other purpose.

Thanks! Too bad, getting a comparable reference of how bright my PJ is on my particular screen with my screensize is one of the things I was REALLY looking forward to when I ordered the D2 .. :(

Daniel,

If you know the gain of the screen you can calculate the pj lumens from measured cd/m2.

Formulas (http://www.practical-home-theater-guide.com/projector-screens.html)

(A footlambert equals 1 / pi candela per square foot, or 3.4262591 candelas per square meter)

OK I'm officially confused; I am not really interested in gain or amount of lumens output by my PJ, all I want to know is how close I am to that recommended 14 foot-lambert figure that is said to be ideal for digital cinema. I want to test 14, knowing I'm at 14, and then see if I'm happier at 10 or 18. And then I want to be able to talk about it on forums. I want to be able to understand how bright 22ftL is, when for instance someone with a high power screen says they're at 22ftL.

Can I or can I not do this with the D2? Can the Spyder2 do it? The i1 pro?

OK I'm officially confused; I am not really interested in gain or amount of lumens output by my PJ, all I want to know is how close I am to that recommended 14 foot-lambert figure that is said to be ideal for digital cinema. I want to test 14, knowing I'm at 14, and then see if I'm happier at 10 or 18. And then I want to be able to talk about it on forums. I want to be able to understand how bright 22ftL is, when for instance someone with a high power screen says they're at 22ftL.

Can I or can I not do this with the D2? Can the Spyder2 do it? The i1 pro?

Foot-lambert is a luminance measurement. Lumens is illuminance. You need an instrument that reads luminance directly if you don't want to fiddle with some calculations. Luminance instruments will not have a cosine corrector while an illuminance meter will. Examples of what you can use to read fL directly are the GMB Eye-One Pro, Minolta LS-100/110 and Sekonic L-558Cine.

The D2 and Spyder2 have cosine correctors and will not read luminance even though they report readings in cd/m2 or fL. Their calibration method works as long as screen gain is 1.0, but it will be wrong in the presence of gain because it is a cosine corrected measurement and cannot see gain.

Greetings,

There seems to have been some confusion regarding the question. Lumens is not measured off of the screen and would require an instrument such as the AEMC CA813; however, all light measurments in regards to viewing are taken off of the screen as the screen plays an important role in achieving standard light output values.

There is no significance to the actual lumen measurment as the screen gain, if it is perforated, the amount of perforations, the size of the perforations, the size of the screen, and many other factors come into the equations when trying to achieve a specific light level.

Image light output is measured off of the screen; whether it is a CRT monitor, an LCD, a front projector, or a rear-projection television. In doing so a meter such as the Minolta LS-100/110 is accurate and provides repeatable results - it is not photographic.

The issue with using most color analyzers and spectroradiometers for this purpose is that they calculate the luminance readings from the spectral measurments that they are taking. While this can be close, it should only be used as a guideline and not absolute.

Regards,

Dave

I was interested in measuring the Lumen output directly from the projector.... well, these (CA811/813) measure in footcandles or lux... hmmmm... gotta figure out how to convert this.... meanwhile, my boss agreed to buy the CA811 since we need one here to measure lighting on the workbenches (lux). Thanks for the help so far, guys!

:cool:

I was interested in measuring the Lumen output directly from the projector.... well, these (CA811/813) measure in footcandles or lux... hmmmm... gotta figure out how to convert this.... meanwhile, my boss agreed to buy the CA811 since we need one here to measure lighting on the workbenches (lux). Thanks for the help so far, guys!

:cool:
The difference between them is square feet vs. square meters. The lux measure will be approximately 11x the fc measure.

Bill

Thanks Jim for starting this thread. It has been the most educational one for me (thanks to inputs from the experts: umr and David Abrams). I know what "cosine correction is supposed to be, but was not quite sure how it applied to the subject at hand (luminance vs. illuminance measurement).

I googled the subject and found this information on the "hansatech-instruments" web site (after correcting 3 spelling errors):As a beam of light is tilted so that it strikes a surface at an angle, the area illuminated by the beam increases (see diagram opposite). Since the number of quanta of light in the beam remains constant, the illumination on the surface decreases. This decrease is proportional to the cosine of the angle, which the incident beam makes with a perpendicular to the surface.

The reading of a fully cosine-corrected light meter represents the true ilumination of the light receiver regardless of the angle from which the incident light approaches. With the light detector at an angle, the reading on the light meter should equal the reading that would occur if the light meter were held perpendicular to the rays of light, multiplied by the cosine of the angle of incidence on the meter detector. Cosine correction is accomplished by using a specially designed white acetyl diffuser receptor. The diffuser provides cosine correction to a high degree of refinements; thus the detector correctly displays the incident light intensity regardless of angle of incident light emission.It summarizes the subject of cosine correction, without going into the gory scientific details.

