OK, I got a new toy, a budget light meter .... now I'd like some info on how best to use it...
I think I understand most of the caveats... its probably best for measuring relative changes, can't really be extended to published lumens of commercial pjs, the sensitivities to rgb(individually) are different....yadda yadda ... anyway, I still thinkthere is some value in measuring and comparing between pjs.... but I'd like to get some ideas on standardizing the measurements... I believe the ansi standard is 9 points of measurement on a white screen?
if anyone else has measurements I'd be glad to see them, info should include mean lumens of bulb, setup,reflector type, size of projection in square inches (or meters since thats the standard but we can convert ;-) ... anway my first very rough measurements and info:
average 21 lumens at the screen, (58 lux to 12 varying on a 1.1 square meter screen), totally untweaked standard pj(a haas style so I ahve some serious tweaking to do..I think)
so..anyone have suggestions? any measurements?
edit: my light meter is a LX1010B
edit2: added measurements
MEASUREMENTS:
-----------------------------------------------------------------------------------
Mikyd:
250 Watt MH venture ED28, Ikea reflector, standard lenses, 1.1 m2 projection 4:3
36|50|29
37|78|41 = 46lumens
27|49|34
------------------------------------------------------------------------------------
Ozstang65:
Osram HQIBT400WD 32000Lumens, 5200K, CRI group 1a - Ra 90-100
17" Magview LCD
Spun 2mm Aluminium reflector ,2.1168m² projection 4:3
30|49|30
50|109|56 = 113 Lumens
37|71|45
--------------------------------------------------------------------------------------
MichaelJ:
8" Hami, 250w MH, polished aluminium cylindrical reflector
160/305mm split fresnels,300mm triplet, 1.8m2 projection 4:3
18.0 23.7 16.1
26.0 34.6 23.9 = 46 Lumens
27.0 34.6 25.6
-------------------------------------------------------------------------------------
Simul8r:64"x84" screen,3.47 s/meter screen, 400 watt ushio s400dd, 15" lcd
first his bulb horizontal:
18 37 20
27 57 28 = 31 avg *3.47 = 107 lumens
23 46 23
now his bulb vertical:
23 34 22
37 59 37 = 34 avg *3.47 = 117 lumens
28 40 26
now "Vertical with 4.5" plano convex precondenser using 317mm rear fresnel on same size screen" :
32 33 30
42 48 39 = 37.89 avg * 3.47 = 131
37 44 36
notice how much more even his precondenser/vertical values are and about 20% higher than the horizontal/no precondenser ..nice work Sim!
...and Sim has made more improvements...
Using the Pro reflector settled within my 4" polished stainless steel ladle and using the 317mm standard fresnel front and rear, 4.5" precondenser lens from Surplus Shed and perpendicular to 5.7 transmissive LCD panel that had it's A/G removed and the panel's rear TAC removed; 500:1 contrast, Dukane Starfire OHP SF5510 triplet....oh, and panel turned on....whew.

50 60 51
65 79 65
56 64 55

60.55 average at 3.10sq/mtr projection = 187.70 ANSI
----------------------------------------------------------------------------------------------------------------------

well, plugged some fairly egregious light leaks the last couple of days and tweaked the bulb/reflector position, amazing how you can tell when you hit the "sweet spot"... much more eveness and overall brightness.... took some new measurements:

at the first fresnel location(took fresnel out of the pj and placed meter at that location) I got 16000 lux, very evenly distributed, lux readings only varied by 30 lux from center to corner ....
projected on screen(totally white screen with boxes to indicate where to place he light meter:
16|25|17
21|47|22
11|24|17
the lower left seems lacking .. not sure what to tweak ... so on my 1 square meter screen that works out to 22.2 lumens though it seems much brighter than the 20 lumen avg I got the other night and the corners are brighter, though the center is down.. but a very watcheable picture

anyone got any ideas on how to calibrate lightmeters? ie, when you nuy a bulb that says "800" lumens, what is the standard they measure it? a sphere 1 meter away? any ideas?

next measurement now that I am happy with the brightness is after the rear fresnel, still using unsplit iptics so I'll see what it is after the collector fresnel too, should give me some idea of exactly how transmissive fresnels are in reality(as opposed to in theory)
then of course after the LCD which should give me an idea of how transmissive the lcd itself is... I'm guessing its in the range of 8%, first rough test a few days ago showed 10% when it was turned off, but I was just goofing around...