Basically, as I think I understand it, the CA813 has a Cosine Correction filter attached (as does the D2/LT & Spyder 2) so they include illumination arriving at the meter from all angles. The result being that it "sees" the total output of the PJ and reduces the effect of "off-angle" sensor positioning.

As I understand light meters, without a Cosine Corrector, the meter is a spot meter and with it, it is an incident light meter. In the photographic world, spot meters are used to measure the amount of light coming towards the camera, and incident meters are used to measure the total light falling on the subject.

What I don't quite understand, is why the spot meter correctly measures the light coming off the screen, including screen gain (which multiplies the light output from the PJ) and the incident meter (spot meter with Cosine Corrector, like the CA813) won't. Of course, the CA813, or equivalent, would have to be masked so it only "sees" the screen and not any stray light (coming from the PJ). When used that way it would read the light "emitted" (reflected light) by the screen (a psuedo light source). Shouldn't this include the screen gain also (assuming that the only light source is the PJ)? Jeff, could you enlighten me (if it's too involved, just send me a PM).

BTW, The reason that photographic light meters can't be used is that they are calibrated (the good ones) for photographic film response to light (the inexpensive ones just depend on the raw response curve of their sensor). Meters like the CA813 and the 6x more expensive LS-100/110, are calibrated for the CIE photoptic (human eye) response. These two responses are not the same.

Jeff, could you enlighten me (if it's too involved, just send me a PM)

NO! NO!..... keep it open.... we're ALL learning a lot here !!!

:D

What I don't quite understand, is why the spot meter correctly measures the light coming off the screen, including screen gain (which multiplies the light output from the PJ) and the incident meter (spot meter with Cosine Corrector, like the CA813) won't. Of course, the CA813, or equivalent, would have to be masked so it only "sees" the screen and not any stray light (coming from the PJ). When used that way it would read the light "emitted" (reflected light) by the screen (a psuedo light source). Shouldn't this include the screen gain also (assuming that the only light source is the PJ)? Jeff, could you enlighten me (if it's too involved, just send me a PM).
You've basically got it. Short of posting Lambert's law, the issue in practical terms is that the cosine-corrected meter would over-correct the measurement if the projector was the ONLY light source in the room. Given that a CA813 only measures down to the whole number (fc or lux), you can generally assume away some of the ancillary effects from VFDs or other low-output, off-axis light sources.

Bill

...What I don't quite understand, is why the spot meter correctly measures the light coming off the screen, including screen gain (which multiplies the light output from the PJ) and the incident meter (spot meter with Cosine Corrector, like the CA813) won't. Of course, the CA813, or equivalent, would have to be masked so it only "sees" the screen and not any stray light (coming from the PJ). When used that way it would read the light "emitted" (reflected light) by the screen (a psuedo light source). Shouldn't this include the screen gain also (assuming that the only light source is the PJ)? Jeff, could you enlighten me (if it's too involved, just send me a PM)...

The thing to remember is that screens with gain are non-Lambertian reflectors. There is no light created. The trick is that light is not reflected in a uniform pattern from the screen. Since a spot meter will image the screen it can see this effect. A cosine corrector will negate this and will show the total light reflected negating the effect of the non-Lambertian reflector since no light was created in the process. I have attached a couple diagrams that illustrate the difference in these screens.

... Given that a CA813 only measures down to the whole number (fc or lux), you can generally assume away some of the ancillary effects from VFDs or other low-output, off-axis light sources.

Bill

The CA813 will measure fractional fc or lux. You just need to press the range button. It will display values as low as 0.01 fc or lux.

Thanks Bill & Jeff. I guess I'll have to brush up on my optics & Lambert's law. Thanks for the CA813 manual, Jeff - I was looking for that and wasn't able to find it (I didn't look all that hard though).

I'm not really concerned in actually measuring screen gain anyway, since I adjust the PJ so that I get a satisfying picture that is not too bright. I am more concerned with getting the grayscale, color temperature, etc. correct (when watching movies on DVDs). For watching TV (SD & HD), I use the DVD performance as a baseline and eyeball adjust for the closest match (recognizing that no 2 channels are the same).

The fact that I may be looking at 10 ft-L or 25 ft-L is immaterial, if the picture is satisfying and not headache or eyestrain inducing. I recognize that lumen outputs and screen gains are only 10% numbers, at best; being paranoid about the numerical amount of light coming off the screen isn't significant (it's strictly ballpark information). Spending $XXX for a CA813 or 6 times that for a LS-100, just to be able to measure the ft-L off the screen is not justifiable (unless that was my job).

I guess I'm the type who is more interested in getting the best PQ, rather than being paranoid about numbers. However, I appreciate the AVS members, who are picky and into detailed measurements. They are an educational information source, and provide useful information on PJ performance. That allows me to make more intelligent purchase choices (increases my "bang for the buck").

... Spending $XXX for a CA813 or 6 times that for a LS-100, just to be able to measure the ft-L off the screen is not justifiable (unless that was my job)....

I agree. Super accurate measurements in this area are not very useful. The key thing is what looks good. The human eye is also not very sensitive to light level differences making even +-25% differences in maximum luminance not very important. The big thing is getting the light levels in a nominal range for a given application. I have owned the Sekonic I mentioned above and it compared well to a Lightspex and my Eye-One Pro. I would consider it or the Eye-One Pro if you want to measure luminance at a lower cost. However, the CA-813 with a factored screen gain is probably good enough.

What is the difference between the CA813 and CA811? Will both work equally well for measuring the light output of a projector?

For our use the CA813 is more accurate with common light sources +/- 11%, the CA811 is +/- 18%. You can also get the CA813 NIST calibrated if you require even more accuracy.

http://www.aemc.com/products/pdf/2121.21.pdf

Thanks for the info and the link.

I'm not really concerned in actually measuring screen gain anyway, since I adjust the PJ so that I get a satisfying picture that is not too bright. I am more concerned with getting the grayscale, color temperature, etc. correct (when watching movies on DVDs). For watching TV (SD & HD), I use the DVD performance as a baseline and eyeball adjust for the closest match (recognizing that no 2 channels are the same).

Well, I finally broke down and got my meter, a Spyder2PRO, so I've been playing with software and reading this forum more lately. Like CT_Wiebe, I primarily wanted the meter to optimize my color/greyscale performance, but at the same time, I'd also like to know (ballpark) how many Lumens my PJ is throwing and how many ftL I'm actually seeing.

And that's where I'm a little confused by this thread, or maybe worrying about it too much. If I read it right (and I'm going to drastically oversimplify here), both types of meters (cosine corrected and not) read the light coming at them from the source, then they apply (or not) a correction factor assuming an "even" distribution of light off the surface.

Thus the problem with cosine corrected meters is that if you have a screen with gain, the correction overcompensates and gives false readings.

If I'm understanding that all correctly, then cosine corrected or not should not be a big deal for ~1 gain screens, but as you increase gain, the measurement will be increasingly inaccurate? Is that about right? Or is a cosine corrected meter always wrong, and if so, by how much?

Basically what I'm asking, is if I've got a Spyder2 and a ~1.2 gain screen, would my readings be 1% off, 10%, are the useful or are they just totally wrong?

Hm. No bites huh. Well I've been playing with my Spyder2 and ColorHCFR, Basically I confirmed what I already knew, the IN76 is pretty much right on out of the box. But I have managed to nudge the color temperature a little closer.

But to the topic of this thread, last night I finally made the effort to try and measure illuminance. So, looking at the measurements in xyY with the Spyder2 pointed at the screen from ~2 ft, I get a Y of ~34, which I assume is in cd/m^2, but I can't seem to confirm that.

Using Lyckman's formua above, that comes out to be right about 10 ftL. So given my 1.16 gain screen, that's 41x110 or ~31 sqft, that gives me about 270 lumens. My lamp has >750 hrs on it.

Does that sound reasonable? Seems rather low to me. Without having anything to base this on, I'd have guessed I was >12ftL, especially considering new, the IN76 puts out 500-700 lumens IIRC.

The issue with using most color analyzers and spectroradiometers for this purpose is that they calculate the luminance readings from the spectral measurments that they are taking. While this can be close, it should only be used as a guideline and not absolute.

Spectroradiometers are calibrated both spectrally and radiometrically the brightness level is not derived from the spectra itself. To derive accurate color measurements the brightness must track accurately as well or the chromaticity values will not be correct! Brightness (spectral radiance) determines the amplitude of the spectra at any given wavelength. Any instrument which is calibrated and certified against an NIST reference standard can be considered absolute to the extent that it will be accurate within its given specifications.