Take your meter to the Movie Theater next time you go and do a few tests on the screen during the previews.

It would be interesting to see the comparisons. If you don't get thrown out or beat up by the other patrons.

hey, cool idea, maybe i'll take my son with me...kids love when their parents do embarassing things.... hmmm.... wonder how I can estimate the screen size..oh, I suppose I could pace it off ....

I guess it's all about going to the right theater for the right movie. I went to Curse of the Were-Rabbit last night at the local budget theater. No one was there but my family and me so I could have tested all I wanted, oh well. (But I have walked off distances and checked out speaker placement with very curious people watching me before) And I was thinking about this exact subject, or at least pretty close. When one of the green preview screens popped up, it was way dark. I guess color description is about as scientific as I can get right now. Anyway, the theater's green was close to a forrest or hunter green and the screen on my pj has always been brighter (after placing the reflector with the then saturation levels it was acid green) but tends to be a subdued lime or a high-key kelly.

Granted this is an older theater with what I'm sure is a very old projector, but if I can be totally subjective the film was very watchable, but after playing FFX later in the evening I would say that my pj is brighter. Dunno.

Are the light meters people are using the same as photographic light meters? Cuz if they are, I'll snag my little brother's and head off to the local theater again.

From the information that I was able to look up on your light meter, i'm not sure it is calibrated to the CIE (photopic) curve. I believe most lumen output measurements by commercial PJ's are done using a light meter which is calibrated to this CIE curve. The CIE curve is used as it closely resembles the human eye response. I don't believe that you can directly compare your lumen measurements with any stated lumen outputs of commercial PJ's. When I first started looking at lux meters I was looking at a similar unit until SupraGuy pointed out the CIE curve. Measuring ANSI lumens requires a light meter calibrated to the CIE curve, and averages the 9 points (as you have done), based on screen size. As I have no practical experience using a light meter, it would be interesting to see what a CIE calibrated lux meter reads and how it compares to the readings with your current lux meter.

yeah, I didn't really want to compare to commercial pjs, this is more for just LL comparative purposes, I figure most people are going to have the cheaper lux meters like mine..did you ever find an inexpensive one that was CIE?

did you know the typical cinema brightness on the screen is 100 lux?

no....hmmm definitley have to go to movies...I imagine it needs to be a white screen projected? I can run around the screen looking for white sections ;-) ..hey rox, you have a light meter?

yes, white screen, as we are doing with ansi.

I have one, it is color corrected and cosine corrected as well, I think I read CIE standar once when looked at the instructions... I remenber 4% accuracy as well.

I figured you'd have an excellent meter ;-) so what kind of readings have you done? you still working on elliptical?

what bulb do you have? it occurs to me if you know the specs on your bulb(initial and mean lumens) you could hold your light meter 1 meter from your bulb and I could hold mine and I could work out a correction factor or at least see if mine is close to yours(taking into account the difference in mean/initial lumens)

yes i am still workin on the eliptical...

mmm, to calibrate any kind of tool, you need an alreally calibrated one, or a "known output lamp" instead. I don´t have any of those, neither you. I just trust my luxmeter.

My lamp is 150W G12, I know the stated initial lumens, but i don´t think it is enough to calibrate my luxmeter propperly. You have to consider the light distribution on the arc (angle dependent) and light spectrum... as well as K value and CRI.

it is typical to calibrate the luxmeters under a known lamp (tungsten lamp type “A” 2850 K ±10K) they have very high CRI, but I would never modify it myself.

Maybe you could ask a photographer shop to show you some readings with his profesional luxmeter... then you set yours to that value... If you did that, I would take a 100 lux reading on the profesional one as refernce, the lineality on this area is what cares to us. (A reference set at 10.000 lux is stupit for us since your readings will be more like 100lux or less area)

good suggestion, thanks, its more comlicated than I thought, I figured if we both had MH bulbs same temp/spectrum at least I'd see if mine was in the same ballpark as yours...I'll see if I can find a photo shop somewhere....

doing the anti-glare stripping the next day or so...should have complete set of measurements by monday!
1: before collimating fresnel
2: after fresnel
3: after collector fresnel (unsplit optics)
4: after lcd (should show transmissiveness of lcd
5: before triplet
6: after triplet
7: at screen

I may do one without the reflector too, been thinking of making an adjustable mount for it... anyone else have any measurements they've done?

heres the pics I'll be projecting(1024x768) gonna try both and see how they compare as to averages...
Click to view attachmentClick to view attachment

well, sickness(bronchitis of the wife ;-) has stalled things around my house but I managed to get in a quick spot reading on my screen before I dismantled the pj for alterations to a 17" lcd... did a quick spot reading of the center of my projection after I took the lcd out(so both fresnels and the triplet were still in) ..pretty sure I hadn't moved pj so it should be in the ball park anyway(at least within an order of magnitude ;-) anyway got a 1900 lux reading ...since my center reading was 40 or so in the center with the fully loaded pj and the average was 22 lets just cut the 1900 in half to get an idea of the average, so figure 950 lumens(still right around a square meter projection) ,,so typical lcd transmissivieness is about 6% so thats...ummm...lets see....carry the one.... works out to be 56 lumens .... more than twice what I measured ...hmmm.... admittedly a quick and dirty method, but probably in the same yard .... hmmm, I'll have to cogitate on that for a while.... obviously will need to redo all the measurements when I get the big lcd in there... but it got me to thinking....
It seems a pretty solid 16000 average before the first fresnel... so I decided to walk thru the losses from there theoretically...
16000 lux before fresnels (unsplit) - measured
* .8 (for 10% loss thru each fresnel)
-------
12800 lux hitting the lcd - 9 x 12 inches = 108 sq. inches which is around .07 sq. meters so
about 896 lumen hitting the lcd (or the screen if you eliminate the lcd and ignore the triplet which is probably within the margin of error for my measurements I'd guess ;-)...
so really the 950 "guesstimate" I made with the 1900 lux at the screen is pretty close actually.... so how in heck did I get 22 lumens when the lcd was in place... it would have had to be about (oh say 800 lumens hi the lcd and I measured 20 lumens- fudge factors ;-)
sigh ..more math..... about 2.5% transmissive? ....

thinking about it some more...I've seen transmissivity figures for lcds can range from 3 to 9% ... so I guess its possible it was on the low end of the spectrum.... ah well, hopefully I'll get back home saturday and whip out the new sled and remeasure everything from the get go...

possibly not
5% seems like the low end of lcd transmission (hami 8", mva panels...)
the triplet may be losing more lux than you think...

well, I've seen figures for coated optics as low as 2%, the average lcd tras is 6%...I'll hopefully have more accurate measurements with the new lcd this weekend

well I made a new sled last night to hold the 17" lcd... still need to mask it off and set up powerstrip to use only 1024x768 pixels since I have the standard fresnels right now...anyway did a quick measurement before I left town, on a white screen(just a white bitmap pic) the center measured out at 75 lux, down to 40 at the side , 25 at the corner...can't really compare it to earlier measurements too much since this was just a quick and dirty and the triplet was definitely not in the optimum place(off at least .5 inch) but considering the highest reading I got before was 58...its looking good.. I think a large part of it may be due to a repositioning of my reflector .....

come on, I know someone out there has a light meter and has made readings too ;-), anyone?

I do, I have.... but I forgot where I posted them


lemme have a look for them....

cool, thanks, also, if you find them can you tell the lcd size, projection diagonal and the bulb/reflector?

ok....back in town and managed to get down to the lab...tweaked, cut a new front piece with a niceley centered triplet hole, then spent 30 minutes getting rid of the "spiderwebs" caused by misaligned frenesls ;-) then tweaked bulb again so....
new measurements:

9points.gif:
36|50|29
37|78|41
27|49|34

9points2.gif:
19|18
48|47
77
47|54
17|15

ok...let me get the ol' abacus out...
9points average: 42.3
9points2 average:38

projection size: 36" by 48" = 1728 sq. inches * 645 (square mm per sq. inch) = 1114836 = 1.1 sq. meters
so assuming any relation to reality:
9points.gif = 42 * 1.1 = 46 lumens
9points2.gif = 38 * 1.1 = 41.8 lumens ...
using a 250 W venture bulb(ed28) ikea reflector (14000 mean,mmm...18000 initial lumens)

which doesn't sound like much but its way up from my earlier measurements... I may tweak the placement of the bulb again, this time just to make sure it is in the center - width wise in the box...

working on a new light engine so the pj is apart for the moment....good time for someone else to do some measuring?

yeah, I've done some measuring...
nothing to bragg about though

8" Hami, 250w MH, polished aluminium cylindrical reflector
160/305mm split fresnels
300mm triplet
on matt white screen 155*116cm (1.8m²)

22,1 24,5 18,7
24,6 31,8 22
23,5 29,7 21,3

44 lumens average

Might try a pre condensor, or reducing lamp-lcd distance with shorter frezzies...

cool..... which bulb you using? interesting that we have the roughly the same lumens on screen, but I'm essentially using a 13" lcd, mine should be brighter I think.... thanks! gives me more food for thought

hmm... depends what fresnels you're using too
bulb is an osram/radium HQI/HRI UVS 250W daylight, rated for 20000 lumens
these measurements are quite normal from what I've seen

well with the hami you've got about 60% less area to "gather" lumens than I do although thats a pretty short FL rear fresnel, which helps your light quite a bit I would guess.... that may be why even with such disparate lcd sizes we're fairly close in lumens..where'd you get the 160 fresnel?

www.exclusiv-online.com do the 160/305 fresnels
they're only 200mm square (8") though...
Short focal lenses definitely makes sense for small panels

Out of curioisity, have you tried measuring your black levels?

no, haven't measured the black levels yet...what kind of reading do you get? just use like a black bitmap?

Lamp: Osram HQIBT400WD 32000Lumens, 5200K, CRI group 1a - Ra 90-100
Ballast: Atco OGS400-100
Ignitor: Atco AZRM6ES
Capacitor: Atco CP45.0-06
Base: E40 Ceramic
Silicon High Temp (205deg C) wiring
5 Amp circuit breaker

17" Magview LCD
Spun 2mm Aluminium reflector (could have been done better!, not polished up much either)
75" Diagonal 4:3 image




I'll try to get some more recent lux measurements soon.

zoomplayer gives me a nice black screen on startup

meter reads 0.3-0.5lux which is like 60:1 contrast ratio!

I really dont know how meaningful that is considering the luxmeter's resolution and accuracy at those levels though...
this is for a 300:1 lcd with absolutely no light leakage (the projector is behind a wall)

bump....been busy working on new light engines so my pj has been down...when I reassemble it I have a few ideas for better light....an adjustable triplet for one(meaning vertically and horizontally) .....

there is a problem on the luxmeter's resolution (and non lineality for near 0 values). I thought about a posible solution;

place the luxmeter at somewhere "close" to the projection lens, the readings (white and black) will both grow up (somewhere 4000lux for the white one...) then the readings are much closer to the reality since the luxmeter is not working on it's limits. Work out the ratio and you would have the contrast.

This is only for contrast meassuring, do not try working out the lumens output that way.

Hi Rox,
Yes, luxmeter measurements for 0.1 or 0.2 lux are not very useable
However, at the projection lens, I measured contrast as 26000:455 (57:1)

Interestingly, after removing my reflector, contrast went up to 89:1
Replaced reflector with pre-condensor, now measuring 30000:310 (97:1)
so much better contrast, and slightly more light

Somebody suggested I should perhaps try an optical grade spherical reflector...

did you do it by projecting a totally black screen, then a totally white one? never thought of that I'll have to try it...good suggestion rox.... would be interesting to test a commercial pj too the same way.....
how different did the picture look?

mmm, interesing...

is it somewhere 100 lumens the light output on your projector? did you ever meassured it propperly? (ansi)

close enough to ansi...
latest lux readings measured on a 3x3 grid:
18.0 23.7 16.1
26.0 34.6 23.9
27.0 34.6 25.6
image base is 1.8m², so average lumens would be 46...