Is there a way to use a photo camera for light measurements from a projector?
With corresponding F-stop/shutter speen/ASA you can find out the EV value and it should not be too hard to go from EV to ftc. Anythoughts would be appreciated

Is there a way to use a photo camera for light measurements from a projector?
With corresponding F-stop/shutter speen/ASA you can find out the EV value and it should not be too hard to go from EV to ftc. Anythoughts would be appreciated
By the time you were done doing this, and had done it on a pixel-by-pixel basis, then you would have invested an enormous amount of time. High-end meters do use CCDs, but the development effort is non-trivial to get absolute, rather than relative, accuracy out of them -- if the CCD is of a quality that it can be accurate.

Note: this topic comes up periodically. I suggest asking William Phelps how long it took him to develop his custom camera-based calibration software.

Bill

This is something I want to do twice a year and have dedicated meter is always an option not a must for me but I realize the ease of use of a meter.
I also have hard time to believe that CCD in a meter that sells for $150 is more accurate than a Nikon pro SLR. Don't you think so?
I can go from readout in the SLR to EV from a chart and then ftc. What is the complications?
I just have not done that and looking for some feedback.

This is something I want to do twice a year and have dedicated meter is always an option not a must for me but I realize the ease of use of a meter.
I also have hard time to believe that CCD in a meter that sells for $150 is more accurate than a Nikon pro SLR. Don't you think so?
You pay for the optics and firmware in an SLR. The CCD is not that much of the price (compare to a P&S with more pixels). A colorimeter is not that complex (diodes, really) and a light meter is even simpler, but calibrating them and putting good filters in them is the cost. Not the electronics.

I can go from readout in the SLR to EV from a chart and then ftc. What is the complications?
I just have not done that and looking for some feedback.
What's so complicated about a computer chip? It's just some sand? ;) Seriously, it is the software you are going to have trouble with. What is the assumed white point in your camera? You can get foot-candles from it? What does that really represent? Are variations in what is registered from one pixel to another real variances in the source or the camera? What about reading in the dark? What is the S/N ratio of the electronics? How sensitive is the measurement to variations in shutter speed?

Et cetera.

As I said, you can have a big task to ahead of you if you want to do it right. If your time is worth the development effort, then go for it. If you want some vague approximation of a white level, then why even worry about it? You can use ftc, assume your camera is consistent and accurate, and move on with your life. :)


Bill

This is something I want to do twice a year and have dedicated meter is always an option not a must for me but I realize the ease of use of a meter.
I also have hard time to believe that CCD in a meter that sells for $150 is more accurate than a Nikon pro SLR. Don't you think so?
I can go from readout in the SLR to EV from a chart and then ftc. What is the complications?
I just have not done that and looking for some feedback.

I take it you want to use the camera's light meter and not the focal plane CCD to measure luminance. If so to convert from EV to luminance at ISO 100 the formula is L = 2^(EV − 3), I have no idea how accurate this is but give it a try, take the reading off the screen as this is for luminance in cd/m^2. To convert to ftL divide by 3.4

Anyone ever compared this to a calibrated meter?

Zoyd, thanks so much.
I also like to know how it is compared to stand alone meters.
Bear5K thanks man. You tend to complicate a simple question.
Unless you buy a minolta meter you are not going to get a better meter IMO.
A pro SLR does offer a very sensitive accurate meter and they are sold w/o optics. But I already have one and want to put in use.
Now if it can be used effectively on this task.

Foot-lambert is a luminance measurement. Lumens is illuminance. You need an instrument that reads luminance directly if you don't want to fiddle with some calculations. Luminance instruments will not have a cosine corrector while an illuminance meter will. Examples of what you can use to read fL directly are the GMB Eye-One Pro, Minolta LS-100/110 and Sekonic L-558Cine.

The D2 and Spyder2 have cosine correctors and will not read luminance even though they report readings in cd/m2 or fL. Their calibration method works as long as screen gain is 1.0, but it will be wrong in the presence of gain because it is a cosine corrected measurement and cannot see gain.

Just get a meter that measures footcandles and multiply it by the area and you will have the lumens. I own an extech 403125 with a removable head and it works fine

Why not Meterman LM631 which reportedly (mfg spec) goes to 0.01 lux instead of the 0.1 lux of CA813. Both are priced roughly the same. BTW, I don't know who manufactures LM631, but it appears to be very similar (identical) to Extech Light ProbeMeterT.

However, I can't find these models for sale with NIST certificate.

A question for those with experience using light meters for measuring lux, such as the CA813. These meters have a strong peak sensitivity well away from the video standard of 6500K. For example the CA813 peak sensitivity is in the green range of the spectrum and from their published curve it looks like at the wavelength that corresponds to 6500K the meter is only about 20% as sensitive as at the peak. So my question is when you measure the lux from a video display or projector how are you deciding what factor to use for the compensation to account for the meter's very peaky sensitivity - or - do you just take the reading from the meter and use that assuming that the light level from the display will be constant across the red-green-blue spectrum and a measurement centered at any point in that spectrum is valid?

So, I've been recently looking into getting a new projector, and wanted to measure the difference in lumens between my old and new. I know this thread is old, but I found it with a simple google search about using i1Display to measure lumens, so thought I would post my findings here for future surfers :)

I have an i1 Display 2, and I also have a friend who works at a flight simulator shop for a major airline, and he has access to a Minolta LS-100. It was even recently calibrated. So I decided to see what could be done with my i1d2. I tried various methods of measurement, such as pointing the i1d2 at the projector, using the white diffuser, and pointing the i1d2 directly at the screen (my screen is a DIY painted with Behr Ultra Pure White at 121 inches). Turns out pointing the i1d2 directly at the screen gave the best results.

The Results:
It turns out the i1d2 is somewhat reliable and repeatable! I used the free ColorHCFR software, and measured in foot-lamberts on both the i1d2 and the LS-100. I pointed the i1d2 at the screen at about a 45 degree angle and about 1 foot away. The readings would vary slightly, but generally in the +- 0.5 ft-L. I found that I could vary both the angle from about 20 degrees to 70 degrees as well as the distance from 1 foot to 4 feet, and the reading from the i1d2 really didn't change. This was surprising, as I expected the i1d2 to 'see' a lot less of the screen at 4 feet away. Anyway, I did all my measurements at 45 degrees and 1 foot with the i1d2, and 45 degrees and 2 feet with the LS-100 (I did verify the LS-100 did not vary based on angle or distance either. I got the same readings from the LS-100 at 15 feet away as I did at 2 feet away).

I used test patterns on the AVS HD 709 disc, and used from 70% to 109% (the full 255 white one). I also used different modes on my projector, testing the calibrated 6500K up to the 9000K torch mode. The range of ft-L during all these tests according to the LS-100 was 2-13 ft-L. The average of the readings from the i1d2 were consistently 7% higher than the LS-100 readings at all brightnesses and color temperatures.

Now that I know this, I can use my i1d2 and just factor in the 7%. It won't be as accurate or repeatable as the LS-100, but it seems close enough to tell within 1 ft-L what the screen's illumination is, and get a good approximations of projector lumens (I think the Behr paint is considered 1.0 gain, but don't remember for sure).

Lastly, the i1d2 seems good at measuring relative brightness changes, but I don't know if all i1d2s are the same or what the range is. It's possible someone else's i1d2 may only read 3% higher, or it could be worse as well. But my guess would be, based on the reports of color accuracy tolerances between different units, is that other ones would be pretty close.

Thanks for that info, very interesting. I've got an i1 Display LT (which I think is the same hardware) and having a lot of hours on my PJ, I've been curious just how bright it is, and was wondering if there was a way to tell with the equipment I've got. So I guess I'll have to fire up HCFR and my i1 and see. I don't really care if it's a 10% error, I'm more curious if I'm getting 6 or 16 ftL sort of thing.

I should also add this method probably won't work on a screen with any sort of gain. Because you have to point the i1 Display at an angle (in order to prevent it from being pointed at its shadow), this would affect the brightness. It might work if you moved it around and played with it, using the highest reading you got, depending on the gain and screen properties.

And to stanger89: yes, the LT and 2 are the same hardware.

Why not Meterman LM631 which reportedly (mfg spec) goes to 0.01 lux instead of the 0.1 lux of CA813. Both are priced roughly the same. BTW, I don't know who manufactures LM631, but it appears to be very similar (identical) to Extech Light ProbeMeterT.

However, I can't find these models for sale with NIST certificate.

AEMC CA813 has display resolution of 0.01 lux measuring from 0.01 lux to 19.99 lux maximum. From 20.0 lux till 199.9 lux has display resolution of 0.1 lux.

I'm Thinking To Do A Grayscale Calibration of My KURO Plasma With AVFoundry VideoEQ PRO using AEMC CA813 for Getting The Target Y for each 21-Step Pattern and use Chroma 5 PRO the same time only for correcting the white balance for each pattern.

Has anyone tried this and saw improvement with this way of calibration?

http://a5.sphotos.ak.fbcdn.net/hphotos-ak-snc6/165357_10150107261652442_584947441_7379582_3060216_n.jpg