I have a second plano/convex lens I may slip in there - double pre-condensor

hmmm.... pretty interesting that we both have very different setups, but both being 250 watts we have the same lumens..... your smaller lcd being balanced by the shorter rear fresnel and precondenser.... maybe thats close to the max we can do with a 250...nah ;-)

46 lumens for 250W lamp.... mm makes me think.

the real measurements on lumenlabs, assumed somewhere 200 lumens on average. do you think it is overstated? maybe we should think somewhere 150 as the max goal for a lumenlab? (when i talk about the lumenlab, I am assuming 400W lamp).

well, I know I have seen people state that they have measured 200 lumen..however, if I just measured the center of my screen I'd say I have 75 lumens...so...thats one reason I started the lux/lumens thread.... and if you look at irongeckos' calculator I'd have to say you're closer to the mark rox.... if you look at posts from a year or more ago, people were talking about "1400" lumens being standard..... and I'm not saying I couldn't get 100 lumens(or maybe 90) out of my setup....but I want to know what has actually been acheived ;-) I think I could probably get close to 75 lumens out of mine by more tweaking and precision, I think 200 is tops at best on 400w though and I think everything would have to be perfect, plus you'd need a more transmissive lcd.....we have to remember the differing lcd transmissiveness too..... thats why if we can get a lot of measurements we can start to figure out which parts can be improved on....
edited: for really bad typing

I agree with you. The 1400 lumens was first time claimed by JCB (as far as i remenber). this is somehing to be proved yet.

I would also say 200 lumens is the max achievable too. But conerning to the average, I think 150 is a good starting point (always speaking about ansi, not the center lux reading multiplied by the image area, cheating yourself is the most stupid thing in this world). maybe the PRO setup, will improve the average value to 200-250 lumens... we'll see.

yes, I'd agree...the couple of readings I have seen are in the 110-150 range , also, did you see the images I am testing with?(9points.gif, 9points2.gif) do you think that those are "reasonable", especially the "9points.gif" one... or should the corner measurements be more towards the corner? I don't want to cheat myself(stupid as you say ;-) but I also don't want to understate by too much, I'm not quite sure how to figure a "good" average

well, your 9points.gif is the correct one for ansi lumens meassuring (assuming the white value is 256). I would say the black lines do not change the reading value in a measurable amount).

Is just the center of the 9 equal boxes. If we are going to compare the values, we should all use those points.

edit; i said about the 256 value because it looks like there is some kind of compresion on the images uploaded to lumenlab, i guess the originals are intact).

well, I'm not sure rox... all I did was choose white for the the background color...I never actually looked at the color values, I'll do that and reupload them if its wrong.... yeah I thought 9points was the correct one, though I like the other one just for testing the corners...

edit: I just downloaded the first one and opened it up and according to Paint(all I have at work) the values are 255 for R, G and B..... is that what you meant? if you can come up with a more accurate one , have at it!

its ok, just a note;

i believe when comertial manufacturers state their projectors output, they preset it to a "best measurement preset", nothing to do with "best image preset". So if you are just stating the projector, you can also do this kind of cheat . My advice is to preset the projector (contrast, brightness, gamma....) to best looking image. We do not need to cheat as comertials do.

Another critisiced point is how true comertial specs are. But this is not today's question .

regarding the 250w/400w thing:
it's highly likely that 400w wont gain much over 250 (with the standard optics) due to the larger arc size...
and for that reason optimising a std 400w setup is also likely to be considerably more difficult

Rox, you can have a look at commercial pj ansi measurements in this consumer report:
http://allinbox.free.fr/Programmes/videoprojecteurs_2006.pdf

the lowest, a SANYO PLV-Z1X, is given for 273lumens
commercially rated as 700

I'll be quite happy if mine reaches 100 lumens
150 may be possible with a very transparent panel, or a large triplet(?)

I might be wrong but the arc is pretty much the same size on the 250/400 wat ventures anyway..... don't know about hqi etc...

edit: a 250 watt PS venture ED28 has a 30mm arc legnth, the ushio s400dd(LL stabdard) has 27mm arc length